First published online on April 6, 2005
Endocrine Reviews, doi:10.1210/er.2004-0019
Endocrine Reviews 26 (3): 346-360
Copyright © 2005 by The Endocrine Society
Development of Growth Hormone Secretagogues
Roy G. Smith
The Huffington Center on Aging, Department of Molecular and Cellular Biology and Department of Medicine, Baylor College of Medicine, Houston, Texas 77030
Correspondence: Address all correspondence and requests for reprints to: Roy G. Smith, Ph.D., Huffington Center on Aging, Baylor College of Medicine, One Baylor Plaza, Room M320, Houston, Texas 77030-3498.
 |
Abstract
|
|---|
The GH secretagogues (GHS) were developed by reverse pharmacology. The objective was to develop small molecules with pharmacokinetics suitable for once-daily oral administration that would rejuvenate the GH/IGF-I axis. Neither the receptor nor the ligand that controlled pulse amplitude of hormone release was known; therefore, identification of lead structures was based on function. I reasoned that GH pulse amplitude could be increased by four possible mechanisms: 1) increasing GHRH release; 2) amplifying GHRH signaling in somatotrophs of the anterior pituitary gland; 3) reducing somatostatin release; and 4) antagonizing somatostatin receptor signaling. Remarkably, the GHS act through all four mechanisms to reproduce a young adult physiological GH profile in elderly subjects that was accompanied by increased bone mineral density and lean mass, modest improvements in strength, and improved recovery from hip fracture. Furthermore, restoration of thymic function was induced in old mice. The GHS receptor (GHS-R) was subsequently identified by expression cloning and found to be a previously unknown G protein-coupled receptor expressed predominantly in brain, pituitary gland, and pancreas. Reverse pharmacology was completed when the cloned GHS-R was exploited to identify an endogenous agonist (ghrelin) and a partial agonist (adenosine); ghsr-knockout mice studies confirmed that GHS are ghrelin mimetics.
- I. Introduction
- II. Drug Discovery by Reverse Pharmacology
- A. Selection of a model
- B. Identifying leads by selected functional screening assays
- III. Identification of Substituted Benzolactams as the First Nonpeptide GHS
- A. Chemistry
- B. In vivo studies in animal models
- C. Clinical evaluation of L-692,429
- D. Benzolactams with improved potency and oral bioavailability
- IV. Search for New Structural Leads: Privileged Structure Design Leading to Development of MK-0677
- A. Chemistry
- B. Potency and selectivity of L-163,191 in vivo
- C. L-163,255 accelerates recovery from limb immobilization in dogs
- D. L-163,255 stimulates thymic function in old mice
- V. Clinical Efficacy of MK-0677
- A. Potency and selectivity
- B. Efficacy in catabolic states
- C. Efficacy in obese subjects
- D. Treatment of osteoporosis
- E. Treatment of elderly hip fracture patients
- VI. Central Mechanism of GHS Action
- VII. Characterization and Cloning of the GHS-R
- A. Binding studies with [35S]MK-0677
- B. Identification of the GHS-R by expression cloning
- C. Structural studies and ligand binding to the GHS-R
- D. GHS-R is expressed in brain centers besides those regulating GH release
- VIII. Additional GHS
- IX. Identification of GHS-R Endogenous Ligands
- A. Ghrelin and adenosine
- B. Ghrelin and the GHS-R are expressed in hypothalamic centers that regulate energy balance
- X. GHS-R Is Essential for the Orexigenic and GH-Releasing Properties of Ghrelin
- XI. Summary
|
I. Introduction
|
|---|
HORMONES ARE GENERALLY released episodically as peaks of differing amplitude with frequencies dependent upon biological rhythms (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20). The traditional pharmaceutical approach to hormone replacement involves pharmacological treatment with the hormone. However, this results in a hormone profile in the blood that is not episodic and therefore fails to provide replacement that mimics normal physiology. Upon joining Merck Research Laboratories in 1987, I initiated a project designed to replace hormones in a physiological way by normalizing the function of the underlying regulatory feedback pathways. The GH axis was selected as the initial target because the frequency of episodic GH release is highly conserved across species, and a decline in pulse amplitude during aging is well documented (21). Hence, learning how to manipulate GH pulsatility in animal models gave promise of results that should translate to humans.
When I initiated the project, the mechanism by which the amplitude of GH pulsatility is fine-tuned was unknown. However, it was well accepted that GHRH was the hypothalamic hormone that stimulated GH release from the pituitary gland and that overstimulation of GH release was prevented by negative feedback regulation mediated by somatostatin. Hence, it seemed likely that GH pulse amplitude would be increased by stimulating GHRH release from hypothalamic neurons, by amplifying GHRH signaling in somatotrophs of the anterior pituitary gland, and/or by attenuating somatostatin-mediated negative feedback. Because the concentration of somatotrophs in the anterior pituitary gland is higher than the concentration of GHRH neurons in the hypothalamus, and in our hands the former were easier to culture, we focused on identifying small molecules that would amplify GH release from the pituitary gland. Having decided that the function we sought for the new drug was to increase the pulse amplitude of episodic GH release, the next step was to establish methods that would identify a lead molecule.
|
II. Drug Discovery by Reverse Pharmacology
|
|---|
A. Selection of a model
During an extensive literature search to seek clues to the identity of compounds that might have potential in amplifying pulsatile GH release, we discovered the work of Bowers and Momany (22) describing the synthesis of a class of small synthetic peptides based on C-amidated met- and leu-enkephalins. Their studies culminated in the development of a synthetic hexapeptide, His-D-Trp-Ala-Trp-D-Phe-Lys-NH2 (GHRP-6; Fig. 1
, structure 1) that stimulated GH release in vitro and in vivo by an unknown mechanism (23). The small size of this peptide seemed ideal for the design of a peptidomimetic; therefore, my laboratory investigated the mechanism of action of GHRP-6 to determine whether it might be a suitable prototype of a drug capable of increasing the amplitude of episodic GH release.
Upon investigating its mechanism of action, we determined that GHRP-6 appeared to act through a novel receptor. Its activity was not blocked by the opiate receptor antagonist naloxone; furthermore, it was not a GHRH receptor agonist or a somatostatin receptor (sst) antagonist (24, 25, 26). Subsequently, my laboratory showed that GHRP-6 had two very important properties that made it an ideal prototype for the design of small molecules that would increase the amplitude of endogenous GH pulsatility; remarkably, in pituitary cells GHRP-6 amplified the GHRH signal transduction pathway and behaved as a functional antagonist of somatostatin (24, 27, 28).
We sought a drug candidate with high oral bioavailability and pharmacokinetics suitable for once daily administration. Although GHRP-6 itself had properties consistent with an amplifier of GH release, GHRP-6 had poor oral bioavailability (0.3%) and short in vivo half-life (20 min) in humans (29). Furthermore, as a peptide it did not readily lend itself to optimization of pharmacokinetic properties by medicinal chemistry. Hence, GHRP-6 was selected only as a model structure; our objective was to design a nonpeptide mimetic. The probability of identifying a nonpeptide mimetic of GHRH was considered low because native GHRH is a 44-amino acid peptide, and the smallest known homolog to exhibit biological activity was a 29-mer (30, 31).
The GHRP-6 template also seemed ideal because it had been demonstrated that nonpeptide antagonists of small peptides could be designed based on a benzodiazepine template (32). However, an issue to be overcome at this time (1989) was the perceived difficulty of designing nonpeptide agonist mimetics. Fortunately, despite considerable resistance from some quarters, with the help of a few enthusiastic medicinal chemists, our efforts met with early success (33). Indeed, the design of GHRP-6 mimetics was considered a milestone achievement, because once we showed that such nonpeptide mimetics could be made, many more examples were forthcoming. For example, nonpeptide agonists of the peptides cholecystokinin, angiotensin, somatostatin, and melanocortin were developed (34, 35, 36, 37, 38).
B. Identifying leads by selected functional screening assays
We sought lead structures that would amplify the GH-releasing capacity of GHRH on somatotrophs and/or would functionally antagonize somatostatin. However, the conventional approach for identifying lead structures through high-volume screening of chemical libraries was not feasible. Although our model was GHRP-6, the receptor activated by GHRP-6 was unknown. Furthermore, the binding of radiolabeled GHRP-6 did not yield the high-affinity, limited capacity binding to pituitary membranes that was characteristic of a specific GHRP-6 receptor. We were also frustrated in our attempts to identify established cell lines that transduced a signal in response to GHRP-6. It was therefore necessary to screen for function using primary cultures of rat pituitary cells, using GH secretion as the endpoint.
To determine the specificity of GH-releasing lead structures by discriminating between pituitary cell membrane depolarizing agents, GHRH mimetics, and GHRP-6 mimetics, it was necessary to develop secondary and tertiary counterscreens. The counterscreens were based on the use of GHRH and GHRP-6 antagonists, GHRH/GHRP-6 cross- desensitization assays, and characterization of signal transduction pathways. Hence, the selectivity of prototype GHRP-6 nonpeptide mimetic lead structures was determined according to the following: 1) GH-releasing activity on primary rat pituitary cells should be blocked by antagonists of GHRP-6 but not by GHRH antagonists (33); 2) like GHRP-6, but in contrast to GHRH, signal transduction should be mediated by phospholipase C, resulting in production of inositol 1,4,5-triphosphate (IP3) and diacylglycerol (25); 3) lead structures should produce should produce a synergistic effect on both GHRH-stimulated cAMP production and release of GH from pituitary cells (24); 4) activity of lead structure on pituitary cells should be blocked by preincubation with GHRP-6 but not GHRH (26); and 5) a lead structure should behave as a functional antagonist of somatostatin by causing depolarization of somatotrophs (27, 39).
|
III. Identification of Substituted Benzolactams as the First Nonpeptide GHS
|
|---|
A. Chemistry
To model the GHRP-6 structure, potential nonpeptide lead structures were selected by focusing on a benzodiazepine-like template containing aromatic substitutions. Based on structure activity relationships derived from the GHRPs, it was clear that a basic amine at position 1 was critical for GHRP stimulation of GH release. Aromatic amino acids were preferred at positions 2, 4, and 5, and the location of D-Trp at position 2 transformed the original opioid peptide to a GHS (22). These parameters were used to select nonpeptide compounds from the Merck & Co. chemical sample collection for evaluation in the rat pituitary cell assay. From directed screening of approximately 100 compounds, a substituted racemic benzolactam (Fig. 2, structure 2) was identified that over a series of concentrations increased GH secretion from rat pituitary cells to yield an EC50 of 3 µM. An increase in potency was afforded by substituting tetrazole for the carboxylic acid side chain. Synthesis of the R-enantiomer furnished the first potent nonpeptide GHS (L-692,429; Fig. 2, structure 3), which displayed an EC50 of 60 nM. The S-enantiomer was inactive, suggesting that the biological activity of L-692,429 was receptor mediated. The discovery of L-692,429 was considered a seminal breakthrough because it showed that a small molecule peptidomimetic agonist for GHRP-6 could be designed (33).
View larger version (21K):
[in this window]
[in a new window]
|
FIG. 2. Benzolactam nonpeptide GH secretagogue lead structures. (EC50 reflects the concentration of the compound required to stimulate 50% of maximal release of GH from rat pituitary cells.)
|
|
B. In vivo studies in animal models
In vivo potency was generally evaluated in dogs because of the ease of sequential blood sampling and the ability to conduct crossover dose-response studies in single animals. Intravenous administration of L-692,429 (0.1–1 mg/kg) showed dose-dependent stimulation of GH release to a maximum of 90-fold that of basal levels within 30 min after administration (40). After 120 min, GH levels returned to baseline. Besides GH, relatively moderate increases in cortisol and ACTH were observed, but no changes in prolactin, LH, insulin, T3, or T4 were noted.
C. Clinical evaluation of L-692,429
L-692,429 underwent preclinical safety and toxicology studies to test the concept that nonpeptide mimetics of GHRP-6 would be effective in humans. After demonstration of its safety, L-692,429 was administered iv to the targeted clinical population: young adults, elderly adults, obese subjects, and patients treated with glucocorticoids (41). A dose-dependent increase in GH release was observed in all subjects after single iv administration of L-692,429. Elderly subjects, obese individuals, and those dosed with prednisone were less responsive than healthy young adults, but all groups produced a robust response. To evaluate the potential of a long-acting GHS, L-692,429 was infused for 24 h, and serum GH levels were measured every 15 min to assess effects on the amplitude of episodic GH release (42). L-692,429 increased GH-pulse amplitude during the infusion. Based on these encouraging results, the medicinal chemistry effort was increased and focused on developing a compound with improved oral bioavailability and extended half-life.
D. Benzolactams with improved potency and oral bioavailability
In parallel with clinical evaluation of L-692,429, investigations of structure activity relationships around structure 2 (Fig. 2
) continued because, although the oral bioavailability of L-692,429 and serum half-life were superior to GHRP-6, the overall pharmacokinetic properties were not suitable for once daily oral administration. Enhanced potency was demonstrated by N-terminal derivatization with the (R)-2- hydroxypropyl moiety (Fig. 3, structure 4, L-692,585; EC50 = 3 nM). The in vivo potency of L-692,585 was much greater than L-692,429, with a minimum effective iv dose of 0.01 mg/kg. Transient increases in IGF-I were observed 6 h after a single dose, returning to baseline within 24 h (43). L-692,585 was also infused into guinea pigs for 12 h, and GH levels were measured at 10-min intervals (44). Increased pulse amplitude of GH release was observed; interestingly, L-692,585 administration initiated GH pulsing, showing that GHS are capable of resetting the GH self-entraining 3-h pulse frequency.
View larger version (15K):
[in this window]
[in a new window]
|
FIG. 3. Evolution of benzolactam GHS illustrating improved potency (structure 4) and markedly improved oral bioavailability (structure 5). (EC50 reflects the concentration of the compound required to stimulate 50% of maximal release of GH from rat pituitary cells.)
|
|
Although with new designs the in vitro and in vivo potency of the benzolactams continued to improve, optimization of oral bioavailability in dogs proved challenging. On the hypothesis that limited oral bioavailability might be due to zwitterion formation, the tetrazole group was replaced by a carboxamide moiety (Fig. 3, structure 5; EC50 = 3 nM), which significantly improved oral bioavailability and furnished a highly potent drug candidate (45).
|
IV. Search for New Structural Leads: Privileged Structure Design Leading to Development of MK-0677
|
|---|
A. Chemistry
In parallel with developing structure activity relationships for the benzolactams, alternative structural leads were sought. Evans et al. (32) had suggested that a useful approach to designing receptor agonists and antagonists was to derivatize frequently occurring units. These recurring structural units were termed "privileged structures" and had been recognized earlier by Ariens et al. (46) as hydrophobic double ring systems that contributed to receptor binding of many antagonists of biogenic amines.
It was hypothesized by Patchett et al. (47) that derivatizing privileged structures with amino acids and small peptides might furnish agonists or antagonists for peptide receptors. The spiroindanylpiperidine was speculated to be a privileged structure, and one of the first new lead structures that screened positive in the rat pituitary cell GH release assay was structure 6 (Fig. 4). Although this molecule was a mixture of four diastereomers, it had relatively good potency (EC50 = 50 nM) (47). Structure 6 lacked high oral bioavailability, but this was improved in a D-tryptophan analog where the ureidoquinuclidine was replaced by the aminoisobutyric acid moiety (Fig. 4, structure 7; EC50 = 14 nM).
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 4. Spiroindanylpiperidine privileged structure leads; improved oral bioavailability of structure 6 was obtained by replacing the ureidoquinuclidine by aminoisobutyric acid moiety in a D-tryptophan analog (structure 7). (EC50 reflects the concentration of the compound required to stimulate 50% of maximal release of GH from rat pituitary cells.)
|
|
Potency and oral bioavailability were optimized by replacing D-Trp by O-benzyl-D-serine and incorporating a methanesulfonylamido group to furnish L-163,191 (Fig. 5, structure 8; EC50 = 1.3 nM) (47). As a spiropiperidine, L-163,191 represented a new structural class of GHRP-6 mimetics. Although L-163,191 evolved from a privileged structure, it was highly selective even at concentrations above 10 µM (28, 47).
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 5. Optimization of privileged structures focusing on providing high oral bioavailability and prolonged serum half-life (t1/2 in dogs = 4.7 h) produced the spiropiperidine clinical candidate L-163,191, which subsequently entered the clinic as MK-0677 (structure 8). (EC50 reflects the concentration of the compound required to stimulate 50% of maximal release of GH from rat pituitary cells.)
|
|
Piperazine analogs of L-163,191 (48) and substituted spiroindanes (49) were also developed and had excellent activity in the rat pituitary cell assay; two examples of these are shown in Fig. 6 (structure 9, EC50 = 6.3 nM; and structure 10, EC50 = 1.9 nM). Compound 10 had comparable oral bioavailability to L-163,191, but shorter serum half-life in dogs (1.7 h vs. 4.7 h). Subsequent design of high-potency GHS involved replacement of the privileged structure component with 3,3-substituted piperidine (50).
View larger version (20K):
[in this window]
[in a new window]
|
FIG. 6. Potent piperazine GHS (structure 9) and a substituted spiroindane (structure 10) with high oral bioavailability and short serum half-life (t1/2 in dogs = 1.7 h). (EC50 reflects the concentration of the compound required to stimulate 50% of maximal release of GH from rat pituitary cells.)
|
|
B. Potency and selectivity of L-163,191 in vivo
In dogs, L-163,191 proved to have oral bioavailability of more than 60% and released GH reproducibly at an oral dose of 0.125 mg/kg (47). Selectivity was excellent and was similar to that previously observed for GHRPs and the benzolactam GHS. Chronic studies in dogs (once daily for 14 d) showed a significant attenuation of the GH response on d 2. Although attenuated, the increase in GH was maintained from d 2–14 and was characterized by increased amplitude of the pulsatile profile. IGF-I increased by up to 126% and was sustained during treatment with L-163,191; the daily predose IGF-I levels were always higher than in placebo-treated dogs (51). Interestingly, the modest increase in cortisol that was observed after acute dosing was not evident during chronic dosing. By contrast, when dogs were dosed with L-163,191 on alternate days, neither the GH response nor the cortisol response was attenuated, which correlated with a lack of sustained increase in IGF-I (51). Subsequent studies indicated that IGF-I feeds back negatively on GHRH neurons to attenuate GHRH release (51). Having established excellent potency, selectivity, and oral bioavailability, as well as appropriate pharmacokinetics suitable for once daily oral dosing, safety assessment studies were initiated. L-163,191 was found to have an excellent safety profile and entered clinical development as MK-0677.
C. L-163,255 accelerates recovery from limb immobilization in dogs
An analog of MK-0677, L-163,255 (Fig. 7, structure 11; EC50 = 1.5 nM), was used extensively in animal studies to evaluate the potential clinical utility of MK-0677. A study was designed to evaluate whether L-163,255 treatment would provide benefit during and after limb immobilization (52). The right hind-limbs of beagles were pinned for 10 wk, after which the pins were removed to allow the dogs to move freely for 5 wk. Throughout this experiment, the dogs were treated orally with placebo or L-163,255 once daily. IGF-I levels decreased in the placebo group, and the decrease was accompanied by weight loss (1.13 ± 0.19 kg). In the L-163,255 group, IGF-I increased by 60%, and weight loss was marginal (0.21 ± 0.20). Muscle strength in the immobilized limb was compared in each group by isometric torque. After 10 wk of immobilization, an equal decline in strength was observed in both groups. At wk 15, muscle strength had increased by 16% in the placebo group and by 43% in the group treated with L-163,255 (52). The improved strength correlated with the diameter of the vastus lacteralis muscle, indicating that the GHS benefited rehabilitation.
View larger version (27K):
[in this window]
[in a new window]
|
FIG. 7. L-163,255 (structure 11) a close analog of MK-0677 that was used extensively in animal models to validate potential clinical targets for MK-0677. (EC50 reflects the concentration of the compound required to stimulate 50% of maximal release of GH from rat pituitary cells.)
|
|
D. L-163,255 stimulates thymic function in old mice
The effect of the MK-0677 analog on thymic function in old mice was investigated because GH is known to enhance immune responses, either directly or through IGF-I (53, 54, 55, 56, 57). Koo et al. (58) reported beneficial effects of restoring GH levels on the immune system of old mice by administering L-163,255. They showed that chronic treatment was accompanied by partial reversal of the age-related shrinkage of the thymus and of reduced T cell production. The advantage of rejuvenating the GH/IGF-I axis was illustrated by implanting aggressively growing tumors into the mice. Treatment with the GHS reduced the rate of growth and metastases of the tumors and increased longevity of the mice (58). L-163,255 did not inhibit growth of the tumor cells in vitro; therefore, we concluded that its stimulatory effect on the mouse immune system caused inhibition of tumor growth.
|
V. Clinical Efficacy of MK-0677
|
|---|
A. Potency and selectivity
Human studies with acutely administered MK-0677 revealed that GH secretion was stimulated dose dependently with a threshold oral dose of 5 mg. Transient increases in cortisol were also observed, similar to that seen with L-692,429. However, during chronic administration, as had been observed in dogs (59), IGF-I levels increased, resulting in attenuation of the GH response together with an absence of an increase in cortisol (60).
To determine efficacy and specificity in older subjects, elderly men and women were treated orally with placebo or MK-0677 (10- or 25-mg doses) once daily for 14 d (60). Before dosing and again on d 14, GH concentrations were measured in serum at 20-min intervals for 24 h to determine pulse amplitude and frequency of release. On d 14, increased peak amplitude and 24-h GH AUC was observed without changes in pulse frequency; IGF-I was increased 40 and 60% by 10- and 25-mg doses, respectively. Serum samples collected at 20-min intervals before dosing and on d 14 were also assayed for cortisol and prolactin. Cortisol pulse amplitude, frequency, and 24-h AUC were unchanged by either 10- or 25-mg MK-0677 treatment. At the 25-mg dose, prolactin increased by 24%; small dose-dependent increases in mean fasting glucose levels were also noted; however, these changes were within the normal range. It was therefore concluded that MK-0677 had acceptable selectivity.
B. Efficacy in catabolic states
MK-0677 was considered to have potential for treating catabolic states. When a catabolic state is induced by dietary caloric restriction, nitrogen loss is accompanied by a decrease in IGF-I and an increase in GH (20, 61). The increase in GH is due to reduction in IGF-I-mediated negative feedback on GH release, and the decrease in circulating IGF-I is probably caused by reduced sensitivity of IGF-I producing cells in the liver to GH stimulation. However, GH resistance is not complete because treatment with exogenous GH increases IGF-I and promotes nitrogen retention (62, 63).
MK-0677 increases GH and IGF-I, but IGF-I-mediated negative feedback prevents MK-0677 from producing major increases in GH and IGF-I that are possible by administering exogenous GH. However, by increasing the pulse amplitude of endogenous GH release to mimic normal physiology, it was believed that benefits would be achieved by administering MK-0677. To test for efficacy, a once daily oral dose (25 mg) of MK-0677 was administered to healthy young men subjected to short-term diet-induced nitrogen wasting (64). Treatment with MK-0677 for 7 d produced a sustained increase in serum concentrations of GH and IGF-I as well as reversal of nitrogen wasting. Hence, GHS might prove useful for treating wasting associated with HIV and cancer.
C. Efficacy in obese subjects
Obesity is associated with refractoriness to stimulation of GH release by classical stimuli; however, we had shown that acute iv administration of L-692,429 was capable of eliciting a robust GH response in obese subjects (65). Therefore, it was important to determine whether chronic oral treatment with MK-0677 would have sustained effects on the GH/IGF-I axis. Healthy obese males were treated once daily with 25 mg MK-0677 for 8 wk. Increased basal metabolic rate was observed after 2 wk, but not after 4 wk of treatment. Sustained increases in serum levels of GH, IGF-I, and IGF binding protein-3 occurred, as well as a 3-kg gain in weight. Dual energy x-ray absorptiometry scanning revealed that weight gain was explained by an increase in fat-free mass. Neither total nor visceral fat was affected. Although food intake was not measured, the sustained increase in fat-free mass is likely explained by increased appetite and deposition of calories into lean tissue. The finding that MK-0677 increased weight as lean mass, coupled with reversal of nitrogen loss observed in the diet-induced catabolism study described in Section V.B. (64), suggest that long-acting GHS, like MK-0677, would provide benefit to cancer cachexia patients.
D. Treatment of osteoporosis
Because GH is known to stimulate bone turnover and activate osteoblast activity, it was postulated that administration of MK-0677 together with an inhibitor of bone resorption would increase the rate of bone formation. Hence, the combination should provide enhanced benefit in women with postmenopausal osteoporosis. The effects of MK-0677, alone and in combination with the bisphosphonate alendronate, were evaluated during 12 and 18 months of treatment in a multicenter, randomized, double-blind, placebo-controlled study. A total of 292 women (ages, 64–85 yr) with low femoral neck bone mineral density (BMD) participated in the study (66). To evaluate efficacy, serum IGF-I, biochemical markers of bone formation (osteocalcin and bone-specific alkaline phosphatase), and a marker of bone resorption [urinary N-telopeptide cross-links (NTx)], as well as BMD were measured. Women were randomly assigned to one of four daily regimens for 12 months: MK-0677 (25 mg) plus alendronate (10 mg); alendronate (10 mg); MK-677 (25 mg); or double placebo (66). Those patients receiving MK-0677 alone or placebo through month 12 received MK-0677 (25 mg) plus alendronate (10 mg) for months 12–18. Other patients remained on their assigned therapy.
MK-0677, with or without alendronate, increased IGF-I levels by 39 and 45%, respectively. Similarly, osteocalcin and urinary NTx increased by 22 and 41% relative to placebo (P < 0.05). The combination of MK-0677 and alendronate increased bone formation compared with alendronate alone, reduced the effect of alendronate on resorption (NTx), and increased BMD at the femoral neck by 4.2% compared with 2.5% with alendronate alone (P < 0.05). However, the combination did not enhance BMD of the lumbar spine, total hip, or total body beyond that achieved with alendronate alone. Before interpreting these results, it is important to remember that the study was discontinued after 18 months, which is too short a time to see optimal effects of stimulating the GH/IGF-I axis. The beneficial effect of GH treatment on BMD occurs slowly because GH stimulates bone remodeling. Indeed, GH administration to GH-deficient individuals leads first to a reduction followed by an increase in BMD, and the increase is not usually seen until after 2 yr of therapy (67). The observed increase in BMD at the femoral neck after 18 months of treatment with MK-0677 should be considered encouraging because this may translate to prevention of fractures at a sight of frequent occurrence in the elderly.
E. Treatment of elderly hip fracture patients
Based on results derived from studies showing that dogs treated with a close structural analog of MK-0677 recovered more quickly than placebo-treated dogs after immobilization of their hind limb, the potential benefit of MK-0677 treatment on recovery in elderly hip fracture patients was evaluated (68). A placebo-controlled, randomized, double-blind trial was implemented at 13 acute care hospitals and rehabilitation centers in England, Sweden, Denmark, Belgium, Switzerland, Canada, and the United States. A total of 161 hip fracture patients were recruited at 3 to 14 d after surgery or not more than 18 d after fracture. The entry criteria included consenting patients aged 65 and older who were ambulatory before their fracture, medically stable, and mentally competent. Patients with multiple fractures, severe trauma, diabetes mellitus, cancer, uncontrolled hypertension, congestive heart failure, or total hip replacement in the involved extremity were excluded.
Patients were randomly assigned to daily treatment with MK-0677 or placebo for 6 months and then followed for an additional 6 months after completing therapy (69). IGF-I levels were measured, and each patient was evaluated from 6 to 26 wk using a panel of functional performance measures. These evaluations included outcome measures such as changes in the Sickness Impact Profile for Nursing Homes (SIP-NH) and ability to live independently. MK-0677 treatment caused an 84% increase in serum IGF-I. There were no significant changes between MK-0677 and placebo in improving functional performance measures or in the overall SIP-NH score, but MK-0677 treated patients showed greater improvement compared with placebo in three of four lower extremity functional performance measures, in the physical domain of the SIP-NH, and in the ability to live independently. Although it is uncertain whether clinically significant outcomes on physical function were achieved, making clear conclusions from functional studies in clinical trials with hip fracture patients is complicated by the lack of validated outcome measures and absence of a baseline assessment. Indeed, present functional performance measures may not be sufficiently responsive to be used as an endpoint in small intervention studies.
|
VI. Central Mechanism of GHS Action
|
|---|
Besides direct stimulation of the anterior pituitary gland resulting in increased GH release, the GHS also act on the central nervous system (41, 69). It was shown that in common with GHRP-6, both the benzolactam and spiropiperidine GHS increased the activity of GHRH and neuropeptide-Y (NPY) neurons in the arcuate nucleus as well as nonadrenergic neurons in the area postrema (70, 71, 72, 73, 74, 75, 76, 77). A series of studies established that depending on the species, GHS stimulate GHRH release or inhibit somatostatin release from arcuate neurons (28, 78, 79, 80). Hence, GHS-induced amplification of GH pulsatility is explained by induction of GHRH release from the hypothalamus, by synergistic stimulation of GH release from somatotrophs by the joint actions of GHRH and GHS, and by attenuation of hypothalamic somatostatin release and antagonism of somatostatin action on somatotrophs (28).
To investigate the potential of GH-mediated negative feedback on MK-0677 activation of arcuate neurons and the mechanisms involved, we generated mice with the somatostatin receptor subtype 2 (sst2) gene deleted (81). Treatment of wild-type mice with MK-0677 caused activation of c-Fos in the arcuate nucleus. However, pretreatment with GH activated c-Fos in the periventricular nucleus (PeN) but prevented MK-0677-induced activation of c-Fos in arcuate neurons. In sst2 –/– mice, GH pretreatment again increased c-Fos expression in the PeN but failed to inhibit activation of c-Fos by MK-0677. These results are consistent with GH-mediated negative feedback of GHS action being regulated by GH stimulation of somatostatin neurons in the PeN that inhibit activity of arcuate neurons through sst2 (Fig. 8). Intriguingly, hyperstimulation of the GH/IGF-I axis by high doses of GHS is prevented by IGF-I-mediated negative feedback (41). These collective observations, coupled with the demonstration that both the amplitude of GH pulsatility and IGF-I levels in elderly subjects are restored to those of young adults, but not higher, indicate that a preexisting uncharacterized physiological pathway is activated by GHS (41).
View larger version (30K):
[in this window]
[in a new window]
|
FIG. 8. GH-mediated negative feedback on GHS-activated GHRH neurons is regulated by somatostatin (sst) through somatostatin receptor-subtype-2 (sst2) (81 ). GHSR-L, GHS receptor agonist; ArcN, arcuate nucleus; ME, median eminence. [Reproduced with permission from H. Zheng et al., Mol Endocrinol 11:1709–1717, 1997 (81 ). © The Endocrine Society.]
|
|
|
VII. Characterization and Cloning of the GHS-R
|
|---|
A. Binding studies with [35S]MK-0677
The striking results obtained in the clinic with MK-0677 emphasized the need to understand the regulatory mechanisms involved by identifying a receptor for the GHS. Characterization of the MK-0677 receptor was accomplished by substituting 32S with 35S in MK-0677 to produce a radiolabeled molecule with specific activity of 1100 Ci/mmol (82). Incubation of porcine pituitary gland membranes with [35S]MK-0677 showed high-affinity (Kd = 140 pM), limited-capacity binding (Bmax = 8 fmol/mg protein) (83). This concentration of binding sites is extraordinarily low; the concentration is even lower in rat pituitary membranes (2 fmol/mg protein). Binding was also characterized in hypothalamic membranes where the concentration of binding sites was approximately 3-fold higher. Neither GHRH nor somatostatin decreased [35S]MK-0677 binding. The dependence of binding upon Mg2+ and allosteric inhibition of binding by GTP-
S suggested that the unknown GHS-R was a G protein-coupled receptor (GPCR) (83).
To determine whether MK-0677 binding to pituitary membranes was specific to somatotrophs, a biotinylated homolog of MK-0677 (Fig. 9, structure 12), was synthesized to determine localization by fluorescence/immunofluorescence. To investigate whether this bulky analog of MK-0677 was biologically active, it was evaluated in binding and signal transduction assays. Despite its increased bulk, the biotinylated derivative was a potent inhibitor of [35S]MK-0677 binding (IC50 = 0.2 nM) and an efficient stimulator of GH release from rat pituitary cells (EC50 = 2.5 nM) (41). To investigate the cell specificity of binding, cultured rat pituitary cells were incubated (3 min at 37 C) with compound 12 and then treated with avidin-Texas red. GH-containing cells were observed with fluorescein-tagged GH antibodies. Dual fluorophor and confocal microscopy showed that only cells containing GH were labeled with Texas red, indicating that MK-0677 selectively binds to somatotrophs.
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 9. Biotin-substituted MK-0677 (structure 12) used to identify GHS-R-expressing cells in the pituitary gland. (EC50 reflects the concentration of the compound required to stimulate 50% of maximal release of GH from rat pituitary cells.)
|
|
B. Identification of the GHS-R by expression cloning
The extraordinarily low abundance of MK-0677 binding sites in the pituitary gland frustrated efforts to clone the GHS-R. Therefore, an expression-cloning strategy using Xenopus oocytes was adopted based on evidence that MK-0677 binds to a GPCR that couples through phospholipase C (83). A cDNA expression library was prepared from mRNA isolated from porcine pituitary gland. Pools of cRNA prepared from the pituitary library were coinjected into Xenopus oocytes with cRNA encoding G
11 and cRNA encoding aequorin (84). Most importantly, reproducible detection of positive pools by activation of aequorin bioluminescence after addition of MK-0677 was absolutely dependent upon coinjection of G11 cRNA; other G-subunits such as Go, Gq, Ga13, G16, Gi1, and Gi3 given singly or in combination, were unable to reproducibly restore MK-0677 responsiveness. By reducing the complexity of each positive pool, a single clone activated by MK-0677 was identified. Sequencing of the clone showed that the receptor was a new orphan GPCR, which we named the GHS-R. The closest homology of the open reading frame with other GPCRs was to neurotensin (35%) and TRH (29%) receptors (84). Southern blotting indicated that a single highly conserved gene was present in the human, chimpanzee, bovine, rat, mouse, and pufferfish genomes (84, 85). Fluorescence in situ hybridization showed that the human GHS-R mapped to band 3Q26.2 (41).
Cloning of the porcine GHS-R was followed by cloning the human and rat homologs from the respective cDNA libraries. In each case, two mRNA species were identified; one encoded a full-length GPCR with seven transmembrane (TM) domains, and the other lacked TM6 and TM7 (84, 86). The former was designated GHS-R1a and the latter GHS-R1b. The swine cDNA pituitary library also contained a GHS-R1b (84). The nucleotide sequences of the human and swine GHS-R1a and -1b are identical from the methionine translation initiation codon to Leu265 where the GHS-R1b sequence diverges from 1a and is fused to a short conserved reading frame of 24 amino acids followed by a stop codon. Inspection of the genomic sequences showed that GHS-R1a is encoded by two exons. TM1–5 are encoded by exon-1 and TM6 and 7 by exon-2; the intron contains a stop codon explaining the production of GHS-R1b mRNA by alternative processing of pre-mRNA. The human and swine GHS-R1a are 93% identical at the amino acid level, and the rat GHS-R1a is 96% identical to human (86). [35S]MK-0677 binding studies on membranes from cells expressing GHS-R1a and 1b showed that only GHS-R1a bound MK-0677 with high affinity. Similarly, studies designed to measure MK-0677 or GHRP-6 activation of GHS-R1b through aequorin bioluminescence in transfected cells or injected oocytes indicated that only GHS-R1a was an active receptor. The function of GHS-R1b remains to be defined.
C. Structural studies and ligand binding to the GHS-R
Antibodies selectively raised to the extracellular and intracellular domains established the topological orientation of the GHS-R expressed in HEK293 cells (Fig. 10). By molecular modeling and site-directed mutagenesis of the GHS-R, combined with binding and activation data obtained for each mutant with GHS of different structures, the ligand binding pocket was mapped. All of the synthetic GHS share a common binding domain in TM3, which is based on mutation E124Q that eliminates the counter-ion to a shared basic amine present in all the GHS. Confirmation of this essential interaction was demonstrated by rescue of function of the E124Q mutant by modifying MK-0677 through replacement of its side chain -NH2 with -OH (87). Analysis of data generated with other GHS-R mutants revealed contact points in TM2 (D99N), TM5 (M213K), TM6 (H280F), and extracellular loop 1 that were specific for different peptide, benzolactam, and spiroindane GHS. Hence, activation of the GHS-R does not require that the agonist binds to an identical pocket.
View larger version (57K):
[in this window]
[in a new window]
|
FIG. 10. Topography of the GHS-R illustrating extracellular and intracellular peptide sequences used to generate antibodies and key residues for site-directed mutagenesis. [Reproduced with permission from S. D. Feighner et al., Mol Endocrinol 12:137–145, 1998 (87 ). © The Endocrine Society.]
|
|
D. GHS-R is expressed in brain centers besides those regulating GH release
Tissue specificity of GHS-R expression could not be determined by Northern blot analysis; 10 µg of polyA+ RNA isolated from pituitary glands failed to produce a positive signal. The abundance of the GHS-R cDNA in a nonamplified swine pituitary gland library was estimated by PCR to be approximately 1 in 300,000; because of the low concentration of GHS-R transcripts, localization of expression was determined by ribonuclease protection assays and in situ hybridization (84, 88). GHS-R expression was predominant in the anterior pituitary gland and specific regions of the brain. In agreement with electrophysiology studies and c-Fos activation induced by GHS, GHS-R expression is observed in the arcuate nucleus. More interesting is that the GHS-R is also expressed in regions besides those generally considered to be involved in GH release. For example, expression is seen in the anterior hypothalamus, suprachiasmatic nucleus, supraoptic nucleus, ventromedial hypothalamus, dentate gyrus, CA2 and CA3 regions of the hippocampal structures, tuberomamillary nucleus, pars compacta of substantia nigra, ventral tegmental area, dorsal raphe nuclei, and median raphe nuclei (88). Recently, GHS-R expression has also been measured in T cells and in the thymus (89).
The sites of expression of the GHS-R in the brain and thymus could have profound significance to aging. Clinical studies in humans have shown that GHS activation of the GHS-R rejuvenates the GH/IGF-I axis. It therefore seems reasonable to speculate that specific central nervous system functions might also be restored in elderly subjects. The hippocampus is enriched with neurotransmitter systems that affect memory and learning, and the substantia nigra and ventral tegmental areas are centers for the dopaminergic systems in the midbrain that affect motor control and reinforcement behavior. Furthermore, the dorsal and median raphe nuclei are enriched with serotonergic neurons that project to areas implicated in nociception and affective behaviors. GHS-R expression in the thymus and on T cells suggests that in elderly subjects immune function might be restored by treatment with GHS. Indeed, we have already seen that L-163,255 benefited thymic function in old mice (58).
|
VIII. Additional GHS
|
|---|
After disclosure of the peptidomimetics L-692,429 and MK-0677, as well as cloning of the MK-0677 receptor at Merck, extensive medicinal chemistry studies were initiated around GHRP-6. These were the subject of a comprehensive review (90). In more recent developments, additional structures were described (91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101). Hansen et al. (93) described NN703 (Fig. 11, structure 13), which has oral bioavailability of 30% in dogs and a plasma half-life of 4.1 h. NN703 dose-dependently increases GH release in both swine and dogs, has a high degree of GH specificity, and significantly increases weight gain. Subsequently, NN703 was shown to have oral activity in humans, but with markedly reduced potency compared with MK-0677 (98). CP-424,391 (Fig. 11, structure 14) is a novel, orally active pyrazolidinone-piperidine GHS that shows excellent in vitro and in vivo potency (101). Chronic oral administration to dogs induced dose-dependent increases in GH, IGF-I, and weight gain. CP-424,391 was subsequently evaluated in humans for treating clinical conditions that could benefit from augmentation of GH and IGF-I levels.
Nagamine et al. (96) described oxindole derivatives as a distinct new structural class of orally active GHS (Fig. 11, structure 15). In the rat pituitary cell GH release assay the oxoindole SM-130686 stimulates GH release with an EC50 of 6.3 nM. Despite its structural difference, SM-130686 is a potent inhibitor of [35S]MK-0677 binding to the GHS-R (IC50 = 1.2 nM). Twice daily oral administration to rats for 9 d produces significant increases in GH, IGF-I, body weight, and fat-free mass. The increase in fat-free mass in rats is remarkable, and evidence was presented suggesting that SM13086 behaves as a partial agonist.
|
IX. Identification of GHS-R Endogenous Ligands
|
|---|
A. Ghrelin and adenosine
Cloning of the GHS-R allowed the engineering of cell lines stably expressing the GHS-R, which were essential for identification of endogenous GHS-R ligands. Two endogenous ligands were identified in fractionated tissue extracts; ghrelin was found in stomach extracts (102) and adenosine in hypothalamic extracts (103, 104). In contrast to ghrelin and the synthetic GHS-R agonists, adenosine failed to stimulate GH release from pituitary cells.
The GHS-R signal transduction pathways activated by ghrelin and adenosine are distinct (105). Ghrelin is a full agonist of the GHS-R and triggers intracellular second messengers coupled to a heterotrimeric G protein complex involving G11, which results in activation of phospholipase C signaling. Adenosine is a partial agonist of the GHSR-1a, acting through a binding pocket distinct from that of ghrelin (104). Analysis of the pathways involved in the regulation of GHS-R signaling showed that adenosine, in a dose-dependent manner, induces calcium mobilization from IP3-sensitive intracellular stores, but does not affect the formation of inositol phosphates. The calcium-mobilizing activity is blocked when the GHS-R-expressing cells are preincubated with cholera toxin, with MDL-12,330A, an inhibitor of adenylate cyclase, and with the protein kinase A blocker H-89. Adenosine also stimulated cAMP production in GHS-R- expressing cells. Based on these data, it was proposed that adenosine activates a GHS-R signaling pathway involving adenylate cyclase and protein kinase A resulting in phosphorylation of the IP3 receptor. Hence, depending on the agonist, the GHS-R is capable of activating different intracellular second-messenger systems. However, the synthetic ghrelin mimetics like MK-0677 and GHRP-6 share the same signal transduction pathway as ghrelin. Ghrelin, as a closer biochemical and biological mimic of the synthetic GHS-R ligands, became the focus of subsequent research.
The production of ghrelin in the stomach focused attention on ghrelin’s potential role in obesity. Plasma ghrelin levels are influenced by nutritional status and are believed to regulate GH, appetite, and fat deposition (106, 107, 108, 109, 110). Intriguingly, low circulating ghrelin levels correlate with sustained weight loss and reduced appetite in obese humans after gastric bypass surgery (111). However, causality remains to be determined; neither ghrelin- nor ghsr- knockout mice have a lean phenotype, and neither is resistant to diet-induced obesity (112, 113). The association with obesity in humans was surprising because administration of the synthetic GHS-R ligand, MK-0677, to obese subjects increased lean mass rather than fat deposition (114).
B. Ghrelin and the GHS-R are expressed in hypothalamic centers that regulate energy balance
We generated ghrelin-knockout mice and in collaboration with Cowley and Horvath addressed the question of whether ghrelin was expressed in hypothalamic areas involved in regulating energy balance (112, 115). Intriguingly, immunohistochemistry showed that ghrelin production was localized to a previously uncharacterized group of neurons adjacent to the third ventricle between the dorsal, ventral, paraventricular, and arcuate hypothalamic nuclei (Fig. 12) (115). Notably, this unique distribution does not overlap with known hypothalamic cell populations.
View larger version (36K):
[in this window]
[in a new window]
|
FIG. 12. Illustration of the interaction between ghrelin and hypothalamic circuits (115 ). GABA, -Aminobutyric acid; GHS-R, GH secretagogue receptor (ghrelin receptor). [Reproduced with permission from M. A. Cowley et al., Neuron 37:649–661, 2003 (115 ), with permission from Elsevier.]
|
|
The results of these studies are consistent with ghrelin-expressing neurons sending efferents onto key hypothalamic circuits that include those producing NPY, agouti-related protein (AGRP), proopiomelanocortin (POMC), and CRH (115). Binding studies with biotinylated ghrelin showed localization of receptors in the arcuate nucleus, lateral anterior hypothalamus, and paraventricular nucleus that are mainly associated with presynaptic boutons. Axon terminals that bound ghrelin were frequently found to contain NPY. Electrophysiological recordings from neurons showed that ghrelin stimulated the activity of arcuate NPY neurons and indeed mimicked the effect of NPY in the paraventricular nucleus of the hypothalamus. We proposed that release of ghrelin stimulates the release of orexigenic peptides and neurotransmitters, establishing a novel regulatory circuit controlling energy homeostasis (Fig. 12). The binding data showing localization of expression of ghrelin in axons adjacent to presynaptic nerve terminals, and results of electrophysiology studies support the notion that ghrelin modulates neurotransmission. These experiments describe an anatomical basis for pre- and postsynaptic interactions between ghrelin and NPY/AGRP, POMC, and CRH circuits (Fig. 12). In collaboration with Chen et al. (116), we confirmed the involvement of NPY/AGRP neurons by showing that like Ghsr-knockout mice, agrp/npy-double knockout mice are insensitive to the orexigenic effects of ghrelin.
The observation of ghrelin-secreting neurons in the hypothalamus was controversial because many laboratories had failed to demonstrate ghrelin mRNA expression in brain sections by in situ hybridization. We addressed this issue by performing RT-PCR on RNA isolated from the brains of wild-type and ghrelin –/– mice using two sets of primers: one directed to amplify ghrelin transcripts; the other designed to amplify lacZ transcripts expressed in the disrupted ghrelin locus of the knockout mice. In wild-type mice, a weak signal corresponding to ghrelin transcripts was produced by the ghrelin primers, but not by lacZ primers. In contrast, RT-PCR on RNA from the brains of ghrelin –/– mice failed to produce a signal with ghrelin primers but produced a weak signal with lacZ primers. For comparison, the expression of ghrelin mRNA in the brain is at least 100-fold lower than in the stomach. We concluded that ghrelin is expressed at very low levels in the brain, but that mRNA abundance is too low to be detectable by in situ hybridization.
|
X. GHS-R Is Essential for the Orexigenic and GH-Releasing Properties of Ghrelin
|
|---|
Having discovered ghrelin using cells engineered to express the GHS-R, it was assumed that the GHS-R was the physiologically relevant receptor for ghrelin. However, this assumption was based on indirect evidence, and suggestions had been made for existence of unknown ghrelin receptor subtypes (117). Like the synthetic ligands, ghrelin stimulates GH release and appetite. However, as a 28-amino acid peptide containing a unique octanoyl modification (102), ghrelin is structurally different from MK-0677 and the synthetic peptide agonists used to characterize and clone the GHS-R. Furthermore, molecular modeling studies comparing structural features assigned from proton nuclear magnetic resonance spectroscopy of GHS-R ligands illustrate similarities, but these are not predictive of the known binding characteristics of the GHS (118). Accordingly, direct evidence was needed before the GHS-R could be named the ghrelin receptor.
To determine directly whether the ghsr encoded the ghrelin receptor that mediated ghrelin’s in vivo orexigenic and GH-releasing properties (113, 115), we generated ghsr-knockout mice to investigate their phenotype and to determine their response to a ghrelin challenge. The ghsr –/– and wild-type mice are visibly indistinguishable in physical appearance and activity; therefore, it seems unlikely that ghrelin plays a dominant role in determining growth and body composition. This conclusion is subject to the caveat that an alternative pathway(s) might be compensating for the lack of ghrelin signaling. However, IGF-I levels are slightly lower and body weight is modestly reduced in ghsr –/– mice, which supports our hypothesis that the GHS-R is a subtle enhancer of function (41). In contrast to wild-type mice, when ghsr –/– mice were treated with ghrelin, ghrelin stimulated neither appetite nor GH release. These results showed unambiguously that the GHS-R is the ghrelin receptor that: 1) regulates the activity of GHRH neurons and GH release (113); and 2) maintains normal IGF levels during aging of young adults (113).
The characteristics of ghsr-null mice added credence to our initial hypothesis that ghrelin mimetics were modest amplifiers of biological function (41). We had speculated, based on evidence from electrophysiology studies (77), that an endogenous GHS-R agonist regulates function by controlling the "gain" or "setpoint" of hypothalamic neurons. During aging, the gain appears to be reduced, which correlates with a decline in ghrelin levels (119, 120). Evidently, the gain can be restored by administering GHS-R agonists exogenously (121). Indeed, chronic treatment with GHS does not produce major increases in GH and IGF-I levels in young animals; nevertheless, in old rodents and elderly humans, chronic treatment is sufficient to restore GH pulse amplitude and IGF-I to levels seen in young adults (58, 60). Restoration of the GH/IGF-I axis in old mice increases the cellularity of the thymus, inhibits tumor growth and metastases, and increases longevity (58). Therefore, modest effects on neuronal activity translate to significant functional benefits on overall physiology.
|
XI. Summary
|
|---|
The reverse pharmacology approach for drug discovery is a unique example of the process of "bench to bedside" that was based on first establishing what function the drug should have and then understanding mechanistically how that function could be implemented by a small molecule. In this case I reasoned that amplification of endogenous GH pulsatility could be established by amplifying GHRH action and/or attenuating the negative feedback properties of somatostatin. Remarkably, the resulting research led to discovery of a small molecule that had both of these properties, as well as having two additional desirable properties: first, the property of stimulating GHRH release by modifying the membrane potential of GHRH neurons; and second, that of antagonizing somatostatin release from the hypothalamus.
To complete the circle of reverse pharmacology, after a potent amplifier of pulsatile GH release was identified and tested in the clinic, we used this molecule to characterize and clone the receptor involved (GHS-R). The GHS-R was shown to be highly specific for GHS and is a new orphan GPCR that had little homology to any known GPCRs. To close the loop, endogenous GHS were sought, which was made possible by the availability of GHS-R cDNA clones. Using cells engineered to stably express the GHS-R, tissue extracts were screened and fractionated, which led to the isolation of the new hormone, ghrelin, in stomach extracts by Kojima and co-workers (102), and identification by my laboratory of adenosine as a partial agonist in hypothalamic extracts.
|
Footnotes
|
|---|
First Published Online April 6, 2005
Abbreviations: AGRP, Agouti-related protein; BMD, bone mineral density; GHRP
Top
Abstract
I. Introduction
