INTRODUCTION

Growth hormone (GH) has a complex regulation with two antagonistic hypothalamic hormones, growth hormone releasing hormone (GHRH) and somatostatin, as well as the liver-derived hormone IGF-I. Perhaps the old name of somatotrophic hormone (STH) is more coherent than GH, as this hormone is tightly regulated by the metabolic milieu; additionally, this regulation appears to be superimposed over the classical regulation by peptide hormones. For example, metabolic signals such as glucose, amino acids, free fatty acids and their by-products, such as keto-acids, as well as the energy balance status regulate the secretion of GH in a very relevant form. In turn, GH causes complex actions on the general metabolism of a given individual. The upshot of this picture is of one hormone whose actions are implicated in a dual action on somatic growth and in the regulation of general metabolism, and which is in turn, regulated by the energetic homeostasis of the individual.1 The recently discovered hormone, ghrelin, may well be the bridge connecting somatic growth with general metabolism.

HISTORICAL BACKGROUND

Ghrelin is the result of the so called "reverse pharmacology", which started with the development of artificial compounds named growth hormone secretagogues (GHS), followed by the cloning of their receptor and finally the identification of the natural hormone. In fact, in the late 1970s the first highly potent GH-releasing hexapeptide, (GHRP-6) was developed. This was followed by other GHS compounds such as hexarelin, or MK-0677.2 These GHSs were found to be potent releasers of GH in vitro and in vivo, by acting on specific receptors at the pituitary level not related to GHRH or somatostatin. Furthermore, they were seen active by any route of administration, including oral, and active in all the species tested. Later GHSs were used for the cloning of the GHS-receptor.3 The GHSs were not discovered, but invented, as no similar compounds existed in nature. Obviously, the new receptor must have a natural endogenous ligand. The orphan-receptor strategy was then employed by the group of Kojima and Kangawa4 to screen different tissue extracts. The highest expression of GHS-receptor activating factor was found in the stomach. This endogenous ligand was named ghrelin. Ghrelin was found to be a potent releaser of GH and in addition, actively participate in controlling energy balance and the regulation of food intake.5 Reverse pharmacology permitted identification of this natural ligand, ghrelin.

DISTRIBUTION OF GHRELIN-SECRETING CELLS

Two cellular areas in the body were found to be relevant in the production of ghrelin. One was an area in the gastric fundus where ghrelin is predominately expressed and secreted. Specifically, plasma ghrelin originates in the oxyntic gland where A-like cells exist.6 Lower concentrations have also been reported in the remainder of the bowel including the colon. Ghrelin positive cells are positioned close to the capillaries and have no contact with the lumen of the oxyntic gland, indicating that secretion occurs into the plasma and not into the intestinal tract.

The second area was found in the central nervous system where neuronal cell groups release ghrelin in a synaptic transmission. Since ghrelin was determined to be implicated in the regulation of appetite, it was not surprising to find abundant ghrelin in the arcuate nucleus of the hypothalamus which also is a region rich in GHRH neurons. 4 Elsewhere, in the CNS, ghrelin was also present. Immunoreactive neurons were observed in a continuum filling the internuclear space between the paraventricular, arcuate, ventromedial, and dorsomedial hypothalamic nuclei, the perifornical region, and the ependymal layer of the third ventricle.7 Interestingly, these novel cell groups of ghrelin immunoreactive neurons did not overlap with any of the known cell populations implicated in energy homeostasis, thus suggesting new functions. In addition to their role in the regulation of energy balance, whether these neuronal groups also participate in the regulation of GHRH or somatostatin neurons is an open question.

Ghrelin has also been identified in the placenta,8 an organ that contains all the main regulatory components of the somatotrope axis, i.e., GH, GHRH, SST, IGF-I, and ghrelin. Although, placental expression of ghrelin changes significantly throughout pregnancy,8 and is involved in the decidualization of human endometrial stromal cells,9 the physiological function of this new hormone in the placenta is unknown. The pituitary, heart, kidney, endocrine pancreas, gonads, lungs, and lymphocytes all express ghrelin in low amounts.10-15

MOLECULAR BIOLOGY

Operation of ghrelin gene in humans Ghrelin is encoded in the two first exons, and it is unique in the hnRNA processing that by alternative splicing two mature mRNAs are derived, one for ghrelin and other for des-Gln-14-ghrelin. The asterisk marks the boundary between the first intron and the second exon where the alternative splicing occurs. Before secretion, a n-octanoyl acylation occurs in Ser3 through a novel and still undefined enzymatic mechanism. In addition other related molecules are also secreted in minor quantities. In rodents, a testis specific ghrelin gene-derived transcript is encoded in the third intron. Reprinted from: Casanueva FF, Dieguez C. Rev Endocrine Metab Dis 2002;3:325-338.

The human ghrelin gene is located in chromosome 3. It is made up of 4 exons and 3 introns. The mature protein is encoded in exons 1 and 2 (Figure 1).16 The genetic structures of the ghrelin genes in the human and rat are identical and very similar to that gene in the mouse. The 5'-flanking region of the gene contains a non functional TATATAA box, as well as a ghrelin promoter which is activated by glucagon and c-AMP, although no AP1 site or CRE element is present.17 Some gastric tumor cell lines express the promoter, however others do not, suggesting that human ghrelin promoter may have cell-specific activity. The hnRNA of the gene transcript is processed by alternative splicing to yield two different mature mRNAs; one produces the ghrelin precursor and the second yields des-Gln 14-ghrelin.18 Ghrelin provides the first example of the production of two different mature biologically active peptides resulting from the alternative splicing of a peptide coding region.

The human ghrelin precursor (prepro-ghrelin) is composed of 117 amino acids, and the ghrelin sequence of 28 amino acids immediately follows the 23-residue signal peptide. Before being secreted, the ghrelin molecule undergoes an enzymatic process at the cytoplasm, an n-octanoyl addition at Ser 3. This esterification by n-octanoic acid, which is essential for the biological activity of ghrelin, yields the finally secreted peptide of 3315 mw. This process of acylation has no precedent in cell biology either, being the first example of acylation in a secreted protein. 10

The main product of that original synthesis process is mature ghrelin. The production of des-Gln14-ghrelin is minor. In addition, the human stomach releases small quantities of other related molecules.19 The active binding core of the molecule consists of the first 4-5 amino acids including the acylated Ser3, short peptides containing this sequence efficiently bind to the GHS receptor, although they are devoid of GH secretory capability.10 It is interesting to speculate how the fatty acid residue changes the physical properties of ghrelin to facilitate its coupling in the biomembrane-receptor structure.

GHRELIN SECRETION AND ACTION

Circulating ghrelin originates mainly in the stomach, and circulates at plasma concentrations of 200-600 ng/L. However, close to 80% of the total content is deamidated ghrelin, i.e., devoid of biological activity. Current RIAs mostly measure total ghrelin. Precaution is needed in the interpretation of data as bioactive ghrelin does not have a fixed ratio in relation to total ghrelin. Interestingly, the integrated secretion of ghrelin during 24 hours correlates significantly with the values obtained in the basal state20 making it possible to use a single determination in clinical situations. No significant differences occur between serum and plasma concentrations; total ghrelin is resistant to repeated thawing, however warm temperatures for prolonged times should be avoided. 21

Regulation of gastric-derived ghrelin by different signals SST= somatostatin CCK = cholecystokinin

Controversy exists whether ghrelin crosses the blood-brain barrier (BBB) to act as the afferent loop controlling either energy homeostasis or GH secretion. Human ghrelin has been reported to cross the BBB, but rodent ghrelin reportedly does it with less effectivity.22 No doubt exists that ghrelin administration activates fos and Egr-1 proteins in neurons of the arcuate, paraventricular and dorsomedial nuclei, and the area postrema of the hypothalamus, while deamidated ghrelin in these studies was devoid of action.23 The debate is whether peripheral ghrelin acts by directly activating CNS receptors located inside or outside the BBB, or if these actions are mediated peripherally through activation of vagal nervous structures.24 The latter point is of extraordinary interest as several reports state that in rats, vagotomy abolishes ghrelin-induced feeding and GH discharge. This suggests that the gastric vagal nerve is the major afferent pathway conveying ghrelin's signals to the brain.24 Regardless, direct neuronal activation occurs after the activation of the ghrelin receptors, which are located on GHRH and NPY neurons, as well as in additional neurons, as was previously demonstrated for GHRP-6.23 Somatostatin, cortistatin, thyroid hormones and insulin powerfully reduce gastric ghrelin secretion,25,26 while cholecystokinin (CCK) and gastrin stimulate it (Figure 2).27 There is no information on the regulation of ghrelin discharge by hypothalamic neurons.

Ghrelin activates the GH secretagogue receptor called GHSR-1a, a G protein coupled receptor. It activates the phospholipase C signaling route leading to an intracellular Ca2+ rise.9 An active cross-talk at the somatotrope cell is maintained between the GHRH and the ghrelin receptors in order to coordinate and potentiate the ulterior cell response.28 There is an ongoing controversy about whether the cloned secretagogue receptor is truly the receptor or just one of the receptors for that family of compounds. GHSs have specific receptors in a wide range of endocrine and non-endocrine human tissues. Most probably different receptor subtypes exist for GHSs, with different tissue distributions.29

GHRELIN ROLE IN THE REGULATION OF SOMATOTROPE CELL FUNCTION AND GH SECRETION

GH secretion GH secretion in normal subjects after the administration of ghrelin, the GHS hexarelin, and GHRH (all at 1 µg/kg intravenously). Redrawn from: Arvat E, et al. J Clin Endocrinol Metab 2001;86:1169-1174. Reprinted from: Casanueva FF, Dieguez C. Rev Endocrine Metab Dis 2002;3:325-338.

Ghrelin is a potent GH releaser in humans (Figure 3). No side-effects have been reported after the administration of large doses of this compound.30 The potency of ghrelin as measured by its GH releasing capability is higher than for GHRH and comparable to synthesized GHS.9 Thus, for ghrelin to be operative, the normal functioning of the GHRH receptor is necessary, as GHRH antagonists prevent or diminish the GH releasing possibilities of ghrelin.31 Ghrelin is able to release GH in vivo when administered intravenously (IV), as well as when infused directly via the intracerebroventricular (ICV) route; 27 since it is able to enter the CNS from the periphery,22 it is possible that stomach-derived ghrelin may physiologically participate in GH regulation, although this has not yet been demonstrated. An important point is that ghrelin's mechanism of action is route dependent, as the vagus nerve and the arcuate nucleus are in the loop when ghrelin is administered peripherally, but not when administered ICV. 24 Ghrelin-mediated GH secretion is partially insensitive to the inhibitory action of somatostatin and of metabolic compounds such as glucose or free fatty acids.25 Ghrelin and GHRH showed a strong potentiation of their GH secretory capability when injected together in humans. 30 This peculiar activity occurs due to a simultaneous ghrelin activation of pituitary and hypothalamic structures.31 There is some evidence suggesting that hypothalamic ghrelin may participate in the physiological regulation of pulsatile GH secretion.32 Contrasted with the in vitro data, ghrelin in vivo, administered in what were probably pharmacological doses, induced a significant secretion of prolactin and ACTH/cortisol without altering the secretion of LH, FSH or TSH. 9,30 It remains to be determined what happens in respect to these responses when more physiological ghrelin doses and long-term administration are tested.

Ghrelin expression and content Ontogenetic changes in ghrelin expression and content in mice gastric tissue. Redrawn from: Liu JL, LeRoith D. Endocrinology 1999;140:5178-5184.

To show that IV pharmacological doses of ghrelin raise GH levels suggests, but is not proof, that ghrelin participates in the physiologic regulation of GH. A negative point is that rodents with knockout of the GHSR-1a did not show significant alterations in somatic growth, although a compensatory mechanism during fetal development may explain the lack of such results. Inferential evidence favoring a regulatory role for ghrelin, are from one side, the report of a simultaneous increase in GH and ghrelin in states of negative energy balance, and from the other the simultaneous decrease in GH and ghrelin in states of positive energy balance and obesity.9 In the fetus, ghrelin mRNA is undetectable, but starts rising progressively after delivery to reach a peak at 3 weeks post-partum and it decreases thereafter.33 The general pattern of ghrelin changes reminds one of similar patterns of growth rate, and GH and IGF-I secretion. Furthermore, ghrelin mRNA level increases rapidly during the early phase of rapid growth (in the 2-3 first weeks of life), a phase which is GH insensitive,34 and a high level is maintained prior to and during the pubertal growth spurt which is GH sensitive (Figure 4).

In trying to understand the participation of this new hormone in the regulation of the somatotrope axis, it is worth mentioning that adult patients with GH deficiency or GH excess (i.e. acromegaly) have ghrelin levels similar to control subjects.35,36 However, it may be that ghrelin plays a contributing role in the gender based differences in the pattern of GH secretion, as women in the late follicular stage have higher ghrelin levels than men. 36 In addition to its regulatory role on GH secretion, ghrelin has recently been reported to activate pit-1 expression in anterior pituitary cells, an action that appears to be developmentally regulated as it is observed only in infant rats but not in adult rats.37

GHRELIN AND THE REGULATION OF ENERGY HOMEOSTASIS

Ghrelin administration in humans powerfully induces a sensation of hunger in 75% of the subjects tested. 30 In rodents, ghrelin stimulates food intake while reducing fat utilization by a metabolic switch that increases the consumption of carbohydrates.38 Different mechanisms than those involved in GH regulation38-40 control the activity of ghrelin over food intake. Its action seems to be the exact opposite of leptin. Ghrelin is the most powerful appetite stimulant of all the known peptides; it is the unique gastrointestinal peptide that stimulates food intake. All other peptides affecting appetite are anorexigenic. Ghrelin also stimulates food intake in rodents when administered either centrally or peripherally. Other orexigenic peptides are devoid of action with periphery administration. CNS peptides such as NPY, orexin, and agouti-related protein (AGRP) partially mediate the ghrelin action.41,42

Relevant changes in plasma levels of ghrelin appear to endorse the hypothesis that gastric derived circulating ghrelin regulates central appetite mechanisms. For example in rodents, ghrelin mRNA in stomach and ghrelin levels in plasma are increased by fasting and reduced by feeding, actions unrelated to gastric volume changes.38,43 Passive immunoneutralization with ICV ghrelin antibodies inhibited starvation-induced as well as natural food intake in rodents, clearly indicating a tonic ghrelin action at hypothalamic receptors.44 However, as blockade of the vagus nerve inhibits ghrelin-induced feeding in rodents,24 perhaps peripheral ghrelin does not need to cross the BBB to activate central structures. These data do not preclude that the CNS neuronal groups secreting ghrelin may play a role, perhaps one even more relevant in the physiological regulation of appetite.

Ghrelin levels are decreased in obese subjects while elevated in states of malnutrition such as cachexia and anorexia nervosa. In the latter, weight recovery normalizes ghrelin plasma values.45 In respect to the etiology of human obesity, no solid information supports its association with polymorphisms in the ghrelin gene. Circulating ghrelin undergoes relevant changes in relation to food intake, it is elevated before and decreased after feeding in a reciprocal pattern with insulin, and with intermeal changes that are in phase with leptin. 20 Such results suggest that the preprandial ghrelin rise has a role in initiating meal consumption in humans. Interestingly, obese subjects who lose weight show an increase in plasma ghrelin. This fact may explain the facility of obese individuals to recover weight after dieting on the classic low-calorie diets. 46 Patients who have undergone bariatric surgery as treatment for obesity show a reduced ghrelin level, probably due to the absence of direct food stimulation on the gastric fundus(Figure 5). 46 It is a well known fact that bypass bariatric surgery is more effective over the long-term than other techniques, and that patients often refer to an absence of appetite after the surgical intervention.

Circulating ghrelin levels in controls and in obese subjects The action of gastric bypass surgery induces a reduction in ghrelin levels. Adapted with permission 2004 from: Cummings DE , et al. N Engl J Med 2002;346:1623-1630. Copyright © 2002 Massachusetts Medical Society. All rights reserved.

Although they need to be replicated by different groups, the above results open new ways of understanding the regulation of energy homeostasis. Furthermore, the linear correlation in humans between hunger sensation and ghrelin levels, and the supranormal levels of plasma ghrelin in patients with uncontrolled hunger, such in Prader-Willy patients, 47 directly links ghrelin with hunger control.

GHRELIN ACTION ON OTHER HORMONAL SYSTEMS AND NON ENDOCRINE STRUCTURES

Ghrelin also may be involved in the neuroendocrine and behavioral response to stress, and in reducing LH secretion.49 Ghrelin and its functional receptor have been shown in testicular tissue to inhibit testosterone secretion, as well as in both the rat and human ovary, suggesting that ghrelin may be responsible in part for the energy homeostasis associated with control of reproduction. 17,50

Ghrelin mRNA and ghrelin receptor mRNAs are expressed in gastric, thyroid, breast and lung neoplasias. 15,51 This opens potential new routes of treatment. Also recent data suggests that ghrelin may be an endogenous factor to promote sleep.52

In a totally different perspective, a most promising report is that both ghrelin and des-acyl ghrelin inhibit cell death in cardiomyocyte and endothelial cells.53 These data support the protective actions of ghrelin on the cardiovascular system, and possibly more importantly, that there may be biological actions for the deacylated molecule.

SUMMARY AND SPECULATION

As ghrelin anticipates the initiation of meals and releases GH, one could share the teleological view that ghrelin integrates anabolic changes in the body. In catabolic situations, raised ghrelin levels may induce a combination of enhanced food intake, increased gastric emptying and food assimilation coupled with GH levels which promote a prompt nutrient incorporation into muscles and to fat. These actions of ghrelin are the opposite of leptin which reduces food intake and selectively eliminates fat mass. Thus, both peptides may act as physiological regulators of energy balance. Interestingly, each comes from a peripheral organ (stomach and white adipose tissue respectively). Furthermore, with conceptual incorporation of ghrelin into the group of physiological regulators of GH (i.e., GHRH, somatostatin, IGF-I), we may be on the verge of understanding better aspects of the regulation of secretion of GH that previously were not understood.

The clarification of these and other speculations are eagerly awaited. For example, it is not known if ghrelin participates in a physiological way in regulating GH secretion and energy homeostasis. If it does, it needs to be clarified whether stomach-derived circulating ghrelin and/or neuron secreted ghrelin regulate CNS food intake and GH secretion. Similarly, it is unknown whether circulating ghrelin acts after crossing the BBB, or alternatively through an unexpected mechanism related to the structure of the vagus nerve. Finally, the part played by the scattered neuronal systems which secrete ghrelin at both hypothalamic and extrahypothalamic sites have been largely ignored for both food intake and regulation of GH secretion. Such studies will provide better knowledge of the intricate regulation of GH secretion and appetite. It can be foreseen that important new physiological insights and contributions will be provided in the future.

ACKNOWLEDGMENTS

The technical collaboration of Ms. Mary Lage is gratefully acknowledged. The results presented were supported by research grants from the Fondo de Investigación Sanitaria and the Instituto de Salud Carlos III, Ministerio de Sanidad y Consumo, Red de Grupos RGTO (G03/028), Red de Centros RCMN (C03/08), Secretaría Xeral de Investigación e Desenvolvemento (PGIDIT02BTF91801PR), Xunta de Galicia, and the Ministerio Español de Ciencia y Tecnología.

References - (linked to )

1. Casanueva FF, Dieguez C. Trends Endocrinol Metab 1999;0:30-38.
2. Bowers CY. In: Ghigo E, et al (eds). Growth hormone secretagogues. Elsevier North Holland , Amsterdam 1999;5-18.
3. Howard AD, et al. Science 1996;273:974-977.
4. Kojima M, et al. Nature 1999;402:656-660.
5. Dieguez C, Casanueva FF. Eur J Endocrinol 2000;142:413-417.
6. Dornonville de la Cour C, et al. Regul Pept 2001;99:141-150.
7. Cowley MA, et al. Neuron 2003;37:649-661.
8. Gualillo O, et al. Endocrinology 2001;142:788-794.
9. Tanaka K, et al. J Clin Endocrinol Metab 2003;88:2335-2340.
10. Casanueva FF, Dieguez C. Rev Endocrine Metab Dis 2002;3:325-338.
11. Korbonits M, et al. J Clin Endocrinol Metab 2001;86:881-887.
12. Date Y, et al. Diabetes 2002;51:124-129.
13. Tena-Sempere M, et al. Endocrinology 2002;143:717-725.
14. Volante M, et al. J Histochem Cytochem 2002;50:1013-1021.
15. Hattori N, et al. J Clin Endocrinol Metab 2001;86:4284-4291.
16. Wajnrajch MP, et al. J Endocr Genet 2000;1:231-233.
17. Kishimoto M, et al. Biochem Biophys Res Comm 2003;305:186-192.
18. Hosoda H, et al. J Biol Chem 2000;275:21995-22000.
19. Hosoda H, et al. J Biol Chem 2003;278:64-70.
20. Cummings DE , et al. Diabetes 2001;50:1714-1719.
21. Groschl M, et al. Clin Chem 2002;48:1114-1116.
22. Banks WA, et al. J Pharmacol Exp Ther 2002;302:822-827.
23. Hewson AK , Dickson SL. J Neuroendocrinol 2000;12:1047-1049.
24. Date Y, et al. Gastroenterology 2002;123:1120-1128.
25. Broglio F, et al. J Clin Endocrinol Metab 2002;87:4829-4832.
26. Dalkjaer AL, et al. J Clin Endocrinol Metab 2003;88:853-857.
27. Kojima M, et al. Trends Endocrinol Metab 2001;12:118-122.
28. Cunha SR, Mayo KE. Endocrinology 2002;143:4570-4582.
29. Papotti M, et al. J Clin Endocrinol Metab 2000;85:3803-3807.
30. Arvat E, et al. J Clin Endocrinol Metab 2001;86:1169-1174.
31. Popovic V, et al. J Clin Endocrinol Metab 2003;88:3450-3453.
32. Tannenbaum GS, et al. Endocrinology 2003;144:967-974.
33. Liu YL, et al. Mol Cell Endocrinol 2002;189:97-103.
34. Liu JL, LeRoith D. Endocrinology 1999;140:5178-5184.
35. Janssen JA, et al. Eur J Endocrinol 2001;145:711-716.
36. Barkan A, et al. J Clin Endocrinol Metab 2003;88:2180-2184.
37. Garcia A, et al. Mol Endocrinol 2001;15:1484-1495.
38. Tschop M, et al. Nature 2000;407:908-913.
39. Shintani M, et al. Diabetes 2001;50:227-232.
40. Wren AM, et al. Endocrinology 2000;141:4325-4328.
41. Kamegai J, et al. Endocrinology 2000;141:47974-800.
42. Yamanaka A, et al. Neuron 2003;38:701-713.
43. Toshinai K, et al. Biochem Biophys Res Commun 2001;281:1220-1225.
44. Asakawa A, et al. Gastroenterology 2001;120:337-345.
45. Otto B, et al. Eur J Endocrinol 2001;145:669-673.
46. Cummings DE , et al. N Engl J Med 2002;346:1623-1630.
47. Delparigi A, et al. J Clin Endocrinol Metab 2002;87:5461-5464.
48. Asakawa A, et al. Neuroendocrinology 2001;74:143-147.
49. Furuta M, et al. Biochem Biophys Res Commun 2001;288:780-785.
50. Gaytan F, et al. J Clin Endocrinol Metab 2003;88:879-887.
51. Ghe C, et al. Endocrinology 2002;143:484-491.
52. Weikel JC, et al. Am J Physiol 2003;284:E407-E415.
53. Baldanzi G, et al. J Cell Biol 2002;159:1029-1037.