I. INTRODUCTION

Top Previous Next References

The secretion of somatotrophin, or growth hormone (GH), like other anterior pituitary (AP) hormones, is regulated through a complex neuroendocrine control system that comprises two main hypothalamic regulators, GH-releasing hormone (GHRH) and somatostatin (SRIH or SS), exerting stimulatory and inhibitory influences, respectively, on the somatotrope cell. Both GHRH and SS, which are subject to modulation by other hypothalamic peptides and by complex networks of neurotransmitter neurons, are the final mediators of metabolic, endocrine, neural, and immune influences for GH secretion in the pituitary.

In the last decade, the rather static scenario of GH control has been shaken by a host of advances, some resulting from molecular biology techniques. Of particular interest is cloning of the GHRH receptor isoforms in the rat, mouse, and human pituitary; discovery of distinct genes encoding a family of structurally related G protein-coupled SS receptors in brain and the pituitary and synthesis of new SS analogs with subtype specificity; identification of a novel class of fairly specific, orally active, hexa-hepta peptides, the GH-releasing peptides (GHRP) and their congeners, whose GH-releasing activity is greater than GHRH; better insight into the neurotransmitter control of GH; and the involvement of somatotropic dysfunction in a host of diseases and in the catabolic processes related to aging.

This review updates and focuses on aspects of the neural control of GH secretion, especially in the light of these advances. For comprehensive surveys of GH secretory control, the reader is referred to Frohman et al. (377), Müller (749), and Cronin and Thorner (274).

	II. GROWTH HORMONE SECRETION PATTERN

Top Previous Next References

The secretion of GH is pulsatile in all species studied so far, with its pattern not dissimilar in humans and male rodents. In the former, an ultrasensitive immunoradiometric assay (IRMA) (limit of detection, 20 ng/l) given to healthy men and women showed an ultradian rhythm with an interpulse frequency of ~2 h (1101), which compared favorably with an interpulse frequency of ~3 h in the male rat (1014) and plasma GH detectable at all time points. In these subjects, GH levels ranged over three orders of magnitude in both sexes.

Women of reproductive age had significantly higher overall GH levels, a higher pulse amplitude, and a higher baseline, but the pulse frequency was the same as in men (mean secretion rates: 327 µg/24 h for men and 392 µg/24 h for women; Ref. 1101). These figures are ~50% lower than previously assumed (see Ref. 1030). A new ultrasensitive GH chemiluminescence (CL) assay (limit of detection, 5 ng/l) detects serum GH concentrations in all samples collected at 10-min intervals for 24 h.

The Cluster method to estimate GH pulsatility and deconvolution analysis, which simultaneously resolves the endogenous hormone secretory rate and clearance kinetics (1063), showed that pulsatile GH secretion continued in all subjects in the awake and fed state, despite undetectable serum GH in daytime samples analyzed by a conventional IRMA, run for comparison. Interestingly, the CL assay also revealed for the first time a low rate of basal (constitutive) GH secretion mixed with pulsatile and nyctohemeral variations (508).

Further studies were then done in healthy lean and obese subjects aimed at ascertaining how age, steroid sex hormones, and obesity, singly or jointly, influence the basal and pulsatile modes of GH release. Basal and pulsatile GH release were found to be regulated differently by adiposity and steroid hormones. Thus basal GH secretion rates were negatively correlated with the circulating estrogen (E₂) concentration and with an interactive effect of age and body fat, whereas the GH secretory burst mass was strongly positively related to serum testosterone concentrations.

Concerning the biological consequences of GH secretion, basal and pulsatile GH release may influence serum insulin-like growth factor I (IGF-I), its main binding protein (BP), i.e., IGFBP-3, and IGFBP-1 in distinct ways (1064). In a study in healthy men, IGFBP-3 correlated negatively to basal GH secretion, whereas IGFBP-1 concentrations were positively correlated with this measure. Measurements of serum IGF-I concentrations indicated that pulsatile GH release was strongly positively correlated with the mean serum IGF-I level. Conversely, there was no association between basal GH release rates and serum IGF-I levels (1064). This observation is consistent with the theory that a pulsatile GH signal is important in acting on one or more major target tissues (e.g., the liver) to stimulate IGF-I production in the healthy human. These findings have major implications for the evaluation of GH control and secretion in different experimental and clinical conditions.

Present knowledge thus indicates that the secretory pattern of GH depends on the interaction between GHRH and SS at the somatotroph level, although we cannot overlook the existence of anatomic connections and functional peptide interactions within the hypothalamus (see sect. IIIA3) and the contribution of a still elusive endogenous GHRP system (see sect. IV).

The basis of the pulsatile release of GH, however, remains unknown. A host of nutritional, body composition, metabolic, and age-related sex steroid mechanisms, adrenal glucocorticoids, thyroid hormones, and renal and hepatic functions all govern pulsatile release in adults (see sect. II, A-C), but spontaneous elevations in GH secretion are more likely to result from the ability of hypothalamic neuroendocrine cells themselves to promote and sustain phasic bursts of action potentials (34). Basal GH release may also be subject to neurophysiological regulation; slow removal of a pool of GH in plasma from trapping with a GH-binding protein seems less likely (508).

Hypothalamic GHRH and SS functions both appear essential for pulsatile secretion. In male rats, anti-GHRH serum inhibited spontaneous GH pulses (1090), reduced growth rate (207), and suppressed stimulated GH secretion (207, 548). Initial studies in male rats utilizing an anti-SS serum bolus injection showed that pulsatile GH secretion was preserved despite an increase in basal GH levels (350). However, when a highly potent anti-SS serum was infused for 6 h, it did inhibit pulsatile GH secretion (375). The most likely explanation for the different effects of anti-SS serum bolus injection versus continuous infusion is that the latter but not the former penetrates the region of the arcuate nucleus (ARC), where the inhibitory effects of SS on GHRH secretion occur (see sect. IIIA3).

Supporting this view, electrolytic lesions in the medial preoptic area (MPOA) or anterolateral hypothalamic deafferentiation in male rats, two techniques which eliminate SS immunostaining in the median eminence (ME), resulted in elevated plasma GH levels and no pulsatile GH secretion. Under these experimental conditions, basal and K⁺-stimulated GHRH release from the hypothalamic ME-ARC complex in vitro was significantly increased, although the hypothalamic GHRH content was decreased, suggesting an increased turnover of the peptide (548).

Overall, these results suggest that SS in the hypothalamus plays a tonic inhibitory role in the regulation of GHRH release and that GH hypersecretion in lesioned rats is due to a combination of secretion changes of both SS and GHRH. They are in line with the general model proposed by Tannenbaum and Ling (1013), in which GHRH release during troughs in a sinusoidal pattern of SS release induces the episodic release of GH, and the rise in SS suppresses baseline release. Direct measurement of GHRH and SS in the rat portal blood would validate this view by showing that each GH secretion episode started by pulsatile release of GHRH, which is preceded by or concurrent with a moderate reduction of the SS tone (853). These portal blood measurements in the rat, however, would require confirmation, since other studies have not been able to maintain GH pulsatility under the conditions reported by Plotsky and Vale (853) (see Ref. 895 for references). Moreover, this model would explain patterns of secretion in the male rat. Female rats show much more continuous irregular GH secretion and much more regular responses to GHRH (248, see also sect. IIA).

The proposed model of SS-GHRH interactions cannot be extrapolated to all the animal species so far investigated. In the sheep, portal GHRH and SS levels suggest regulation is more complex, although the majority of GH pulses coincided with or occurred immediately after a GHRH pulse, ~30% were completely dissociated (378). This variability was also seen in SS measurements, in the sheep, where pulses were not necessarily asynchronous with those of GHRH and did not correlate with GH troughs, nor SS troughs with GH pulses. In addition, active immunization of rams against SS did not change basal and pulsatile GH secretion, whereas in the anti-GHRH group, plasma GH levels were very low, and the amplitude of GH pulses was strikingly reduced (652). The redundant GHRH and SS pulsatility in the portal blood may reflect neuropeptide actions beyond pure GH release, e.g., trophic effects on the target somatotrophs (895).

It must be recalled, in addition, that the simple rat GHRH-SS model ignores all other alleged hypophysiotrophic factors, including the endogenous GHRP system (see sect. IV), additional modulators such as pituitary GHRH and SS receptors, feedback effects of GH and IGF-I, and other target gland hormones.

In humans, the neuroendocrine control of pulsatile GH secretion is also poorly understood. Because portal blood cannot be sampled directly in humans, most of our understanding of the regulation of GH secretion comes from either indirect human studies or extrapolation of animal data. In healthy young men, a single dose of a competitive GHRH antagonist blocked 75% of nocturnal GH pulsatility (521), supporting the hypothesis that pulsatile GH secretion in humans is driven by GHRH. The persistence of pulsatile GH secretion in normal and GH-deficient subjects during continuous GHRH infusion (473, 1056) and the ability of exogenous GHRH to release GH in the face of grossly elevated GHRH levels in the ectopic GHRH syndrome (74) are also in keeping with this suggestion.

Periodic falls in SS secretion might account for the persistence of pulsatile GH secretion in the above conditions, a proposal in line with GH secretory profiles determined by deconvolution analysis (460). Plasma GH secretory patterns were measured in normal adult men at baseline and during submaximal intravenous boluses of GHRH every 2 h for 6 days (522). No evidence was found of either acute or chronic desensitization; GH responses were seen at every GHRH bolus, and circadian patterns of GH secretion were similar at baseline and during the GHRH schedule, with maxima occurring during the early nightime hours. This pattern was consistent with a model for the ultradian secretion of GH in which circadian SS secretion (reduction at night?) is superimposed on more rapid GHRH pulses.

Sex-related differences in GH control by GHRH and SS have been found particularly in the rat. They include different 1) responsiveness to exogenously administered GHRH (248, 1085), 2) endogenous content of hypothalamic immunoreactive (IR) GHRH and SS (386), 3) hypothalamic mRNA levels of SS (238, 1044) and GHRH (40, 653), 4) pituitary GHRH receptor expression (799), and 5) GH autofeedback mechanism(s) (169, see also sects. IIIA8B and VII). Particularly noteworthy was the higher SS mRNA content in the periventricular nucleus (PeVN) of male rats than in proestrus females and the finding that in castrated males administration of testosterone prevented the significant reduction in SS mRNA after surgery (238).

It can be inferred from these results that were the amount of SS available for release from the ME, sex differences in GH secretory patterns (see also below) and perhaps also sex differences in growth rates might be partially attributable to a difference in the ability of PeVN SS neurons to produce the neuropeptide.

In male and female rats, passive immunization with specific antisera to both peptides was used to assess the physiological roles of GHRH and SS in the dimorphic secretion (802). An acute dose of an anti-SS serum to males raised the GH nadir, as expected, but did not alter other parameters of GH secretion, whereas in females it markedly increased plasma GH levels at all time points of the 6-h sampling period and there was, in addition, a significant increase in GH peak amplitude, GH nadir, and overall mean 6-h plasma GH levels. Anti-GHRH serum had no effect on the GH nadir in males but markedly raised the GH nadir in females.

These findings led to the conclusion that the secretory pattern of SS undoubtedly plays an important role in this sexual dimorphism, the female pattern of SS secretion possibly being continuous rather than cyclical, its level intermediate between the peak and troughs of SS release suggested for males. The episodic GHRH bursting, which does not appear to follow a specific rhythm, as in the male, superimposed on the fairly continous SS release, would cause the erratic GH secretory profile, present in the peripheral plasma of female rats.

This pattern of low-amplitude GH release is reminiscent of that after continuous GH infusion and may be more effective in inducing specific biological changes, such as GH-binding protein in plasma and hepatic cytochrome P-450-containing enzymes (955). For the role of sex steroids in imprinting the sexually differentiated GH secretory pattern, see also section IIIA8B.

The pattern of GH secretion in rodents seems to be of considerable importance for optimal growth. In female rats, GHRH pulses delivered every 3 h changed the pattern of pituitary GH release toward that of the male and stimulated body growth, whereas continuous infusions of GHRH were much less effective (248). Interestingly, a similar pulsatile GH secretory pattern was achieved with the intermittent infusion of SS, which also stimulated body weight gain and increased pituitary GH stores (250). This strongly supports the proposal (246, 249) that in the rat the pulsatile pattern of GH is optimal for growth, irrespective of whether this pattern is achieved by a stimulator or inhibitor of endogenous GH secretion. The pulsatile delivery of SS would be vital in the generation of GH pulsatility not only for its opposing action to GHRH but also because it enables GHRH to be effective, because unopposed exposure to the releasing hormone leads to downregulation of GH responses.

Sex differences in the pattern of GH secretion are not so clear-cut in humans, but there are noticeable quantitative differences, with serum GH concentrations being definitely higher in women than in men (221, 474, 1101). In a small group of healthy middle-aged men and premenopausal women, highly sensitive immunofluorometric assays (sensitivity 0.011 µg/l) and blood sampling at 10-min intervals showed the 24-h GH secretion was three times higher in the women, the GH secretory episodes were larger and more prolonged, more GH was secreted per burst, and maximal plasma GH concentrations were higher than in men. The frequency of pulses, timing during the day/night estimates of baseline GH secretion, and the endogenous clearance rates for GH were the same in both sexes (1059) (Fig. 1).

View larger version (20K): [in this window] [in a new window]   Fig. 1. Profiles of 24-h serum growth hormone (GH) in human female (top panels) and male (bottom panels) subjects. Left panels indicate measured serum GH concentrations over time and curve fitted by deconvolution analysis. Right panels depict GH secretory rate calculated from this analysis. [From van den Bergh et al. (1059); Copyright The Endocrine Society.]

A sex-related difference was also found in the serial regularity of GH release in a tissue series (848). Application of novel regularity statistics (approximate entropy) showed that the serial regularity of GH release was lower in female humans and rats (847).

Although E₂ is an obvious candidate for the mechanism that increases the mass of GH secreted per burst in women, it is not yet known at what level(s) in the GH axis it may act. Another question is whether the much higher daily GH output in women has biological significance for GH action, and what action this might have in middle age. In GH-deficient men and women, small but frequent intravenous boluses and continuous intravenous infusion of GH were equally effective in increasing circulating IGF-I (532), which implies that women do not secrete more GH than men simply to maintain the same IGF-I levels.

The 24-h pattern of spontaneous GH release changes with age in experimental animals and humans. Plasma GH in the fetus is elevated compared with postnatal levels in a number of species, including humans and sheep (431). The rat, an altricial species, which is an animal model for the midgestational hypothalamic differentiation in humans, has a rise in plasma GH immediately before birth and a decline to near adult levels by ~18-20 days postnatally (1078). This is coupled with striking changes in pituitary susceptibility to GHRH, which at doses of 1-50 ng/100 g given subcutaneously in 10-day-old pups induces clear-cut plasma GH rises. However, in 25-day-old pups, the dose of 500 ng/100 g given subcutaneoulsy has this effect but not lower doses (205) (see also sect. IIIA9). Moreover, in rat pups but not in young adult rats, an in vivo and ex vivo 5-day GHRH treatment stimulated pituitary GH biosynthesis (205, 272). Consonant with these results, in monolayer cultures of rat pituitary cells, the stimulatory effects of GHRH and dibutyryl cAMP on GH secretion were inversely related to age (281).

Although differential developmental regulation of pituitary GHRH receptor gene expression is very likely involved in these events (572), maturational changes in the somatotroph responsiveness to SS may contribute to the pattern of circulating GH levels characteristic of early development. In fact, pituitaries from newborn rats are relatively resistant to the GH-suppressive effect of SS, and sensitivity increases with age (281, 889), although in rats brain SS mRNA is detectable by day 7 of embryonic life (1130a) and IR-SS has been detected in the fetal rat as early as 18.5 days of gestation (1130a).

In the human fetus, plasma GH peaks at mid-gestation, then GH levels fall in the third trimester and into the neonatal period (540). Excessive GHRH or deficient SS release from and function in the fetal hypothalamus, excessive somatotroph responsiveness to GHRH, deficient negative feedback control by IGF-I, or stimulation of GH secretion by paracrine factors may account for GH hypersecretion at mid-gestation.

Functional hypersomatotropism is still present in small for gestational age (SGA) newborn twins as revealed by higher baseline GH levels and stronger GH responses to GHRH than in appropriate gestational age (AGA) twins (302, 633). Interestingly, in these studies mean serum IGF-I was higher in SGA than in the AGA newborns, which would rule out any underlying effect of a defective IGF-I feedback mechanism or functional GH resistance (see also sect. VI). Growth hormone-releasing hormone appears in the rat and human hypothalamus by 18-20 days and 18-22 wk of gestation, respectively (292, 514, 748), and at late gestation and mid-gestation, respectively, rat and human fetal pituitary cells respond in vitro to GHRH or SS by increasing or reducing GH secretion (559, 748). Miller et al. (725) studied changes occurring after birth in pulsatile GH release of male and female term infants. There were significant differences between infants studied at ~28 h and later at ~80 h postnatally. Spontaneous pulsatile release was present in all infants; within the first 1-2 days of life, GH release was enhanced through the amplitude and frequency of pulses, then as the infant matured, there was a decrease in GH frequency, amplitude, and nadir. The decrease in GH levels in older infants very likely resulted from decreased pituitary secretion, secondary to enhanced IGF-I hypothalamo-pituitary feedback or, as shown in the postnatal rat, to changes in the pituitary sensitivity to SS or GHRH.

Considering that rats are born at the equivalent of 100 gestational days in the human (559), the ontogeny of neuroendocrine control and GH secretion in the human is chronologically consistent with that in the rat.

In initial studies of the 24-h pattern of spontaneous GH secretion, a small sample of prepubertal children lacked waking GH peaks and secreted GH only during sleep (356). This observation, however, contrasted with many subsequent reports of GH peaks during waking hours as well as during sleep (726, 978). Deconvolution analysis depicted the pattern of GH secretory events subserving the changes in serum GH concentrations in normal boys at various stages of puberty and young adulthood (660). Late pubertal boys secreted more GH per 24 h than boys in any other pubertal group and triple that of prepubertal boys (1,800 vs. 600 µg/24 h). After normalization of GH secretion for body surface area, late pubertal boys still had higher GH release than any other pubertal group. The primary neuroendocrine alteration responsible for the enhanced GH secretion was an increased mass of GH released per secretory burst, with no change in GH burst frequency, burst duration, or serum GH half-life.

In girls, GH responses to GHRH did not change throughout pubertal development, and the responses in boys at midpuberty were somewhat lower than either prepubertal or adult men (395). It cannot be excluded, however, that periodic release of SS might have obscured the genuine GH response to GHRH at puberty.

In young adults, total GH secretion was almost one-half that in late pubertal boys (900 µg/24 h), but the 24-h burst frequency was the same. This figure compares favorably with reports based on other techniques (574, 694). However, absolute GH secretion rates estimated by deconvolution analysis are dependent on the specific assay system, and the values reported here tend to overestimate GH secretion compared with the more recent and sensitive assays. Studies in which a large number of prepubertal and pubertal girls and boys were evaluated for GH secretion using the Pulsar program (713) found no correlation between GH secretion and age in the prepubertal children, supporting the proposal that gonadal hormones are not important for the regulation of growth and GH secretion during childhood, whereas at puberty there was a marked increase in GH secretion rates in both sexes (23). However, the change was sex specific, since the increase in GH secretion rate was an early event (stages 2-4) in puberty for girls but a late event (stages 3-4) in puberty for boys, parallel to the height velocity curves of girls and boys in puberty (1019). Shortly after cessation of linear growth in boys, the overall peripheral GH pulse pattern returned toward prepubertal levels so that the concentration profiles in young men were remarkably similar to those in prepubertal boys despite the rising testosterone concentrations (661).

The gonadal hormones, particularly circulating androgens, appear to play a central role in the GH pulse amplitude changes (619, 661). However, androgens alone do not seem to account entirely for the pubertal increase in GH pulse amplitude in humans, in view of the pattern in young men (see above), and although ample data in rats support it (see Ref. 970), the evidence that testosterone is involved in the control of GH secretion in humans is not so clear. To explain these discrepant findings, it would seem, however, that testosterone itself has little or no effect and that it is the estradiol derived from aromatization of testosterone that increases GH secretion (1065). This is suggested by the potent action of physiological amounts of E₂ in stimulating pulsatile GH secretion and the fact that nonsteroidal anti-E₂ inhibits it (see Ref. 557 for review).

However, apart from conversion to E₂, a role for androgens cannot be dismissed. In prepubertal boys with constitutional delayed growth, both testosterone and the nonaromatizable androgen oxandrolone raised the daily GH secretory rate by specifically increasing the total mass of GH released per pulse, with no effect on GH burst frequency or the half-life of endogenous GH (1045). Supporting these findings, an increase in GH in response to GHRH was suggested in one study of boys with constitutional growth delay treated with oxandrolone for 2 mo (633), and increased plasma IGF-I concentrations have also been reported (980). In boys with isolated hypogonadotropic hypogonadism, progressively higher testosterone doses, which affected plasma E₂ levels, increased the GH secretory burst number and amplitude and GH response to combined GHRH and arginine (427).

The effect of endogenous estradiol concentrations on total and pulsatile GH release in humans also emerges from a study that evaluated the separate and combined effects of age and sex in young adults and aged women and men (474). The integrated GH concentration was significantly higher in women than men and in the young than the old. A major finding was that estradiol but not testosterone concentrations correlated with indices of total and pulsatile GH release, i.e., integrated GH concentration (IGHC), pulse amplitude, number of large pulses, and fraction of GH secreted per pulse (FGHP) so that on removing the variability due to free estradiol, neither sex nor age influenced IGHC or mean pulse amplitude. The finding, however, that aging still affected the residual variability, i.e., large pulses and FGHP, indicated that other factors in addition to E₂ contribute.

That aging lowers GH secretion in mammals is almost a tenet of neuroendocrinology (754). In male and female rats, the amplitude of the GH pulses is significantly smaller, and in male rats, this results in lower mean secretion of GH to about one-third that in young rats. In both female and male old rats, the frequency of GH pulses does not change with age (970). In humans, the mean concentration of GH over 24 h is similar after the fourth decade of life to that in prepubertal children and is about one-quarter of that in pubertal youngsters (1120). In addition, approximate entropy analysis in healthy aging men revealed a progressive increase in the entropy of 24-h GH profiles, which signifies less orderliness or regularity of GH release with age (1064).

These findings complement earlier reports of substantial changes in GH secretion in the aging human (169). Despite evidence of a preserved pituitary GH pool (see sect. VB3) and the strong belief that, as in old rodents, the primary disturbance is hypothalamic, a low GH response has been clearly shown to many secretagogues, including GHRH (406, 478). In agreement with the GH hyposecretion, decreased plasma IGF-I levels have also been found (see sect. IVD1). However, investigations on this topic are not unanimous in reporting a decline in GH release with age, and it has been suggested that a significant decline occurs only in women (see above and Ref. 474). In a study in 142 healthy elderly people of both sexes, aged 60-90 yr, basal GH levels showed a very weak positive correlation with age, whereas IGF-I showed a highly significant negative correlation; basal GH and IGF-I did not correlate with each other (518).

It is likely that other counfounding variables, such as institutionalization, body mass, nutritional status, and sex steroid hormone concentrations, exert either independent or combined effects on total and pulsatile GH release. To control some of the confounding variables, a group of normal adult males was studied using deconvolution analysis to clarify the specific features of GH secretion and clearance related to age and adiposity, alone or jointly (509). Age was a major negative statistical determinant of GH burst frequency and endogenous GH half-life; body mass index (BMI), an indicator of relative adiposity, correlated negatively with GH half-life and GH secretory burst amplitude. Age and BMI both correlated negatively with the daily GH secretion rate and together accounted for >60% of the variability in 24-h GH production and clearance rates. On average, it was estimated that for a normal BMI, each decade of increasing age reduced the GH production rate by 14% and the GH half-life by 6%. Each unit increase in BMI, at a given age, reduced the daily GH secretion rate by 6%.

Thus a logical corollary is that age, sex, pubertal status, BMI or other indices of adiposity, quality and quantity of sleep, medication use, and physical exercise are but a few of the confounding factors that must be considered in the design and interpretation of clinical studies.

For patterns of GH secretion in different pathophysiological contexts such as fasting, exercise, type I diabetes mellitus, obesity, see the corresponding sections.

It has long been recognized that GH release increases during the night (498, 1003) and nocturnal blood sampling for GH has been proposed as a diagnostic alternative to pharmacological testing of GH secretory capacity (470, 560). Early studies indicated a consistent relationship between slow-wave sleep (SWS) and GH secretion during early sleep and later during the night (503). Subsequent studies, with 30-s sampling of plasma GH during sleep, showed mean GH concentrations and secretory rates were significantly higher during stage 3 and 4 sleep than during stages 1 and 2 and rapid-eye-movement (REM) sleep. Growth hormone secretory rates and peripheral GH concentrations were maximally correlated with sleep stage in a fashion suggesting that GH release is maximal within minutes of the onset of stage 3 or 4 sleep (486). The temporal correlation between pituitary GH secretion and SWS is consistent with a model in which altered cortical activity is transmitted through GHRH and SS centers to the hypothalamus, and further stimulation by these centers is conveyed to the pituitary.

These findings supported the concept that the daily GH secretory output is dependent on the quality and occurrence of sleep and were consistent with results of sleep reentry and SWS deprivation studies (923). However, a series of reports, e.g., pronounced nocturnal GH surges divorced from the presence of SWS (526), selective SW stage deprivation failing to suppress or delay the sleep-onset GH pulse (126), temporal disconnection between the effect of GHRH on SWS and on GH release (984), challenged this concept and suggested that GH and SWS are not causally related but are two independent outputs of a common hypothalamic GHRH mechanism. However, studies in human African trypanosomiasis (sleeping sickness) have shown that the association between GH secretion and SWS persisted despite disrupted circadian rhythms (871), further suggesting that sleep and stimulation of GH secretion are outputs of a common mechanism.

A careful study in healthy young men examined how normal and delayed sleep and time of day affected pulsatile GH secretion, analyzed by the deconvolution method and blood sampling at 15-min intervals (1054). During normal waking hours, the GH secretory rate was similar in the evening and morning, double during wakefulness at times of habitual sleep, and tripled during sleep even when it was delayed until 0400 h. Interestingly, the amount of GH secreted in response to sleep onset was closely correlated with the levels during wakefulness, with the hourly GH output in these young men being two to three times greater when asleep than when awake.

Although these results demonstrated that SWS facilitates GH secretion, they did not show that SWS is obligatory for nocturnal GH secretion to occur. Approximately one-third of the nocturnal GH pulses occurred in the absence of SWS, and approximately one-third of the SWS periods were not associated with increased GH secretion. Thus circadian timing, regardless of sleep, influences somatotrophic activity, although under normal conditions sleep occurs at the time of maximal propensity for GH secretion, with circadian and sleep effects being superimposed.

In a sequel to these studies, GHRH or saline was administered intravenously to young healthy men at various times of day and stages of sleep. When injected during the waking hours, GHRH elicited a response whose magnitude was directly related to the amount of spontaneous GH secretion, negatively correlated with circulating levels of IGF-I and was not influenced by time of day. The response was greater when GHRH was given during SWS, whereas when GHRH was given during REM, it was similar to that observed during waking. This suggests that the two sleep stages may be associated with reduced and enhanced SS secretion, respectively. Awakenings during sleep consistently inhibited the secretory response to GHRH, which reappeared with resumption of sleep (1053).

Because the intersubject variability in GH responses was very wide but the response was quite reproducible for each subject, it would seem that there is no diurnal variation in the sensitivity to exogenous GHRH and that the GH response to GHRH is dependent on the sleep or waking condition, circulating levels of IGF-I and, possibly, genetic and life-style factors.

For the electrophysiological effects of GHRH and GHRP, see sections IIIA10 and IVC.

	III. HYPOPHYSIOTROPIC HORMONES

Top Previous Next References

A.  Growth Hormone-Releasing Hormone

1.  General background

The regulation of GH secretion has been the focus of research since Reichlin's pioneering observation in 1960 (879) in the rat that surgical destruction of the ventromedial hypothalamus slows growth velocity. Parallel to the identification and sequencing of most of the other hypothalamic regulatory hormones at the end of the 1960s and during the 1970s was the discovery of the GH release inhibiting hormone SS, the principal modulator of GH secretion (141).

Although the possibility had been advanced that changes in SS secretion might account for all the hypothalamic influences on GH secretion, a hypothalamic stimulating hormone was needed. Growth hormone-releasing factor or GHRH was only identified and sequenced ~10 years after SS.

The identification of GHRH represented an exception to the rule, since this neuropeptide is unique among regulatory hormones in that it was first isolated (1121) and sequenced (447, 892) from human and not animal tissue, and, secondly, not from the hypothalamus but from pancreatic islet tumors, in which the ectopic secretion was associated with hypersecretion of GH and acromegaly (376).

The discovery of GHRH has been reviewed extensively (see Refs. 274, 749) and is only briefly alluded to here. Growth hormone-releasing hormone was isolated in 1982 from two patients in whom pancreatic tumors had caused acromegaly. From one of these tumors a 40-amino acid peptide designated human pancreatic growth hormone-releasing factor hpGRF-(1---40) was isolated (892),¹ and three GHRH peptides, GHRH-(1---44)-NH₂, GHRH-(1---40)-OH, and GHRH-(1---37)-OH, were sequenced from the other tumor (450). Structure-activity relationships showed that the NH₂-terminal 29 residues of GHRH-(1---40)-OH have biological activity and indicated that although the NH₂-terminal portion of the molecule is involved in binding to the receptor, the COOH-terminal portion is critical in regulating the potency of GHRH.

Two of the three GHRH forms found in tumors were subsequently identified in human hypothalamus [GHRH-(1---44)-NH₂ and GHRH-(1---40)-OH] and differed only by the absence in the latter of the four COOH-terminal amino acid residues (613, 617). Identification of the GHRH sequence in pigs, goats, sheep, and cattle has shown the existence of the 44-amino acid form amidated at the COOH terminal (617). Rat hypothalamic GHRH (rGHRH) contains 43 amino acids with a free acidic group. In addition, rGHRH is structurally different from the other GHRH peptides [14 amino acid substitutions or deletions compared with GHRH-(1---44)] (see Refs. 756, 979).

Growth hormone-releasing hormone belongs to an expanding family of brain-gut peptides that includes glucagon, glucagon-like peptide I (GLP-I), vasoactive intestinal polypeptide (VIP), secretin, gastric inhibitory peptide (GIP), a peptide with histidine as NH₂ terminus and isoleucine as COOH terminus (PHI), and pituitary adenylate cyclase-activating peptide (PACAP) (160).

The availability of GHRH-(1---40)-OH and GHRH-(1---44)-NH₂ means the topography of GHRH neurons could be established in humans and a number of animal species (see Ref. 756 for review). There are slight species differences in GHRH distribution. Growth hormone-releasing hormone IR is present mostly in the basal hypothalamus and is appropriate anatomically for release into the pituitary portal vasculature. In early studies, GHRH antisera stained neuronal cell bodies in the ARC (infundibular nucleus) of squirrel monkeys and humans with fibers projecting to the ME and ending in close contact with the portal vessels (114). Studies in macaques and squirrel monkeys revealed GHRH-IR neurons also in the ventromedial nucleus (VMN) (114), where GH release may be increased by electrical stimulation (664). These neurons were also seen in the rat hypothalamus (708 and see below).

In the rat, GHRH-positive perikarya were located in the following regions: ARC from the rostral portion as far caudally as the pituitary stalk, medial perifornical region of the lateral hypothalamus, lateral basal hypothalamus, paraventricular nucleus (PVN) and dorsomedial nucleus (DMN), and the medial and lateral borders of the VMN (708). Immunocytochemical studies in humans showed a similar distribution of GHRH-positive perikarya, mainly within, above, and lateral to the infundibular nucleus. In addition, some GHRH staining cells were found in the perifornical region and in the periventricular zone (114, 115).

Careful charting of the distribution of GHRH-IR in the brain of adults rats was consistent with at least two distinct GHRH-containing systems, one responsible for delivery of the peptide to portal vessels in the ME and one whose relationship to hypophysiotropic function was less direct (see sect. IIIA10).

Growth hormone-releasing hormone neurons innervating the external layer of the ME were centered in the ARC on each side of the brain, whereas the other GHRH-stained cells, very likely contributing to the extrahypophysiotropic GHRH-stained projections, were in the caudal aspect of the VMN. From here, GHRH fibers could be traced projecting to the periventricular region of the hypothalamus including the preoptic and anterior part, where SS-containing neurons are clustered (see below and sect. IIIB1). Other important terminal fields were found in discrete parts of the DMN, PVN, suprachiasmatic nucleus (SCN), and premammillary nuclei and the MPOA and lateral hypothalamic area. Beyond the hypothalamus, sparse projections were traced to the lateral septal nuclei, the bed nucleus of the stria terminalis, and the medial nucleus of the amygdala (926). Growth hormone-releasing hormone projections in the amygdala, a part of the visceral brain involved in emotion and responses to stress, may account for GH responses elicited by electrical stimulation of this area in rats (663) or humans (639) and suggest the amygdala in humans may be part of the pathways responsible for stress-induced GH release.

Immunohistochemical (292, 514) and RIA (525) measurements indicated that GHRH is first detectable at embryonic days 18-20 in the rat, reaching adult levels at postnatal day 30 (525), and in situ hybridization has detected GHRH mRNA 1 day earlier (896). In the human fetus, ontogenetic studies detected GHRH neurons between 18 and 29 wk of gestation (115, 143). These figures are on the whole consistent with the appearance of fetal pituitary somatotropes and GH storage and release in plasma (see sect. IIB).

Colocalization studies in the rat using radioimmunologic, immunohistochemical, and in situ hybridization techniques indicated that a subset of GHRH neurons in the ventrolateral part of the ARC contains a large number of molecules including tyrosine hydroxylase (TH), the key enzyme of catecholamine (CA) biosynthesis, GABA, choline acetyltransferase (ChAT), the ACh synthesizing enzyme, neurotensin (NT), and galanin (GAL) (see Refs. 697, 756 for review) (Fig. 2). From a functional viewpoint, particularly important colocalizations are with NT, ChAT, and GAL, in view of the clear involvement of these compounds in GH release (see sect. VB). Other studies have shown colocalization of GHRH with neuropeptide Y (NPY), in neurons which do not project to the external layer of the ME but constitute a tubero-extra-infundibular system, not involved in direct control of the AP, but presumably having a neuromodulatory function (see Ref. 243 and sect. VB6).

View larger version (22K): [in this window] [in a new window]   Fig. 2. Schematic drawing illustrating distribution of neuroactive compounds in different parts of arcuate nucleus as well afferents to median eminence (ME). Anatomic abbreviations: DM, dorsomedial; MO, medulla oblongata; MPOA, medial preoptic area; Pe, periventricular nucleus; pVN, parvocellular paraventricular nucleus; VM, ventromedial; VL, ventrolateral; 3V, third ventricle. Abbreviations for neurotransmitters and peptides: CCK, cholecystokinin; ChAT, choline acetyltransferase; CRF, corticotropin-releasing factor; DA, dopamine; DYN, dynorphin; E, epinephrine; ENK, enkephalin; GAL, galanin; GRF, growth hormone-releasing factor; LHRH, luteinizing hormone-releasing hormone; NPK, neuropeptide K; NPY, neuropeptide Y; NT, neurotensin; NE, norepinephrine; POMC, proopiomelanocortin; SOM, somatostatin; SP, substance P; TRH, thyrotropin-releasing hormone; VP, vasopressin; TH, tyrosine hydroxylase. [From Meister et al. (697), with permission from Elsevier Science.]

Direct proof that cells in the ventrolateral ARC are likely to be the main source of GHRH-containing projections to the ME has been provided using a combination of retrograde tract-tracing and immunocytochemistry in normal and GH-deficient dwarf mice (902). Inferential evidence had been provided with the use of monosodium glutamate (MSG), a neurotoxic substance which destroys about 80-90% of the cell bodies in the ARC (794), without significantly injuring other hypothalamic regions or axons of passage. Injection of MSG into rats resulted in the virtually complete disappearance of GHRH-stained fibers in the ME (115), stunted skeletal growth, and severe obesity. Interestingly, these changes were combined with marked decreases in ME-stained fibers containing GAL, the GABA synthesizing enzyme, glutamic acid decarboxylase (GAD), dynorphin, and enkephalins, but not those containing SS and NPY, meaning these peptides do not project to the ME (NPY) or to the ME from another source (SS) (696, see also sect. IIIB1).

All in all, the localization of GHRH-IR neurons mainly in the ARC and in the capsule surrounding the VMN is consistent with the results of stimulation and ablation studies that implicate the medial basal hypothalamus (MBH) in the stimulatory control of GH secretion (see Ref. 749).

The organization of hypothalamic SS-containing neurons is essentially similar in all mammalian species studied so far, including humans (see sect. IIIB1). Most hypothalamic areas contain an extensive network of SS-containing fibers, and there are IR-SS perikarya in the anterior periventricular area and the parvocellular portion of the PVN and in the SCN, ARC, and VMN. Combined retrograde tracing and immunohistochemistry showed that up to 78% of periventricular neurons project to the external layer of the ME, and other hypothalamic neurons do not innervate this neurohemal organ (555). The SS axons and perikarya in the ARC and VMN, two areas from which GHRH neurons originate, indicate the existence of anatomic and functional peptide interactions. Immunocytochemical double-labeling studies located SS-producing neurons in the dorsomedial part of the rat ARC and found SS-IR axons in the lateral part of the nucleus, an area containing GHRH-synthesizing cells (622). Several somatostatinergic axon varicosities were clustered around single GHRH synthesizing cells, suggesting synaptic associations. In the external layer of the ME, axonal terminals immunolabeled for GHRH and SS had the same pattern of distribution and close proximity.

More morphological proof of SS-GHRH interactions in the ARC was provided by autoradiography studies showing a selective association of ¹²⁵I-labeled SS binding sites with a subpopulation of cells, concentrated rostrocaudally in the lateral portion of the ARC; their distribution was remarkably similar to that of GHRH-IR neurons (338). That at least some SS receptors may be directly linked to GHRH-containing cells was demonstrated in colocalization autoradiography and immunohistochemistry studies in which ~30% of the IR-GHRH cells labeled with ¹²⁵I-SS in sections adjacent to GHRH cells in extra-ARC regions did not show SS binding (690). Interestingly, most of the GHRH-SS colabeled cells were detected in the ventrolateral region of the ARC, where there is a dense fiber network of SS-nerve terminals (see above).

An ultradian pattern has also been reported in SS binding to ARC neurons in relation to the peaks and troughs of the GH rhythm (1010), although the changes were in the direction opposite those expected. Further direct evidence of cellular colocalization of SS receptors and GHRH-producing cells within the ARC was provided by in situ hybridization studies showing that in the ventrolateral portion of the ARC GHRH mRNA labeled neurons had the same distribution as ¹²⁵I-SS labeled cells (97) (Fig. 3). For the SS receptor subtypes involved, see section IIIB2.

View larger version (28K): [in this window] [in a new window]   Fig. 3. Schematic drawings of rat arcuate nucleus (ARC) comparing distribution of prepro-growth hormone-releasing hormone (GHRH) mRNA-containing neurons (left) and 125I-SRIH-labeled cells (right) on adjacent 5-µm-thick cryostat coronal sections of rat mediobasal hypothalamus at 3 anatomic levels (interaural line: 6.70, 6.20, and 5.70 mm according to Paxinos and Watson, 1986). ARC was separated in three parts [lateral (L), ventrobasal (VB), and periventricular (PV)]. , GHRH-producing cells coidentified as 125I-SRIH labeled; , 125I-SRIH-labeled cells; , 125I-SRIH-labeled cells coidentified as GHRH-synthesizing cells. Each symbol represents 3 labeled cells. [From Bertherat et al. (97).].

Two other cell types could be identified in these studies: GHRH perikarya, around the perimeter of the VMN not associated with pericellular SS binding sites, and very likely containing substance P and enkephalin-8 binding sites (282), and ¹²⁵I-SS labeled cells not associated with GHRH perikarya, in the periventricular zone along the dorsal part of the third ventricle, probably associated with -endorphin- (558) or TH-producing (484) cells.

With regard to GHRH inputs to SS-IR neurons, only a few of these neurons were closely approached by GHRH-IR fibers, indicating that SS neurons are only scantly innervated by GHRH cells (1098) and suggesting the relative independence of SS from GHRH.

The data reported provide strong morphological support for a direct hypothalamic control of GHRH neurons by SS and for the concept that the two hypophysiotropic peptides interact within the central nervous system (CNS) to modulate GH secretion (see also sect. II). A wealth of physiological evidence bears witness to the interaction, including a rapid increase in plasma GH after intracerebroventricular administration of SS to anesthetized (2) or awake rats (641), depending at least in part on hypothalamic GHRH (762); increased GHRH concentrations in hypophysial portal blood after intracerebroventricular administration of SS antiserum (853); elevated basal GH levels, which can be reduced by anti-GHRH serum after lesions of the hypothalamic MPOA (548); and triggering of GH secretion through a mechanism involving GHRH, after acute withdrawal of endogenous (722) or exogenous SS (210, 244, 964). For the GHRH-SS interactions in ultra-short feedback mechanisms, see section IIIB2.

Like other neuropeptides originally described as primarily of hypothalamic origin, GHRH has also been found in peripheral tissue, although its physiological role as an extrahypothalamic hormone, in addition to stimulating GH secretion, is far from clear. Activity of GHRH outside the brain was initially only found in neoplastic tissues producing GHRH (see sect. IIIA1). However, the hormone was then seen in the upper intestine, particularly in the jejunum, with smaller amounts in the duodenum. None was found in the ileum or colon. The peptide was confined to the epithelial mucosa, suggesting a possible endocrine role. In addition to the gastrointestinal tract, mRNA for GHRH was found in leukocytes, where GHRH inhibits human natural killer cell activity (819) and also affects human leukocyte migration (1126), in the testis (94), ovary (65), and placenta (657). In the rat testis, the estimated size of GHRH-like immunoreactivity (GHRH-LI) was 3.7 times that of synthetic GHRH (831), and it was found in the interstitial tissue (740) and in mature germ cells (831), its function being to amplify the action of GH on Leydig cells (steroidogenesis) and of follicle-stimulating hormone on Sertoli cells (cAMP formation) (240). In view of the existence of a blood-testis barrier and the abundance of GHRH in rat testis, it presumably acts as an autocrine or paracrine factor.

In the rat placenta, GHRH mRNA is distinct from that described from the hypothalamus, and placental GHRH is synthesized in the cytotrophoblast (657). Cells expressing GHRH are located differently in the mouse placenta (992). The GHRH peptide and mRNA have also been identified in the human placenta (95). Placental GHRH may have a role in regulating fetal pituitary somatotrophs (445) and placental growth factors in an autocrine and/or paracrine fashion (476). Initial studies in the somatotrophs found IR-GHRH in secretory granules and the nuclei (738), and others found it internalized in another intracellular compartment (705).

Pancreatic tumors from which the GHRH peptide was purified provided the mRNA source for the construction of libraries from which cDNA encoding the human GHRH precursor could be isolated (444, 685). A rat GHRH cDNA was isolated directly from a hypothalamic library (681). Using cDNA to screen human genomic libraries resulted in the elucidation of the human GHRH gene (682), and the rat and mouse GHRH genes have also been identified in the genome (991).

Like most small peptides, GHRH is generated by the proteolytic processing of a larger precursor protein. This is a single copy in the human genome, and it has been mapped to the p12 band of chromosome 20 in the human (682) and chromosome 2 in the mouse (433). The gene includes five exons spanning ~10 kb of DNA, with clear segregation of the functional domains of the GHRH precursor into the five exons of the genes (for further details, see Refs. 377, 683).

The GHRH precursor proteins identified in the human, the mouse, and the rat range in length from 103 to 108 amino acids; their sequences are fairly homologous, with the amino acids being closest in the NH₂-terminal portion of the mature GHRH peptide, the region crucial for GHRH binding and biological activity (160). The rat and mouse proteins differ completely from the human GHRH precursor near the COOH terminus. In the rat, although most of the 104 amino acids of the GHRH precursor protein are similar to the human precursor, the last 13 residues in the COOH-terminal domains are all different (678).

The cDNA probes for GHRH made it possible to examine gene expression in the mammalian hypothalamus. In situ hybridization analysis of GHRH mRNA in the rat brain showed without doubt that GHRH-expressing neurosecretory cells, as revealed by immunocytochemistry (see sect. IIIA2), are an actual site of GHRH synthesis (89, 678, 1125). In agreement with RIA and immunohistochemical findings, most of the GHRH-expressing cells were on the ventral surface of the hypothalamus within the ARC, with additional cells extending into the DMN. No substantial expression was observed in brain regions outside the hypothalamus.

It is now clear that the hormonal and metabolic status of the animals dictates appropriate changes of GHRH mRNA expression in ARC neurons. The GHRH mRNA levels are higher in male than female rats, and testosterone is reported to increase the expression of GHRH mRNA in males and to prevent the decline due to castration. This latter effect is shared by dihydrotestosterone, suggesting direct activation of the androgen receptor as the underlying mechanism (1124) and supporting a role of androgens per se to stimulate GH secretion (see sect. IIB). Thyroid hormone deficiency also increases GHRH mRNA levels; also this is mediated by the decrease in GH secretion (see below), since treatment with GH restores normal GHRH mRNA despite persistent hypothyroidism (288, 324).

Metabolic state is also important, and in genetically obese Zucker rats (258, 354) as well as in caloric deprivation (149), where GH secretion is decreased, GHRH mRNA levels also drop, indicating effects divorced from GH secretion and the primary hypothalamic origin of the disturbance (see sect. VIC1). Streptozotocin-injected diabetic rats have diminished GH secretion, which can be restored by an anti-SS serum (1007). However, GHRH mRNA levels are low, and SS mRNA levels are high, indicating a metabolic regulation independent of GH (793) (see also sect. VI for discussion).

Hormonal feedback regulation through GH appears to be the most important factor in GHRH mRNA expression. Growth hormone deficiency induced by target organ removal (hypophysectomy) or selective abrogation of GHRH function is associated with increased GHRH mRNA levels, whereas GH treatment lowers them (237, 290, 678), although not always (202) (see sect. VII for discussion).

Levels of hypothalamic GHRH do not always change in parallel with GHRH mRNA (237, 386) because they also reflect a different rate of peptide increase and impairment of GHRH mRNA translation. Thus proper interpretation of changes in GHRH content requires independent information on GHRH mRNA levels and/or GHRH secretion.

The expression of GHRH mRNA in the placenta and other tissues is discussed above.

In vitro studies with the use of both long-term cultures of fetal rat hypothalamus and short-term static incubations of adult rat hypothalami with perfusion have been used to evaluate GHRH secretion. Such systems showed the releasing effects of K⁺-induced depolarization (547) and the increased release after hypophysectomy (237) and thyroidectomy (324) in response to low glucose or intracellular glucopenia (64) and ₂-adrenergic stimulation (536). Conversely, IGF-I (959), SS (1114), and activation of GABAergic function (63) impaired GHRH release. Multiple second messenger pathways underlie these effects, i.e., Ca²⁺, phosphatidylinositol-protein kinase C, and adenylate cyclase-protein kinase A (63, 277).

In vivo picomolar to nanomolar concentrations of GHRH have been reported in the rat and the sheep portal venous blood. In anesthetized, hypophysectomized rats, portal plasma levels of GHRH were ~200 pg/ml with synchronized pulses up to 800 pg/ml (853). In freely moving, ovariectomized ewes, pulsatile GHRH secretion reached a mean of 20 pg/ml, with peak values from 25 to 40 pg/ml and a secretion rate of 13 pg/min. About 70% of spontaneously occurring GH peaks could be attributed to GHRH pulses, as indicated by simultaneous measurement of peripheral GH levels (378, see also sect. II).

Unstimulated plasma levels of GHRH were in the range of 10-70 pg/ml in normal adults (827, 1032), with no apparent sex-related differences (924). Comparable values have been reported in the cerebrospinal fluid (CSF) of children and adults of both sexes with no endocrine diseases (541). Growth hormone-releasing hormone is high in the cord blood from term human newborns (39), as are levels of GH (see sect. IIB). Normal children of both sexes between ages of 8 and 18 had levels five times higher at mid-puberty than prepuberty in girls and double in boys (41), supporting the central role of gonadal hormones (see also sect. IIB).

The source of GHRH-LI in human plasma is not known, but the comparable levels in normal subjects and patients with hypothalamic lesions (232) suggest that the major source is not the hypothalamus (see also sect. IIIA4). Food (972) or glucose injection (542) can raise GHRH-LI levels, and this also occurs in patients with hypothalamic lesions or in GH-deficient subjects. These observations add weight to the idea that GHRH produced in the periphery (stomach, pancreas, small intestine) (239) may also act as a paracrine or autocrine factor. The hypothalamic component of circulating GHRH-LI, however, is specifically stimulated by insulin-hypoglycemia (542, 909; but see also Ref, 1022) or levodopa (231, 542, 914), but not by clonidine or arginine (1085).

Other conditions that increase circulating GHRH levels include the initial slow stage of sleep (914) and the sauna bath in young subjects (605). For other details on GHRH levels in GH hypo- and hypersecretory states, see Reference 749.

Growth hormone-releasing hormone is metabolized in human plasma by dipeptidyl peptidase type IV (DPP-IV) and trypsinlike endopeptidases (375, 380). Degradation comprises an initial step whereby DPP-IV removes the NH₂-terminal dipeptide (Tyr-Ala) leaving metabolites [GHRH-(3---44)-NH₂ and GHRH-(3---40)-OH] with no bioactivity. Independent cleavage by trypsinlike endopeptides at residues 11-12 also inactivates the peptide. Because the half-life of GHRH-(3---44)-NH₂ in vitro is considerably longer (55 min) than the native molecule (15 min), most of the plasma GHRH-IR is in the biologically inactive form. Studies with bovine GHRH suggest metabolic disposal is similar to human GHRH (579).

Growth hormone-releasing hormone exerts its primary control of somatotrophic function by increasing GH secretion, GH gene transcription and biosynthesis, and somatotroph proliferation. A host of intracellular second messenger systems including the adenylate cyclase-cAMP-protein kinase A, Ca²⁺-calmodulin, inositol phosphate-diacylglycerol-protein kinase C, and arachidonic acid-eicosanoic pathways are activated after binding to specific GTP-linked receptors in the plasma membranes (see also Ref. 683).

Earlier evidence that cAMP was involved as a second messenger in the stimulation of GH secretion rested on the demonstration that crude ectopic GRF preparations, cAMP analogs, and theophylline, a phosphodiesterase inhibitor, stimulated GH release (822, 938, 997). With GHRH in pure form, it was then shown that GH release was associated with a dose-dependent rapid stimulation of adenylate cyclase activity and cAMP production in the somatotrophs (102). Within 5 min of the addition of 1 nM GHRH to cultured rat pituitary cells, intracellular cAMP levels rose 6-fold, with a maximal response at 30 min; 10 times more GHRH was required to obtain half-maximal stimulation of cAMP production than GH secretion, suggesting that only partial receptor occupancy is required to elicit a maximal GH response (1066).

Growth hormone-releasing hormone-induced stimulation of cAMP production would occur in the manner typical of receptors coupled to GTP-binding proteins (106). The occupied GHRH receptor activates a heterotrimeric stimulatory G protein (G_s), composed of -, -, and -subunits, by catalyzing the binding of the -subunit to GTP. Guanosine 5'-triphosphate binding leads to dissociation of the -subunit complex from the G_s -subunit, and the latter then activates the catalytic subunit of adenylate cyclase. Thus adenylate cyclase stimulation by GHRH is dependent on GTP (769, 974), but nonhydrolyzable GTP analogs or a high concentration of GTP reduced GHRH receptor binding (988), and stimulation of adenylate cyclase (974), consistent with existing models for receptor-G protein interactions. Like GHRH, agents that bypass the receptor, such as cholera toxin, which decreases the GTPase activity of G_s (182) or forskolin, which directly activates the catalytic subunit of adenylate cyclase (944), potently stimulated cAMP accumulation and GHRH release (140, 275).

In the absence of GHRH or other hypothalamic peptides, somatotropes in culture have either a stable basal intracellular Ca²⁺ concentration ([Ca²⁺]_i) level (487) or asynchronous spontaneous [Ca²⁺]_i transients (487) that result from Ca²⁺ entry through Ca²⁺-permeable channels (487).

Extracellular Ca²⁺ is required for GHRH to stimulate cAMP accumulation in pituitary cell preparations (140), suggesting that Ca²⁺ is also a second messenger for GHRH and that Ca²⁺ acts beyond, or independently of, cAMP. On purified preparations of normal rat somatotrophs, GHRH elicited a rapid increase in [Ca²⁺]_i detectable within 5 s (487, 643); this first phase was followed by a decline in [Ca²⁺]_i to a plateau that was always higher than baseline and lasted 5 min (643).

In addition to GHRH, other GH secretagogues rapidly raised intracellular Ca²⁺ (371, 576). Incubation in low Ca²⁺ medium or in the presence of Ca²⁺ channel blockers or antagonists blocked or greatly diminished both the GHRH-dependent increase in [Ca²⁺]_i and GH release, with no changes in the cytosolic Ca²⁺ levels (102, 140, 487, 644); this was taken to demonstrate that GHRH increases [Ca²⁺]_i in somatotrophs by stimulating Ca²⁺ influx through L-type voltage-operated channels (VOCC). These effects are highly dependent on external Na⁺ and result from a cascade of voltage- sensitive currents (581). The inhibitory action of SS on basal and GHRH-induced GH release resulted from its ability to lower [Ca²⁺]_i by inhibiting Ca²⁺ influx (643, 644). Somatostatin also inhibited the Ca²⁺ influx induced by other GH secretagogues, the only exception being high K⁺.

These studies led to a model in which a net influx of extracellular Ca²⁺ through VOCC is the primary source of the first phase of the GHRH-dependent increase in [Ca²⁺]_i (222). Alternatively, the increase in cAMP accumulation might increase Na⁺ conductance directly or through protein kinase A-dependent phosphorylation, leading to depolarization and opening of the L-type VOCC (643). A more important role of Ca²⁺ mobilized from intracellular stores has also been suggested, based on a high correlation between Ca²⁺ efflux and GH release in perfused somatotrophs or the ability to stimulate GH secretion and Ca²⁺ efflux despite the absence of extracellular Ca²⁺ (343, 789).

With regard to the interaction between the two second messenger functions, there is some evidence of a direct relationship between cAMP and Ca²⁺. W-7, an antagonist of the Ca²⁺ binding protein calmodulin, inhibited GHRH stimulation of adenylate cyclase, cAMP accumulation, and GH release, whereas the calcium ionophore A-23187 stimulated GH secretion and cAMP production (714, 935).

Admittedly, Ca²⁺ effects occur together with a metabolic reaction, the hydrolysis of membrane phosphoinositides (PI), and these two processes are functionally connected (780). It was initially reported that GHRH increased ³²P incorporation into PI of rat AP cells (165); however, this would be entirely separate from the PI catalysis involving the production of inositol 1,4,5-trisphosphate (IP₃) and diacylglycerol, since GHRH did not enhance IP₃ (647). Other investigators too have found no early effect of GHRH on PI hydrolysis (260, 370, 875). Although GHRH does not induce PI hydrolysis, synthetic diacylglycerols and phorbol esters, which mimic the effect of the endogenous substrate by activating protein kinase C (184), dose-dependently stimulate GH secretion (535, 790). These compounds bind to the cell membrane and then induce translocation of the inactive enzyme from the cytosol to the cell membrane, resulting in a large increase in its affinity for Ca²⁺. Protein kinase C appears to act synergistically with Ca²⁺ mobilization to increase cell product release in various systems because low levels of Ca²⁺ ionophore can potentiate the protein kinase activation (535). This synergism may operate through several mechanisms, and no firm conclusion has been reached (535).

Whatever the underlying mechanism is, it is noteworthy that increased Ca²⁺ mobilization in the pituitary enhances the sensitivity of the somatotroph to protein kinase C activators (535). It would seem, however, that this mechanism is only marginal to the secretagogue action of GHRH and may better account for the action of GHRP and their analogs (see sect. IVC2), and protein kinase apparently increases GH release by mechanism(s) different from those used by GHRH (790).

The fact that GHRH does not enhance IP₃ production and may thus be unable to mobilize intracellular Ca²⁺ from the IP₃-sensitive Ca²⁺ pool implies only a minimal effect on fractional Ca²⁺ efflux. However, GHRH may affect Ca²⁺ efflux through the formation of arachidonic acid, derived from the direct activation of phospholipase A and hydrolytic cleavage of the fatty acid from membrane phospholipids or, alternatively, by hydrolysis of diacylglycerol by diacylglyceride. Arachidonate and its metabolites are known to mobilize Ca²⁺ (567, 765). Growth hormone-releasing hormone and Ca²⁺ efflux are discussed above.

Supporting a GHRH-arachinodate interaction, arachidonic acid elicits a dose-dependent increase in GH secretion and GHRH stimulates arachidonic acid release in cultured rat pituitary cells (166, 534).

Prostaglandin (PG) E₁, PGE₂, and prostacyclin (PGI₂) enhance GH secretion in the rat in vitro (716, 800) and in vivo (551, 800) as well as the production of cAMP (716, 938), probably by stimulation of adenylate cyclase through a G_s protein-coupled PG receptor (966). Growth hormone-releasing hormone stimulates PGE₂ release from incubated rat pituitaries, and indomethacin and aspirin, two cycloxygenase inhibitors, reduce GHRH-induced GH and PGE₂ release (347).

8.  GHRH receptor

A) CHARACTERIZATION, CLONING, AND STRUCTURE. A better understanding of GHRH action in promoting GH secretion and linear growth requires knowledge of the GHRH receptor (GHRH-R) structure and function. Previous studies in intact cells or cell membrane preparations of the rat described high-affinity low-capacity specific GHRH binding sites, using GHRH or GHRH analogs (5, 947, 1066). Two studies reported an additional class of low-affinity high-capacity sites (5, 6). Binding sites were also evidenced in bovine pituitaries (1066) and human pituitary tumor tissue (502). Binding of GHRH resulted in the activation of adenylate cyclase, whereas guanine nucleotides inhibited hormone binding, suggesting the involvement of G_s protein (see sect. IIIA7 for details).

View larger version (60K): [in this window] [in a new window]   Fig. 4. Schematic structure of GHRH receptor. Amino acid sequence shown in a 1-letter code is that of rat GHRH receptor. Seven membrane-spanning domains are illustrated as cylinders crossing lipid bilayer. Solid circles represent amino acids that are conserved in related receptors for hormones secretin, glucagon, glucagon-like peptide I, gastric inhibitory peptide, vasoactive intestinal polypeptide, and pituitary adenylate cyclase-activating peptide. Open arrow indicates approximate site of signal sequence cleavage. Branched structures represent 2 consensus sites for N-linked glycosylation. Solid arrow indicates site of an insertion generated by alternative RNA processing in a variant form of GHRH receptor. [From Mayo et al. (683); Copyright The Endocrine Society.]