Ghrelin in Central Neurons

Ghrelin, an orexigenic peptide synthesized by endocrine cells of the gastric mucosa, is released in the bloodstream in response to a negative energetic status. Since discovery, the hypothalamus was identified as the main source of ghrelin in the CNS, and effects of the peptide have been mainly observed in this area of the brain. In recent years, an increasing number of studies have reported ghrelin synthesis and effects in specific populations of neurons also outside the hypothalamus. Thus, ghrelin activity has been described in midbrain, hindbrain, hippocampus, and spinal cord. The spectrum of functions and biological effects produced by the peptide on central neurons is remarkably wide and complex. It ranges from modulation of membrane excitability, to control of neurotransmitter release, neuronal gene expression, and neuronal survival and proliferation. There is not at present a general consensus concerning the source of ghrelin acting on central neurons. Whereas it is widely accepted that the hypothalamus represents the most important endogenous source of the hormone in CNS, the existence of extra-hypothalamic ghrelin-synthesizing neurons is still controversial. In addition, circulating ghrelin can theoretically be another natural ligand for central ghrelin receptors. This paper gives an overview on the distribution of ghrelin and its receptor across the CNS and critically analyses the data available so far as regarding the effects of ghrelin on central neurotransmission.

Ghrelin is a 28-amino acid motilin-related peptide [ 4 ], originally purified from the rat stomach [ 81 ]. The peptide is characterized by the presence of an n -octanoylation on the hydroxy group of serine in position 3. This acylation is a post-transcriptional modification that is essential for binding to GHS-R [ 81 , 145 ] to a point that ghrelin was originally supposed to be biologically active only in acylated form [ 81 ]. However, about 80-90% of circulating ghrelin is not acylated (des-acyl-ghrelin), and it still remains unclear whether or not des-acyl-ghrelin represents a precursor or a degradation product of the acylated peptide [ 65 , 83 ]. Moreover, des-acyl-ghrelin does not replace radio labeled ghrelin at pituitary and hypothalamic binding sites, nor it seems capable of inducing growth hormone (GH) release. Therefore its biological role, if any, remains puzzling, and the possibility that des-acyl-grelin is a biologically active molecule acting through a specific, but yet uncharacterized receptor still remains a matter of debate. In support of this hypothesis, several in vitro studies have demonstrated that radio labeled ghrelin and des-acyl-ghrelin bind to the membranes of PC-3 prostate tumor cells, H9C2 cardiomyocytes and isolated adipocytes, none of which expressed the GHS-R [ 6 , 24 , 101 ]. In addition, ghrelin and des-acyl-ghrelin, at least in some cases, exhibit similar GHS-R independent biological activities, such as the inhibition of cell proliferation of breast carcinoma cell lines [ 23 ], the ionotropic effect on guinea pig papillary muscle [ 9 ], the promotion of bone marrow adipogenesis [ 138 ], the control of glucose output by primary hepatocytes [ 47 ]. Circulating ghrelin is mainly produced by X/A-like cells of the oxyntic stomach mucosa [ 3 , 34 , 41 , 120 , 154 ]. However, expression of the peptide has also been demonstrated in many other organs such as testis [ 136 ], ovary [ 19 ], placenta [ 55 ], kidney [ 100 ], pituitary [ 85 ], small intestine [ 34 ], pancreas [ 147 ], lymphocytes [ 61 ] and brain [ 33 , 40 , 90 , 140 ]. In CNS, the main site of ghrelin synthesis (albeit at much lower levels than the stomach) is the hypothalamus. Expression of ghrelin in brain was initially established in the seminal paper by Kojima and co-workers [ 81 ]. Later, by using a combination of RIA and HPLC, Sato and co-workers clearly identified hypothalamic ghrelin [ 123 ]. By immunocytochemical techniques and colchicine pre-treatments, ghrelin expression was demonstrated in the internuclear space between the lateral hypothalamus, the arcuate nucleus (ARH), the ventromedial nucleus (VMN), the dorsomedial nucleus (DMN), the paraventricular nucleus (PVN) and the ependymal layer of the third ventricle [ 33 , 66 ]. In these areas, ghrelin was localized in axon terminals innervating the ARH, VMN, PVN, DMN and the lateral hypothalamus. These axons made synapses with neurons expressing NPY/AgRP and pro-opiomelanocortin (POMC) [ 33 ]. However, according to other immunocytochemical studies, ghrelin is also synthesized by ARH neurons and these ghrelin-producing neurons display synaptic interactions with POMC, NPY and other ghrelin-containing nerve cells [ 57 , 58 , 63 , 90 ]. These findings were confirmed by RT-PCR experiments [ 98 ] and, very recently, by the use of transgenic mice where the transcription regulatory regions of the ghrelin gene have been engineered to drive the expression of enhanced green fluorescent protein (EGFP) [ 74 ]. Finally, expression of the hormone in hypothalamus was also detected in human samples [ 95 ]. Outside the hypothalamus, ghrelin-immunopositive staining was observed in pyramidal neurons of layer V in the sensorimotor area and in the cingulate gyrus of the cerebral cortex, and the ghrelin mRNA was found in the sensorimotor cortex and in the dorsal vagal complex (DVC) of the medulla oblongata [ 66 ]. Localization in spinal cord and dorsal root ganglion (DRG) neurons remains to be ascertained with certainty, albeit after tyramide intensification we observed a limited number of positive medium-to-large DRG neurons and occasional cell bodies in laminae IV-IX of the spinal gray matter (unpublished data).

According to a recent definition, a neurotransmitter is a molecule, released by neurons or glia that physiologically influences the electrochemical state of adjacent cells [ 129 ]. In this respect, the first evidence that ghrelin may act as neurotransmitter in CNS was provided by studies on the hypothalamic neurons involved in the control of feeding [ 72 ]. The physiological significance of the biological functions of ghrelin on nutritional homeostasis and metabolism has been authoritatively reviewed [ 32 , 72 , 108 ]. Besides to these widely acknowledged functions, initial electrophysiological studies in hypothalamus strongly suggested that ghrelin also modulates neuronal excitability and synaptic transmission by acting on GHS-R type 1a [ 33 , 117 , 140 , 144 ]. The main site of action of the hormone was found in ARH (Fig. 1A ), where ghrelin-positive terminals innervate a population of GHS-R type 1a expressing neurons [ 33 , 125 , 164 ]. Within ARH, ghrelin was shown to increase the firing rate of a population of neurons that were inhibited by the anorexigenic peptide leptin [ 117 , 140 ]. The effect was partly blocked by the GHS-R type 1a antagonist (D-Lys3)-GHRP-6 [ 139 ]. The mechanism and the circuitry involved were elucidated in the interesting study of Cowley and coll. [ 33 ]. These authors used an acute slice preparation from hypothalamus obtained from two lines of transgenic mice in which NPY- or POMC-expressing neurons were genetically engineered to express a reporter fluorescent label. By this approach, ghrelin was shown to directly increase the firing rate of NPY/ AgRP neurons and to indirectly inhibit POMC neurons by facilitating the pre-synaptic release of γ-amino butyric acid (GABA) and NPY. Furthermore, in PVN, the main projection site of ARH neurons, ghrelin reduced the inhibitory tone of corticotropin-releasing hormone (CRH) neurons through activation of pre-synaptic NPY receptor Y1 and Y5, thus disinhibiting an important orexigenic pathway mediated by these neurons. Taken together, the above observations demonstrate that the effects of ghrelin on hypothalamic feeding circuitry are due to modulation of transmitter release from ARH neurons expressing NPY. Further confirmation was subsequently obtained after molecular genetic studies [ 28 ], where the effect of peripheral ghrelin on feeding behavior resulted to be reduced after NPY deletion (but not AgRP) and completely blocked in NPY/AgRP double knock-out animals. As observed by Cowley and Grove [ 32 ], the loss of ghrelin effect by simply switching off the NPY and AgRP genes excludes that the inhibitory effect of ghrelin onto POMC neurons is due to a direct facilitation of GABAergic transmission, but rather suggests that is largely mediated by peptide release. Interestingly, by coupling patch-clamp recordings and single cell RT-PCR, van den Top and coll. [ 144 ] demonstrated that ghrelin induces regular bursts of action potentials with underlying oscillation of membrane potentials in NPY/ AgRP-expressing ARH neurons. This ghrelin-induced pacemaker activity is driven by low-threshold T-type Ca 2+ channels and the bursting frequency is modulated by transient outwardly rectifying K + currents. As demonstrated in the seminal paper of Poulain and Wakerley [ 114 ], burst activity is particularly effective in inducing peptide release from hypothalamic neurons. Therefore ghrelin should more efficiently modulate peptide release than amino acid release in ARH. In a series of calcium imaging experiments on isolated ARH neurons, the activation of N-type Ca 2+ channels via protein kinase A (PKA) was also proposed as a mechanism by which ghrelin may rise cytosolic Ca 2+ levels in NPY expressing neurons [ 78 ]. Within the same population of neurons, ghrelin also elevates intracellular Ca 2+ concentration by acting via phospholipase C and AMP-activated protein kinase [ 79 , 80 ]. The increase of cytosolic Ca 2+ , besides changing membrane excitability, is generally known to activate many intracellular pathways that regulate protein synthesis and metabolism. In keeping with this notion, several lines of evidence suggest that ghrelin may regulate gene expression and de novo synthesis of other neurotransmitters [ 51 , 103 , 125 ]. In particular, intracerebroventricular (ICV) administration of orexigenic doses of ghrelin increases the mRNA level of AgRP and NPY in ARH neurons [ 103 , 125 ], and a similar effect (even though conditional to the co-administration of corticosteroids) was observed in a hypothalamic organotypic culture model [ 51 ]. The stimulatory effects of ghrelin on NPY gene expression were abolished in the presence of cycloheximide, that blocks protein translation [ 51 ]. Furthermore in several hypothalmic nuclei involved in feeding control (ARH, PVN, DVM, VMN and lateral hypothalamic nuclei), ICV injections of ghrelin were shown to increase expression of Fos, a marker of neuronal activation [ 87 ]. Interestingly, the ghrelin-induced Fos expr

Hippocampus, amygdala and dorsal raphe nucleus (DRN) are the main brain areas involved in learning and memory mechanisms. Initial evidence suggesting that ghrelin may have a role in mnestic functions was provided by the early work of Carlini and coll. [ 20 , 21 ]. First, by using a behavioural test these authors showed that ICV injections of ghrelin increase memory retention [ 20 ]. To more precisely define the site of peptide action, the experiment was repeated after intraparenchimal injections of increasing concentrations of the hormone in hippocampus, amygdala and DRN [ 21 ]. A dose-dependent increase of memory retention was observed in each condition, with maximal effect in hippocampus. More recent data from the same laboratory suggest that the effects of ghrelin on memory could depend on the availability of serotonin (5-HT), since a 5-HT uptake inhibitor (fluoxetine) decreases both short and long term memory retention [ 22 ]. Underlying mechanisms have been partly elucidated by Diano and coll. [ 40 ], who showed that peripheral ghrelin injections rapidly rearrange synaptic organization with an increase in spine density in CA1 regions of hippocampus, similarly to hypothalamus [ 111 ] and midbrain [ 1 ]. Interestingly, comparable results were also obtained after analysis of spine density in wild type and ghrelin knock-out animals, giving further support to involvement of endogenous ghrelin in CA1 spine formation. Furthermore, the same authors showed that ghrelin promotes long term potentiation, a phenomenon that has a positive correlation with spatial memory and learning [ 40 ]. Animals that received ghrelin injections showed in fact enhanced performance in several behavioral memory tests that are dependent by hippocampus. Again, there is a lack of knowledge concerning the molecular mechanisms and transmitters involved. As discussed above, ghrelin receptors positively interact with DA and 5-HT receptors [ 22 , 69 ] and D1/GHS type 1a co-expression has been reported in hippocampus [ 69 ]. Since the loss of cognitive functions in aging has been supposed to involve a decline in DA or 5-HT signaling [ 8 , 15 ], ghrelin potentiation of these neurotransmitters in hippocampus may represent an interesting mechanism to intervene on memory impairment due to senescence or Alzheimer disease [ 8 , 40 ]. On the other hand, the still limited number of data available so far leaves open the possibility that the effects of ghrelin on mnestic performances may be primarily related with feeding behavior. The model proposed by Diano and coll. [ 40 ] implies that circulating ghrelin is able to reach significant concentrations in the hippocampus. Furthermore, Carlini and coll. [ 20 ] have shown that injections of ghrelin in the hippocampus and DRN increased food intake in a dose-dependent manner. Altogether these data suggest that a gut-hippocampus axis may facilitate memory retention for the spatial localization of food [ 99 ].

Several studies have shown that ghrelin has anti-apoptotic and protective effects on different types of cells subjected to ischemia/reperfusion injury. The hormone inhibits cell apoptosis in cardiomyocytes and endothelial cells [ 6 , 10 , 163 ], adipocytes [ 76 ], cells of the adrenal zona glomerulosa [ 93 ], osteoblastic MC3T3-E1 cells [ 77 ], pancreatic β cells [ 52 , 53 , 162 ], intestinal epithelial cells [ 110 ], and ovarian follicle cells [ 116 ]. Moreover, it exhibits protective effects against ischemia/reperfusion in gastric mucosa [ 14 ], pancreas [ 39 ], and the isolated rat heart [ 27 , 44 ]. Recently, it has been observed that similar effects are also exerted in CNS, where ghrelin inhibits apoptosis during oxygen-glucose deprivation (OGD) [ 30 ]. Protection of hypothalamic neurons is achieved by inhibition of reactive oxygen species generation, stabilization of mitochondrial transmembrane potential, increase of the Bcl-2/Bax ratio, prevention of cytochrome c release, and inhibition of caspase 3 activation [ 30 ]. Similar effects were demonstrated in rat hippocampal neurons after ischemia/reperfusion injury, with increase of cell survival and reduction of death [ 89 ]. In keeping with the observations in vivo , primary cortical neurons are protected from apoptosis induced by lipopolysaccharide, glutamate, n-methyl-d-aspartate (NMDA) and H 2 O 2 . The anti-apoptotic effect is related to up-regulation of Bcl-2 and heat-shock protein 70 (HSP70), and inhibition of caspases 8, 9, and 3 upon binding to GHS-R type 1a [ 96 ]. Similarly, apoptosis is blocked in rat pheochromocytoma (PC12) cells, where ghrelin reduces apoptosis by inducing the expression of HSP70 that, in turn, inhibits signal-regulating kinase 1 (ASK1) activity and ASK1-mediated caspase 3 activation [ 156 ]. The effect of ghrelin on neuronal survival is not limited to neuroprotection, but it also extends to cell proliferation in both the embryonic and adult nervous system [ 73 , 122 , 159 , 160 , 161 ]. Both the acylated and the des-acylated form of the peptide have been shown to promote embryonic spinal cord development and neurogenesis [ 122 ]. In the adult rat nervous system, ghrelin stimulates in vitro and in vivo neurogenesis in DMN and NST, after cervical vagotomy [ 160 , 161 ]. Finally, the synthetic GHS-R agonist hexarelin and ghrelin itself stimulate the incorporation of (3) H-thymidine in adult rat hippocampal progenitor cells, as an index of increased cell proliferation. In addition, hexarelin, but not ghrelin, also shows a significant inhibition of apoptosis and necrosis [ 73 ]. The effects of ghrelin on cell proliferation appear to be linked not only to GHS-R type 1a activation, but also to other yet uncharacterized peptides [ 73 , 122 ]. All together, these data indicate that ghrelin may also act as a survival factor that promotes neurogenesis, preserves tissue integrity from ischemic injuries, and inhibits apoptosis.

Given that ghrelin synthesized in the stomach and released into the general circulation seems to be the main source of the peptide acting on central neurons, one has to keep clear in mind that most of the effects of ghrelin in brain are likely due to a gut peptide released in the bloodstream under fasting conditions [ 83 ]. Therefore the central effects of the peptide should be placed within a more general framework of adaptive mechanisms facilitating feeding behavior. The role of ghrelin in central neurons has encountered increasing interest in recent years, and a coherent scheme of ghrelin functions is taking shape. It is clear that the latter cannot be confined to the GH-releasing effects or the increase of food intake. The arrays of effects produced by ghrelin seem to converge in concert to put the organism in the condition of recovering from a negative energetic status. Nonetheless, the neuropharmacology of this peptide is still poorly understood, and much more work should be done in this direction. In particular, it will be of interest to establish whether or not des-acyl ghrelin is indeed a neuroactive molecule, if receptors subtypes other than GHS-R type 1a are responsible for ghrelin and/or des-acyl-ghrelin central effects, and how can acylation and des-acylation be regulated in the brain. Heterogeneity of ghrelin receptors has already been documented outside the CNS [ 102 ]. If similar data will be obtained also in CNS, they may add new insights in the complex pattern of ghrelin functions in the brain.