H.-P. Nothacker, Department of Pharmacology, University of California, 354 MedSurge II, Irvine CA 92697-4625, USA Fax: +1 949 824 4855 Tel: +1 949 824 1892 E-mail: firstname.lastname@example.org
The discovery of novel biologically active peptides has led to an explosion in our understanding of the molecular mechanisms that underlie the regulation of sleep and wakefulness. Urotensin II (UII), a peptide originally isolated from fish and known for its strong cardiovascular effects in mammals, is another surprising candidate in the regulatory network of sleep. The UII receptor was found to be expressed by cholinergic neurons of laterodorsal and pedunculopontine tegmental nuclei, an area known to be of utmost importance for the on- and offset of rapid eye movement (REM) sleep. Recently, physiological data have provided further evidence that UII is indeed a modulator of REM sleep. The peptide directly excites cholinergic mesopontine neurons and increases the rate of REM sleep episodes. These new results and its emerging behavioral effects establish UII as a neurotransmitter/neuromodulator in mammals and should spark further interest into the neurobiological role of the peptide.
Urotensin II (UII) is a peptide structurally related to somatostatin/cortistatin peptides. It contains a carboxyterminal cysteine-bridged cyclic hexapeptide sequence that is conserved across species. UII was originally isolated from fish urophysis, a neuroendocrine gland located in the caudal part of the spinal cord, using a trout hindgut contraction assay . For the decade following its discovery UII was regarded as an exclusive product of the teleost urophysis. Contrary to this belief, UII peptides have been shown to have a wide phylogenetic distribution across the vertebrate lineage (reviewed in ) and independent reports touting UII's potent cardiovascular effects in rats have dispelled its mammalian irrelevance [3,4]. These important studies inferred for the first time that a specific mammalian UII receptor must exist and hence postulated the existence of mammalian UII-like peptides. The genes for the mammalian orthologues coding for the preproform of UII were finally discovered in 1998 [5,6]. Shortly after, in the race for the identification of natural ligands for orphan G-protein coupled receptors (GPCRs) several groups including ours reported the identification of the UII receptor, an orphan receptor known before as both GPR14 or SENR (for sensory epithelium neuropeptides-like receptor) [7–12]. When the UII receptor was identified in 1999, all of the immediate research was concentrated on the elucidation of UII's vasomodulatory properties. Indeed, the research direction was warranted by mounting evidence for a role of UII in the pathogenesis of cardiac, renal and hepatic disease (reviewed in [13–16]). In the fervor of cardiovascular research successes it seemed to be forgotten that UII was originally identified as a neuropeptide and that its major expression sites in mammals were confined to certain motor nuclei of the central nervous system (CNS). From the very beginning, these findings pointed to a possible neuromodulatory role for UII.
The aim of this review is to provide an overview of the UII system and to describe the recently accumulated evidence of its association with centrally acting functions in mammals, particularly with regard to arousal and sleep.
Molecular constituents of the UII system
Presently, three molecules are known to form the UII system: the UII receptor and the peptides UII and urotensin-related peptide (URP). In humans, both peptides are biologically synthesized from their own respective 124 and 119 amino acid long secretory preproforms that are encoded by distinct genes (Fig. 1), located on chromosome 1 and 3, respectively. Although both peptides share a considerable amount of identity, the precursor molecules differ substantially in their amino acid sequence. The gene structures, however, have many structural features in parallel, such as the number of exons and placement of the active peptide sequence at the very C-terminal end. These conserved features greatly suggest that both have originated from a common ancestral protein through gene duplication and therefore should be considered paralogue genes.
The precursors are proteolytically cleaved to produce the mature form of the biologically active peptides (Fig. 1). UII's length is heterogenous between different species and ranges from 11 amino acids for human to 14 amino acids for rat. This is probably due to the lack of a highly conserved cleavage site for pro-hormone convertase (PC) enzymes in the upstream region of the mature peptide. This raises the possibility of generating peptides of multiple lengths using atypical processing. On the other hand, the active peptide of the URP precursor is amino terminally flanked by a dibasic processing site and can be processed by typical PC cleavage, thus generating a cysteine bridged octapeptide whose entire sequence is found invariant in all mammals sequenced to date .
The major biologically active part of both molecules consists of a canonical cysteine bridged hexapeptide ring with the sequence CFWKYC, also known as core, which is invariant between species and/or paralogues. Destruction of the cysteine bond by covalent modification leads to an immediate loss of biological activity . Conversely, the amino terminus of both UII and URP from mammals does not seem to carry much biological information, as it can be chemically modified without significant loss of activity [19,20]. It is worthwhile to mention that most of the mature mammalian UII structures have been deduced from genomic sequences and so far only pig UII and rat URP have been isolated and directly sequenced from hypothalamus and whole brain, respectively [9,17].
All UII isoforms identified so far contain an acidic amino acid residue (aspartic or glutamic acid) that directly precedes its cyclic core structure (Fig. 1) . This particular amino acid is absent in URP, but the peptide is still a potent agonist at the UII receptor with a pharmacological profile equal to that of UII [17,20]. Therefore, this residue is not necessary for receptor activation. However, its stringent evolutionary conservation in all known UII isoforms implicates a not yet understood biological function related to this residue. It can be speculated that a receptor subtype exists that is able to distinguish between UII and URP peptides, but in mammalian genomes no homologous sequences can be found that would clearly provide evidence for a putative UII receptor subtype. Pharmacologically, there is, so far, no indication of the existence of a UII receptor subtype, but the possibility can not yet be ruled out. The emergence of a UII receptor selective antagonist offer the tools to tackle this question [16,21].
Alternatively, the presence or absence of the acidic residue may also influence the pharmacokinetic profile of the peptides. In rat spinal cord motor neurons UII and URP seem to colocalize . This apparent redundancy would only make biological sense if the peptides act on different receptor subtypes or with a different pharmacokinetic profile.
Little is known about the actual steps in the biosynthesis and degradation of the peptides. The expression of both precursors in a wide range of tissues ranging from cardiovascular to CNS suggests cell type specific differences in biosynthetic pathways due to the differential expression of proteolytic enzyme in neurons, endothelial and smooth muscle cells. Direct evidence has been presented for a UII converting enzyme activity present in porcine renal tissue . This is an intriguing finding, as differential expression of this or similar enzymes would have dramatic influences on UII functional dynamics in different tissues. The present knowledge of UII biosynthesis has been recently reviewed .
The UII receptor belongs to the large family of GPCRs. It was originally cloned by homology screening methods as an orphan receptor called GPR14 or SENR. Human UII receptor is located on chromosome 17q25 as an intronless gene that codes for a 389 amino acid long polypeptide. The receptor exhibits the prototypical GPCR serpentine structure containing seven transmembrane domains alternately interspersed by intra- and extracellular loops (Fig. 1). Phylogenetic analysis of the receptor's primary sequence shows striking similarities to the somatostatin receptor family, most notably in the transmembrane domains. The pharmacological relatedness of somatostatin receptors is demonstrated by the fact that the UII receptor is activated by micromolar concentrations of somatostatin derivatives and can be blocked by somatostatin antagonists [8,24]. However, it is rather implausible that somatostatin acts at UII receptor sites in vivo under normal physiological conditions, because UII is more than 10 000-fold more potent and acts at picomolar concentrations in a quasi irreversible manner.
Initially the receptor expressed in heterologous expression systems was thought to exclusively interact with the Gαq11 subclass of heterotrimeric G-proteins which leads, via activation of phospholipase C (PLC), to a rapid and short-lasting rise in intracellular calcium ions. Further linkage of UII receptor to PLC was provided by studies performed with isolated rabbit thoracic aorta strips in which UII application increased inositol phosphates, an effect that could be blocked by PLC inhibitors . Besides the short-term effect, a line of evidence points to additional long lasting UII-mediated cellular changes that result in growth stimulation and cellular remodeling. Those effects have been shown to be transduced via extracellular signal regulated kinase 1/2 (ERK1/2) pathways in cardiomyocytes, and in transfected Chinese hamster ovary cells. Those effects are independent of both calcium ions and protein kinase C (PKC), and have been suggested to occur via transactivation of the epidermal growth factor receptor . UII activation of ERK1/2 pathways has been suggested to play a role in cardiomyocytic hypertrophy, a cellular adaptive response that is characterized by an increase in cell size and protein content in the absence of cell proliferation.
In the CNS, indirect evidence exist that UII also activates ERK and Rho-kinase pathways. In rat, centrally administered UII causes cardiovascular responses that were attenuated by Rho and ERK-kinase inhibitors , but the cellular targets of this effect remain elusive. Short-term UII receptor activation examined by Fura-2 imaging in dissociated rat spinal cord motor neurons showed an influx of extracellular calcium via N-type calcium channels that could be prevented by N-type specific channel blockers . Additional experiments suggested an activation of the channels via proteinase kinase A dependent phosphorylation and no involvement of PKC in the process.
While the UII receptor was initially considered to only couple Gαq proteins, the receptor seems to be involved in an array of interactions with other signaling molecules whose full spectrum remains to be determined.
Tissue localization of UII receptor and peptides
Early studies using hybridization techniques revealed strong mammalian UII expression in only restricted areas like the spinal cord, medulla oblongata and kidney [6–8]. Recently, more sensitive RT-PCR techniques have shown a more ubiquitous distribution of UII mRNA in various tissues and blood vessels. A recent survey of both UII and URP transcripts in rat and human showed species specific expression patterns for both genes . The most common denominator of UII expression in rat and human seems to be spinal cord, which constantly exhibits highest expression independent of the detection technique. URP is also ubiquitously found, but in rather low expression levels compared with UII. The exception, where URP clearly exceeds UII expression, is found in human reproductive tissue. In rat reproductive tissue, however, UII exceeds URP expression pointing to species specific differences in the expression levels of these two molecules. Interestingly, URP levels in various human brain tissues seem to be slightly higher than UII, pointing to URP as a possible activator of central UII receptors. URP's role as a centrally acting ligand, at least in rats, is supported by the fact that it is the only peptide that can be isolated from rat brains when monitored with a nonselective UII/URP immunoassay . On the other hand, UII has also been found and sequenced from porcine hypothalamus and purified to advanced degree from human spinal cord [9,29], and therefore can also exist in neuronal tissue, at least in certain species. UII-like immunoreactivity, which might reflect the peptide and/or the precursor molecules, has been localized to cholinergic neurons of rat brainstem motor nuclei . These are the same motor nuclei that show strong expression of the UII transcript, but are absent of URP (H.-P. Nothacker & S.D. Clark, unpublished results). Therefore it is very likely that both peptides coexist in the brain, but are expressed in different areas and at different levels to fulfil diverse neuromodulatory roles. It has been recently reported that UII and URP are both expressed by spinal cord motor neurons, although it is not yet clear if both peptides are also colocalized in the same neurons .
UII receptor expression seems rather ubiquitous when assessed with highly sensitive RT-PCR technology. It has to be kept in mind however, that the UII receptor might be present in microvessels of the investigated tissue and not directly expressed in the tissue specific cell populations. That might be one explanation for the sometimes controversial results using different detection techniques that vary in their sensitivities. For example, the UII receptor expression in rat brain seems to be ubiquitous when assessed by RT-PCR . Those results substantially differ when less sensitive in situ hybridization methods are used. In the latter case UII receptor mRNA exclusively localizes to brainstem cholinergic neurons of the laterodorsal tegmental (LDT) and pedunculopontine tegmental nuclei (PPT), and no other expression sites could be found . Further investigation of the receptors' site of action using in situ binding techniques revealed a much broader expression of the receptor. However, the binding sites matched the projection areas of the LDT–PPT complex leading to the hypothesis that the binding sites might represent presynaptic neuronal terminals. Because cholinergic neurons of the LDT–PPT are functionally associated with sleep and wake states, it led to the hypothesis that UII receptor might be involved in the regulation of sleep-wake cycles. Functional UII receptors are also detected in cholinergic motor neurons of the spinal cord, possibly in the same neuronal population that expresses UII peptide [10,28]. UII's role in spinal cord motor neuron physiology remains unclear, however, in frog sciatic-sartorius nerve-muscle preparations UII has been shown to modulate frequencies of miniature endplate potentials .
In the future more in-depth anatomical studies will be necessary by using combinations of immunohistochemical and in situ hybridization techniques in order to fully understand the extent of the UII system. The involvement of the UII system in sleep also points to a possible circadian regulation of the transcript and/or protein levels that may underlie the reported differences.
Neurophysiology of UII
Studying neurophysiological effects of UII is complicated by the fact that intracerebroventricular injections of UII lead to cardiovascular responses [4,27] that are thought to be caused by a UII activated and centrally located regulator in the brain stem . In general, central administration of UII leads to hypertensive and tachycardic responses [27,35,36] and the hypothalamic paraventricular (PVN) and/or arcuate nucleus have been suggested as anatomical substrates . Parvocellular neurons of the PVN synthesize corticotrophin releasing factor (CRF) , which had been described 50 years ago as one of the most potent secretagogues of adrenocorticotropic hormone (ACTH) [38,39]. Ewes, centrally injected with UII, responded with increased plasma levels of adrenocorticotropic hormone and adrenalin providing the first direct evidence for a role of UII in the activation of the hypothalamic-pituitary axis  although the pathway remains mysterious. An increase of c-fos immunoreactivity, a general indicator of neuronal activity, could not be found in the PVN  two hours after UII injections. Because c-fos expression temporally progresses to downstream neuronal circuitries, and the latter study recorded only one time point, it is possible that activation of PVN neurons could have simply been missed. In preliminary studies also carried out in rats we have seen a significant increase of c-fos mRNA levels in PVN 20 min after intracerebroventricular (icv) administration. The PVN plays an important role in stress and arousal related behavioral responses (reviewed in ) and several stress-related responses have been reported in rodents after central UII injection. Low doses of icv administered UII into rats led to a dose-dependent increase in locomotion in a familiar environment and a dose-dependent increase of ambulatory movements, although the effects were less pronounced as compared with CRF and orexin, which are both much stronger inducers of locomotion and arousal in rats .
In mice central UII injections caused anxiety-like behavior assessed by two different paradigms: the elevated plus maze and the hole-board head-dipping test . UII acted dose-dependently in an anxiogenic fashion the same way as CRF, but UII effects were much smaller and less potent. There were marked differences between the two peptides in regards to locomotion. CRF administration strongly inhibits locomotion in mice, whereas, UII does not show any effects. Therefore, UII's behavioral profile is different from that of CRFs and warrants further investigations.
Investigations of the central actions of UII/URP, as possible interchangeable agonists, have only recently begun and a clear neurophysiological role of UII has yet to emerge. Complicating matters is that, similar to UII's cardiovascular action, species specific differences exist in the UII mediated behavioral responses that need to be diligently addressed. It also has to be kept in mind that UII is a potent vasoconstrictor and icv injected UII causes a significant increase in cortical blood flow . It is therefore conceivable that the behavioral effects seen after icv administration of UII are consequences of both its effects on cortical blood flow and on neuronal activation. The generation of UII receptor knockout mice has been completed, but so far only investigated with regard to the modulation of the cardiovascular system . Once these mice are made available by their industry-funded creators, they will provide valuable information with regard to the behavioral role of the UII system.
Sleep is an essential part of the daily life cycle and as important as food intake for survival. In mammals the identification of sleep and wake is accomplished by polysomnography, the simultaneous recording of electrophysiological potentials measured in the cortex and skeletal musculature of the neck and eye. Sleep exhibits a distinct architecture that can be basically divided into three different states; wakefulness, slow wave and rapid eye movement (REM, also known as paradoxical sleep). In most laboratory animals such as rodents the three stages form a cycle that is repeated many times during the entire sleep period, but occurs less frequently in humans. In waking states, neurotransmitter systems governed by norepinephrine, serotonin, and acetylcholine are all activated in the brainstem, while in slow wave sleep, all are suppressed. Wakefulness is accompanied by fast, low-voltage electrical activity in the cortex and the subcortical structures of the brain, and by a significant amount of tonus in the skeletal muscles. REM sleep represents a paradox sleep state in the sense that electrical activity in the cortex is similar to that seen in wakefulness while electrical activity in muscles has disappeared.
REM sleep is related to the specific activation of cholinergic circuits located in the LDT and PPT of the pons-midbrain transition area (Fig. 2). Lesions of the cholinergic LDT–PPT complex lead to a loss of REM sleep without deficits in wakefulness and cortical activation . The injection of cholinergic agonists into the medial pontine reticular formation, a target region of the LDT–PPT induces a REM-like state [47–50]. These and other studies clearly indicate that cholinergic neurons in the PPT are important regulators of REM sleep (reviewed in ).
The expression of UII receptor mRNA in cholinergic LDT–PPT neurons led to the hypothesis that UII might influence cholinergic LDT–PPT neuron activity and alter REM sleep patterns in rats. Recent data provide evidence that UII in fact acts as a modulator of REM sleep . Local administration of UII into the PPT led to a significant increase in the number of REM sleep episodes and this increase could be blocked by a UII antagonist. Wakefulness and slow wave sleep were not significantly affected by UII indicating a unique profile not seen with other sleep related peptides like hypocretin/orexin that enhances wakefulness and suppresses REM sleep . Similar effects of UII on REM sleep episodes were observed by icv route of administration, but analysis of high-frequency electroencephalogram bands suggest qualitative differences in cortical activation between the two routes of administration. Moreover, icv administered UII also led to an increase in cortical blood flow, which points to the direct activation of cerebral vasculature or the activation of brain areas that are involved in cardiovascular regulation. The noradrenergic A1 area, located in the lower medulla has been identified as a possible neural substrate of UII's central cardiovascular action because microinjections of UII into A1 causes strong systemic cardiovascular responses in anesthetized rats .
At first UII's ability to increase the number of REM sleep episodes may seem at odds with the earlier discussed observations, which point to more anxiogenic and stress related properties. This apparent contradiction may be due to the time course of the effect, and the route of administration. The studies describing increased locomotion and anxiety were measured soon after UII application (within one hour) and utilize icv route of administration that is known to increase cortical blood flow  within the same period of time. Huitron-Resendiz et al. , reported that when UII was applied icv, a significant increase in wakefulness was observed during the first hour postinjection. The amount of wakefulness returned to control levels after two hours, whereas, the increase in the number of REM sleep episodes could be observed for up to five hours. Local UII application into the PPT did not produce any effect on wakefulness, nor was there an increase in cerebral blood flow, suggesting the effect on wakefulness was not due to UII action on the PPT. Therefore, the acute effects that have been measured and interpreted as anxiogenic, or stress related, may be due to UII induced changes in cerebral blood flow or other cardiovascular changes that follow a more rapid time course.
The most direct evidence for UII neuromodulatory properties in the PPT comes from electrophysiological studies . UII evoked membrane depolarization and promoted firing of cholinergic PPT neurons that were identified by immunohistochemical means while neighboring noncholinergic neurons were not affected. These data show for the first time that UII is able to directly activate central neurons and may even act as a neurotransmitter. Together, the results are consistent with the idea that UII plays a role as a modulator of REM sleep and establishes UII as neuromodulatory peptide in addition to its well known cardiovascular properties.
To establish UII as an important player in the modulation of cholinergic LDT–PPT neurons, one central questions remains: where are the UII positive neurons located that project to and activate the cholinergic mesopontine neurons? UII and URP are both candidates that could pharmacologically fulfil this task because both act with similar potencies at the UII receptor [20,53]. Interestingly, URP seems to be the one that exists in rat brains as a peptide, whereas, UII can only be found in the form of mRNA transcripts [5,17]. This suggests that URP might possibly be the endogenous regulator of REM sleep, at least in rat. Preliminary data collected in our laboratory from rats show only diffuse URP mRNA expression in certain areas of the medulla oblongata that are not known to project to the PPT. There will be a need to produce selective tools (antibodies, UII/URP transgenic mouse strains) that distinguish between the two precursors and that can be used to create a comprehensive map of the anatomical circuitry of the UII system. An established UII circuitry will help to build hypothetical models that can be experimentally tested and that will aid the understanding of UII's interaction with other sleep and wake promoting systems.
The discovery of UII as a direct activator of central cholinergic neurons and a modulator of REM sleep features again the power of reverse physiological strategies to gain new insights into little understood physiological processes. It also introduces a novel player in the neurochemistry and the electrophysiology of the complex basis of sleep regulation and offers the opportunity to study the UII system in relation to other sleep associated neuropeptides. More sleep studies will be necessary to establish UII's role in REM sleep modulation. Besides the sleep data, there is an accumulation of evidence that the UII system might have a general role in promoting acetylcholine release and act as an excitatory modulator. The availability of mouse models devoid of UII receptor or its ligands will provide a rational approach for the investigation of the neurophysiological role of the UII system. While our understanding of the molecular constituents of the UII system in mammals seems to be complete, the anatomical distribution of the precursors on the protein level still remains to be established in order to understand the interconnectivity of the UII system with other brain circuitries. Promising UII receptor antagonists  are already in early phase clinical development for diabetic nephropathy. Depending on their pharmacodynamic profiles, these antagonists could provide invaluable pharmacological tools to further extend sleep studies, and will be critical in evaluating the physiological effects of UII in humans.
We would like to thank our many excellent collaborators who have contributed to the studies reviewed here. The work of Hans-Peter Nothacker is supported by Grant MH68396 of the National Institute of Mental Health.