This Article Abstract Full Text (PDF) An erratum has been published Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by van der Lely, A. J. Articles by Ghigo, E. Search for Related Content PubMed PubMed Citation Articles by van der Lely, A. J. Articles by Ghigo, E.

Biological, Physiological, Pathophysiological, and Pharmacological Aspects of Ghrelin

Division of Endocrinology and Metabolism (A.J.v.d.L.), Department of Internal Medicine, Erasmus Medical Center, 3015 GD Rotterdam, The Netherlands; Department of Psychiatry (M.T.), University of Cincinnati, Cincinnati, Ohio 45237; Endocrine Research Department (M.L.H.), Eli Lilly and Co., Indianapolis, Indiana 46285; and Division of Endocrinology (E.G.), Department of Internal Medicine, University of Turin, Turin, Italy 10095

Correspondence: Address all correspondence and requests for reprints to: Aart Jan van der Lely, M.D., Department of Internal Medicine, Erasmus Medical Center, 40 Dr. Molewaterplein, 3015 GD Rotterdam, The Netherlands. E-mail: a.vanderlelij{at}erasmusmc.nl

	Abstract

Top Abstract I. Introduction II. Historical Background III. The Biology of... IV. Control of Ghrelin... V. Physiological and... VI. Ghrelin as an... VII. Pharmacological and... References

 
Ghrelin is a peptide predominantly produced by the stomach. Ghrelin displays strong GH-releasing activity. This activity is mediated by the activation of the so-called GH secretagogue receptor type 1a. This receptor had been shown to be specific for a family of synthetic, peptidyl and nonpeptidyl GH secretagogues. Apart from a potent GH-releasing action, ghrelin has other activities including stimulation of lactotroph and corticotroph function, influence on the pituitary gonadal axis, stimulation of appetite, control of energy balance, influence on sleep and behavior, control of gastric motility and acid secretion, and influence on pancreatic exocrine and endocrine function as well as on glucose metabolism. Cardiovascular actions and modulation of proliferation of neoplastic cells, as well as of the immune system, are other actions of ghrelin. Therefore, we consider ghrelin a gastrointestinal peptide contributing to the regulation of diverse functions of the gut-brain axis. So, there is indeed a possibility that ghrelin analogs, acting as either agonists or antagonists, might have clinical impact.

I. Introduction

II. Historical Background

III. The Biology of a Ubiquitously Expressed Hormone and Its Receptor Family

A. Known and unknown GH secretagogue receptors

B. Known and unknown ligands of the GH secretagogue receptors

IV. Control of Ghrelin Secretion: Indications for Its Importance in Biology

V. Physiological and Pathophysiological Actions of Ghrelin

A. Hypothalamic-pituitary actions

B. Central actions of ghrelin and GHS

C. Peripheral activities of synthetic and natural GHS

VI. Ghrelin as an Important Member of the Survival Kit of Nature

VII. Pharmacological and Clinical Perspectives

	I. Introduction

Top Abstract I. Introduction II. Historical Background III. The Biology of... IV. Control of Ghrelin... V. Physiological and... VI. Ghrelin as an... VII. Pharmacological and... References

 
GHRELIN IS A 28-amino residue peptide predominantly produced by the stomach (Fig. 1). Substantially lower amounts were detected in bowel, pancreas, kidneys, the immune system, placenta, testes, pituitary, lung, and hypothalamus (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12). Ghrelin displays strong GH-releasing activity, which is mediated by the activation of the so-called GH secretagogue (GHS) receptor type 1a (GHS-R 1a) (13). Before the discovery of ghrelin, this orphan receptor had been shown to be specific for a family of synthetic, peptidyl and nonpeptidyl GHS (1, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23). GHS-Rs are concentrated in the hypothalamus-pituitary unit but are also distributed in other central and peripheral tissues (8, 13, 14, 20, 22, 23, 24, 25, 26, 27, 28, 29, 30). Indeed, apart from stimulating GH secretion, ghrelin and many synthetic GHS (Fig. 2): 1) exhibit hypothalamic activities that result in stimulation of prolactin (PRL) and ACTH secretion; 2) negatively influence the pituitary-gonadal axis at both the central and peripheral level; 3) stimulate appetite and a positive energy balance; 4) influence sleep and behavior; 5) control gastric motility and acid secretion; and 6) modulate pancreatic exocrine and endocrine function and affect glucose levels.

View larger version (33K): [in this window] [in a new window]   FIG. 1. Ghrelin is the only known natural peptide in mammalian biology in which acylation of one amino residue is required for at least the majority of its biological activities. Under the influence of a still unknown acyl-transferase, a hydroxyl group of serine at position 3 of the ghrelin molecule is octanoylated. This posttranslational modification of ghrelin is essential for binding and activation of the GHS-R 1a, for the GH-releasing capacity of ghrelin, and most likely also for its action on endocrine axis, energy balance, and glucose homeostasis. Several naturally occurring variants of ghrelin have been reported based on the acylation at the serine-3 position, including nonacylated, octanoylated (C8:0), decanoylated (C10:0), and possibly decenoylated (C10:1) ghrelin. Any other synthetic variant of ghrelin with a chemical modification of either the acyl group or the N-terminal amino residue sequence did not activate or bind the receptor GHS-R 1a. However, ghrelin did still bind and activate the GHS-R 1a in vitro after modification or even significant deletion of C-terminal amino residues. It however remains unclear whether the same modalities are relevant in vivo. Although the major active form of human ghrelin is a 28-amino acid peptide with an octanoylation at the serine-3 position, the vast majority (80–90%) of circulating ghrelin has been found to be nonacylated. This predominant form of serum ghrelin seems to be devoid of any effects on endocrine axes or energy balance, as previously expected based on its inability to bind and activate GHS-R 1a, which is still the only identified ghrelin receptor. However, nonacylated ghrelin does have cardiovascular and antiproliferative effects, and it seems tempting to speculate that these activities are mediated by yet to be identified receptor families or subtypes (1 ). In the absence of further information on the tissue specificity, reversibility, balance, and enzyme kinetics of the (des-) octanoylation process, the information one can possibly gain from plasma ghrelin quantification is very limited but should include the analysis of both total and acylated ghrelin (17 131 132 133 ).

View larger version (49K): [in this window] [in a new window]   FIG. 2. Known biological activities of ghrelin. Some of the effects of ghrelin shown here are believed to be indirectly mediated via pituitary hormones or hypothalamic neurocircuits and their efferent pathways; others, such as the effect on the cardiovascular system, appear to be direct. Depending on the origin of the hormone, which is mainly derived from the stomach but also expressed in the pancreas, the hypothalamus, the pituitary, the duodenum, and other organs, these effects may have endocrine, paracrine, or autocrine character. AVP, Arginine vasopressin.

	II. Historical Background

Top Abstract I. Introduction II. Historical Background III. The Biology of... IV. Control of Ghrelin... V. Physiological and... VI. Ghrelin as an... VII. Pharmacological and... References

 
The gastric hormone ghrelin was identified as an endogenous ligand for the former orphan receptor GHS-R 1a (1, 13, 14); the discovery of this receptor followed by 20 yr that of synthetic GHS, which specifically binds it (1, 14, 17, 19, 20, 21, 28, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72). This makes the discovery of ghrelin an example of reverse pharmacology, which in this case means that it started with the synthesis of analogs and it ended with the discovery of a natural ligand via the discovery of a natural receptor.

Synthetic GHS are a family of ligands, including peptidyl and nonpeptidyl molecules (Table 1) (8, 14, 20, 68, 69, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82). The first synthesized molecules were nonnatural peptides [GH-releasing peptides GHRPs)] that were designed by Bowers and Momany (59, 60, 68, 73) in the late 1970s. They were metenkephalin derivatives devoid of any opioid activity (Table 2).

Very recently, another new peptidomimetic GHS with potent and selective GH-releasing activity was synthesized and called EP1572 UMV1843 [Aib-D-Trp-DgTrp-CHO]). EP1572 shows binding potency to the GHS-R in animal and human tissues similar to that of ghrelin and peptidyl GHS and induces marked GH increase after sc administration in neonatal rats. Preliminary human data show that acute iv EP1572 administration (1.0 µg/kg) induces strong and selective increases in GH levels. Moreover, a single oral EP1572 administration strongly and reproducibly increases GH levels even after a dose as low as 0.06 mg/kg (106).

MK-0677 resulted in the discovery and cloning of the GHS-R. The existence of this GHS-R, as shown by mRNA expression, had been indicated already by binding studies (14, 24, 26, 29, 30, 47, 73, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119). Studies focusing on the distribution of the identified GHS-Rs showed a particular concentration of these receptors in the hypothalamus-pituitary area. However, GHS-R expression and/or the presence of specific binding sites has been found in other brain areas and peripheral, endocrine, and nonendocrine animal and human tissues (1, 2, 3, 4, 5, 6, 7, 8, 9, 22, 24, 25, 26, 30, 120, 121). Actually, both the concentration of binding sites and the displacement of peptide-radioligand by ghrelin suggest that the majority of binding in peripheral tissues is not specific for ghrelin or MK-0677. The studies using labeled GHRPs overestimate the concentration of specific ghrelin binding because they exhibit high capacity and low affinity binding in addition to limited capacity ghrelin-specific binding. Anyway, hypothalamopituitary and peripheral distribution of GHS-Rs probably explains both the GH-releasing effect of GHS and their other endocrine and nonendocrine biological activities (2, 8, 9, 14, 16, 20, 25, 27, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 122).

As discovered by Kojima et al. (1) in 1999, ghrelin appeared to be a 28-residue peptide that is predominantly produced by the stomach but is also expressed in many other tissues (1, 2, 3, 4, 5, 6, 7, 8, 9, 12, 123). Ghrelin is produced in the stomach by the enteroendocrine X/A-like cells that represent a major endocrine population in the oxyntic mucosa. The hormonal product of these cells had not been previously clarified (1, 3, 124, 125, 126, 127). Ghrelin production has also been reported in neoplastic tissues as gastric and intestinal carcinoids (126, 128) and in medullary thyroid carcinomas (129).

Ghrelin is the first natural hormone to be identified in which the hydroxyl group of one of its serine residues is acylated by n-octanoic acid (1). This acylation is essential for binding to the GHS-R 1a, for the GH-releasing capacity of ghrelin, and most likely for its other endocrine actions also (1, 17, 21, 130) (Fig. 1). Nonacylated ghrelin, which is present in human serum in far greater quantities than acylated ghrelin, seems to be devoid of any endocrine action. However, it is able to exert some nonendocrine actions including cardiovascular and antiproliferative effects, probably by binding different GHS-R subtypes or receptor families (3, 27).

There is another endogenous ligand for the GHS-R 1a that can be isolated from the endocrine mucosa of the stomach. Des-Gln14-ghrelin has undergone the same process of acylation at its Ser3 residue and is homologous to ghrelin except that it lacks one glutamine. Des-Gln14-ghrelin is the result of alternative splicing of the ghrelin gene, and it seems to possess the same hormonal activities as ghrelin (1, 131). Studies with several analogs of ghrelin with various aliphatic or aromatic groups in the side chain of residue 3 and several short peptides derived from ghrelin as well showed that bulky hydrophobic groups in the side chain of residue 3 are essential for maximum agonist activity. In addition, short peptides encompassing the first four or five residues of ghrelin were found to functionally activate GHS-R 1a about as efficiently as the full-length ghrelin. Thus, the entire sequence of ghrelin is not necessary for activity; the Gly-Ser-Ser(n-octanoyl)-Phe segment appears to constitute the active core required for agonist potency at GHS-R 1a (17, 132). Hosoda et al. (133) isolated, in the course of purification of ghrelin from the stomach, human ghrelin of the expected size, as well as several other ghrelin-derived molecules that could be classified into four groups by the type of acylation observed at the serine-3 position. These peptides were found to be nonacylated, octanoylated (C8:0), decanoylated (C10:0), and possibly decenoylated (C10:1). The major active form of human ghrelin is a 28-amino acid peptide octanoylated at the serine-3 position, as was found for rat ghrelin. Both ghrelin and the ghrelin-derived molecules were found to be present in plasma as well as stomach tissue. Del Rincon et al. (134) pointed out that identification and characterization of the novel gastric peptide hormone, named motilin-related peptide by Tomasetto et al. (127, 135), were strictly connected to ghrelin. Motilin-related peptide shows the same amino acid sequence as ghrelin, reflecting that the same gene encoding for this peptide was discovered by two different groups and was given two different names.

That the scientific work on the GHS also led to the discovery of the motilin receptor is not by chance; the motilin receptor is a member of the GHS-R family having a 52% identity (112, 117, 118, 127, 134, 135, 136, 137, 138, 139). This former orphan G protein-coupled receptor was isolated based on its high homology with the GHS-R, and through ligand screening assays motilin was identified as its endogenous ligand. However, unlike ghrelin, acylation of motilin is not needed for activation of its receptor (112). Prepromotilin, which is also produced by the enteroendocrine cells of the stomach, is almost identical with human preproghrelin, except for the serine-26 residue that is not octanoylated in the prepromotilin-related peptide; however, human ghrelin and motilin show only 36% homology (1, 112, 117, 127, 134, 135, 136, 137, 138, 139, 140). Motilin and motilin receptors have been well characterized in humans and dogs, whereas rodents do not have a motilin receptor. The ability of motilin to exert some stimulatory effect on GH secretion and some orexigenic effect after central administration cannot be mediated by the GHS-R 1a, because motilin does not activate GHS-R 1a (112, 137, 140, 141). On the other hand, ghrelin does not activate motilin receptors (142). Therefore, we consider both ghrelin and motilin as representatives of a novel family of gastrointestinal peptides contributing to the regulation of diverse functions of the gut-brain axis. This in itself is a remarkable turn of a story that started as a field of research focused on GH secretion alone.

	III. The Biology of a Ubiquitously Expressed Hormone and Its Receptor Family

Top Abstract I. Introduction II. Historical Background III. The Biology of... IV. Control of Ghrelin... V. Physiological and... VI. Ghrelin as an... VII. Pharmacological and... References

 
A. Known and unknown GH secretagogue receptors
In addition to the physiological stimulation by hypothalamic GHRH, the release of GH from the pituitary is stimulated by small synthetic peptidyl and nonpeptidyl molecules called "GH secretagogues" (for reviews, see Refs. 48 and 74). They act through a specific G protein-coupled receptor (13), the GHS-R, for which the ligand was unknown until a Japanese group of scientists led by M. Kojima (1) isolated an endogenous ligand specific for the GHS-R, ghrelin, from the stomach. The discovery of this novel gastric hormone, ghrelin, which consists of 28 residues containing an n-octanoyl modification at serine 3, has been recently reviewed by Bowers (70), Kojima et al. (143), Inui (52), and Muccioli et al. (8).

The GHS-R is expressed by a single gene found at human chromosomal location 3q26.2 (113, 114). Two types of GHS-R cDNAs, which are presumably the result of alternate processing of a pre-mRNA, have been identified and designated receptors 1a and 1b (13, 14, 114) (for reviews, see Refs. 112 ,144 ,145 ; also see Ref. 146). Their sequences do not show significant homology with other known receptors; the closest receptor relatives are the neurotensin with 34.9% identity and motilin 1a with 51.6% identity (113). cDNA 1a encodes a receptor, named GHS-R 1a, of 366 amino acids with seven-transmembrane regions and a molecular mass of approximately 41 kDa. The 1b cDNA encodes a shorter form, named the GHS-R 1b, which consists of 289 amino acids with only five-transmembrane regions (13). GHS-R 1b is derived by readthrough of the intron, which produces an in-frame stop codon so that the potential translation product has an identical N terminus with transmembrane domains 1–5 but lacks transmembranes 6 and 7. Although this process does not seem to end in the transcription of a protein, GHS-R 1b expression, however, is widespread in many endocrine and nonendocrine tissues, but its significance remains to be determined (13, 22, 112).

The human GHS-R 1a shares 96 and 93% identity with the rat and pig GHS-R 1a, respectively, and the existence of this receptor can apparently be extended to pre-Cambrian times because amino acid sequences strongly related to human GHS-R 1a have been identified in teleost fish (117). These observations strongly suggest that the GHS-R 1a is highly conserved across the species and probably does have an essential biological function.

The binding of ghrelin and synthetic GHS (such as the peptidyl GHRP-6 and the nonpeptidyl derivative MK-0677) to the GHS-R 1a activates the phospholipase C signaling pathway, leading to increased inositol phosphate turnover and protein kinase C activation, followed by the release of Ca²⁺ from intracellular stores (14, 147). GHS-R activation also leads to an inhibition of K⁺ channels, allowing the entry of Ca²⁺ through voltage-gated L-type, but not T-type channels (148, 149). Unlike the GHS-R 1a, the GHS-R 1b failed to bind GHS and to respond to GHS (13), and its functional role remains to be defined. Synthetic GHS and ghrelin, as well as des-Gln14-ghrelin, a natural isoform that has the same GH-releasing activity as ghrelin (131), bind with high affinity to the GHS-R 1a; their efficacy in displacing [³⁵S]MK-0677 or [¹²⁵I][Tyr⁴]ghrelin binding to pituitary membranes correlates well with concentrations required to stimulate GH release (14, 21, 150). The n-octanoyl group at serine 3 of the ghrelin molecule seems to be essential for the binding and bioactivity of the hormone, at least in terms of GH release. In fact, the nonacylated ghrelin, which circulates in amounts far greater than the acylated form (131), does not displace radiolabeled ghrelin from its hypothalamic or pituitary binding sites (21) and has no GH-releasing or other endocrine activities in rat (1, 70). In man also, the administration of nonacylated ghrelin does not induce any change in the hormonal parameters or in glucose levels, indicating that at least in humans nonacylated ghrelin does not possess endocrine activities of human acylated ghrelin (151).

Interestingly, it has been reported that ghrelin binds a species of high-density lipoprotein (HDL) associated with the plasma esterase, paraoxonase, and clusterin. Both free ghrelin and paraoxon, a substrate for paraoxonase, can inhibit this interaction. Some endogenous ghrelin is found to copurify with HDL during density gradient centrifugation. This interaction links the orexigenic peptide hormone ghrelin to lipid transport and a plasma enzyme that breaks down oxidized lipids in low-density lipoprotein (see Fig. 4). Furthermore, the interaction of the esterified ghrelin with a species containing an esterase suggests a possible mechanism for the conversion of ghrelin to des-acyl ghrelin (152).

View larger version (39K): [in this window] [in a new window]   FIG. 4. Pathways by which ghrelin may influence chronic energy balance. Ghrelin produced by the stomach or the gut can be transported by the bloodstream to specific neuronal circuits situated in hypothalamus or the brainstem that are regulating food intake as well as energy expenditure. It still remains uncertain whether or not ghrelin has to cross the blood-brain barrier (BBB) to influence these central structures. During transport via circulating blood, serum HDLs and presumably other proteins such as albumin bind ghrelin. Ghrelin however may also signal the brain by activating the afferent vagal nervous system as either an endocrine or a paracrine signal directly at the stomach level. Ghrelin-responsive GHS-Rs are expressed at gastric vagal nerves, and vagotomy prevents some of ghrelin’s effects on energy balance. Incoming information represented or triggered by ghrelin is, however, believed to be constantly sensed and analyzed in hypothalamus and the brainstem, independent from its origin or afferent pathway used. Based on constant integration of this and other afferent information about the status of acute and chronic changes in energy balance, an efferent response seems to involve several pathways to balance energy stores and adipose tissue mass. These mainly include the sympathetic nervous system (SNS), the hypothalamic-pituitary adrenal (HPA) axis, and the hypothalamic-pituitary thyroid (HPT) axis. In addition, ghrelin is thought to be produced in brain centers of energy balance control, and, although present there in very small amounts, brain-derived ghrelin might play an additional role in the regulation of energy homoeostasis (51 123 152 198 372 ).

Expression of the GHS-R 1a was shown in the hypothalamus and anterior pituitary gland (13, 30, 156, 157), consistent with its role in regulating GH release. The GHS-R 1a is largely confined to somatotroph pituitary cells and to the arcuate nucleus (13, 14, 158), a hypothalamic area that is crucial for the neuroendocrine and appetite-stimulating activities of ghrelin and synthetic GHS (120, 159). This is supported by the demonstration that ghrelin, as well as synthetic GHS, effectively stimulates the expression of some markers of neural activity (c-fos and early growth response factor-1) in arcuate nucleus neurons (160, 161). The activated hypothalamic cells include GHRH-containing neurons, but also cells expressing the appetite-stimulating neuropeptide Y (NPY) (71, 158) and the endogenous melanocortin receptor inverse agonist, agouti-related protein (AGRP) (31). Detectable levels of GHS-R 1a mRNA were also demonstrated in various extrahypothalamic areas such as the dentate gyrus of the hippocampal formation, CA2 and CA3 regions of the hippocampus, the pars compacta of the substantia nigra, the ventral tegmental area and dorsal and medial raphe nuclei and Edinger-Westphal nucleus, pons and medulla oblongata (24, 30, 162), possibly indicating its involvement in as yet undefined extrahypothalamic actions. More recent localization studies have demonstrated GHS-R expression in multiple peripheral organs, although the RT-PCR primers were generally not selected to differentiate GHS-R 1a from 1b (1, 2, 3, 4, 5, 7, 8, 9, 22, 24, 25, 26, 30, 120, 121, 163). mRNA was shown in the stomach and intestine (3), pancreas (30), kidney (4), heart and aorta (38, 164, 165), as well as in different human pituitary adenomas (6, 163) and various endocrine neoplasms of lung (166), stomach (126), and pancreas (7, 121, 163). These data are in accord with the reported observations indicating, for ghrelin and synthetic GHS, broader functions beyond the control of GH release and food intake (see Section V.C).

Ghrelin and GHS compounds exhibit a high binding affinity to the cloned GHS-R 1a. However, there is evidence, indicating that there are additional binding sites for GHS. Specific binding sites for Tyr-Ala-hexarelin [Tyr-Ala-His-D-2Methyl-Trp-Ala-Trp-D-Phe-Lys-NH₂], other peptidyl GHS (GHRP-2 [D-Ala-D-ßNal-Ala-Trp-D-Phe-Lys-NH₂], GHRP-6, and hexarelin [His-D-2Methyl-Trp-Ala-Trp-D-Phe-Lys-NH₂]) with a receptor density that at least equals the density that was found in the pituitary, have been found in rat and human heart (24, 25, 107, 167, 168), as well as in a wide range of other nonendocrine peripheral human tissues such as lung, arteries, skeletal muscle, kidney, and liver (20, 26). These binding sites are presumably not ghrelin receptors, because they show a very low binding affinity for ghrelin (26). As reported by Bodart et al. (167), the cardiac GHS-R has a molecular mass larger (84 kDa) than that of GHS-R 1a and shows no homology with this receptor. The predicted amino acid sequence of the GHS-R expressed in heart muscle is similar to that of CD36, a multifunctional receptor also known as glycoprotein IV (49).

The functional significance of receptors for peptidyl GHS in peripheral nonendocrine tissues is still unknown. Some findings in the cardiovascular system suggest that these binding sites could mediate GH-independent cardioprotective activities exerted by peptidyl GHS, but not by ghrelin (see Section V.C).

Recently, we have found ghrelin receptors with a binding profile different from the GHS-R 1a ghrelin receptor in human thyroid and breast tumors, as well as in related cancer cell lines (27, 169). In fact, binding of acylated ghrelin to these receptors is surprisingly inhibited by nonacylated ghrelin, as well as by some synthetic GHS (27, 169); a receptor with the same binding profile has been demonstrated at the cardiovascular level (8, 170). It has to be emphasized that nonacylated ghrelin, although unable to bind to the classical GHS-R 1a and to show any endocrine activity, exerts antiproliferative (27) and cardioprotective effects (171). This is illustrated in the experiment of Fig. 1 that compares the ability of unlabeled ghrelin and nonacylated ghrelin to displace [¹²⁵I][Tyr⁴]ghrelin binding to membranes from cultured pituitary explants, H9C2 cardiomyocytes, and MCF-7 mammary carcinoma cells.

B. Known and unknown ligands of the GH secretagogue receptors
Taking into account the GHS-R 1a as receptor of reference, acylated ghrelin and Des-Gln14-ghrelin are its natural ligands; in fact, both molecules possess the same endocrine activities (150).

There are also other natural ligands of the GHS-R 1a. Besides adenosine that binds and activates the receptor (112, 154, 155), it has been demonstrated that CST, a neuropeptide, binds with high affinity the GHS-R 1a in human hypothalamus and pituitary tissues (19, 112, 154, 172). CST is a recently described neuropeptide showing high structural homology with SS (173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188) that binds to all SS receptor subtypes with an affinity (1–2 nM) close to that of SS (103, 189, 190). In fact, in humans as well as in animals, CST and SS exhibit the same endocrine activities (191, 192). The existence of specific receptors that selectively bind SS or CST has been hypothesized (189, 190), based on evidence that CST possesses an action profile different from SS (189, 193, 194) and on the fact that CST and SS are often coexpressed in the same neurons but are regulated by different stimuli (189, 195, 196). Given these findings, the ability of CST to bind the GHS-R 1a is of particular relevance because SS and its fragments do not bind the same receptor (19, 112, 154, 172). Interestingly, the classical synthetic SS analogs, i.e., octreotide, lanreotide, and vapreotide, bind the GHS-R 1a with an affinity lower than that of CST (19, 112, 154, 172). These findings have generated the hypothesis that CST could play a potential role in the control of somatotroph secretion via both SS and GHS-Rs. Where this is the case, CST would represent the link between ghrelin and the SS/CST system that had not previously been demonstrated.

On the other hand, the GHS-R 1a is unlikely to be the only GHS-R (see Section III. A). It has already been demonstrated that a GHS-R subtype able to bind nonacylated as well as acylated ghrelin exists and likely mediates biological activities (27). This report might provide another explanation besides the existence of different pockets within the GHS-R 1a to explain the observation that different molecules are able to bind, but not activate it (112).

Another GHS-R subtype likely mediates the influence of ghrelin on insulin secretion and glucose metabolism, because this effect is not shared by synthetic peptidyl GHS that generally mimic ghrelin actions (35). The cardiovascular receptor that only binds peptidyl GHS is unlikely to be a GHS-R, because it does not bind ghrelin (25, 26, 49, 49, 167, 197).

Given this complexity, it is clear that further studies are required to clarify whether ghrelin is the sole ligand or one of a number of ligands activating the GHS-R 1a and whether that receptor used for ghrelin isolation is the sole receptor or one of a group of receptors for such ligands.

	IV. Control of Ghrelin Secretion: Indications for Its Importance in Biology

Top Abstract I. Introduction II. Historical Background III. The Biology of... IV. Control of Ghrelin... V. Physiological and... VI. Ghrelin as an... VII. Pharmacological and... References

 
Although the majority of circulating ghrelin originates from the stomach and the small bowel (3, 124), ghrelin is expressed in a variety of tissues that include the stomach, the intestine, the pituitary, the placenta, lymphocytes, the testes, the lungs, the kidney, the pancreas, and the hypothalamus (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12).

The activation level of the ghrelin receptor(s) is the parameter that is decisive for ghrelin action. Regulation of the extent and magnitude of ghrelin action therefore involves several mechanisms that are, at least in part, independent. These mechanisms include: 1) the regulation of transcription and translation of the ghrelin gene; 2) the level of enzymatic activity of the putative acyl transferase that is responsible for the posttranslational octanoylation of the ghrelin molecule; 3) secretion rates of the bioactive ghrelin molecule; 4) putative enzymatic processes deactivating circulating ghrelin; 5) possible influence of ghrelin binding proteins on the hormone’s bioactivity; 6) variable accessibility of target tissue (i.e., blood-brain barrier transport); 7) clearance or degradation of ghrelin by kidney or liver passage; 8) circulating concentration of additional endogenous ligands or other possibly cross-reacting hormones; 9) the amount of expression of ghrelin receptor(s) in target tissue; and 10) their sensitivity to the level of intracellular signaling mechanisms. Most studies to date have focused on changes in gastric ghrelin mRNA expression or variation of circulating ghrelin concentration as quantified by immunoassay measurements from plasma samples. The search for a ghrelin activating acyltransferase, for possible additional endogenous ligands to ghrelin receptor(s), and for specific ghrelin binding proteins is still ongoing. However, a few studies have started to shed light on important issues such as blood-brain barrier transport of ghrelin (198) and the regulation of ghrelin receptor expression, i.e., in the hypothalamus (199). Analysis of the regulation of ghrelin expression levels in tissues other than stomach is hardly possible due to the small amounts of peptide present in these organs.

The measurement of ghrelin immunoreactivity involves technical difficulties, which imply that all results based on the concentration of circulating ghrelin as quantified by immunoassays should still be interpreted with much caution. Although commercially available ghrelin immunoassays are very likely to reflect total ghrelin peptide concentration, reliable methods to routinely quantify individual ghrelin species are still not available. An ideal tool would be a sensitive and specific sandwich immunoassay based on two monoclonal antibodies recognizing an epitope associated with the octanoyl side chain and another one at the C-terminal end of the 28-amino residue peptide. Existing assays targeting the C-terminal end of the molecule miss potential crucial changes in the percentage of circulating octanoylated ghrelin. Immunoassays targeting the octanoyl side chain of the molecule might suffer from interference from other octanoylated molecules, which are likely to exist. Therefore, plasma ghrelin levels as described by several research groups vary according to the antiserum used and are influenced by varying techniques, such as the use of an additional extraction step. Furthermore, there are controversial data on the stability of ghrelin in plasma samples, the influence of storing time and thaw/freeze cycles, pH changes, or the necessity for enzyme-blocking additives to plasma samples before measuring ghrelin. Although absolute plasma ghrelin levels and ghrelin reference standards still have to be determined, it appears reasonable to investigate ghrelin regulation and physiology by measurement of relative differences of circulating total ghrelin levels using available immunoassays. In the following section, existing knowledge regarding the regulation of ghrelin expression and secretion is summarized, although this current model might have to be revised substantially when details of the unknown mechanisms and open questions mentioned above become available.

Ghrelin mRNA expression as well as ghrelin peptide have been localized most impressively in the oxyntic glands, specifically the X/A-like cells of the gastrointestinal tract. These cells represent about one fourth of all endocrine cells in the oxyntic mucosa, whereas other cells within these glands, such as histamine-rich enterochromaffin-like cells (70%) and D-(SS) cells (10%), are not ghrelin positive (3, 124). Ghrelin is found from the stomach to the colon with caudally decreasing density of expression, which is in agreement with the fact that X/A-like cells are not strictly confined to oxyntic mucosa (3, 124). Ghrelin-containing enteroendocrine cells mostly have no continuity with the lumen, probably respond to physical and/or chemical stimuli from the basolateral side, and are closely associated with the capillary network running through the lamina propria (3, 124). A recent study shows that ghrelin-secreting cells occur as open- and closed-type cells (open or closed toward the lumen) with the number of open-type cells gradually increasing in the direction from the stomach to the lower gastrointestinal tract (200).

Although a classical endocrine role for ghrelin as a peptide hormone that is secreted into this capillary network is evident, local paracrine activities of ghrelin might play an additional role (3, 124). Removal of the stomach or the acid-producing part of the stomach in rats reduces serum ghrelin concentration by approximately 80%, further supporting the view that the stomach is the main source of this endogenous GHS-R ligand (3, 124). However, in a recent study, plasma levels of ghrelin after total gastrectomy gradually increased again (133), suggesting that the stomach is the major source of circulating ghrelin but that other tissues can compensate for the loss of ghrelin production after gastrectomy (201). Cummings and co-workers (202, 203, 204) showed that total plasma ghrelin is hardly detectable after gastric bypass surgery, a phenomenon that is interpreted as a shutdown of gastric ghrelin-secreting cells due to a lack of contact with ingested nutrients. On the other hand, no evidence for Roux-en-Y gastric bypass surgery per se having an effect on ghrelin levels, independent of weight loss, was obtained (205).

Small concentrations of ghrelin are found in the pancreas (206), where ghrelin immunoreactivity was localized in a subgroup of endocrine cells that are also immunopositive for pancreostatin. Neither colocalization of ghrelin and enterochromaffin-like cells nor colocalization of ghrelin and D cells was found in the pancreas. Therefore, it was concluded that the ghrelin-positive cells must belong to the third subpopulation, the A cells (124). We have recently shown that ghrelin is produced by a fraction of endocrine pancreatic cells, namely the insulin-producing H cells, as confirmed by double immunofluorescence studies (169). Ghrelin mRNA and ghrelin peptide have also been detected in rat and human placenta in which they are expressed predominantly in cytotrophoblast cells and very sporadically in syncytiotrophoblast cells. A pregnancy-related time course, represented by an early rise of ghrelin expression in the third week and decreasing in the latest stages of gestation, but still detectable at term, was found in rats. In human placenta, ghrelin is mainly expressed in the first half of pregnancy and is not detectable at term (5). Involvement of ghrelin in fetal-maternal interaction via autocrine, paracrine, or endocrine mechanisms is discussed (5). In addition to the presence of GHS-R in pituitary cells, ghrelin mRNA expression and ghrelin immunopositive cells were detected in normal pituitary cells as well as in pituitary tumors (6, 22, 163). This suggests a possible autocrine or paracrine role for hypophyseal ghrelin, although only about 5% of the detected ghrelin peptide derived from the pituitary has been found to be octanoylated (6, 163); Ghrelin synthesis has been shown by the use of real-time PCR and direct sequencing of the PCR product in corticotroph, thyrotroph, lactotroph, and somatotroph cells of the pituitary (6, 163), whereas the highest levels of ghrelin expression were found in nonfunctioning adenomas, moderate ghrelin levels were found in GH-producing adenomas and gonadotropin-producing adenomas, and the lowest level was found in prolactinomas (6, 163). The same group detected small amounts of ghrelin in the adrenal glands, esophagus, adipocytes, gall bladder, muscle, myocardium, ovary, prostate, skin, spleen, thyroid blood vessels, and liver using real-time PCR (22). By combinatory use of reverse-phase HPLC and RIA of purified aliquots, production of ghrelin in mouse kidney was shown in greater abundance than in mouse plasma. In addition, preproghrelin production was shown in rat mesangial cells and mouse podocytes, indicating the production of ghrelin in kidney, glomerulus, and renal cells and suggesting possible paracrine roles for ghrelin in the kidney (4). In several regions of the brain, ghrelin was detected by means of immunohistochemistry. However, the location of the few ghrelin-positive neurons that were identified depends on the recognized epitope of the ghrelin antiserum used (51, 207). Very recently, ghrelin expression has been demonstrated in a previously uncharacterized group of neurons adjacent to the third ventricle. These neurons send efferents onto NPY, AGRP, proopiomelanocortin (POMC), and CRH neurons, suggesting that local ghrelin would represent a novel regulatory circuit controlling energy homeostasis (Fig. 3) (123). However, ghrelin found in the hypothalamus still has to be considered as possibly derived from the periphery, and the participation of hypothalamic ghrelin in neuropeptidergic energy balance control mechanisms remains questionable (51). Human ghrelin as well as GHS-R mRNA expression was shown by real time-PCR and confirmed by DNA-sequencing in human T-lymphocytes, B-lymphocytes, and neutrophils from venous blood of healthy volunteers. Large interindividual differences in ghrelin mRNA expression levels were described; however, cell type and maturity of the cells did not seem to have an influence on ghrelin production in immune cells (9, 208). Interestingly, it has recently been shown that small-molecule GHS have a considerable immune-enhancing effect (9, 208).

View larger version (38K): [in this window] [in a new window]   FIG. 3. Within a complex neuroendocrine network, afferent signals from the periphery are continuously indicating acute and chronic changes of energy balance, whereas integrative regulatory circuits in the CNS are modulating efferent pathways to adjust orexigenic drive, energy expenditure, and nutrient metabolism accordingly. Ghrelin is thought to be significantly involved in this neuroendocrine network regulating energy balance in at least two ways: 1) as a peripheral hormone from the stomach that, along with other signals such as insulin or leptin, informs the central energy balance control when energy stores diminish, to increase orexigenic drive and decrease energy expenditure; and 2) as a hypothalamic neuropeptide expressed in previously unidentified population of neurons adjacent to the third ventricle between the ventromedial hypothalamus, the dorsal hypothalamus, the paraventricular nucleus, and the arcuate nucleus. Efferents of ghrelin-expressing neurons project to key circuits of central energy balance regulation and may balance the activity of orexigenic NPY/AGRP with anorectic POMC neurons to modulate a resulting efferent message, which is believed to be mediated in part by TRH and CRH. Dotted lines indicate indirect effects or actions, whereas question marks indicate unproven actions (51 233 351 352 353 ).

Only a few determinants of circulating ghrelin concentration have been identified to date. Spontaneous ghrelin secretion is pulsatile in rats (44), and 24-h ghrelin variation is reported in humans by some (202, 209), but not by others (210). It is unclear whether aging is a determinant of serum ghrelin concentrations. Ghrelin secretion is reported to be sexually dimorphic in humans, with women in the late follicular stage having higher levels than men (210). Among determinants of ghrelin secretion, the most important appear to be insulin (211, 212, 213, 214, 215, 216, 217), glucose (33, 218, 219, 220, 221, 222), and SS (210, 223, 224, 225, 226, 227). Possibly, GH (224, 228, 229, 230, 231, 232), leptin (204, 233, 234, 235, 236, 237), melatonin (238), thyroid hormones (239), glucagon (240), and the parasympathetic nervous system (32, 241) also play a role in ghrelin metabolism. In mice, rats, cows, and humans, ghrelin mRNA expression levels or circulating ghrelin levels are increased by food deprivation and appear to be decreased postprandially (33, 137, 202, 209, 242, 243, 244, 245, 246). This phenomenon, which has been confirmed by several study groups in the recent past, further supports the emerging concept of ghrelin as an endogenous regulator of energy homeostasis. In addition to fasting, ghrelin expression can be stimulated in rats by insulin-induced hypoglycemia, leptin administration, and central leptin gene therapy (233, 243). Ingestion of sugar suppresses ghrelin secretion in rats (33). These observations indicate a direct inhibitory effect of glucose/caloric intake on ghrelin-containing X/A-like cells in the oxyntic mucosa of the rat stomach rather than an exclusively insulin-mediated effect. That insulin is an independent determinant of the circulating ghrelin concentration has recently been shown by several study groups using hyperinsulinemic euglycemic clamps in humans (211, 212). These findings add further evidence connecting ghrelin to mechanisms governing energy balance and the regulation of glucose homeostasis.

Further insight into the apparently complex mechanisms regulating ghrelin secretion is based on studies showing an increase of circulating ghrelin levels in rats after surgical interventions such as vagotomy and hypophysectomy (32, 228). Human GHD, however, is not associated with increased plasma ghrelin levels (247). On the other hand, administration of synthetic GH to rats decreases circulating ghrelin levels, and therapeutic interventions causing normalization of GH levels in patients with acromegaly increase ghrelin levels (228, 229, 230). These partial, somewhat contradictory, observations could be due to species-specific differences between rodents and humans, or they could indicate that an acute, but not chronic, change of GH levels modulates ghrelin concentration. An increase in circulating ghrelin levels in rats with age, up to 90 d (248), has not been confirmed as yet for human populations. Early studies seem to indicate, however, that human ghrelin secretion decreases with age during childhood (249). A pathophysiological factor that might increase circulating ghrelin levels is the production of ghrelin by certain endocrine tumors of the stomach and the intestine such as carcinoids (126). A recent, very intriguing series of clinical studies by Cummings et al. (202, 209) indicates that each daily meal is followed by decreases of circulating ghrelin levels, most likely reflecting acutely reduced ghrelin secretion from the gastrointestinal tract. The authors speculate in addition that an observed premeal rise of circulating human ghrelin levels might reveal a role for ghrelin in meal initiation. This theory fits well with the observation that ghrelin administration in healthy volunteers causes hunger (15, 16, 35). Ghrelin might also reflect the acute state of energy balance, signaling to the central nervous system (CNS) in times of food deprivation that increased energy intake and an energy-preserving metabolic state are desirable (33, 51). In addition, one biological purpose of these multiple roles of ghrelin might be to ensure the provision of calories that GH requires for growth and repair.

In summary, ghrelin expression and ghrelin secretion are mainly influenced by changes in energy balance and glucose homeostasis, followed by alterations of endocrine axes such as increasing GH concentrations. Based on the currently available data, ghrelin therefore seems to be part of a molecular regulatory interface between the energy homeostasis, glucose metabolism, and physiological processes regulated by the classical endocrine axes such as growth and reproduction.

	V. Physiological and Pathophysiological Actions of Ghrelin

Top Abstract I. Introduction II. Historical Background III. The Biology of... IV. Control of Ghrelin... V. Physiological and... VI. Ghrelin as an... VII. Pharmacological and... References

 
A. Hypothalamic-pituitary actions
1. GH-releasing activity.
Ghrelin and synthetic GHS possess strong and dose-related GH-releasing activity that is more marked in humans than in animals (1, 14, 15, 16, 18, 20, 250, 251, 252). Natural and synthetic GHS stimulate GH release from somatotroph cells in vitro (1, 62, 65, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265), probably by depolarizing the somatotroph membrane and by increasing the amount of GH secreted per cell (258). A stimulatory effect of GHS on GH synthesis has also been reported by some authors (260). In vitro, the GH-releasing activity of GHS is lower than that of GHRH (62, 65, 254, 266). Under this condition, an additive or a true synergistic effect of GHS on GHRH-stimulated GH has been reported (65, 254, 255, 256, 262, 263). At the pituitary level, the stimulatory effect of GHS on GH secretion from somatotroph cells is abolished by specific GHS antagonists but not by GHRH antagonists (65, 255, 258). SS inhibits the stimulatory effect of GHS on GH secretion from pituitary (253, 254, 266, 267). However, there is evidence suggesting that GHS could act by antagonizing the inhibitory activity of SS on GH release by counteracting its hyperpolarizing effect on somatotroph cell membranes (258).

The GH-releasing activity of GHS is clearly greater in hypothalamic-pituitary preparations than in pituitary preparations (268), in agreement with evidence that their stimulatory effect on GH secretion is greater in vivo than in vitro (65, 269). Indeed, in vivo, GHS show synergistic effects on GHRH-stimulated GH release (65, 270) and prevent the normal cyclic refractoriness to GHRH (269). To confirm that the most important action of ghrelin and synthetic GHS to release GH takes place at the hypothalamic level, the GH-releasing effect of GHS is markedly reduced, although not abolished, in animals with lesions of the pituitary stalk (271, 272, 273).

At the hypothalamic level, ghrelin and GHS act via mediation of GHRH-secreting neurons as indicated by evidence that passive immunization against GHRH, as well as pretreatment with GHRH antagonists, reduces their stimulatory effect on GH secretion (65, 269, 274, 275, 276). An increased release of GHRH in portal blood of the pituitary after GHS administration has also been shown in sheep (277). In terms of GH release, GHS are active in dwarf mice (278) but not in the lit/lit mouse, which has no pituitary GHRH receptors (279). However, in both GHD rats and lit/lit mice, systemic administration of GHS activates a subpopulation of hypothalamic arcuate neurons where the highest density of GHRH-secreting neurons is present. Furthermore, because the lit/lit mouse pituitary does not release GH after GHS administration, the finding that the central actions of GHS remain intact in these animals suggests the possible existence of two subpopulations of putative GHS-Rs (280, 281).

At the hypothalamic level, ghrelin and GHS do not inhibit SS secretion in vitro in rats; however, some inhibition of hypothalamic SS secretion after exposure to GHS was observed in vivo in pigs (137, 227, 282, 283, 284). Interestingly, GHS likely act as functional SS antagonists at either the hypothalamic or the pituitary level (71, 258). In vitro and in vivo, GHS and GHRH induce homologous but not heterologous desensitization (61, 65, 78, 254, 255, 262, 263, 266, 269). Prolonged administration of GHS in animals increases IGF-I levels (61, 87, 93, 267, 285, 286), indicating that they are able to enhance the activity of the GH/IGF-I axis.

As anticipated, ghrelin and synthetic GHS show their most potent GH-releasing activity in humans (1, 14, 15, 16, 18, 20, 250, 251, 252) and in animals in vivo because GHS and GHRH are synergistic, indicating that they act, at least partially, via different mechanisms (14, 20, 71, 120). Nevertheless, GHS require GHRH activity to fully express their GH-releasing effect (14, 20, 71, 120). In humans, the GH response to GHS is strongly inhibited, although not abolished, by GHRH receptor antagonists as well as by hypothalamopituitary disconnection (272, 287, 288, 289). This is in agreement with the assumption that the most important action of GHS takes place at the hypothalamic level (14, 20, 61, 63, 65, 120). Moreover, patients with a GHRH receptor deficiency show no increase in GH secretion in response to GHS stimulation (290, 291, 292) but maintain their ability to increase PRL as well as ACTH and cortisol secretion after GHS stimulation (290, 291, 292).

There is evidence, both in humans and in animals, that ghrelin and synthetic GHS can also act as functional SS antagonists at both the pituitary and hypothalamic levels (20, 71, 191, 218, 227, 258, 293). In fact, in humans the GH response to ghrelin and GHS is not enhanced by inhibition of SS release (induced by indirect cholinergic agonists or arginine), whereas it is partially refractory to the inhibitory effect of substances acting via stimulation of hypothalamic SS secretion (such as acetylcholine receptor antagonists, ß-adrenoreceptor agonists, glucose) (20, 191). Indeed, ghrelin and GHS are even partially refractory to the inhibition of substances that act on the pituitary somatotroph cells, such as free fatty acids and even to exogenous SS (20, 191, 293, 294). GHS are also partially refractory to the negative feedback of GH itself and to the negative feedback of IGF-I action (193, 295, 296).

In humans, as in animals, there is evidence that GHS and GHRH induce homologous, but not heterologous, desensitization (64, 218, 261, 297, 298, 299, 300, 301, 302). Homologous desensitization to the activity of GHS has been shown during GHRP infusion (297, 299, 300), but not after intermittent oral or intranasal daily administration of the peptide for up to 15 d (303, 304). On the other hand, prolonged administration of GHS by the parenteral, intranasal, or oral route enhances spontaneous GH pulsatility over 24 h and increases IGF-I levels in normal young adults, as well as in short children and elderly subjects (11, 85, 86, 299, 300, 304, 305, 306).

The GH-releasing effect of GHS undergoes marked age-related variations, increasing at puberty. It plateaus in adulthood and decreases during further aging (20). The mechanisms underlying these variations differ by age. The enhanced GH-releasing effect of GHS at puberty, for instance, is caused by the positive influence of increased serum estrogen levels, which increase GHS-R expression (110, 196, 307, 308, 309, 310, 311). However, estrogen insufficiency does not fully explain the reduced GH response to GHS in postmenopausal women (20, 196, 312, 313, 314). In agreement with the reduction in hypothalamic GHS-Rs in human aging, the GH response to hexarelin in elderly subjects is further increased, but not restored, by supramaximal doses (20, 24, 307). The most important mechanism accounting for reduced GH-releasing activity of GHS in aging is probably represented by age-related variations in neural control of somatotroph function, including GHRH hypoactivity and somatostatinergic hyperactivity (20, 314). On the other hand, it has also been hypothesized, but not proven, that the age-related decline in GH secretion might reflect a decrease in activity of the endogenous GHS system (i.e., ghrelin release and/or receptor expression) (24, 70, 314, 315). As with GHS, the GH-releasing effect of ghrelin is independent of gender but undergoes age-related decrease. Again, the effect of ghrelin on lactotroph and corticotroph secretion is age and gender independent (316).

Despite the strong GH-releasing effects of ghrelin and GHS, whether GH release is the most important physiological action of ghrelin has been recently questioned. In fact, GHRH antagonist strongly inhibits 24-h GH secretion, whereas it does not affect circulating ghrelin levels (210). Moreover, ghrelin does not mediate the GH response to provocative stimuli such as insulin-induced hypoglycemia (213, 214), as well as GH rebound after withdrawal of SS infusion (223). These observations are in agreement with evidence from animal studies showing that ghrelin secretion is pulsatile and is associated much more with food intake than with GH pulses (44).

Theoretically, ghrelin or GHS could have diagnostic and therapeutic implications based on the strong and reproducible GH-releasing effects of orally active GHS.

Particularly when combined with GHRH, ghrelin and GHS can be used for one of the most potent and reliable stimulation tests to evaluate the capacity of the pituitary to release GH for the diagnosis of GHD (252, 317, 318, 319, 320). Provided that appropriate cut-off limits are established, these tests using GHS for the diagnosis of GHD are as sensitive and specific as an insulin tolerance test, the gold standard test for the diagnosis of GHD (252, 318, 319). Long-acting and orally active ghrelin analogs might represent an anabolic treatment in frail elderly subjects. This potential treatment modality is based on the rationale that age-related reduction in the activity of the GH/IGF-I axis probably accounts for changes in body composition, structure functions, and metabolism in normal elderly subjects that are remarkably similar to those in GHD adults (321, 322). Also, the potential pituitary GH release in aged subjects is still remarkably intact, given the fact that the appropriate stimuli are used (322). Finally, GH-releasing substances would represent a more physiological approach to increase endogenous GH pulsatility than a single daily dose of recombinant human GH (321, 322).

At present, there is no definite evidence that shows the therapeutic efficacy of ghrelin analogs as anabolic agents acting via rejuvenation of the GH/IGF-I axis in elderly subjects, although some benefits in osteoporosis have been reported (99, 102).

2. PRL- and ACTH-releasing activities.
Activity of both ghrelin and synthetic GHS is not fully specific for GH, because it also includes stimulatory effects on both the lactotroph and corticotroph system (16, 18, 20, 63, 250, 323, 324). However, some synthetic GHS that exclusively stimulate GH secretion have been reported (106). The stimulatory effect of ghrelin and its analogs on PRL secretion in humans is far less age and gender dependent than the effect on GH secretion (316).

The stimulatory effect of GHS on the activity of the hypothalamus-pituitary-adrenal axis in humans is remarkable and similar to that of the administration of naloxon, vasopressin, and even CRH. Interestingly, the effect of ghrelin on ACTH secretion is even more pronounced than that elicited by synthetic GHS (16, 18, 69, 325, 326, 327, 328, 329, 330). However, this ACTH release after GHS administration appears to be an acute neuroendocrine effect that attenuates during prolonged treatment (14, 20, 70, 85, 86, 285, 331).

The GHS-induced ACTH release is independent of gender but shows peculiar age-related variations (331). It increases at puberty, then shows a reduction in adulthood and, again, a trend toward an increase in aging when the GH response to these molecules is clearly reduced (191, 316, 325, 331).

Under physiological conditions, the ACTH-releasing activity of GHS is entirely mediated via the CNS (20, 272, 331, 332). These mechanisms via the CNS not only include CRH- and/or vasopressin-mediated actions (20, 137, 325, 327, 328, 330, 331, 333) but also via NPY and/or -aminobutyric acid (GABA) (160, 283, 334). The ACTH response to natural and synthetic GHS is generally sensitive to the negative cortisol feedback mechanism (20, 105, 331, 334). However, the stimulatory effect of ghrelin and GHS on corticotroph secretion is exaggerated and higher than that of human CRH in patients with pituitary ACTH-dependent Cushing’s disease, probably reflecting a direct action of ghrelin and GHS on the pituitary ACTH-secreting tumor cells (20, 121, 331, 335, 336, 337). Interestingly, the administration of CRH to humans does not induce any significant increase in ghrelin secretion (230). In agreement with the presence of ghrelin and GHS-R expression in ectopic ACTH-secreting tumors, exaggerated ACTH and cortisol response to GHS has also been observed in patients with ectopic ACTH-dependent Cushing’s syndrome (121, 163, 166, 331). These observations, however, reduce the potential use of GHS in testing ACTH secretion to distinguish patients with pituitary from ectopic ACTH-dependent hypercortisolism.

B. Central actions of ghrelin and GHS
1. Effects on food intake.
Years before ghrelin was discovered, sporadic and greatly neglected reports on observations in rodents indicated that some GHS might possess orexigenic activity (338, 339, 340, 341). Rumors about GHS-induced ravenous hunger attacks in children with idiopathic short stature occurring within the framework of clinical studies on GHS have never been officially confirmed. A research group led by S. Dickson had, however, gathered a substantial amount of very intriguing data during the last decade showing GHS-induced neuronal activity in hypothalamic areas that are currently considered the central processing unit controlling energy balance (160, 199, 279, 280, 342, 343, 344, 345). A dense expression of the G protein-coupled receptor GHS-R 1a has been shown on those neurons, and it is bound and activated by ghrelin as well as by other GHS and GHRP (13, 113, 114, 158). Still, it was a surprise to many when ghrelin, the endogenous ligand of the GHS-R (1, 143), emerged as one of the most powerful orexigenic and adipogenic agents known in mammalian physiology (33, 34, 52, 70). At first, it was puzzling to link adipogenic effects to a hormone that had originally been discovered as a potent secretagogue of a lipolytic hormone, GH (346, 347). However, a rapidly growing body of data reflecting a previously unidentified interface between energy balance regulation, glucose homeostasis, and hypothalamic neuropeptides started to make sense as an evolving mosaic drawn together by ghrelin. Similar to the discovery of the satiety effects of leptin that indicated adipocytes as endocrine organs, the observation of ghrelin’s involvement in energy balance regulation is pointing to an additional endocrine role for the stomach as well (1, 124). Very recently, ghrelin expression was found in a previously uncharacterized group of neurons adjacent to the third ventricle between the dorsal, ventral, paraventricular, and arcuate hypothalamic nuclei. These neurons send efferents onto key hypothalamic circuits, including those producing NPY, AGRP, POMC, and CRH. Within the hypothalamus, ghrelin mostly stimulated the activity of arcuate NPY neurons in the paraventricular nucleus, so the release of ghrelin may stimulate the release of orexigenic peptides and neurotransmitters, thus representing a novel regulatory circuit controlling energy homeostasis (Fig. 3) (123, 348,