INTRODUCTION

TOP ABSTRACT INTRODUCTION OVERVIEW OF OREXINS AND... CELLULAR RESPONSES TO OREXINS DISTRIBUTION OF OREXINERGIC... SYSTEMIC EFFECTS OF OREXINS FUTURE PERSPECTIVES REFERENCES

OUR KNOWLEDGE of the orexinergic system was initiated in 1998 by two independent groups and approaches (reviewed in Refs. 26, 157). In January 1998, the group of Sutcliffe published a study (37) in which they predicted that a particular rat hypothalamic mRNA species ("clone 35"; Ref. 50) would code for a precursor peptide 130 amino acids (aa) in length, each molecule of which would generate two separate peptides of 39- and 29-aa lengths. The peptides were named hypocretins, on the basis of their hypothalamic localization and their proposed sequence similarity to the secretin family of peptides; however, only orexin B bears any significant resemblance to secretin. The precursor peptide would thus be preprohypocretin and the two final peptides hypocretin (hcrt)-1 and hcrt-2. Simultaneously with the work of the group of Sutcliffe, the group of Yanagisawa (154) isolated a hypothalamic 33-aa peptide that activated the orphan receptor HFGAN72. This receptor was found also to respond to another peptide, 28 aa in length. A second receptor was cloned on the basis of its sequence similarity to HFGAN72, and this receptor also responded to both isolated peptides with Ca²⁺ elevation. On the basis of the peptide sequences, Sakurai and coworkers (154) cloned the cDNA for the precursor peptide (130 aa in the rat). The peptides were named orexins, because they increased food intake in nonfasted rats when injected into the lateral ventricle. The precursor peptide thus became preproorexin, the 33- and 28-aa final peptides orexin A and orexin B, respectively, and the receptors OX₁ and OX₂ receptors. Both preprohypocretin and preproorexin were found preferentially in the rat brain and more specifically in the hypothalamus (37, 154). Cellular and systemic responses to the peptides were observed in the rat. Application of hcrt-2 increased postsynaptic current frequency in rat hypothalamic neurons in culture (37), and orexin A and -B increased food intake in nonfasted rats (154).

It soon became clear that the peptides isolated by the two groups were, in principle, identical. However, there was an error in the peptide sequences deduced by de Lecea and coworkers (37), so that most authors refer to the accepted sequences by Sakurai et al. (154) and not to de Lecea et al. (37) even when they are using the hypocretin nomenclature. Therefore, the current praxis makes preproorexin = preprohypocretin, orexin A = hcrt-1, orexin B = hcrt-2, OX₁ receptor = hcrtr-1, and OX₂ receptor = hcrtr-2. There is an ongoing argument concerning which nomenclature should be used. To avoid confusion with the sequences we use the orexin nomenclature throughout this review.

Today there is a wealth of information concerning the anatomic architecture of the orexinergic system as well as the systemic effects of orexins. The most prominent and well-demonstrated effect of orexins is undoubtedly regulation of sleep/wakefulness (reviewed in Refs. 74, 172, 190). Although many other responses have been shown, their physiological significance is less clear. Effects on feeding behavior have been demonstrated in a multitude of studies, but the results are somewhat contradictory. The current view is that orexins may rather regulate short-term appetite, energy metabolism, and feeding-associated processes (e.g., gastrointestinal functions, mastication). Orexins also affect hormonal secretion, the most well-demonstrated effects being those on the hormones involved in the stress response and energy metabolism such as glucocorticoids and norepinephrine. At the same time, information concerning the regulatory mechanisms of orexinergic systems and their targets on a cellular level is lagging behind. In this review, we try to create a comprehensive picture of orexins and their proven and putative physiological role. The main emphasis is placed on functional properties and interactions between the orexin-producing and orexin-responding cells.

	OVERVIEW OF OREXINS AND OREXIN RECEPTORS

TOP ABSTRACT INTRODUCTION OVERVIEW OF OREXINS AND... CELLULAR RESPONSES TO OREXINS DISTRIBUTION OF OREXINERGIC... SYSTEMIC EFFECTS OF OREXINS FUTURE PERSPECTIVES REFERENCES

Genetics and Chemistry of Orexins

Cleavage of one molecule of preproorexin and further modification leads to production of one molecule each of orexin A and orexin B; however, HPLC detection of orexins in the human brain suggests two- to fivefold higher levels of orexin B than orexin A in several areas of the central nervous system (CNS) (34, 35, 125, 126). Mammalian orexin A in rat is a 33-aa peptide with two intrachain disulfide bridges, and orexin B is a 28-aa linear peptide (Fig. 1A). There is a substantial sequence identity in the COOH termini of these peptides. In addition, orexin B and secretin contain an identical stretch of seven amino acids. Orexin A and -B are likely to be COOH-terminally amidated, and the NH₂-terminal glutamine residue of orexin A may be modified to a pyroglutamoyl residue (154); however, the posttranslational modifications have been experimentally verified only in the rat. Orexin A in human, pig, dog, rat, and mouse are identical (Fig. 1B), whereas orexin B in pig and dog differs by one amino acid and orexin B in rat and mouse by two amino acids from human orexin B (Fig. 1C). The three-dimensional structure of orexin A is not known, but orexin B has been determined to consist of two -helices at a 60-80° angle to each other (100).

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Alignment of orexin A and orexin B peptide sequences from different species (A-C). Gray boxes indicate dissimilar amino acids (as compared with human orexin variants). Cystine bridges in orexin A are shown in A with black lines. D: OX1 receptor antagonist SB-334867. hcrt, Hypocretin.

Orexin A seems to have much higher stability than orexin B in the physiological milieu (89), which may explain why orexin A is more readily detected in cerebrospinal fluid (CSF) than orexin B (148). Orexin A also displays much higher lipid solubility than orexin B, probably making orexin Ain contrast to orexin Bblood-brain barrier permeant (89). Orexin immunoreactivity has been found in putative endocrine release sites of hypothalamus, neurohypophysis, and pancreas (see below) offering the interesting possibility that orexins could also act as hormones between the CNS and the periphery. Yet indirect evidence suggests that orexin A in plasma originates from peripheral sources (32), and no brain penetration of intravenous orexin A has been seen in the rat (11).

Orexins from Xenopus laevis have also been described (163). In this species, preproorexin is rather similar to that in humans (overall identity = 56%). Xenopus orexin A is 31 aa long, lacking two NH₂-terminal amino acids seen in human orexin A, and it contains 6 aa substitutions compared with human orexin A (Fig. 1B). Xenopus orexin B is 28 aa long, and it contains 7 aa substitutions compared with human orexin B (Fig. 1C). Both Xenopus orexins are supposed to be posttranslationally modified in the same way as the rat orexins (disulfide bridges and COOH-terminal amidation) except that Xenopus orexin A does not contain the NH₂-terminal pyroglutamoyl residue.

Orexin Release

Orexin Receptors

View larger version (45K): [in this window] [in a new window]   Fig. 2.   Human orexin receptor sequences in putative 7-transmembrane plots. Sequences are from the SwissProt and from Ref. 154. Putative transmembrane domains are represented as in the GPCRDB database (Ref. 68; http://www.gpcr.org/7tm). The putative N-glycosylation sites were calculated with NetNGlyc (R. Gupta, E. Jung and S. Brunak, unpublished; http://www.cbs.dtu.dk/services/NetNGlyc) and the putative phosphorylation sites for protein kinase A (PKA) and C (PKC) and Ca2+/calmodulin kinase II (CaMKII) with PhosphoBase (Ref. 95; http://www.cbs.dtu.dk/databases/PhosphoBase). Possible palmitoylated cysteines are also shown. Amino acids that may represent polymorphisms are marked with *, , or  (see Orexin Receptors and Narcolepsy, a Disorder of the Orexinergic System).

	CELLULAR RESPONSES TO OREXINS

TOP ABSTRACT INTRODUCTION OVERVIEW OF OREXINS AND... CELLULAR RESPONSES TO OREXINS DISTRIBUTION OF OREXINERGIC... SYSTEMIC EFFECTS OF OREXINS FUTURE PERSPECTIVES REFERENCES

In the original studies on orexin receptors two cellular responses were identified: recombinantly expressed orexin receptor strongly elevated Ca²⁺ in Chinese hamster ovary (CHO)-K1 cells (154), whereas in the hypothalamic neurons increased action potential frequency was observed (37). Many systemic responses were described thereafter but the molecular mechanisms of these have not in most cases been investigated in detail.

Ca²+ Elevation

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Extracellular Ca2+ dependence of the orexin-stimulated Ca2+ response in rat medial hypothalamic neurons in culture. The data demonstrate the extracellular Ca2+ dependence of the orexin-stimulated Ca2+ response. Cells were stimulated with 1 µM rat hcrt-2 (H) in the presence of 1 mM or ~0 mM (0 Ca2+/EGTA) extracellular Ca2+ concentration ([Ca2+]o), and intracellular Ca2+ levels were monitored with fura 2 methodology. As shown, the response to hypocretin 2 was completely blocked by removal of extracellular Ca2+. The experiment was performed in the continuous presence of 1 µM tetrodotoxin (TTX). Rat hcrt-2 is identical to rat orexin-B2-28. Figure is reproduced from Ref. 183 with permission (Copyright 1998 by the Society for Neuroscience).

What then is the general role of Ca²⁺ influx in orexin receptor-mediated cellular responses? In neuronal cells, Ca²⁺ influx has been observed by fura 2 measurements (181, 182-184) and is also suggested by electrophysiological measurements (44, 70). Because orexins may depolarize neurons via block of K⁺ channels (see Depolarizing actions), some Ca²⁺ influx may even result from an indirect activation of voltage-gated Ca²⁺ channels. The situation may be more complex at other sites as discussed below.

In CHO cells, Smart et al. (168) and Lund et al. (107) discovered that removal of extracellular Ca²⁺ caused a significant (8- to 100-fold) drop in the potency of orexin A for the OX₁ receptor (Fig. 4A). In other respects, the results of Smart et al. (168) suggest qualities expected from G_q-dependent signaling, that is, sensitivity to the phospholipase C inhibitor U-73122, rapid and transient spike, and extracellular Ca²⁺-dependent secondary phase (capacitative Ca²⁺ entry). In contrast, a more detailed analysis of Ca²⁺ signaling in CHO cells indicated that the dependence of the Ca²⁺ response on extracellular Ca²⁺ is due to the activation of a Ca²⁺ influx pathway (107). The Ca²⁺ response was abolished when the membrane potential was held close to the reversal potential of Ca²⁺ (Fig. 4B), and a Ca²⁺ elevation could be observed by returning to a negative holding potential (107). In addition, orexin A activated an influx of Mn²⁺ ions in conditions where no intracellular release occurred. The primary response to OX₁ receptor stimulation in CHO cells is thus receptor-operated Ca²⁺ influx (107). This Ca²⁺ influx amplifies phosphatidylinositol-specific phospholipase C, so that in the absence of Ca²⁺ influx, IP₃ production and Ca²⁺ release occur with a potency 100 times lower than in the presence of Ca²⁺ influx. Ca²⁺ elevation alone is not sufficient to activate phospholipase C in these cells, suggesting that Ca²⁺ influx acts in concert with a G protein-mediated mechanism. The identity of this Ca²⁺ influx pathway and its activation mechanism by the orexin receptors remains unresolved; however, the potency of pharmacological blockers suggests that it is a different molecular entity than the store-operated Ca²⁺ channel in CHO cells (96). It thus seems that both in neuronal and nonneuronal cells orexin receptors activate both Ca²⁺ influx and the phosphatidylinositol-specific phospholipase C pathway, a view that is supported by a study in which orexin receptors are heterologously expressed in a variety of different cell types (Holmqvist, Åkerman, and Kukkonen, unpublished observation). In a recent study, orexin A and -B equipotently increased norepinephrine release from rat cerebrocortical slices. When extracellular Ca²⁺ was removed, the maximum responses to both peptides were lowered by 50% and the EC₅₀ values shifted to >10-fold higher concentrations (66). Although the response mechanisms have not been clarified with direct Ca²⁺ measurements, the overall pattern is very similar to what is seen with recombinant CHO cells (107).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Extracellular Ca2+ dependence of the orexin-stimulated Ca2+ response in CHO cells expressing human OX1 receptors (CHO-hOX1). A: reduction of [Ca2+]o to 140 nM shifts the orexin A concentration-response curve 2 orders of magnitude to the right. B: orexin A-mediated Ca2+ response can be blocked by reducing the driving force for Ca2+ entry. A CHO-hOX1 cell was patch-clamped in the whole cell configuration to regulate its membrane potential, and therefore the driving force for Ca2+ entry, and Ca2+ transients were measured with fura 2. Gray boxes and black bars indicate presence of 10 nM orexin A in the bath. Under the light gray box, the cells were exposed to orexin A at the holding potential +60 mV, and under the dark gray box the cells were exposed to orexin A at the holding potential 50 or 40 mV. Inset: voltage application profile. Ca2+ response to orexin A is only seen at the negative holding potential, which indicates that Ca2+ influx is required for the response. Figure adapted from Ref. 107.

Functional experiments indicate that in addition to Ca²⁺ release and activation of store-operated Ca²⁺ channels, there are several types of Ca²⁺ influx pathways activated by G protein-coupled receptors (reviewed in Ref. 7). Although in most cases the mechanisms have not been clarified, these pathways may be activated directly by G protein subunits or second messengers. These pathways may be identical or related to the transient receptor potential (TRP) ion channel family (reviewed in Ref. 29). The TRP genes encode a family of over 20 proteins. Many of the TRP gene products form Ca²⁺-permeable nonselective cation channels, and some members of the TRP (super)family even have inherent protein kinase or ADP-ribose pyrophosphatase domains. TRP channels are widely expressed in mammalian cells, and they are activated after G_q-coupled receptor stimulation but the activation mechanisms are in many cases unknown. It will be interesting to see whether any of the TRP channels can be held responsible for the orexin receptor-mediated Ca²⁺ influx.

Electrophysiological Responses

Methodological considerations. Characterization of receptor-mediated electrical responses in neuronal preparations is performed with some common electrophysiological techniques. In many situations in which intracellular recordings are difficult, extracellular electrodes can be more practical to use and they are likely to induce much less cellular damage. On the other hand, these response measurements are usually limited to the registration of action potential frequency. Measurements with intracellular electrodes (normal or microelectrodes) can be performed with feedback amplification allowing voltage or current clamp or in a "free-running" mode, in which case voltage changes are monitored. These latter methods, most importantly, offer better possibilities for investigation of the mechanisms underlying the electrical activity than the extracellular electrodes.

Increased synaptic activity. The most common electrophysiological response to orexins appears to be increase in spontaneous or evoked action potential frequency, as observed with extracellular or intracellular electrodes. Increased action potential frequency has been measured in the hypothalamus [rat tuberomamillary nucleus, magnocellular preoptic nucleus, lateral hypothalamic glucose-sensitive, -responsive, and -indifferent neurons (for definition, see Responses of orexinergic cells to feeding stimuli), arcuate nucleus, and paraventricular nucleus; Refs. 8, 42, 44, 103, 147, 159, 164, 165, 195], brain stem (rat locus ceruleus , substantia nigra pars reticulata, and dorsal motor nucleus of vagus and mouse laterodorsal tegmental nucleus; Refs. 15, 19, 56, 70, 75, 93, 170), spinal cord (rat preganglionic sympathetic neurons of intermediolateral cell column of thoracic and lumbar cord; Ref. 3), and peripheral neurons (guinea pig ileal S-type submucosal neurons; Ref. 92). In contrast, orexin A suppresses the firing rate of glucose-responsive neurons in rat ventromedial hypothalamic nucleus, antagonizing the effect of glucose (164). Action potentials and their enhancement with orexins are blocked by TTX (3, 44, 54, 70, 75, 165, 170, 195) but usually not by synaptic isolation with low Ca²⁺-high Mg²⁺ (8, 42, 93, 147) or voltage-gated Ca²⁺ channel block with Cd²⁺ (170). The latter data therefore suggest that orexins would directly affect the postsynaptic neurons at some sites, with reservations for the methodological shortcomings (see above). It should also be noted that the putative presynaptic block has only been applied in a minority of studies. Interestingly, a TTX-insensitive response is observed in rat locus ceruleus. Hypocretin 2 (orexin B) depolarizes and increases spontaneous firing frequency in locus ceruleus slices (70). In the presence of TTX these spontaneous spikes vanish, but orexin B still causes a small depolarization (3.4 mV) and electrical discharges, although these spikes are slower than the spontaneous spikes. These spikes may be caused by Ca²⁺ channels or TTX-insensitive Na⁺ channels.

Depolarizing actions. Depolarization or inward current has often been measured in response to orexin application. This depolarization also results, when investigated, without exception in increased action potential frequency, and often the depolarization has been studied in the presence of TTX to block this "disturbance" (Fig. 5, A and B). Depolarizations of a maximum magnitude of 3-10 mV have been observedmainly with free-running electrodesin the hypothalamus (rat tuberomamillary nucleus, paraventricular nucleus, and lateral hypothalamic glucose-sensitive neurons; Refs. 8, 44, 103, 159, 165), in the brain stem (rat locus ceruleus and dorsal motor nucleus of vagus; Refs. 56, 70, 75, 80, 170), in the spinal cord (rat intermediolateral cell column; Ref. 3), and in the periphery (guinea pig ileal S-type submucosal neurons; Ref. 92). Inward currents have been observed in rat locus ceruleus (150 pA; Ref. 170), dorsal motor nucleus of vagus (30 pA; Ref. 75), and superficial dorsal horn (35 pA; Ref. 54) and in mouse laterodorsal tegmental nucleus (20 pA; Ref. 19).

View larger version (49K): [in this window] [in a new window]   Fig. 5.   Orexin-mediated increase in action potential frequency (A) and depolarization (B) in rat hypothalamic tuberomamillary nucleus neurons in a slice preparation. Whole cell recordings were performed with intracellular microelectrodes in free-running mode. Cells were stimulated with 300 nM orexin A where indicated. Experiment in B was performed in the presence of 300 nM TTX. Figure reproduced from Ref. 44 (Copyright 2001 by the Society for Neuroscience).

Summary of orexin effects on electrical activity of neurons. Thus orexins most often increase neuronal activity via presynaptic effects (increased transmitter release via unknown mechanisms) and via postsynaptic depolarization (block of K⁺ channels, activation of cation channels and electrogenic transporters?). Some of the effects such as the putative K⁺ channel inhibition are typical for G_q-coupled receptors (see, e.g., Refs. 59, 65, 112, 187). Other mechanisms may be secondary to the Ca²⁺ elevation or may depend on the activation of nonselective cation channels like the TRP channels. On particular neurons, orexins may also cause inhibition via increased inhibitory transmitter release or via decreased postsynaptic effects of excitatory transmitters. It is unclear which direct or indirect roles Ca²⁺ transients play in electrical responses to orexins, and not much effort has been put into investigation of this to date. Independent of the method of measurement of electrical activity, the effects of orexins usually have a slow onset of action (10 s-4 min) and the recovery is even slower (see, e.g., Refs. 3, 9, 19, 44, 54, 75, 93, 103, 159, 165, 195). Most studies have been performed with slices, a preparation in which perfusion of orexins in and out may be slow. However, because rather superficial neurons are patched, there is no obvious reason why the perfusate would not reach the neurons immediately. Furthermore, similar slow onset (10- to 20-s lag, 1 min to maximum) and delayed offset (>2 min to complete decay after removal) are seen in dissociated neurons (195). Therefore, it is possible that orexins affect ion channels via, e.g., phosphorylation rather than direct G protein interaction. Some results indicate involvement of protein kinase A and -C in orexin responses (93, 181, 183). Orexin B has been reported to reduce N-methyl-D-aspartate (NMDA)-induced current amplitude by 19% and to enhance GABA-induced current amplitude by 48% in dissociated rat nucleus accumbens cells (113); this could also possibly be a phosphorylation-dependent event.

Transmitter Release

Other Responses

Orexin receptors may also couple to protein kinase cascades involved in cell growth, differentiation, and death. These effects are clearly seen when orexin receptors are expressed in CHO cells (S. Ammoun, L. Korhonen, L. Lindholm, K. E. O. Åkerman, and J. P. Kukkonen, unpublished observation). One of the most immediate indications is the rapid activation of mitogen-activated protein kinase (MAPK) pathways seen on OX₁ receptor stimulation (Fig. 6).

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Mitogen-activated protein kinase (MAPK) activation in response to OX1 receptor stimulation in CHO-hOX1 cells (S. Ammoun, L. Korhonen, L. Lindholm, K. E. O. Åkerman, and J. P. Kukkonen, unpublished observation). Extracellular signal-regulated kinase (ERK)1 and -2 activation was measured by blotting against the respective phosphorylated species with specific antibodies. Both 3 and 100 nM orexin A lead to rapid (3 min) activation of mainly ERK2, a response that is more clearly decreased at the 30 min time point at the lower concentration.

Pharmacology

Molecular determinants required from orexin peptides for binding to and activation of orexin receptors have been investigated in CHO cells. Comparison of structures and activities of orexin A and -B suggests that the NH₂-terminal amino acids are less important for receptor binding and activation than the COOH-terminal amino acids (33). Truncation of orexin A to orexin A_15-33 reduces its potency for both OX₁ and OX₂ receptors 20- to 60-fold (33, 138), and further truncation successively abolishes the Ca²⁺ response, at least for OX₁ (33). Mutation of cystine-forming cysteines to alanines in orexin A reduces the potency of the mutant peptides for both OX₁ and OX₂ receptors ~10-fold (138). However, it is difficult to judge whether this effect is due to the lost disulfide bridges or the exchanged cysteines themselves. One-by-one replacement of each amino acid in orexin-A_15-33 with alanine (alanine scan) has revealed regions of particular importance for orexin A activation of the OX₁ receptor, in particular leucines 19 and 20 and the most COOH-terminal aa 26-33 seem to be of great importance (33). Recently, we observed (S. Ammoun, T. Holmqvist, R. Shariatmadari, R. Oonk, M. Detheux, M. Parmentier, K. E. O. Åkerman, and J. P. Kukkonen, unpublished observations) that OX₁ and OX₂ receptors are approximately equally sensitive to NH₂-terminal truncation and alanine scan of the orexin peptides. The amino acids of importance are largely the same as those described in Ref. 33. Change/removal of these amino acids affects both OX₁ and OX₂ receptors, but some interesting differences are seen in the potency of the response. Interestingly, none of the mutated peptides without activity are antagonists.

Xenopus orexin B has been proposed to have 10-fold higher affinity and potency than Xenopus orexin A for human OX₂ receptor expressed in CHO cells (163). Both Xenopus orexin A and -B were equipotent on human OX₁ receptor. Synthetic peptides (hcrt-1 and hcrt-2) based on the originally described hypocretin sequences (37) have been shown to be ~1,000-fold less potent than the corresponding orexin peptides with respect to Ca²⁺ elevation in recombinant CHO cells (167).

In more than 50% of the studies on cellular responses, rather high concentrations (>100 nM) of orexins were used to evoke response. Even the EC₅₀ values, when determined, can be rather high [e.g., 200 nM for orexin A in rat locus ceruleus (increased firing); Ref. 170]. On the other hand, even subnanomolar concentrations evoke some responses (e.g., Ca²⁺ response in rat ventral tegmental neurons; Ref. 130). To resolve the mechanical basis for this, experiments could be performed in recombinant expression systems at different expression levels with measurement of different responses. Another reason for the high concentrations required in some tissues may be the use of orexin B at OX₁ receptor or use of the hypocretins based on the original sequences, if the weak potencies reported in recombinant systems hold true in tissues (see above).

At the moment, no reliable in vivo selectivity can be obtained with known orexin receptor agonists. Recently, synthesis of an orexin receptor subtype-selective antagonist was reported (145, 169). This antagonist, SB-334867 (Fig. 1D), displays a 100-fold higher affinity for OX₁ than OX₂ receptor. This selectivity has been successfully utilized in in vivo experiments to elucidate the role of OX₁ receptors in some systemic effects of orexins (see SYSTEMIC EFFECTS OF OREXINS). Unfortunately, no antagonist able to block OX₂ receptors is available. Conclusions on the role of OX₁ receptors in some physiological processes should not be based solely on SB-334867, because orexin receptor subtypes may have overlapping and interacting roles. However, SB-334867 is a uniquely useful tool in orexin receptor investigations. From this perspective, it is unfortunate that SB-334867 is not widely available for orexin receptor researchers.

Recently, a variety of peptide transmitters, including secretin and NPY variants, were reported to inhibit orexin A binding to membrane preparations from endogenous and recombinant cells expressing OX₁ receptors, even with a nanomolar affinity (87). These results were interpreted to indicate that interaction between orexinergic and NPYergic systems (see Connections of orexinergic pathways with other feeding-regulating transmitters/hormones) could occur at the orexin receptor level. These results on binding could not, however, be verified in functional experiments or in binding experiments with intact cells (67, 169). Other peptides [secretin, pituitary adenylate cyclase-activating polypeptide (PACAP), NPY variants] do not display any detectable affinity or activity on OX₁ or OX₂ receptors (67, 169). As suggested in Ref. 67, it is more likely that the other peptides affect some intracellular process, for instance G protein binding to the receptor, which is required for high-affinity agonist binding, and therefore reduce orexin binding in some particular preparations. Therefore, secretin should not be used to define specific orexin A binding, because this can to lead to possibly erroneous conclusions (86, 188).

	DISTRIBUTION OF OREXINERGIC CELLS AND OREXIN RECEPTORS

TOP ABSTRACT INTRODUCTION OVERVIEW OF OREXINS AND... CELLULAR RESPONSES TO OREXINS DISTRIBUTION OF OREXINERGIC... SYSTEMIC EFFECTS OF OREXINS FUTURE PERSPECTIVES REFERENCES

Orexinergic Cells in CNS

Orexinergic Cell Bodies in Rat CNS

Orexinergic neurons are variable in size (diameter of cell body = 15-40 µm) and shape (spherical, fusiform, multipolar) (24, 31, 36, 131), and they have been assumed to number from 1,100 to 3,400 in the whole rat brain (61, 143). Because orexinergic cells have been shown to project to other orexinergic cells within the hypothalamus (69) there might be different populations of orexinergic cells within the hypothalamus.

Preproorexin mRNA and protein have even been detected in the ependymal cell layer (97, 143), and orexin A has been found in choroid plexi (31). It would be interesting to find out whether this expression occurs in the ependymal cells or, for instance, in neuronal stem cells.

Orexinergic Projections in Rat CNS

Despite the low number of orexinergic cell bodies in the hypothalamus, orexinergic fibers project widely in the CNS. The most important orexinergic projection areas are within the hypothalamus, two thalamic nuclei, brain stem, and the whole length of the spinal cord (Table 1; see also Fig. 7). Interested readers are recommended to consult the original papers, some of which are very detailed (24, 31, 34, 36, 37, 61, 69-71, 97, 125, 126, 154, 182).

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Orexinergic pathways (according to Refs. 24, 31, 34, 36, 37, 61, 69-71, 125, 126, 154, 182) and orexin receptor mRNA distribution (according to Refs. 111, 180) in the rat. Only a minority of the sites are shown. The density of the hatching (3 different levels) indicates OX1 and OX2 receptor mRNA expression levels. It should be noted that there are subregions within the marked regions that display different expression (e.g., in the hypothalamus), which cannot be shown in the figure. When data in 2 references were diverging or even contradictory, an average was taken. Arcn, arcuate nucleus; CA1-3, areas of hippocampus; cC, cingulate cortex; CMn, centromedial nucleus; dR, dorsal raphe nucleus; LC, locus ceruleus; mE, median eminence; mR, median raphe nucleus; nST, nucleus of solitary tract; olfB, olfactory bulb; olfT, olfactory tubercle; PVn, paraventricular nucleus (in the thalamus and hypothalamus); Rm, nucleus raphe magnus; Ro, nucleus raphe obscurus; SCn, suprachiasmatic nucleus; SOn, supraoptic nucleus; STn, spinal trigeminal nucleus; TMn, tuberomamillary nucleus; VAMn, ventral anteromedial nucleus.

Orexins in CNS of Other Species

Orexin Receptors in CNS

Similar amounts of OX₁ and OX₂ mRNA are found in the rat brain as a whole, and OX₁ and OX₂ receptor mRNAs mostly show similar distribution, although some differences are seen (Fig. 7). Most interesting are the areas in which essentially only one subtype is expressed (111, 180). However, mRNA expression may not directly correspond to receptor protein expression, as discussed by van den Pol et al. (184). The correlation of receptor distribution with different functions of the orexinergic system is shown in Table 3. While this article was under review, studies on the expression of OX₁ and OX₂ receptor protein in the rat CNS were published (30, 64). A particularly interesting finding is the expression of OX₂ receptor protein in the molecular and granular layers of the cerebellum (30); earlier studies had failed to detect any orexin receptors or their mRNA in the cerebellum although orexinergic fibers were projecting to the cerebellum. Otherwise, the immunohistochemical studies suggest a distribution generally similar to that from the mRNA measurements, although some differences are seen. For instance, greater overlap of the expression of OX₁ and OX₂ receptors is suggested on the basis of immunological studies than of in situ hybridization. As discussed by the authors, these differences may in some degree be related to the fact that mRNA measurements visualize the expressing cell bodies whereas immunohistochemistry indicates the actual cellular localization of the protein. Both receptor mRNA and protein measurements are semiquantitative rather than quantitative so they do not allow absolute comparison of subtype expression within any site.

Coexpression and Connections of Orexins With Other Transmitters

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Some connections between lateral hypothalamic orexinergic neurons and other transmitters/hormones. Orexinergic cells also express dynorphin, and at least some of them also express Ob-Rb (long form of leptin receptor) and the leptin receptor-activated promoter element STAT3. Subpopulations of orexinergic cells may release galanin and glutamate (for references, see text). Thin black arrows indicate input to the orexinergic cells, and thick gray arrows indicate orexinergic output. Anatomic loci: Arcn, LC, laterodorsal tegmental nucleus (LDTn), lateral hypothalamus (LH), hypothalamic periventricular nucleus (PeVn), preoptic area (POA), PVn, raphe nuclei (Raphe n), SCn, substantia nigra (SN), TMn, ventral tegmental area (VTA). ACh, acetylcholine; AGRP, agouti gene-related peptide; AVP, (arginine-) vasopressin; CART, cocaine/amphetamine-regulated transcript; CRF, corticotropin-releasing hormone; DA, dopamine; 5-HT, 5-hydroxytryptamine (serotonin); Glu, glutamate; HA, histamine; MCH, melanocyte-concentrating hormone; NE, norepinephrine; NPY, neuropeptide Y; POMC, proopiomelanocortin; SSt, somatostatin; VIP, vasoactive intestinal peptide.

View larger version (14K): [in this window] [in a new window]   Fig. 9.   Some connections between orexins and other appetite-regulating transmitters/hormones in the hypothalamus. Arrows marked + and  indicate stimulation or inhibition, respectively, of neuronal activity or release or synthesis of the hormones in the affected site; ± indicates either stimulatory or inhibitory effects under different conditions or toward similar cells in different anatomic loci. Unmarked arrows indicate connections with unknown effects. As indicated in the text, orexinergic neurons may either themselves respond to glucose or have reciprocal connections with glucose-sensitive (GSN) or glucose-responding (GRN) neurons. Thus it is not clear whether the effect glucose on neurons in LH or Arcn is direct or mediated by other "sensor neurons"; therefore, glucose, GSN, and GRN are grouped in the figure with a dotted line. Solid lines indicate the anatomically defined hypothalamic areas.

Vasoactive intestinal peptide (VIP)-, vasopressin-, and NPY-containing neurons innervate orexinergic neurons in the hypothalamus (Refs. 1, 69, 71; Figs. 8 and 9). On the other hand, orexinergic nerves from lateral hypothalamus innervate neuronal circuits, which utilize many different transmitters, for instance, norepinephrine, dopamine, serotonin, histamine, acetylcholine, vasopressin, VIP, somatostatin, corticotropin-releasing hormone (CRF), NPY/agouti-gene-related peptide (AGRP), proopiomelanocortin (POMC), cocaine/ amphetamine-regulated transcript (CART), GABA, MCH, and glutamate (Figs. 8 and 9; for the anatomy, see, e.g., Refs. 5, 31, 36, 131, 143), and even functional responses to orexins have been shown at some of these sites (e.g., Refs. 9, 18, 19, 42, 44, 49, 56, 70, 80, 93, 113, 130, 159, 165, 170, 183, 195; see CELLULAR RESPONSES TO OREXINS). Orexinergic neurons even make contact with other orexinergic neurons within the hypothalamus (5, 69).

Connections of orexinergic pathways with other feeding-regulating transmitters/hormones. Orexins were first isolated as appetite-increasing peptides, although this view was later disputed. Several hypothalamic sites involved in the regulation of feeding, such as the lateral hypothalamic area, parvocellular paraventricular nucleus, ventromedial nucleus, and arcuate nucleus, are innervated by orexinergic fibers from the lateral hypothalamic area (see e.g., Ref. 36). Both OX₁ and OX₂ receptors may be expressed at these projection areas (111). There are abundant connections between orexin and other feeding-regulating transmitters/hormones such as leptin, NPY, MCH, POMC, CART, galanin, and AGRP (Figs. 8 and 9). Almost all the orexin-positive cells also express leptin receptor, and even STAT3the leptin receptor-activated promoter elementand orexin are coexpressed (Refs. 58, 69, Fig. 8; see also Effects of orexins on gastrointestinal tract and other organs of feeding/energy metabolism). On the other hand, not all the leptin receptor-positive cells in the lateral hypothalamic areas where orexinergic cell bodies lie are orexin positive (58). Other cells that express leptin receptors are the MCHergic neurons in the lateral hypothalamus and the NPY-/AGRP-, POMC- and CARTergic cells in the arcuate nucleus (Refs. 69, 57, 58, 119; Fig. 9). Orexinergic fibers project to NPY- and POMCergic cells in the arcuate nucleus (69). On the other hand, NPY/AGRP- and POMCergic fibers, most likely from the arcuate nucleus, project to orexin- and MCHergic cells of the lateral hypothalamus (17, 43, 55, 69, 71). Orexinergic cells of the lateral hypothalamus also contain other feeding-regulating transmitters such as dynorphin and galanin (Refs. 10, 27, 58; Fig. 9), which may enhance the effects of orexins on feeding.

Responses of orexinergic cells to feeding stimuli. Hypoglycemia induced by various means has been shown to increase c-fos (see below), orexin, and preproorexin mRNA expression in a subpopulation of the orexinergic cells of the lateral hypothalamus (16, 20, 22, 53, 99, 105, 127; see also Effects of orexins on gastrointestinal tract and other organs of feeding/energy metabolism). Fasting also abolishes the circadian variation in CSF orexin levels (47). The lateral hypothalamic area, as well as some other hypothalamic regions, contains neurons that are activated by decreased levels of glucose ("glucose-sensitive neurons") and neurons that are activated by elevated levels of glucose ("glucose-responsive neurons") (Ref. 101; Fig. 9). In the study of Muroya et al. (129), almost half of the histochemically defined orexinergic neurons of the rat lateral hypothalamus directly responded to lowering of extracellular glucose with an increase in cytosolic Ca²⁺; thus some of the glucose-sensitive neurons in lateral hypothalamus may be orexinergic. Liu et al. (103) identified glucose-sensitive, -responsive, or -indifferent populations among rat hypothalamic neurons that responded to orexin stimulation with depolarization and increased firing rate. In contrast, orexin A suppressed the firing rate of glucose-responsive neurons in rat ventromedial hypothalamic nucleus (164). Orexinergic neurons may also make contact and/or be contacted by glucose-sensitive or -responsive neurons in other brain areas such as lateral hypothalamus, ventromedial hypothalamus, arcuate nucleus, and the nucleus of the solitary tract (Ref. 101; see also anatomic studies on orexinergic projections).

Orexins and Orexin Receptors Outside CNS

Orexin and orexin receptor immunoreactivity has been found in the gastrointestinal tract and pancreas (rat, guinea pig, human; Refs. 92, 133; Tables 4 and 5). Substantial levels of both orexin A and -B and OX₁ and OX₂ mRNA and protein are found in the rat pituitary (13, 35, 84), and Jöhren et al. (84) found preproorexin and OX₁ mRNA in the rat testis. Both OX₁ mRNA and OX₂ mRNA or immunoreactivity are found in the rat and human adrenal glands (12, 84, 104, 108, 109), although the subtype expression in different layers is somewhat disputed and there may be differences between species (see, e.g., Ref. 117). Only OX₁ mRNA has been found in the kidney and thyroid and only OX₂ mRNA in the lung (84). Orexinergic fibers extend from the CNS toward the rat pineal gland, and OX₂ mRNA is detected in the pineal gland (121).