| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Section (A.G.), Department of Internal Medicine, University of Brescia, 25125 Brescia, Italy; and Division of Endocrinology (J.D.V.), Department of Internal Medicine and National Science Foundation Center in Biological Timing, University of Virginia, Charlottesville, Virginia 22908
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
). After the
isolation and characterization of GHRH and the cloning of its receptor,
as well as the more recent availability of molecular probes for
somatostatin and its recently cloned receptor family, more detailed
studies of the pathophysiological neuroregulation of GH secretion have
been performed. In this review, we will update earlier discussions in
the Journal (1, 2) by presenting recent developments in our
understanding of neuroregulatory mechanisms and their relevance to
clinical alterations in GH control. Primary diseases of the pituitary
gland or hypothalamus (e.g., tumors originating at these
loci) resulting in excessive or deficient GH secretion and peripheral
actions of GH or insulin-like growth factor-I (IGF-I) on target tissues
will not be reviewed here, but are discussed recently elsewhere (3, 4, 5, 6, 7, 8, 9).
Where possible, major interspecies differences in neuroregulation of
the GH axis (rat, sheep, and human) will be distinguished to limit any
confusion on this basis.
|
|
II. Contemporary Tools for Neuroendocrinological Investigation of the GH Axis |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
). Several
valuable molecular models, namely, genetically impaired GH secretion in
the little (lit) mouse, the dwarf (dw) rat, the fatty (fa) rat, and the
high-growth (hg) mouse, were recently reviewed in detail by Frohman
(10), to which the reader is referred. Conversely, the GHRH-transgenic
mouse exhibits excessive GH secretion, pituitary somatotrope
hyperplasia, and increased plasma IGF-I levels (11) that are
antagonized by potent GHRH-receptor inhibitors (12). In addition,
transgenic (rodent) models of suppressed secretion have been developed,
e.g., utilizing hypothalamically targeted GH or GHRH
transgenes (13, 14, 15). The experimental notion of targeting GH gene
expression selectively to the hypothalamus, and thereby producing
deficiency of endogenous GH, has disclosed alterations in
non-rapid-eye-movement sleep in the induced hyposomatotropic state
(16). Targeting of the GH gene in the mouse to the hypothalamus was
accomplished via use of either the tyrosine hydroxylase or GHRH
promoters to drive topographically localized GH excess (13, 14, 15).
This approach thus allows study of hypothalamic-regulatory peptide
responses to localized overexpression of a selected gene. In the case
of GH gene overexpression targeted to the hypothalamus, the
pathophysiological responses include increased hypothalamic
somatostatin and decreased GHRH accumulation, as predicted by other
experiments of GH autofeedback (17, 18, 19). Conversely, the genetic model
of the Ames dwarf mouse, which has a recessive defect with a
hypocellular anterior pituitary gland lacking somatotrophs,
lactotropes, and thyrotrophs (20), shows, as predicted from
physiological principles, increased expression of GHRH peptide and mRNA
in the hypothalamus (21). This follows from the premise that GH
autofeedback normally suppresses GHRH and stimulates somatostatin
expression (19). This thesis has recently been corroborated via
implantable GH-secreting (GC) cells in primary genetic vs.
transgenic GH-deficient rat models (22).
|
A recent novel gene knockout model, STAT5b gene disruption, which inactivates a specific signal-transduction pathway evokes loss of multiple sexually differentiated growth and cellular responses to GH, akin to those expected if the sexually dimorphic (male) pattern of pituitary GH secretion were abolished (32). This reflects the presumptive intracellular transcriptional factor-signaling role of STAT5b in mediating the cellular and nuclear (gene-transcriptional) actions of pulsatile (but not continuous) GH delivery. Thus, the STAT5b knockout mouse, albeit not measurably GH deficient, appears to be GH pulse resistant as reflected by dwarfism, low plasma IGF-I concentrations, and obesity. These features also are recognized in human GH receptor molecular defects causing tissue resistance to GH actions, e.g., Laron-type dwarfism (6, 33, 34, 35), and in the GH receptor knockout mouse (36)
Another transgenic strategy, transgenic expression of a GH receptor antagonist, has confirmed an important role of GH in antagonizing insulin action (37). Moreover, coexpression of the bovine GH gene and a human GH-receptor antagonist minigene in mice shows that the GH receptor antagonist will oppose the development of severe renal glomerulosclerosis and the increase in body growth otherwise driven by excess GH (38). In contrast, IGF-I overexpression augments body growth without inducing glomerulosclerosis, thus distinguishing certain tissue actions of GH and IGF-I. Indeed, another selective role of GH, but not IGF-I, is to increase motor neuron size in the lumbar spine, as inferred in transgenic mice overexpressing GH vs. IGF-I peptide (39).
In relation to pituitary developmental mechanisms, transgenic ablation
(via coupling a relevant glycoprotein 
-subunit cDNA to diptheria
toxin A chain) produces loss of gonadotrophs and thyrotrophs, as
distinct from somatotrope and corticotroph cell lineages (40). Other
recent experiments indicate the potential to "trap" developmental
progenitor cells by using transgenic overexpression of regulatory
regions responsive to cell-specific transcription factors,
e.g., the homeodomain protein, GHF-1. The latter is believed
to be required for the developmental generation of somatotropes and
lactotrophs, and hence GH and PRL gene expression (41). This technique
of so-called developmental entrapment can evaluate the roles of GH
progenitor cells by constitutive overexpression of GHF-1, and hence
"immortalization" of GH or PRL progenitor cells, which results in
dwarfism. A pituitary-targeted transgenic mouse model expressing the
leukemia inhibitory factor has suggested the possible role of
altered pituitary gene transcription and cell replication in the
pathogenesis of Rathke’s cysts (42).
Another informative molecular strategy is experimentally controlled transgene expression. For example, exogenously triggered activation and inactivation of the GH transgene in mice will reversibly reduce and induce obesity in this species (43). GH antisense RNA-transgene transfer in rats has achieved dose-dependent inhibition of GH gene expression, thus reducing GH secretion to varying degrees in heterozygous and homozygous transgenic animals (44). Conversely, expression vectors designed to transfer the GH gene as directed by tissue-specific promoters (e.g., muscle) provide an exciting prospect for targeted delivery of GH to, and for regulated expression of GH in, specific cell types, such as myoblasts in vitro (45) or myocytes in vivo (46). Indeed, in the latter, in vivo context or single injection of a myocyte-targeted GHRH-minigene expression vector can stimulate GH secretion and increase serum IGF-I levels by 3- to 4-fold or more for 2 weeks.
In brief, a rich and expanding repertoire of genetic, molecular, and transgenic and knockout models of receptors, regulatory peptides, IGF-I or GH itself, and/or signaling molecules has emerged and continues to unfold. This resource will make more specific and compelling studies possible of the molecular mechanisms of neuroendocrine regulation of the hypothalamo-somatotrophic axis in experimental animals. In some circumstances, a clinical counterpart is identifiable, in which a mutation of the corresponding gene is discernible in an individual or family. For example, IGF-I gene deletion, GH receptor, and GHRH receptor mutations with dwarfism all are recognized in the human (6, 28, 33, 34, 35, 47, 48, 49). However, GHRH receptor mutations causing isolated GH deficiency in the human are not common (50).
B. Human GH rhythms
1. Introduction. In conjunction with the development of
increasingly specific pharmacological agents to block or activate
individual regulatory receptors, new neuroendocrine tools have been
developed that allow improved quantitative measures of hormone release
over short intervals (ultradian rhythms or episodic peaks), as well as
over 24 h (circadian rhythms), and in relation to the patterned
orderliness of the release process (e.g., approximate
entropy estimates). These novel technologies often aid in dissecting
the neurohormonal mechanisms that underlie pathophysiological states or
molecular models (above). Biomathematical advances have also enhanced
clinical neuroendocrinological investigation, since in the human direct
portal catheterization techniques implemented in experimental animals
to monitor hypothalamo-pituitary secretion of GH, GHRH, and
somatostatin individually and concurrently, e.g., in the rat
and sheep (29, 51, 52, 53, 54, 55, 56), are not practicable ethically.
2. Background. Even relatively direct studies are not always
free of controversy. For example, a portal vein sampling study in the
anesthetized male rat indicated that both somatostatin withdrawal and
GHRH release tend to coincide with a GH pulse (56). Other repetitive
portal blood sampling investigations in the unrestrained and
unanesthetized (ovariectomized or intact male) sheep indicate that GHRH
increments typically precede GH pulses, but somatostatin decrements do
not necessarily do so (51, 57, 58, 59, 60) (Fig. 2). Moreover, in sheep, somatostatin
release can actually rise before a GH pulse or remain unchanged (see
Table 2). Consequently, GHRH-somatostatin
interrelationships in this species may be variable and quite complex.
For example, stress appears to stimulate both GHRH and somatostatin
release into portal blood, although the net result is increased GH
secretion (61). In addition, technical differences (e.g.,
anesthesia, sampling frequency, etc.) may confound a simple mechanistic
model originated in the adult male rat consisting of reciprocal
somatostatin withdrawal and GHRH release in initiating a GH pulse.
Indeed, taken as a whole, available studies clearly allow for other
input into GH secretory-burst timing, such as 1) extra- or
intrapituitary feedback by IGF-I (62, 63, 64); 2) intrapituitary paracrine
factors; 3) other hypothalamic coregulators [e.g.,
endogenous GH-releasing peptides (GHRPs), galanin, neuropeptide Y
(NPY), etc.]; for example, either NPY or galanin can colocalize with
GHRH in the hypothalamus (65, 66); and/or 4) other as yet unexplained
technical factors (e.g., hydraulic variations in blood flow
or sample collection; inconsistencies in GHRH, somatostatin, or GH
immunoassays; etc.).
|
|
Considerable clinical effort and numerous studies have appraised the
endocrine, metabolic, and pharmacological regulation of episodic GH
release (82, 83, 84, 85, 86, 87, 88, 89). For example, in the case of the GH axis, a burst of
GH release presumably reflects an appropriate stimulus such as
endogenous GHRH secretion (with or without other putative
secretagogues) in the context of responsive somatotropes
(i.e., at least not maximally inhibited by somatostatin),
and adequate releasable GH pools within the anterior pituitary gland.
Discrete pulse analysis provides information about the frequency and
amplitude of serum hormone concentration peaks (67, 68, 70) and can
disclose insights into neuroendocrine mechanisms that control this
ultradian pituitary release activity. For example, sampling blood every
30 sec during sleep in young men has revealed a strongly correlated
pattern of episodic GH release and the onset of slow-wave (stages III
and IV) sleep (90). Significantly, in some species such as the rat,
there is abundant evidence that the time profile of (pulsatile) GH
release strongly conditions the target tissue response anticipated,
e.g., induction of the LDL receptor and GH receptor (GHBP)
genes by a "feminized"/continuous GH release pattern, rather than
by a "masculinized"/pulsatile GH profile (91) (see Table 3). Hence, substantial physiological
effort has been applied to understand the neuroendocrine mechanisms
that supervise and dictate such pulsatile (GH) release patterns.
|
). This general class of analytical technology
thereby affords insights into the neuroendocrine control of the
underlying secretory events per se (104), which may be
regulated in principle by way of shape, duration, maximal amplitude,
and/or mass (106), with or without any attendant changes in hormone
half-life or distribution volume (73, 93, 94, 107). Indeed, altered GH
half-lives are likely only in subjects with severely impaired hepatic
or renal function [prolonged GH half-lives (108, 109, 110, 111)], and in
obesity [reduced GH life-life (112, 113, 114)]. Most recently, a
stochastic differential equation, random effects, maximum-likelihood
methodology based on feedback concepts within a neuroendocrine axis,
ultradian pulsatility coupled to circadian variations, and
biexponential kinetics has emerged to capture secretory dynamics more
fully (115, 116, 117).
|
0.020–0.05 µg/liter). This assay also revealed a pulse amplitude-
(and mass-) dependent mechanism for the gender distinction between GH
release in healthy middle-aged men and women (118). An even more
sensitive chemiluminescence-based GH assay (sensitivity of
0.002–0.005 µg/liter) confirmed small amounts of (<15%) basal
GH release in healthy men and women and illustrated that glucose
ingestion reduces serum GH concentrations typically to less than 0.7
µg/liter in women and to less than 0.07 µg/liter in men (119, 120, 121).
Thus, earlier "normal" glucose suppression of serum GH
concentrations assessed by RIA or IRMA has been grossly misestimated
and has also overlooked a strong gender difference. A high-sensitivity
enzyme-linked immunosorbent assay (ELISA) also corroborated very low
rates of basal GH release in both normal and hypopituitary adults
(122, 123, 124). How such basal secretion is generated or regulated is not
known (125). To our knowledge, detectable interpulse basal GH release
has not yet been measured in the male rat or other nonhuman species. Statistically correct treatment of assay data variability, especially at the low end, is also essential (126). Variability arises from uncertainties in the zero-dose tubes, among other replicates in the standard curve, due to the fit of the standard curve itself, and from the replicates of the unknown samples (127). Methods have been created recently to address these four joint/combined sources of within-assay experimental uncertainty (126, 128).
Clinical studies now indicate that visceral obesity, age, and other states of relative hyposomatotropism can be attributed mechanistically to decrease in GH-secretory burst mass, as estimated by the combined techniques of deconvolution analysis and ultrasensitive GH assays (121). Notably, as many as 97% of daytime samples for serum GH concentrations in older and/or obese subjects can be undetectable by conventional RIA or IRMA methods (125). Immunofluorometric, ELISA, and chemiluminescence GH measurement techniques, combined further in some cases with improved statistical assay analyses at the low end (126), have overcome some earlier limitations in studying dynamics of the human GH axis (125). For example, an ultrasensitive human GH assay has revealed that somatostatin infusion in young men suppresses both GH pulse amplitude and frequency (123). This is consistent with somatostatin’s putative inhibition not only of pituitary GH secretion but also of hypothalamic GHRH release (22, 129, 130). The reciprocal relationship between somatostatin and GHRH has also been inferred recently in in vitro studies in bovine hypothalamus (131).
Beyond and complementary to the pulsatile and 24-h (nyctohemeral) modes
of GH secretion, the subordinate (nonpulsatile) regularity or pattern
orderliness of hormone release over time can be quantified now by a
so-called approximate entropy statistic (132, 133). This is a
model-free and scale-invariant measure different from deterministic
chaos (nonlinear dynamical measures). An intuitive notion of the
entropy statistic is shown in Fig. 3B
. The orderliness of GH release
over time so quantified is markedly altered in certain conditions of
health or disease. For example, acromegalic patients secrete GH with a
high degree of quantifiable disorderliness, i.e., there is a
profound loss of or deterioration in regularity or reproducibility of
point-by-point subpatterns recurring across the 24-h GH profile,
whether defined by the approximate entropy statistic or a network-based
predictability measure (134, 135, 136, 137, 138). The distinctions between
GH-secretory patterns in active and remitted acromegaly as achieved via
the approximate entropy statistic are illustrated in Fig. 4.
|
). Unlike deterministic
chaos measures, which require 105 or more observations (and
theoretically noise-free data), approximative entropy values can be
calculated with good statistical replicability in hormone series
containing as few as 50–300 samples (118, 140, 141, 142). Such a small
sample size confers significant utility to this new measure.
|
In addition to the above available tools, other recent studies have begun to model the network or feedback-control linkages within the GH and other axes, to provide a more quantitative basis for articulating, testing, and revising specific neuroendocrine hypotheses (76, 154, 155). Considerable additional research will be required to refine the correct conceptual basis and dynamic features of the neuroendocrine component of the GH axis in appropriate biomathematical models.
|
III. Neuropeptide Regulation of the GH Axis: Somatostatin and GHRH |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
A. Mechanism of somatostatin actions and its receptors
Somatostatin binds to a family of specific receptors and inhibits
adenylyl cyclase via Gi, with additional actions to reduce
net calcium influx. Somatostatin inhibits GH release but not its
biosynthesis. This concept is important, since it may explicate
"rebound" GH secretion after somatostatin priming and withdrawal in
both the rat and human (157, 160). In addition, somatostatin may play
potentially (dual) inhibitory or stimulatory roles in controlling GH
secretion by acting on two distinct somatotrope cell populatins at
least in porcine pituitary (161).
Five somatostatin receptor subtypes have been cloned and characterized to date (162, 163, 164, 165). The somatostatin receptor is regulated in a subtype- and tissue-specific manner (166, 167). Furthermore, the cloning of somatostatin receptors and the development of somatostatin-receptor scintigraphy (168, 169, 170) hold significant clinical implications for identifying a variety of neuroendocrine and gastroenteropancreatic tumors. The clinical tool of somatostatin receptor scintigraphy has been reviewed in detail elsewhere (168).
Somatostatin peptide and somatostatin receptors are probably important in mediating many feedback and regulatory actions of neurohormones on the GH axis. For example, glucocorticoids likely influence both hypothalamic GHRH and somatostatin activity (171). In addition, GH autonegative feedback at the hypothalamic level involves both the GH receptor and hypothalamic somatostatin expression, inasmuch as treatment with antisense RNA to the GH receptor amplifies GH pulsatility (i.e., by abrogating GH receptor-mediated autonegative feedback) and decreases hypothalamic somatostatin gene expression in the rat (130). Other neuronal pathways may also participate in GH autonegative feedback, such as neuropeptide Y and galanin (172, 173). Indeed, the GH receptor also is expressed in NPY neurons in the arcuate nucleus (174). Moreover, the GH receptor gene in the hypothalamus is modulated by sex steroid and glucocorticoid hormones under several conditions (175).
Somatostatin receptor subtypes are expressed in individual rat pituitary cells, as defined by double-labeling in situ studies. For example, somatostatin subtypes II and V are present in all five major pituitary cell types in the rat (176), with the somatotropes expressing especially subtype V and to a lesser degree II, while thyrotrophs predominantly express subtype IV (177). In the rat hypothalamus, somatostatin receptor subtypes I and II may modulate GHRH and somatostatin release (178).
Somatostatin itself regulates somatostatin receptor subtype expression in some pituitary cell lines, e.g., GH-3 cells with differential up-regulation of subtype I vs. subtype II. In addition, food-deprived and diabetic rats show differential pituitary and hypothalamic somatostatin subtype gene expression/regulation (179). The molecular pharmacology of somatostatin receptors is reviewed further in Ref. 180 .
Somatostatin receptors also colocalize with GHRH-secreting neurons in the rat arcuate nucleus (181, 182). This important synaptic connectivity allows for intrahypothalamic neurohormonal interactions in GH regulation (58, 183), as reviewed further below (see GHRH). Somatostatin receptors in the locus ceruleus of male rats, seemingly paradoxically, stimulate GH secretion (184), thus illustrating the role of central nervous system (CNS) topography of somatostatin receptor distribution in mediating differential actions.
B. Human somatostatin receptor
Various somatostatin receptor subtypes are expressed within
pituitary adenomas, e.g., subtypes II and V in acromegalic
tumors that are responsive to somatostatin (185, 186, 187). This observation
has significant clinical implications, since available long-acting
somatostatin analogs (e.g., octreotide and lanreotide) are
known to interact especially with the subtype II somatostatin receptor
(187).
All five human somatostatin receptor subtypes are expressed in pituitary tumors as well as in normal (fetal) pituitary tissue (13, 180). In culture, the human pituitary gland predominantly expresses subtypes II and V whether in tumoral or normal pituitary cells (176, 188, 189). The human somatostatin receptor subtype genes are localized on different chromosomes, e.g., 14, 17, and 22, with simple tandem repeat DNA polymorphisms in subtypes I and II (190). All five cloned human somatostatin receptors subtypes are functionally coupled to adenylyl cyclase (191), and subtype I additionally stimulates inositol phosphate accumulation (192). The type V receptor also can mediate inhibition of GH secretion from acromegalic tumor cells in vitro (189). Consequently, the development of highly specific somatostatin receptor subtype agonists by structure/activity-based methods, such as a lanthionine octapeptide with high affinity for this receptor subtype, may have clinical application (193). Other studies of pituitary adenomas confirm expression of multiple subtypes, e.g., as assessed by PCR or other molecular methods (185, 194, 195).
Somatostatin itself is expressed in the brain and in the periphery in two principal forms; namely, somatostatin-14 and somatostatin-28 (196) (the latter preferentially binds to the subtype V receptor) (168). Somatostatin influences not only GH secretion but also that of numerous other hormones, as well as cognitive and behavioral processes, and impacts the gastrointestinal tract, the cardiovascular system, and tumor growth (156, 158). Thus, somatostatin and its receptor subtypes are widely distributed and regulated throughout the body, rather than solely in the hypothalamus and pituitary gland. Considerable complexity exists in the domain of somatostatin receptor subtype regulation in health and disease. This should represent a valuable purview for further incisive clinical and basic studies.
C. GHRH
1. Isolation, actions, and neuronal distribution. Human GHRH
was isolated originally from two pancreatic tumors in patients with
acromegaly from the United States and France as 44- and 40-amino acid
forms (197, 198). These bioactive peptides are derived from either of
two larger polypeptide precursors (pre-pro GHRH 107 and 108) (199, 200). The human GHRH gene resides on chromosome 20. The naturally
occurring variants of GHRH (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40) and the synthetic analog
1–29 are biologically equipotent on a molar basis in their capacity to
stimulate GH release in humans (201). The biological half-life of GHRH
1–44 is about 3–6 min. This is because GHRH is rapidly inactivated by
a plasma dipeptidylaminopeptidase, producing a more stable metabolite,
GHRH 3–44, which is about 1,000 times less potent than the parent
compound (202).
Intravenous administration of GHRH in humans evokes dose-related
release of GH from the pituitary gland. Approximately 1.0 µg/kg is a
maximally stimulating dose (203, 204, 205, 206, 207). GH secretion is detectable
within 5 min following GHRH injection, becomes maximal at 15–45 min,
and returns to baseline by 90–120 min (203). GH release induced by
intravenous GHRH injection in adults is highly variable
interindividually (208) and also in the same volunteer stimulated on
different occasions (207). This may be explained in part by the
diversity of hypothalamic somatostatin and/or cosecretagogue tone that
likely characterizes different individuals, and also by time-variable
somatostatin (or cosecretagogue) release and/or action in the same
subject (209). The relevance of variable hypothalamic somatostatin
secretion per se to the nonuniformity of the GH response to
single-bolus GHRH injection is suggested by experiments showing that a
functional somatostatin antagonist, i.e., the cholinergic
agent pyridostigmine, is able to cause a significant leftward
(increased sensitivity) and upward (increased efficacy) shift of the
GHRH dose-response curve in young men (206) (Fig. 6).
|
Age is a critical factor in influencing the GH response to GHRH in humans. Neonates, children, and young adults have similar GH responses to GHRH, whereas GH release in most studies is reduced in older individuals (212). Cells from the human fetal pituitary gland become responsive to GHRH in vitro before the appearance of either GHRH-containing neurons or the hypophysial portal vasculature (213). Human fetal pituitary responsiveness in vitro to GHRH matures further during gestation (214), which suggests that GH secretion from the human pituitary gland is not dependent on hypothalamic GHRH’s availability from weeks 7–24 of fetal life (213).
Sex steroids may also influence the GH-secretory response to GHRH in the human. However, not all workers have observed consistent differences between men and women (205), in young women after ovariectomy (215), during progestin-opposed estrogen replacement in the postmenopausal woman (216), or at different stages of the menstrual cycle (210). This is discussed further in Section VII.B.2.
Prolonged continuous infusion or repeated intravenous bolus administration of GHRH leads to a modest decrease in the subsequent GH responses to GHRH especially in the rat in vitro (217, 218) and to a lesser extent in the human in vivo (219, 220). Prolonged continuous GHRH infusion in normal humans causes acute maximal GH release; thereafter, pulsatile GH release continues but tends to decline (albeit remaining above basal values) by 5 h despite uninterrupted GHRH administration (220). Continued pulsatility of GH secretion occurs despite an unvarying GHRH stimulus, which suggests intermittent somatostatin action and withdrawal in normal humans or intermittent cosecretagogue release (221, 222, 223).
Treatment with GHRH in the rat leads to a fall in pituitary GH content. Concomitant administration of somatostatin with GHRH in the human or rat largely prevents this attenuation (human) or loss (rat) of somatotrope responsiveness, which occurs in vivo in normal human subjects (224) and in vitro in rat anterior pituitary cells (225). Thus, depletion of a GHRH-sensitive (and somatostatin-antagonized) releasable pool of GH is one factor contributing to the loss of GH responsiveness to GHRH in the rodent. Other biochemical mechanisms are also suggested, however, since prior treatment with GHRH in vitro in the presence or absence of somatostatin causes a decrease in the cAMP response to a subsequent GHRH stimulus as well as an increase in the ED50 for the stimulation of both cAMP accumulation and GH secretion by GHRH (226). Whereas GHRH pretreatment in vitro leads to a 48% decrease in the number of GHRH receptors on somatotropes, maximal GH responses to GHRH can be obtained by occupancy of only 10–20% of the total number of anterior pituitary GHRH receptors. Thus, the physiological relevance of in vitro receptor desensitization to GHRH observed in the rat is not easily translatable to in vivo down-regulation in the human.
In the human, continuous intravenous infusion of GHRH in normal subjects for 2 weeks evokes a marked increase in nocturnal GH pulses and plasma IGF-I levels (227). Moreover, 3 days of pulsatile GHRH administration (0.33 µg/kg every 90 min iv) to men of varying ages and body compositions augments and maintains pulsatile GH secretion and its nyctohemeral rhythm (151, 228). Lastly, GHRH-secreting tumors achieving sustained high serum GHRH concentrations elicit pulsatile GH release (229). These clinical data indicate the robustness of the human (in contradistinction to the rat) somatotrope to unvarying GHRH stimulation and also suggest the feasibility of treating selected patients with GHRH on a continuous basis, e.g., via a repository form of GHRH.
Immunoreactive GHRH exists in high concentrations in neurons of the
median eminence and arcuate nuclei of mammalian and nonmammalian
species. GHRH is also expressed in the anterior hypothalamic region as
well as the dorsomedial and ventromedial nuclei (230, 231). GHRH is
colocalized with other neuropeptides (e.g., galanin or NPY)
in the hypothalamus (232). Moreover, hypothalamic GHRH neurons receive
significant afferents from somatostatin neuronal nuclei (181, 233) that
are hypothesized to be important in the GHRH-somatostatin interplay
that presumptively directs ultradian GH pulsatility (Refs. 22, 129, 130, 131 and Fig. 7). GHRH has also been
identified both in the secretory granules and the nuclei of
somatotropes (234). Appreciable quantities of immunoreactive material
are detected in plasma (235), duodenum (236), and placenta (237). Thus,
circulating GHRH levels are not usually helpful to the endocrinologist
clinically, except in the rare condition of ectopic GHRH secretion with
attendant acromegaly (229).
|
The human GHRH receptor was cloned from an acromegalic pituitary cDNA library (240). The porcine and rat GHRH receptors were cloned independently (245, 246). The GHRH receptor is homologous with that for secretin and vasoactive intestinal polypeptide. Expression of the pituitary GHRH receptor is developmentally regulated in the rat (247). Moreover, albeit initially unexpected, both the GHRH gene and peptide are expressed in the gonad, e.g., rat ovary (248). There is tissue heterogeneity of GHRH receptor expression in the human, e.g., kidney compared with pituitary gland (249).
GHRH receptors are critical for GH release and growth, since point
mutations of the extracellular peptide-binding domain of the GHRH
receptor in the dwarf lit/lit mouse (50) disrupt body growth and
abrogate GHRH receptor function (250). GH deficiency also occurs in
corresponding mutations of the human GHRH-receptor gene (see Table 1).
Moreover, in the human and the rat, overnight GH secretion as well as
GH release stimulated via a number of secretagogues can be blocked by a
selective GHRH antagonist (251, 252), thus supporting a role for GHRH
in endogenous pulsatile GH secretion.
3. GHRH receptor regulation. The GHRH receptor, its signal transduction mechanisms, and mediation of GHRH-stimulated gene expression were reviewed recently (243, 244). In principle (although not yet described), activating mutations of this receptor might lead to constitutive cAMP overproduction, e.g., in patients with GH-secreting tumors and acromegaly. Alternatively spliced mRNA species encoding truncated nonfunctioning GHRH receptors can be identified in human pituitary adenomas (253). The receptor for GHRH is regulated physiologically by glucocorticoids and estrogen (254, 255). Steroid hormone regulation of GHRH receptor activity will thus require further study.
Lesioning of GHRH neurons or GHRH antagonist administration diminishes growth length and weight in the experimental animal, decreases GHRH receptor number, and reduces pituitary GH content (252). The receptor is clearly distinguishable from that of GHRP (below), somatostatin, or pituitary adenylate cyclase-activating peptide (PACAP), etc. GHRH receptor mRNA is expressed in the pituitary gland, as well as in the periventricular, arcuate, and ventromedial nuclei of the hypothalamus, thus suggesting that these areas are sites of intrahypothalamic GHRH action in the rat (256). Moreover, GH autofeedback likely involves, in part, suppression of hypothalamic GHRH expression, since mice expressing a human GH transgene targeted to the hypothalamus exhibit markedly reduced expression of GHRH peptide and mRNA in the hypothalamus, of GH in the anterior pituitary gland, and of IGF-I in blood (15).
D. Feedback regulation of GH secretion
Somatostatin, GHRH, GH itself, and its nearly ubiquitous tissue
mediator, IGF-I, are maintained homeostatically in hypothalamic loci,
the pituitary gland, and the circulation by a complex interplay of
feedback signals involving the 4 (poly-) peptides themselves (see Fig. 7). So-called GH autofeedback can be demonstrated in normal subjects,
since GH injection reduces the subsequent GH secretory response to a
GHRH stimulus (257, 258). This inhibitory effect is observed as early
as 3 h after intraperitoneal GH administration before any rise in
the plasma total IGF-I concentration occurs (259, 260). GH autofeedback
can be relieved by pyridostigmine pretreatment (260), which likely acts
in part to reduce hypothalamic somatostatin secretion and increase GHRH
release (261, 262). Thus, such clinical experiments support the view
that GH autonegative feedback stimulates somatostatin release from the
hypothalamus in vivo. In addition, a less important direct
pituitary inhibitory effect and/or concomitant GHRH withdrawal in
response to a GH stimulus cannot be excluded. In the rat, gender
differences in GH autonegative feedback exist; there is continuing
responsiveness to repeated GHRH stimuli (less sensitivity to GH’s
negative feedback) in the female rat, but not in the male animal (see
below).
After cloning of the GH receptor gene, studies indicate that GH feeds back to suppress the hypothalamic expression of the GH receptor itself (263). It is noteworthy that GH receptor antisense administration centrally [intracerebroventricularly (icv)] in the rat augments spontaneous GH pulsatility and diminishes hypothalamic somatostatin expression. This supports a primary role of the GH receptor and secondarily of somatostatin in mediating physiological GH autonegative feedback (130).
In addition to the network-like feedback actions of GH on hypothalamic
somatostatin, GHRH, and the GH receptor in the rat, greater complexity
arises since hypothalamic GHRH and somatostatin can each negatively
regulate its own secretion and reciprocally control secretion of its
counterpart (Figs. 1 and 7). In particular, GHRH inhibits its own
release but increases somatostatin release in vitro (264).
Conversely, somatostatin inhibits its own secretion and that of GHRH
in vitro (131, 265). Intrahypothalamic interactions are also
indicated by intracerebroventricular administration of somatostatin or
GHRH to rats, which elicits, respectively, an increase or a decrease in
blood GH concentrations (266). Moreover, hypothalamic GHRH and
somatostatin neuronal systems are anatomically coupled (183),
e.g., somatostatin neurons from the periventricular nuclei
synapse on GHRH neurons in the arcuate nucleus (267). Thus, GHRH and
somatostatin may play opposing roles in the control of GH secretion not
only on the pituitary gland but also at the hypothalamic level by
acting as neuromodulators.
Presumptively, functional reciprocal intrahypothalamic linkages between GHRH and somatostatin release also make in vivo studies with GHRH or somatostatin antagonists more difficult to interpret. For example, in the human, a predominant GHRH antagonist with some agonist properties can inhibit spontaneous pulsatile GH secretion (251, 268), as well as that stimulated by various secretagogues, e.g., GHRH, sleep, insulin, pyridostigmine, and L-arginine (269). Such inhibition might be explicable not only if endogenous GHRH mediates the actions of these secretagogue, but also if the GHRH antagonist serves to block GHRH-stimulated somatostatin release.
Feedback actions of IGF-I (and IGF-II) also are inferred at the hypothalamic-pituitary levels in the rat and human (270). Although somewhat controversial (271), a dominant feedback action of IGF-I administered peripherally in the human, sheep, or rat probably occurs directly on the pituitary gland (62, 63, 64), whereas rapid feedback by GH itself is primarily central (except perhaps in the sheep) (272, 273). Evidence also exists for hypothalamic actions of IGF-I (271, 274, 275), and hence this issue is not fully established. In earlier studies in the rat, partially purified preparations of IGF-Is administered centrally initially suggested major inhibitory effects of IGF-I and IGF-II each at hypothalamic loci (270, 276). However, more recent experiments using recombinant human IGF-I or IGF-II indicate that either IGF-I or IGF-II acting alone is not strongly inhibitory of the GH axis when administered centrally or in hypothalamic cultures in vitro (275, 277, 278). Coexposure to IGF-I and -II suppresses the GH axis. Hypothalamically targeted IGF-I gene and/or IGF-I receptor antisense mRNA expression would likely help clarify this important conceptual issue.
GH autofeedback in the rat may also decrease GHRP receptor expression in the arcuate and ventromedial hypothalamic nuclei (279). Whereas a putative GHRP-like endogenous ligand has not been identified, its demonstration would allow the conjecture that the (endogenous) GHRP-pathway participates in GH autonegative feedback.
|
IV. Other GH-Regulating Neuropeptides |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
A. GHRPs
1. Introduction. Potent GH-releasing synthetic oligopeptides
(so-called GHRPs) were developed mostly in the laboratory of Bowers
(280, 281, 282) via conformational energy calculations, solid-phase
synthesis, and screening for biological activity. These compounds were
first synthesized in the early 1980s as enkephalin derivatives and
modified subsequently to nonpeptidyl analogs (283, 284, 285, 286, 287). Although
unproven, GHRP oligopeptides likely mimic endogenous effector molecules
of currently unknown structure and identity, since they release GH via
specific receptors expressed in the hypothalamus and pituitary gland
(below) in a dose-related and specific manner both in vivo
and in vitro in several species (280, 288, 289).
Furthermore, chronic administration of GHRP to rats, dogs, or humans
can promote an increase in body weight, GH secretion, and plasma IGF-I
or BP-3 concentrations (290, 291, 292, 293, 294, 295, 296, 297). The mechanisms of GHRP actions are
at least partially independent of those of GHRH, since the combined
effects of GHRP and GHRH in vivo (but not usually in
vitro) on GH release are typically synergistic at maximal
concentrations (282, 298). GHRPs also release in vivo
hypothalamic GHRH into portal blood in the sheep (299, 300), and act
directly in vitro on functional GHRP receptors in human
(fetal) pituitary gland (214). In the rat pituitary in
vitro, GHRP but not GHRH receptor antagonists block GHRP(-2)
actions (301). Available data thus indicate (plausibly) multiple sites
of GHRP actions (Refs. 289, 302, 303, 304 and see below).
In one study in the rat, estrogen (or aromatizable androgen) augmented the maximal effect of GHRP in stimulating GH release (305), and in older women oral estradiol treatment amplified the steepness of the GH secretory response (sensitivity change) to increasing doses of GHRP-2 (306), suggesting sex steroid regulation of this pathway. The maximal GH-secretory response to GHRP (hexarelin) rises 3-fold in mid-late puberal children and falls in elderly individuals (307). In contrast, postpartum hyperprolactinemia and Cushing’s syndrome are associated with markedly attenuated responsiveness to GHRP (308).
In the human, an intact hypothalamo-pituitary unit is required for
maximal GHRP actions (309, 310). Propranolol and clonidine (see
Section V.B on catecholamines) do not greatly modify
the GH-secretory response driven by GHRP (311). Combined GHRP and GHRH
administration typically evokes marked GH release in the human (312, 313), and GHRP plus GHRH may be a nearly age-independent test of GH
reserve (314). Most recently, novel GHRP mimetics, such as the
intranasally or orally active nonpeptide (spiropiperidine) analog,
MK-0677 (283, 315, 316, 317), can stimulate and maintain pulsatile GH
release and increase plasma IGF-I concentrations during oral treatment
over hours to weeks in young, older, and obese men and women (290),
thereby suggesting possible GHRP mimetic use in GH adjuvant therapies
(see Fig. 8). Intranasal GHRP stimulates
linear growth in children (318). Indeed, both injectable and
noninjectable (intranasal or oral) forms of GHRP agonists and mimetics
hold promise for enhancing GH release, even in certain conditions of GH
axis pathophysiology, e.g., obesity (296), starvation (297),
critical illness (319, 320, 321), or aging (290, 295, 322). Repeated GHRP or
mimetic administration in the human is only sparingly (or, in some
cases, seemingly not at all) down-regulating (290, 293, 296, 297, 323, 324, 325, 326, 327, 328, 329), unlike a proclivity to down-regulation in the rat (330).
|
Further studies on the mechanisms of action of GHRP compounds (283, 287, 324), as well as eventual isolation of the putative endogenous GHRP-like ligands, will clearly be important.
2. GHRP receptor(s) and actions. GHRPs constitute a distinguishable family of synthetic oligopeptides (e.g., tri-, penta-, hexa-, and heptapeptides) that act presumptively at the hypothalamic as well as (probably to a lesser degree) pituitary levels to drive GH release (340, 341, 342). No natural ligand is yet known, but the receptor family has been cloned (340, 342, 343, 344, 345, 346, 347). The type 1a (but not 1b) GHRP receptor is biologically active (345). Stimulation by these novel peptides does not absolutely require endogenous GHRH activity, since they remain active after pituitary desensitization to GHRH (348) and can stimulate GH secretion by cells not expressing the GHRH receptor, such as GH1 cells (349). GHRPs act negligibly on GH secretion in mice with an 80% reduction in somatotrope cell number associated with a mutated GHRH receptor, e.g., lit/lit mouse (289). However, in this mutant animal, the arcuate nucleus c-fos genomic response to GHRP is preserved, indicating that brainstem neuronal activation by GHRP does not require GH or, for that matter, the GHRH receptor (350). GHRP, unlike GHRH or PACAP-38, does not activate the pituitary-specific transcription factor, GHF-1/Pit-1, in rat anterior pituitary cells in vitro (351), further distinguishing its biochemical actions from those of GHRH.
The cloned receptor families for GHRP and GHRH are distinct (352). Purification and later cloning of the GHRP receptor in pig and rat anterior pituitary membranes revealed a magnesium-dependent, GTP-binding, non-GHRH, non-somatostatin receptor that is G-protein coupled and apparently mediates the action of multiple GHRPs, e.g., GHRP-6, hexarelin, and the nonpeptidyl (spiropiperidine) GHRP analog, MK-0677 (340, 342, 343, 353, 354, 355, 356). GHRP receptors estimated by binding and functional reactivity have high affinity [e.g., dissociation constant (Kd) = 0.7 nM for MK-0677 (356)], and are expressed in the normal pituitary gland, pituitary tumors, the hippocampus, and ventromedial and arcuate nuclei in the hypothalamus (214, 340, 342, 344, 345, 345). GH autonegative feedback reduces (and GH deficiency increases) hypothalamic (ventromedial and arcuate nuclear) as well as hippocampal GHRP receptor expression, thus suggesting multisite feedback control mechanisms in regulating the putative endogenous GHRP receptor system (279).
GHRP’s intracellular signaling mechanisms in the rat include the phospholipase C-phosphoinositide pathway (357, 358, 359). GHRP also stimulates delayed calcium influx in rat and sheep somatotropes (360, 361), activates protein kinase C (357, 362), heightens GHRH signaling (the latter via cAMP) (302), and depolarizes somatotrope cells (343, 363). The mechanisms of GHRP action are distinctly nonopiate and non-GHRH receptor dependent (280, 298, 302, 352, 364, 365). However, effects of GHRP are partially antagonized by somatostatin in the rat and human (339, 366, 367, 368, 369), by centrally (icv) infused somatostatin in the rat (370), and (partially) by a GHRH receptor antagonist in the rat and human (371), but not by an opiate receptor antagonist in the human (372). GHRP peptides are active in vitro and in vivo in multiple species, such as the rat, monkey, human, sheep, cow, and chicken (214, 280, 298, 373, 374, 375, 376, 377, 378, 379).
Acute intravenous injection of GHRP-6, GHRP-2, hexarelin, or a nonpeptidyl GHRP mimetic (L-692, 429) rapidly increases serum GH concentrations in humans within 5–15 min, with the peak GH concentration usually observed 15–30 min after infusion (298, 313, 332, 366, 380, 381). The amount of GH released after GHRP is much larger than that secreted after GHRH (313, 332). Maximally effective doses of GHRP-6 and GHRH, when injected simultaneously in man, typically stimulate GH secretion additively or synergistically (298, 382). Both GHRH and GHRPs administered nocturnally individually facilitate slow-wave sleep (332, 336, 383, 384, 385) via mechanisms that are not yet established (289, 336). The effects of GHRH on sleep decline with age (337, 386, 387) and are especially evident after pulsatile infusion (338, 388). Similar data are not yet available for possible age-related actions of GHRP on sheep. Daytime GHRH treatment does not modify nighttime sleep or GH release (389), but sleep deprivation may limit GHRH-stimulated nocturnal GH release (390). This may reflect the increase in somatostatin evoked by sleep deprivation, at least in the rat (391). Analogous data are not yet available regarding GHRP actions or receptors in sleep-deprived animals.
Prolonged infusion of GHRP amplifies pulsatile GH secretion over 24 h in men by increasing GH pulse amplitude (323, 392). Partial attenuation of GH release in response to a subsequent GHRP bolus occurs, and is not caused by absolute depletion of pituitary GH stores, since the effect of GHRH persists at this time. Conversely, desensitization to GHRH does not eliminate responsiveness to GHRP, further indicating the capacity for independent actions of these peptide secretagogues (348).
The synergy between GHRH and GHRP-6 is not enhanced by pyridostigmine, suggesting that somatostatin withdrawal may participate in their synergy (393), or that pyridostigmine also acts via stimulating GHRH release, as suggested in the sheep (299). Similarly, hexarelin and GHRH do not synergize in another human model (type I diabetes), in which somatostatin tone is putatively reduced (394). In combination with GHRH or L-arginine, GHRP evokes substantial GH release in aging animals and humans (314, 379), and combined GHRP and GHRH stimulation will elicit marked GH secretion even in obese adults (312). Thus, GHRPs constitute among the most effective clinical GH secretagogues known, whether administered alone or in combination with physiological or other pharmacological stimuli of GH release.
The effects of GHRP, albeit demonstrably distinct from those of GHRH (374), are also influenced by GH autonegative feedback [see Section III.D]. For reasons not yet established, GHRP’s action may be affected to a lesser degree than that of GHRH by GH feedback inhibition (395, 396). In the rat, GH autofeedback reduces arcuate and ventromedial nuclear GHRP receptor expression (279), thus offering a speculative linkage between this (putative) endogenous effector pathway and GH autofeedback.
3. Mechanisms of neuroregulation by GHRPs. Several plausible
mechanisms of GHRP action have been considered by the laboratory of
Bowers and others, including, for example 1) a direct stimulatory
action on the anterior pituitary gland, although presumptively
relatively minor (281, 340, 349, 353, 397); 2) indirect actions on the
hypothalamus causing release of GHRH (and, less plausibly, a decrease
in central somatostatin activity), the former being relevant in the
sheep and adult rat (299, 300, 340, 398), and possibly human (291, 303); and 3) possible release of an unknown ("U") hypothalamic
factor, which in both the human and rat is postulated to elicit GH
release by acting synergistically with GHRH (302, 399). GHRP may also
reduce somatostatin’s feedback inhibition of GHRH neurons and thus
attenuate GH autofeedback (400). Figure 6 summarizes these plausible
sites of GHRP actions. We caution that what has been inferred from
intravenous or even icv injections of GHRPs may not fully predict the
ultimately identified (if any) activities of endogenous GHRP receptor
ligands. Natural effectors might act inside or outside the blood-brain
barrier at very high local concentrations to exert important (new)
autocrine and/or paracrine effects in the hypothalamus, pituitary
gland, and/or other sites of GHRP reception.
In three of four studies in the adult animal, and in two studies in rat pups, GHRP-6 stimulated GH secretion in vivo even in the presence of antiserum to GHRH, or somatostatin, or both (350, 398, 400, 401, 402, 403). GHRP-6 or mimetics likewise can enhance pituitary GH gene expression, apparently independently of GHRH and/or somatostatin, and directly stimulate GH release in vitro from incubated pituitary glands and cultured rat anterior pituitary cells (284, 301, 341, 350, 352, 357, 358, 367, 404, 405, 406, 407). However, the in vitro stimulatory effect of GHRP on either rat or human dispersed GH-secreting pituitary cells is often somewhat diminutive compared with that of GHRH, except in a recent study of GH-secreting tumors (367). GHRP action in vitro also is typically much less than that observed after in vivo GHRP administration (214, 349). Of note, early studies utilizing in vitro incubations of intact pituitary glands from immature rats showed larger (6- to 10-fold) stimulatory effects of GHRPs on GH release than were later observed in dispersed (monolayer) pituitary cell culture studies derived from mature animals (1.5- to 4-fold effects). At the single-cell level, GHRP-6 acts on both GHRH-responsive and nonresponsive somatotroph subpopulations by increasing the percentage of GH-secreting cells as well as the amount of GH secreted per cell (404). The foregoing experiments collectively suggest that, at least in the (adult) rodent, enhancement of GHRH, as well as, conversely, inhibition of (central) somatostatin, action play modulatory roles in the GH-stimulating properties of GHRPs. However, these roles are only partial, since GHRPs effectively stimulate GH secretion directly in vitro (above), and in three of four studies somatostatin antiserum pretreatment actually augmented the GH response to GHRP in both immature female and adult male rats (400, 401, 402). Augmentation of the action of GHRPs by experimental neutralization of endogenous somatostatin supports strongly non-somatostatin-dependent actions of GHRPs, which are otherwise impeded partially by (endogenous) somatostatin. Indeed, in the male rodent, icv somatostatin administration limits GHRP actions (370).
The actions of GHRP on the CNS are quite likely to be relevant to its
overall effects (408). Brain actions are suggested by the presence of
the GHRP receptor in the arcuate and ventromedial nuclei of the
hypothalamus and in the hippocampus, and by the ability of GHRP to
induce GHRH mRNA expression in the arcuate nucleus even in GH- (and
GHRH receptor)-deficient mice (279, 289, 409). Other CNS actions of
GHRPs are also apparent in the rat, since GHRP stimulates eating
behavior after icv infusion without altering GH release (410),
increases electrophysiological activity in the arcuate nucleus, and
induces brainstem c-fos mRNA expression in NPY and
GHRH-releasing neurons (411, 412, 413). These effects are not mimicked by
GHRH (408). In the conscious sheep, GHRP acutely releases GH in
association with an increased frequency of pulsatile GHRH release (with
no apparent changes in somatostatin secretion) into the hypophysial
portal circulation (299, 300). In the monkey, indirect studies suggest
that GHRH’s involvement may be less conspicuous (374), whereas in the
human recent [but not earlier (414)] studies with a GHRH receptor
antagonist indicate a major (85%) dependence of GHRP-6’s
stimulation of GH secretion on endogenous GHRH actions (415).
In the rat, sex hormones can increase the effect of GHRP in both males
and females (305). In the human, the (single-dose) maximal GHRP-6
effect in early studies was relatively independent of age, gender, or
menstrual cycle stage (416). However, dose-responsive sensitivity of GH
release to GHRP has not been appraised in detail as a function of the
sex-steroid milieu in the human. With respect to aging when sex
steroids decline, GHRP augments the effect of GHRH in both young and
old dogs and potentiates the effects of clonidine, an -2
agonist, at least in young dogs (417).
In the human, the potential relative involvement of the hypothalamus
vis-a-vis the pituitary gland in the GH-releasing action of
GHRP-6 has been assessed using different pathophysiological and
pharmacological strategies (418). Direct pituitary effects have been
evaluated by injecting GHRP in patients with pituitary stalk section
(309, 419) or acromegaly (367) and by in vitro studies
(214). Both of the foregoing categories of patients are presumed to
have a pituitary gland that is partially or completely disconnected
from the hypothalamus, and both show significant, albeit variable,
attenuation of GHRP effects. In acromegaly, "functional"
disconnection likely reflects tumoral autonomy. The potential role of
somatostatin has been appraised in chronically glucocorticoid-treated
men (382), who are considered to have high somatostatin tone (171) and,
conversely, in type 1 diabetic patients, with putatively low
somatostatin tone (420). These clinical paradigms indicate that GHRP-6
can increase the GH response to GHRH in glucocorticoid-treated patients
(382), but to a lesser degree in type 1 diabetics (394). Such
observations suggest indirectly that inhibition of endogenous
somatostatin tone may play only a subordinate role in GHRP’s
stimulatory actions in humans. Conversely, the involvement of
endogenous GHRH in GHRP’s actions in humans was restudied recently via
concurrent GHRH antagonist administration, which significantly
(85%) attenuated the GH-secretory response to GHRP (294). Thus, in
the human, available evidence supports at least the importance of
(hypothalamic) GHRH release in GHRP’s effects (294). The marked
enhancement of GH secretion by combined GHRH and GHRP administration
has led to the suggestion that these joint stimuli afford near-maximal
stimulation of somatotrophs and, when used together, offer a plausible
test for GH deficiency in obesity, children, and older adults (295, 314, 350, 421, 422, 423, 424, 425). These suggestions will require further validation
in larger diverse normal and hypopituitary populations.
B. Galanin
Galanin is a 29-amino acid peptide initially isolated from porcine
small intestine (426). Galanin-like immunoreactivity is widely
distributed in the central and peripheral nervous system (427). The
hypothalamus is particularly rich in cell bodies and fibers containing
galanin-like immunoreactivity, with the highest concentration in the
median eminence (427). Specific binding sites for galanin are also
demonstrable in the mediobasal hypothalamus (428). Taken together,
these observations suggest that galanin may have a role in regulating
anterior pituitary function.
The molecular sequences of porcine, rat and, recently, human galanin have been deduced (429). The three peptide forms are identical with respect to their first 15 residues, but differ at several positions in the C-terminal part (430). Virtually all human experiments have been conducted with porcine galanin, with fewer studies performed using rat (430, 431) or human (432) galanin. Rat galanin, the structure of which is more similar than porcine galanin to that of human peptide, is able to cause significantly greater GH release than porcine galanin in normal humans (430). Therefore, possible species-dependent differences in galanin action should be considered in the interpretation of experiments concerning the physiological role of galanin in man (433).
In the rat, galanin increases GH release when given icv (434, 435, 436), subcutaneously (437), or intravenously (434, 436). Although galanin’s interaction with GHRH is not fully defined, GHRH-overexpressing transgenic mice show 7-fold and 4-fold increases, respectively, in galanin peptide and mRNA expression in the pituitary gland (438).
In the human, both porcine and human galanin elicit GH secretion when
given alone (432, 439) and facilitate the GH-secretory response to
GH-releasing hormone (GHRH) in normal young men (434, 440, 441) (Fig. 9). Although the mechanism underlying
this action of galanin is unknown, evidence in the rat suggests that
galanin may act at the hypothalamic level, as the peptide was effective
when injected into the third ventricle, but does not stimulate
pituitary cells in vitro (435, 436). An interaction between
galanin and GHRH has been proposed, since the treatment of male rats
with GHRH antiserum markedly inhibits the GH response to either
intravenous or central galanin administration (434). The demonstration
of coexistence of galanin and GHRH in the same neurons in the arcuate
nucleus (232) further suggests possible interactions between the two
neuropeptides. Other authors have hypothesized that galanin may also
inhibit somatostatin release from the hypothalamus (442). However,
porcine galanin fails to restore normal GH secretion in various
pathophysiological conditions thought to be characterized by increased
somatostatin tone in man (441, 443, 444). Galanin infusion stimulates
GH secretion more in women than in men, and peak serum GH responses in
the female are proportional to the blood estradiol concentration (445),
consistent with a hypothesis of gender and sex hormone modulation of
galanin action. In this regard, a functional estrogen-response element
has recently been identified within the human galanin gene promoter
(446).
|
Immunoneutralization of endogenous galanin within the CNS significantly disrupts the normal GH-secretory pattern in male rats (448). These results suggest an important physiological role for endogenous galanin in the control of spontaneous pulsatile GH secretion in this species and gender. The major impairment in GH secretion induced by galanin-antiserum injection involved a severe reduction in GH pulse amplitude, a parameter that appears to be critical for optimal growth-promoting actions of GH. Other alterations of the GH pulses include an increase in pulse frequency and a loss of the normal 3-h periodicity. Although the net effect on integrated GH secretion of galanin-passive neutralization is inhibitory, this inhibition is only partial, suggesting that GH secretion can still be maintained by other neuromodulators (448). Alternatively, in these studies, neutralization of endogenous rat galanin within the brain might have been incomplete, e.g., due to imperfect intrahypothalamic penetrance, and/or since the antiserum was raised against the porcine peptide, and the sequences of the two heterologous peptides differ by three amino acids at the C-terminal end (430).
More recent studies have characterized the expression and distribution of the human brain galanin receptor, as well as its molecular and biochemical properties. The human galanin receptor has a size comparable to that of other mammalian galanin receptors and shares trans-species properties such as coupling to a GTP-binding protein and ligand specificity (449). Its abundant expression in human brain suggests an important function of galanin as a neuromodulator in the CNS.
Knockout (450) and transgenic (451) mice deficient in or overexpressing the galanin gene have recently been developed, but data on GH neuroregulation in these animal models are not yet available.
C. Calcitonin
Calcitonin is a 32-amino acid peptide produced in the medullary or
C cells of the thyroid gland, which inhibits osteoclastic activity and
decreases renal clearance of calcium and phosphate (452). Calcitonin
imposes several inhibitory endocrine effects; namely, it inhibits
gastrin secretion (453), basal and stimulated insulin release (454),
and pituitary TSH- and LH-secretory responses to TRH and GnRH,
respectively (455). Moreover, calcitonin’s possible neuroendocrine
role (452) is suggested by the finding of calcitonin immunoreactivity
in the human CNS (456) and by localization of the calcitonin receptor
in the human hypothalamus (456).
In the rat, icv, but not intravenous, administration of salmon calcitonin diminishes 24-h GH secretion and GH responses to GHRH (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29)NH2 (457). In man, intramuscular salmon calcitonin injection blunts GH responses to hypothalamic stimuli such as L-arginine (458) and insulin-induced hypoglycemia (459). Intranasal salmon calcitonin administration also attenuates the pituitary GH response to the (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29)NH2 fragment of human GHRH (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44) in men (460). Moreover, salmon calcitonin inhibits the GH-secretory response to GHRH, even when humans are pretreated with the putative functional somatostatin antagonist, pyridostigmine (461). Therefore, calcitonin may not inhibit the GH response to GHRH in normal humans via stimulating somatostatin release. Alternatively, pyridostigmine may also release GHRH in man as in sheep (299), which would allow for the possibility that calcitonin’s mechanism of inhibition embraces blockade of GHRH action.
Calcitonin’s inhibition of GH secretion in humans could be modulated by several other factors. Acute large increases in serum cortisol concentrations impede the GH response to GHRH (462). A slight rise in serum cortisol levels occurs after calcitonin administration (463). However, cortisol’s acute inhibitory action on GH secretion is blocked by pyridostigmine (464), which is unlike the case for calcitonin action. An alteration in extra- and/or intracellular calcium ion levels in pituitary cells also may be involved in the inhibitory action of calcitonin (465). This may be relevant, since calcium ions participate in mediating the stimulatory effect of GHRH on GH release (466). Thus, available data suggest that calcitonin inhibits GHRH actions but not via somatostatin. Interactions between calcitonin and GHRP have not yet, to our knowledge, been evaluated.
D. PACAP
Pituitary adenylate cyclase-activating polypeptide consists of 38
amino acids that are C-terminally amidated (PACAP 38). This peptide was
first isolated from ovine hypothalamus, based on to its ability to
enhance cAMP accumulation in anterior pituitary cells (467). PACAP-like
immunoreactivity is widely distributed in the CNS and highly
concentrated in the hypothalamus (468). Histochemical studies have
revealed dense PACAP-positive fibers in the internal and external
layers of the median eminence and in the pituitary stalk (469). The
cell bodies are found in the magnocellular region of the
paraventricular nuclei and the suprachiasmatic nuclei (470). Dow
et al. (471) reported that PACAP-like immunoreactivity is
detectable in the hypophysial-portal plasma of
pentobarbital-anesthesized rats in concentrations of 50–100
pM, which are significantly higher than those in the
peripheral plasma. These findings suggest that hypothalamic PACAPs,
mainly PACAP-38, can be released into the hypophysial portal vessels
and reach the anterior pituitary, where they might play a physiological
agonistic role.
Other indirect evidence further suggests that PACAP plays a role in promoting anterior pituitary GH secretion. PACAP stimulates GH release from the rat pituitary gland both in vivo and in vitro (472, 473, 474). The variable increase in GH release in vitro after stimulation with PACAP-38 probably reflects nonuniform experimental procedures (475). The peptide also stimulates pituitary hormone release from rodent clonal pituitary cell lines such as ATT-20 and GH3 cells (474). Hormone release from tumoral GH3 cells is stimulated by PACAP through the type II PACAP receptor (476), which is common to PACAP and vasoactive intestinal polypeptide (VIP). In contrast, GH release from normal rat somatotropes is stimulated by PACAP, but not by VIP, suggesting that the effect is mediated by a PACAP-specific (type I) receptor (477). The action of PACAP on rat somatotropes is also distinct from that of GHRH (473). Thus, a synthetic GHRH antagonist fails to impede PACAP-induced GH release in pituitary cell culture and perifusion experiments.
PACAP-38 is a potent secretagogue of GH in conscious male rats in vivo. In mechanistic studies, Jarry et al. (472) found that PACAP-38 increases plasma GH in hypothalamus-lesioned anesthetized rats. However, the pulsatile pattern of GH release was not affected by a PACAP antagonist (475). If confirmed, the latter observation (assuming adequate effective antagonist action at relevant sites) would speak against a primary physiological role of endogenous PACAP in stimulating spontaneous pulsatile GH secretion in the rat. The role of PACAP in the regulation of GH secretion in man also remains to be elucidated.
E. Opioid peptides
Opioid peptides stimulate GH secretion in both rodents and humans
(478, 479). Opiates can directly inhibit rat mediobasal hypothalamic
release of somatostatin in vitro (480, 481). Opiate effects
also can be mediated via GHRH, since rats pretreated with GHRH
antibodies fail to show the expected opioid-mediated rise in GH levels
(479). While there is no clinical evidence at present to suggest that
endogenous opiates, e.g., acting through naloxone-sensitive
receptors, play a major role in the control of spontaneous basal GH
release, they may participate in some forms of stress-stimulated GH
secretion, such as marathon running and calorie restriction (482). In
healthy individuals, naloxone and naltrexone (opiate receptor
antagonists) exert few acute effects on GH secretion (482).
F. TRH
In the rat, TRH act as a physiological GH secretagogue (483). In
contrast, in the human, TRH stimulates GH secretion only in certain
experimental and pathophysiological conditions, such as acromegaly
(484), type 1 diabetes mellitus (485), and hepatic and renal failure
(486, 487), but not usually in normal unmedicated subjects. Conversely,
GH responses to L-dopa, arginine, and insulin-induced
hypoglycemia (488), but not to GHRH (489), are reduced or blocked
during TRH administration. In normal subjects these divergent effects
of TRH are probably due to two different loci of actions of TRH. TRH
can directly elicit GH secretion from pituitary cells of hypothyroid
rats in vitro (490) and can act via specific, high-affinity,
low-capacity TRH receptors on GH3 and GH1 cells
(491, 492). TRH receptors are thyroid hormone inhibited and
consequently increase (reciprocally) in hypothyroidism. This might
explain the paradoxical GH response to TRH in hypothyroidism (493),
despite the well described reduction in GH responses to other
secretagogues including GHRH in this condition (494). Conversely, the
inhibitory actions of TRH on secretagogue-stimulated GH release are
likely exerted at the hypothalamic level via enhanced somatostatin
release.
TRH paradoxically stimulates GH secretion in type 1 diabetes (484), a condition characterized by low somatostatin tone (see below), when the direct stimulating effect of TRH at the pituitary level may be unmasked. Women receiving 30 µg ethinyl estradiol and a synthetic progestin daily for 6–24 months (495) also show paradoxical GH release after TRH injection, but the mechanism of this acquired responsiveness to TRH is unknown. In another study in men and midluteal phase women, histamine pretreatment unmasked TRH/GnRH-stimulated GH release (496). The mechanism underlying this phenomenon has also not been explained. Overall, in harmonizing the discrepant available data, one can hypothesize that in the human a direct stimulatory effect of TRH that is expected (based on rat data) at the pituitary level can be overcome by TRH-mediated somatostatin release (the latter likely does not occur in the rat).
G. NPY
Studies a decade ago indicated that the orexigenic peptide, NPY,
can inhibit GH secretion in the male and female rat (497, 498, 499). In
particular, in the intact male rat, as well as in the ovariectomized
female rat (with or without sex steroid hormone replacement), icv
infusion of NPY reduces GH secretion, and antibody to NPY increases GH
release in these animals. The suppressive action of NPY appears to be
mediated by both NPY-1 and NPY-2 receptor subtypes expressed in the
mediobasal hypothalamus (500). Most plausibly, NPY stimulates
hypothalamic somatostatin release and thereby inhibits GH secretion, at
least in the male rat (498). In addition, NPY may act, in part,
directly on the anterior pituitary gland, since NPY will reduce human
somatotroph tumor secretion of GH in vitro (501) and can
limit somatotroph cell proliferation in the rat, possibly by way of
gonadotroph-dependent paracrine mechanisms (502). In contrast to
studies in the rodent, in a single clinical study, infusion of NPY
paradoxically evoked GH secretion in 60% of a small group (n =
15) of patients with prolactinomas (503). NPY also stimulated GH
secretion directly from goldfish pituitary in vitro (504).
Thus, the preponderance of evidence favors an inhibitory role of NPY on
the GH axis in the rat, with possible species differences,
e.g., stimulation of GH secretion in the human.
NPY may participate in the negative-feedback actions of GH on its own secretion in the rat (172, 502). For example, GH receptors are expressed on NPY neurons in the arcuate nucleus (174). GH treatment increases c-fos gene expression in NPY neurons (505) and stimulates hypothalamic NPY mRNA expression in hypophysectomized rats (506). Such observations sugggest the thesis that NPY may enlist somatostatin release during GH’s negative feedback and/or limit GHRH release in this autofeedback context. Further definitive studies with NPY neutralization (e.g., both immunological and via antisense molecular methods, etc.) will be required to establish or refute this hypothesis definitively.
The activation of GH secretion in the sheep, and inhibition of GH secretion in the rat, by nutrient restriction might also be mediated in part via NPY. In the ovariectomized ewe, NPY immunoreactivity is increased in the arcuate nucleus by prolonged food withdrawal and accompanying weight loss, which resultant selective amplification of GH secretion and, in this particular paradigm, preservation of secretory activity of the reproductive axis (507). Akin to fasting, neuronal glucoprivation via a metabolic inhibitor of intracellular glucose utilization increases c-fos gene expression in NPY neurons in the rat (508). Such studies suggest the hypothesis that NPY participates in mediating the impact of dietary manipulation on both the GH and reproductive axes, both of which are suppressed in the rat by fasting, as reviewed elsewhere (509). However, considerable additional study will be required to pinpoint the species specificity and neurohormonal mechanisms of NPY’s actions in various nutritionally modified contexts, e.g., how fasting suppresses both LH and GH in the rat, but stimulates GH while suppressing LH in the human and sheep. How NPY might integrate such responses in different species is not yet evident.
As noted in Section IV.A, NPY may also participate in the actions of non-GHRH secretagogues. For example, GHRP-6 in the rat increases c-fos expression in 51% of NPY neurons and 23% of GHRH neurons (413). Additional studies will be required to assess the exact role of NPY in modulating GHRP‘s actions, and to define whether NPY neurons themselves express the receptor for GHRP.
In summary, NPY inhibits GH secretion in the male and female rat, possibly by stimulating hypothalamic somatostatin release, and likely participates in GH autonegative feedback in this species. In addition, in rodents and ruminants, NPY may have a role in mediating nutritional effects on the GH and/or reproductive axes. The ability of the nutritional regulatory peptide signal, leptin, to stimulate NPY neurons in the rat is consistent with this theme, although the acquisition of more detailed knowledge of interactions among leptin, GH, and NPY will be essential (see Section VI.B below).
H. Substance P
In male rats, a high percentage of somatotropes colocalize
immunoreactive substance P and GH. Coexpression of GH and substance P
falls in estrogen-treated ovariectomized (female) rats (510).
Additional studies are required to clarify the role, if any, of
substance P in neuroregulation of the human GH axis, although this
agent will enhance basal and GHRH-stimulated GH release in normal men
(511).
I. Bombesin
In the male rat, bombesin inhibits GH secretion, apparently (based
on antibody studies) independently of somatostatin (512). In the female
rat, bombesin stimulates GH release via mechanisms that are partially
antagonized by GHRH antiserum (513). Limited clinical data are
available indicating this peptide’s absence of an effect on basal GH
release and inhibitory impact on the hypoglycemia-stimulated human GH
axis in young male volunteers (514, 515). Neuromedin C is a
bombesin-like peptide that stimulates GH release from perifused (rat)
pituitary cells (513). Little, if anything, is known about neuromedin
C’s actions on the somatotropic axis in the human.
J. Melatonin
The pineal gland, via melatonin, may also modulate GH secretion.
Oral administration of melatonin to normal subjects increases basal GH
levels and the GH response to GHRH (516), but marginally affects GH
responses to hypoglycemia or apomorphine (517, 518). This suggests that
melatonin might play a minimal (stimulatory) role in baseline GH
secretion, possibly acting at the hypothalamic level via inhibition of
somatostatin. The GH response to L-dopa is reduced in blind
human subjects (519), who presumptively lack both light-mediated
inhibition of melatonin release and the normal slow wave
sleep-associated rise in plasma GH concentrations (520). Thus, it is
possible, but entirely unproven, that variations in the release of
endogenous melatonin could modulate GH secretion in humans.
K. Other GH secretagogues
The cytokine, interleukin-1, stimulates pituitary GH release at
the single-cell level by increasing both the number of somatotropes
secreting GH and the amount of GH secreted per cell (521).
In the fasted male rat, leptin infusion increases GH secretion (522).
However, in estrogen-unreplaced postmenopausal women with varying
degrees of age-related sarcopenia, 24-h serum leptin concentrations
correlate inversely with daily GH secretion rates, GH pulse amplitude,
and mean serum GH concentrations (523) (see Fig. 10). An inverse correlation also exists
for leptin and GH in fed and fasted midluteal-phase young women, in
whom fasting increases GH release and decreases leptin levels over
24 h (524). Thus, akin to the opposite directional responses of
the GH axis to nutritional stressors (e.g., food
deprivation) or diabetes mellitus in the two species, rat and human,
the leptin-GH relationships may also be (inexplicably) species specific
(see Section VI.B).
|
|
V. Neurotransmitter Regulation of the GH Axis |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
). However, due to the lack of highly specific and
nontoxic pharmacological probes (agonists and antagonists) for most of
these neurotransmitters and/or their receptors, as well as the scarcity
of available experimental data in the human, convincing clinical
evidence of major GH-regulating roles exists principally for
acetylcholine and catecholamines, with evident but less compelling
roles to date for serotonin, 
-aminobutyric acid (GABA), histamine,
etc., in GH neuroregulation.
A. Interspecies differences
Species nonuniformities in the neuroregulation of the GH axis have
a potential to confound interpretations in this field. Most molecular
and biochemical studies have been carried out in the adult male rat.
Substantial data also exist in sheep and, to a lesser extent, in the
human. Additional (but considerably fewer) details are available in the
nonhuman primate, dog, mouse, rabbit, hamster, avian species, hedgehog,
goat, goldfish, etc. However, even among the rodent, sheep, and human,
several conspicuous neuroregulatory differences are identifiable (see
summary in Table 4).
|
). Conceptually, the interspecies disparities in the rat and sheep can be summarized as reflecting 1) greater somatostatin dependence in the rat vs. greater GHRH dependence in the sheep (59); 2) more complex and variable GHRH/somatostatin relationships preceding a GH pulse in the sheep; 3) stress inhibition of GH release in the rat vs. stress stimulation in the sheep and human; and 4) a tendency for primarily amplitude-dependent regulation of pulsatile GH release in the sheep (and human) vs. combined amplitude and frequency modulation in the rat. These distinctions should be kept in mind, and the temptation avoided to extrapolate easily across species (59, 60, 183).
B. Acetylcholine and catecholamines
1. Acetylcholine. Cholinergic muscarinic pathways play an
important role in activating pulsatile GH secretion and, presumptively,
do so primarily via hypothalamic somatostatin withdrawal (532, 533, 534).
Indirect muscarinic agonists stimulate GH release acutely and enhance
GH responses to GHRH in the rat, dog, sheep, and human (206, 535, 536, 537).
For example, in clinical experiments, the indirect cholinergic agonist,
pyridostigmine, administered orally repeatedly over 48 h doubles
the daily pulsatile mass of GH secreted in healthy young and older men,
but evokes lesser GH release in obese subjects (261). Conversely,
antagonism of endogenous cholinergic pathways with muscarinic
receptor-blocking drugs, such as methscopolamine, atropine, or
pirenzepine, strikingly reduces sleep-associated GH release (538, 539)
and virtually abolishes GH secretion otherwise triggered by various
secretogogues including GHRH (535, 540, 541). An important exception is
that the GH response to insulin-induced hypoglycemia is spared
significantly during cholinergic muscarinic receptor blockade, although
the neuroendocrine mechanism of this interaction remains to be
elucidated (542). In the dog, pirenzapine (a muscarinic antagonist)
inhibited GHRH-stimulated GH release in both fed and fasted animals,
whereas somatostatin inhibited this response preferentially in the fed
state (543). Such findings in the dog tend to speak against the facile
somatostatin hypothesis (as developed in the rat) of muscarinic
receptor-mediated withdrawal of somtatostatin and suggest secondary
effects of such drugs on GHRH release and/or a plurality of effects on
two or more functionally distinct receptor or neuronal populations
(543, 544, 545).
The inferred mechanism of the GH-releasing action of acetylcholine is based on direct and indirect in vitro and indirect in vivo experimental evidence in the rat (532, 533, 534, 546). Acetylcholine can decrease somatostatin release directly in vitro from rodent hypothalamic slices (547), although other experiments have contradicted this observation in the same species (546, 548).
In the rat, pyridostigmine administration stimulates GH release in the food-deprived animal (262), which is known to have increased hypothalamic somatostatin tone (534, 549). Of mechanistic relevance, pyridostigmine’s effect is diminished by pretreatment of the animal with somatostatin antibodies (262). Pyridostigmine counteracts the inhibitory effect of glucocorticoid treatment for 4 days on the GH response to GHRH in normal male rats, which treatment also is thought to increase somatostatin tone (550). In support of the foregoing somatostatin hypothesis, the inhibitory influence of atropine (cholinergic antagonist) on GH responses to GHRH in normal rats is abolished by pretreatment with somatostatin antiserum (551).
A recent hypophysial portal catheterization study showed that neostigmine, another indirect cholinergic agonist, stimulates GHRH without inhibiting somatostatin release in the sheep (531, 546). These ostensible species differences have not been fully rationalized; however, a plausible notion is that, due to intrahypothalamic somatostatin-GHRH neuronal interconnectivity, central somatostatin withdrawal induced by this drug elicits rebound GHRH release (261, 544). This notion would bring several divergent observations into agreement.
In humans, pyridostigmine partially restores the GH response to GHRH in
physiological states hypothesized to reflect increased hypothalamic
somatostatin release, i.e., 2-h repeated bolus
intraperitoneal GHRH injections (536), as well as GHRH stimulation
given 3 h after a bolus intravenous infusion of recombinant human
GH (259, 260). Moreover, pyridostigmine alleviates the acute inhibitory
effect of glucocorticoid treatment on the GH- secretory response to
GHRH in normal adults (465), and partially counteracts inhibition of GH
by glucocorticoid administration in children (552). These findings are
relevant to the hypothesis that pyridostigmine acts by inhibiting
somatostatin release, since, both in the rat and human, glucocorticoids
appear to increase hypothalamic somatostatin secretion (171) (as
discussed below). Conversely, in human GH hypersecretory states
characterized by putatively low somatostatin tone, e.g.,
type 1 diabetes mellitus, pyridostigmine does not further enhance
exaggerated GH release after GHRH injection (420, 553) (Fig. 11). The relatively good clinical
tolerability and the favorable pharmacodynamic profile of oral
pyridostigmine (compared with some other cholinergic agents) have
favored the administration of this substance, together with GHRH, in
numerous clinical investigations of the somatotropic axis (421).
|
Muscarinic and nicotinic cholinergic pathways may play opposing regulatory roles (facilitative and inhibitory, respectively), since nicotinic receptor blockers enhance GH release during insulin-induced hypoglycemia or in response to sleep (558).
2. Catecholamines.
a. Dopaminergic pathways.
In vitro dopamine reduces
human pituitary GH release stimulated by GHRH (559, 560). Since
dopamine receptors are present on human somatotropes, this in
vitro inhibitory effect of dopamine is likely exerted directly.
However, most human in vitro experimental data derive from
the study of GH-secreting pituitary adenomas; notably, acromegalics
show a paradoxical inhibitory response to L-dopa in
vivo compared with normal subjects. Specifically, in acromegalic
patients, acute or chronic dopamine agonist treatment inhibits GH
secretion (560), whereas in normal subjects acute administration of
dopamine agonists such as L-dopa, apomorphine, dopamine
itself, and bromocriptine causes GH release (561, 562). The latter
results are not in accord with most direct studies of dopamine’s (D2)
stimulation of somatostatin secretion in the rat (see Fig. 1A) (548, 563, 564, 565). Analagously, in bovine brain explants, in vitro
dopamine acting via D1 receptors increases hypothalamic somatostatin
secretion and reduces GHRH release (566). This may reflect species
differences compared with the human.
Estrogen treatment in one clinical study enhanced the GH-releasing
effect of L-dopa in girls with Turner’s syndrome (567). In
prepubertal children, a single oral dose of 20–40 µg ethinyl
estradiol augmented L-dopa-stimulated GH release (568).
Analogously, and consistent with somatostatin withdrawal in the human,
significant increases in the GH response to GHRH occur after
pretreatment with bromocriptine (569, 570). In contrast, dopamine and
dopaminergic drugs fail to enhance the GH-secretory responses to
insulin-induced hypoglycemia and L-arginine (571, 572),
both of which are thought to limit somatostatin release. Thus,
dopaminergic stimulation of GH release in the human is estrogen
sensitive and appears to facilitate GHRH action possibly via
somatostatin withdrawal (573). Somatostatin withdrawal per
se may evoke or accentuate GHRH release (22, 129, 130, 131, 574), which
mechanistically likely reflects relief of inhibitory somatostatinergic
synapses on GHRH-secretory neurons (Fig. 7). Thus, the rat and human
(Fig. 1, A and B) responses to dopamine are apparently opposite, akin
to the major species differences recognized in fasting, diabetes,
stress, and hypoglycemia (Table 4).
Antidopaminergic drugs seem to exert disparate effects on baseline vs. exogenous GHRH-stimulated GH secretion in normal humans. For example, whereas metoclopramide failed to increase baseline GH release in young adults (575), this dopamine antagonist increased (576) or did not alter (577) the GH response to GHRH. Significant differences in the GH response to metoclopramide are observed in men and women (578), suggesting a sex difference in dopaminergic regulation. Understanding the relative roles of dopamine receptor subtypes and their differential topographies in modulating GH release spontaneously and in response to various secretagogues will require more detailed clinical study.
b. -Adrenergic pathways.
Intravenous infusion of
phentolamine, a nonspecific 1- and
2-receptor-blocking drug, reduces the GH response to
many stimuli in humans and the rat, such as insulin-induced
hypoglycemia (579) and GHRH (580, 581). In contrast, prazosin, an
1-selective blocker, does not inhibit GH secretion
effectively. Various 1-agonist drugs (unlike
2, below) also do not significantly influence basal or
insulin-stimulated GH secretion in the human (579). On the other hand,
topographically distinct (afferent) adrenergic systems subserve
opposite 1-adrenergic effects in the sheep, with locus
coeruleus activation serving to stimulate, and paraventricular nucleus
stimulation serving to inhibit, GH secretion (184), thus illustrating
the hypothalamic regional complexity of noradrenergic control of the GH
axis. In the male rat 1-inhibition of GH secretion also
can be evoked via paraventricular nuclear effects (582).
The stimulation of 2-receptors (e.g., with
agonists such as clonidine and guanfacine) induces GH release in man
and animals (583, 584, 585). Relevant 2-adrenergic receptors
are probably accessible to the blood within the median eminence (586).
The postsynaptic 2-adrenergic receptors involved in the
regulation of GH secretion are similar to those mediating hypotensive
effects in the cardiovascular system (587), which thus may confound
experimental interpretations, e.g., in the rat (see below).
Interestingly, the GH-releasing action of clonidine is effectively
blunted in patients with essential hypertension, who may harbor
hypothalamic alterations in adrenergic tone (588).
Many studies in the rat have shown that, when GHRH is inactived by
passive antibody transfer (585), or when the hypothalamic arcuate
nucleus that secretes GHRH is lesioned (589), the GH response to
2-receptor agonists is abolished. In contrast,
pretreatment of animals with antibodies to somatostatin does not alter
the GH response to 2-agonist drugs (590). Some other
experiments suggest that GH secretion induced by
2-receptor stimulation in the rat is not mediated by
GHRH release from the median eminence but rather by reduced
somatostatin secretion (591). On the other hand, marked GH release
after clonidine treatment in the sheep is temporally associated with or
preceded by portal GHRH secretion with no consistent change in
somatostatin release (592). In bovine hypothalamic slides, guanabenz,
an 2-receptor agonist, also releases GHRH but does not
affect somatostatin (566). Caution is required in interpreting some
earlier experimental data, since clonidine (an 2
agonist) is not a very potent GH secretagogue in the rat and induces
potentially confounding hemodynamic, ventilatory, and behavioral
changes in this species (591).
Clinical studies reveal attenuation of clonidine’s stimulation of GH
rlease in patients with presumptively increased hypothalamic
somatostatinergic tone, e.g., individuals receiving chronic
treatment with glucocorticoids (590, 593), or those with
hyperthyroidism (594). However, in humans the mechanism of
2-receptor-mediated GH secretion may be more complex
than that in experimental animals. For example, the hypothesis that
2-adrenergic agonists may act also via non-GHRH
mechanisms, such as somatostatin release, arises from the finding that
pretreatment with GHRH abolishes the GH response to a second GHRH
stimulus without affecting clonidine’s action (595). These
observations are consistent with, but not proof of, an ability of
clonidine to withdraw somatostatin in the human.
The role of estrogen in regulating 2-adrenergic pathways
that impact GH secretion remains unclear. Whereas the magnitude of GH
release after clonidine (0.15 mg) ingestion in midluteal phase women is
proportionate to the serum estradiol concentration (596), conjugated
estrogen administration does not alter clonidine’s stimulation of GH
release in postmenopausal women (597). Thus, the nature and extent of
sex-steroidal interactions with 2-receptor-mediated GH
neuroregulation are not established in the human. In clinical studies
in adults, clonidine is a relatively poor GH secretagogue (598), making
investigations with this compound difficult.
c. ß-Adrenergic pathways.
Experiments performed in the human
and laboratory animals in vivo using several
ß-adrenergic receptor-blocking agents, such as propranolol (599)
or atenolol (600), support the operating hypothesis that ß-adrenergic
receptors mediate significant inhibitory effects on GH release. Studies
with inhibitors to phenylethanolamine N-methyl transferase,
the enzyme converting norepinephrine to epinephrine, indicate that the
neurotransmitter primarily involved in the ß-receptor-mediated
inhibition of GH-release is L-epinephrine (601). If
ß-adrenergic receptors are responsible for the inhibitory action of
catecholamines on GH release in vivo, this effect is not
likely to be due to a direct action on the pituitary gland. Although
rat somatotropes express ß-adrenergic receptors, ß-agonistic agents
actually stimulate GH secretion in rat pituitary cell cultures (602).
Therefore, the inhibitory effect on GH release that follows
ß-adrenergic activation in vivo is probably mediated by
hypothalamic actions on GHRH and/or somatostatin release.
Several experiments suggest that ß-receptors modulate hypothalamic somatostatin tone: 1) Whereas isoprenaline, a ß-agonist, stimulates somatostatin release from the rat pancreas (603), propranolol inhibits somatostatin release from hypothalamic slices in vitro (604); 2) Somatostatin antibodies block inhibition by isoprenaline of GHRH-induced GH release in the rat (605); 3) ß-Adrenergic antagonists disrupt the (negative) autofeedback of GH, which is likely mediated by somatostatin in the human (606); 4) Propranolol enhances GH responses to hypoglycemia, exercise, glucagon, and GHRH in humans (599), and there are ethnic differences in GH responses to this nonspecific ß-receptor blocker (607); and 5) clinical studies indicate that the ß2-adrenergic receptor agonist, salbutamol, inhibits GH secretion and overcomes stimulation by the amino acid, L-arginine, and the indirect cholinergic agonist, pyridostigmine (82). The last two substances are thought to reduce somatostatin release. Thus, ß2-adrenergic receptor stimulation of hypothalamic somatostatin secretion seems able to overcome the ability of L-arginine and acetylcholine to limit somatostatin release.
Acute administration of either of two different
ß2-receptor agonist drugs (salbutamol or broxaterol)
blunts the GH response to physical exercise in man (608). One
mechanistic explanation is that exercise alters catecholaminergic
pathway activities in the hypothalamus so as to favor an increase in
2 adrenergic (GH-stimulatory) tone; pretreatment with
ß2-agonist (GH-inhibitory) drugs may prevent this
phenomenon and thus attenuate the serum GH response to exercise.
Another consideration is that exercise increases (via an unknown
mechanism) hypothalamic GHRH secretion, which is counteracted by
the rise in hypothalamic somatostatin release caused by the
ß2-receptor agonists. Finally,
ß2-adrenergic activation and physical exercise may
exert opposing effects on hypothalamic somatostatin tone, and/or
act on cosecretagogues or modulators. The neuroregulation of
exercise-induced GH release is discussed further below.
C. Other neurotransmitters
1. Serotonin. Serotoninergic pathways are stimulatory to GH
release in the rat, possibly via promoting GHRH release, since
pretreatment with GHRH antiserum inhibits the GH-secretory response to
serotonin agonists in this species (609). In contrast, in the sheep,
intravenous injection of tianeptine, a serotonin uptake enhancer and
inhibitor of serotonin action, evokes a significant, immediate, and
short-lasting (30-min) increase in peripheral GH (+750%) and
hypophysial portal GHRH (+180%) concentrations in conscious unstressed
animals. There is no significant concurrent change in the secretion of
somatostatin (610). Such data suggest that endogenous serotoninergic
input is inhibitory to GHRH and GH secretion in the ruminant,
comparable to indirect inferences in the dog (611), but not in the
rodent. Thus, species differences exist in serotoninergic
neurotransmitter regulation of the GH axis.
In man, administration of 5-hydroxytryptophan, a serotonin precursor (which however also releases catecholamines from synaptic terminals) increases serum GH concentrations (612). On the other hand, neither quipazine, a direct serotonin receptor agonist, nor fenfluramine, a drug that releases serotonin from serotoninergic terminals (and also interferes with dopamine neurotransmission), alters basal GH concentrations (613, 614). Fenfluramine abolishes GH release stimulated by L-dopa and propranolol without affecting the response to L-arginine (615). Administration of the nominal serotonin receptor antagonist, cyproheptadine (which also has anticholinergic actions) or methysergide (which also appears to be a serotonin agonist), can inhibit GH responses to presumptive hypothalamic stimuli, such as hypoglycemia, L-dopa, and clonidine (614, 616). Because of the imperfect specificity of the drugs used in the foregoing clinical studies, the precise role of serotonin pathways in the regulation of GH secretion in man has been difficult to establish.
At least four major subtypes of brain serotonin receptors have been identified recently (617). In this context, a novel selective serotonin-1 (subtype D) receptor agonist, sumatriptan, has been employed to evaluate the relevance of specific serotonin receptors in GH secretion (618). The main advantage of this agent is that, although it also interacts in humans with the subtype B serotonin-1 receptor, the latter is not present in human brain; moreover, this drug does not activate other serotonin receptors, or adrenergic, dopaminergic, muscarinic, or GABA receptors (619). Sumatriptan increases basal GH release in normal adults without affecting cortisol secretion, thus suggesting that this effect is not stress related (620). The actions of this serotonin subytpe-1D receptor agonist have also been assessed in normal subjects in combination with GHRH, as well as with compounds that act as presumed functional somatostatin antagonists, e.g., L-arginine and pyridostigmine. Since sumatriptan potentiates the GH response to a maximal dose of GHRH (206, 618), one can postulate that this agent decreases hypothalamic somatostatin secretion. The lack of additive effect of sumatriptan and pyridostigmine further implies that the two drugs act through the same (final) mechanism, i.e., inhibiting somatostatin release. That sumatriptan increases GH responses to L-arginine is somewhat paradoxical and suggests that these secretagogues may act through at least partially independent mechanisms and/or synergize in inhibiting somatostatin release (618). Thus, available clinical data indicate that subtype-1D serotonin receptors mediate a significant GH-stimulating effect in humans, presumably via inhibition of hypothalamic somatostatin secretion.
2. Histamine. Information regarding the possible role of
histamine in regulating GH release is relatively sparse. The central
histaminergic system seems to mediate inhibition of physiological (621)
and pharmacologically induced (622) GH secretion in adult rats.
Histamine does not alter GH secretion from the pituitary gland in
vitro (563), but exerts an inhibitory effect on GH secretion in
adult rats when injected icv (621, 622). This occurs possibly via
suppression of GHRH release, since histamine does not appear to
increase hypothalamic somatostatin release (623). Reduction of
histaminergic activity by treatment with -fluoromethylhistidine, an
inhibitor of histamine synthesis (624), increased GH release as
stimulated by opioids in both neonatal and adult rats (625). An
explanation is that -fluoromethylhistidine, by removing inhibitory
histaminergic control of GHRH release, may facilitate opioid
stimulation of GHRH secretion.
Administration of histamine in man does not affect GH secretion (496). However, an H1-receptor antagonist reduces the GH response to L-arginine but not to hypoglycemia (626), whereas an H2-receptor antagonist diminishes stimulated GH release to a much smaller extent (627). An H1-receptor agonist elicited paradoxical GH release in response to a TRH stimulus in normal subjects (496). Thus, in humans, histaminergic pathways, acting presumptively through H1-receptors, may play a conditional stimulatory role in GH secretion in man (in contrast to what is observed in the rat). The facilitative effects of histamine in the human are likely mediated at the hypothalamic level, albeit via mechanisms that are unknown at present.
3. GABA. In humans, administration of sodium valproate, a postsynaptically active drug that increases accumulation of endogenous GABA, inhibits diazepam-induced GH release without affecting GH release stimulated by hypoglycemia (628, 629) or GHRH. On the other hand, administration of the GABAB receptor agonist, baclofen, stimulates basal GH release and inhibits the GH responses to L-arginine and hypoglycemia (630). The stimulation by GABA-like agonists of basal GH secretion is not likely mediated via GHRH release from the hypothalamus, since, at least in the rat, stimulation is not abolished by passive immunization with GHRH antiserum (631). On the other hand, GABA may inhibit exogenous secretagogue-stimulated GH secretion via an increase in hypothalamic somatostatin tone (632). Of mechanistic interest, a GABA agonist stimulates both GH secretion and slow-wave sleep in the human (633), which is consistent with, but not proof of, GABA pathways interfacing with somatostatinergic systems (222, 383). Nocturnal withdrawal of somatostatin in the human has been inferred by enhanced GH release in response to GHRH infusions during slow-wave sleep (634).
Although GABA also evokes galanin release from rat hypothalamic fragments in vitro (635), and galanin can stimulate GH release in vivo (above), GABA antagonists actually impede galanin’s stimulation of GH secretion in the rodent (434). Thus, the exact relationship between GABA and galanin’s activation of GH secretion needs to be clarified further. In this regard, GABA can directly stimulate GH release in vitro from neonatal rat pituitaries in a chloride and calcium-dependent manner (636, 637, 638, 639). In the sheep, like the rat, either intravenous or icv injection of GABA agonists will stimulate GH secretion (640). In the human, stimulatory effects of GABA agonists are also evident whether assessed via the use of a GABA-B agonist (baclofen), a GABA-ergic stimulus (valproic acid), or a GABA metabolite (641, 642). The mechanisms subserving GABA-stimulated GH secretion in the human have not been elucidated, although metergoline, a nonspecific serotonin-receptor antagonist, and flumazenil, a benzodiazapine receptor antagonist, both inhibit GABA-stimulated GH secretion (642, 643).
Various clinical conditions are accompanied by diminished GH release (compared with healthy controls) after baclofen or valproic acid stimulation of GH secretion. For example, in depression, whether or not preceded by 1 month’s treatment with tricyclic antidepressants (644), GABA’s stimulation of GH secretion is attenuated (643). Other examples of impaired responsiveness to GABA activation include heroin addicts (645), type I diabetes mellitus (646), Parkinson’s disease (647), and schizophrenia (648). Importantly, there are both sex-specific and age-dependent influences on GABA’s stimulation of GH secretion in the human, with baclofen (a GABA-B agonist) stimulating GH secretion in men but not women (649), and in younger but not older men (650). In addition, in acromegalic patients, valproic acid fails to heighten GH secretion (651), which likely reflects the (partial) autonomy of somatotropinoma cells.
The production of GABA requires the activity of two key enzymes: GABA-T (4-aminobutyrate-2-oxaglutarate aminotransferase) and GAD (L-glutamate decarboxylase). GAD converts glutamate to GABA via a decarboxylation reaction, which is the rate-limiting enzymatic step in the biosynthesis of GABA (652). GAD is a major autoantigen in type I diabetes mellitus and a target of both humoral and cell-mediated autoimmunity in this disease (653). Marked GH hypersecretion is characteristic of type I diabetic patients (654), unlike variably increased or blunted GH secretion in patients with type II diabetes (655). Whether a putative autoimmune process is involved in this derangement of GH neuroregulation is unknown. Against this speculation, circulating autoantibodies to GAD accompany, but do not correlate with, an increased GH response to GHRH in long-standing type I diabetic patients (656). Moreover, neutralizing antibodies or antibodies mirroring destruction of GAD-containing cells would diminish GABA production and hence be expected to reduce GH secretion.
In summary, the GABA pathway facilitates basal GH secretion in the rat, sheep, and human, possibly via mechanisms involving galanin, serotoninergic, or benzodiazepine receptors, albeit no neurotransmitter mechanisms are established unambiguously. Moreover, in the neonatal rat, GABA stimulates GH release directly in vitro from pituitary cells. Numerous clinical conditions associated with either excessive GH secretion (type I diabetes mellitus or acromegaly) or relatively decreased GH production (healthy aging in men) are accompanied by impoverished stimulation of GH secretion by GABA agonists. Further investigation will be required to clarify the pathophysiological relevance of this neurotransmitter pathway in impacting the human GH axis in health and disease.
4. Excitatory amino acids. In the rat and guinea pig,
excitatory amino acids, such as
N-methyl-D,L-aspartate (NMDA) as well as
glutamate, can stimulate the GH axis (657, 658, 659). Although the
mechanisms mediating excitatory amino acids’ stimulation of GH release
are not established, GHRH antiserum blocks excitotoxin-stimulated GH
secretion in vivo in the pig (657), suggesting hypothalamic
mediation. However, smaller stimulatory effects (2-fold and not
always dose dependent) are reported in vitro (657, 660). For
example, NMDA stimulates single-somatotrope GH secretion in the reverse
hemolytic plaque assay, an effect that is blocked by a specific NMDA
receptor antagonist, is additive with GHRH actions, and is suppressed
by somatostatin (661). Among the array of known excitatory amino acids,
both NMDA and kainic acid stimulate GH secretion in vivo in
the rat (659). Much additional investigation will be required to
delineate the express role of excitatory amino acids in the regulatory
pathophysiology of GH secretion in the rat and other experimental
animals, as well as eventually in the human.
5. Role of nitric oxide as a neuromodulator in the GH axis. The amino acid, L-arginine, is the immediate biological precursor of nitric oxide (662). The enzymes converting L-arginine to nitric oxide were purified and cloned in 1991 (663, 664) and named nitric oxide synthase (NOS). This enzyme family releases nitric oxide as a mediator gas from the terminal guanidine nitrogen group of L-arginine, producing L-citrulline as a byproduct. NOS is inhibited by derivatives modified at the terminal guanidino group, such as N-monomethyl-L-arginine (L-NMMA), N-nitro-L-arginine methyl ester (L-NAME) (665, 666). NOS is present in at least three isoforms variously expressed in vascular, neural, and other tissues. In central and peripheral nervous tissues, a constitutive isoform prevails (NOS I) (667). Whereas L-arginine likely promotes GH release via inhibition of somatostatin secretion, neither the neurotransmitter intermediates nor the precise biochemical basis of this stimulatory effect (e.g., whether via nitric oxide) is known (668).
The possible role of the mediator gas, nitric oxide, in modulating GH secretion and/or action will be important to investigate further. GH’s stimulation of target tissues activates a complex intracellular signaling cascade, as reviewed elsewhere recently (5, 669). Whereas GH’s signaling actions may involve, in some measure, nitric oxide, the necessary and/or sufficient roles of nitric oxide in various effects of GH on different target tissues remains to be established. In favor of nitric oxide’s role in GH action, a recent study in GH-deficient patients showed that treatment with recombinant human GH stimulated urinary nitrite (a principal metabolite of nitric oxide) as well as cyclic GMP excretion. This suggests but does not prove that nitric oxide generation might account for some of the vasodilatory and other hemodynamic actions of GH clinically (670).
The neuroendocrine role of nitric oxide in mediating or directing
secretory activity of the GH axis in the rat is unclear, in contrast to
its likely agonistic regulation of neuronal CRH and GnRH release (671, 672). For example, recent studies are controversial, suggesting either
an inhibitory or a stimulatory role for nitric oxide in GH secretion.
In one in vitro study, a nitric oxide scavenger abolished
GHRH-stimulated GH secretion by (prepubertal) rat male pituitary cells
in culture (673), whereas in another in vitro study, a
nitric oxide synthase inhibitor increased GHRH actions (674). In
keeping with the latter inference, nitric oxide seemed to mediate the
inhibitory actions of the cytokine, -interferon, on GH secretion
in vitro (675). In in vivo experiments, an NOS
inhibitor lowered GH secretion in the rat (671). Considerable
additional study will be required to clarify these discrepancies, which
may arise in large part from the nonspecificity of the experimental
probes of the nitric oxide pathway, and the confounding complexities of
in vivo inhibition of the nitric oxide signal.
In the human, no direct evidence exists for a role of nitric oxide in basal or L-arginine-stimulated GH neuroregulation. In negative clinical studies, intravenous infusion of an NOS inhibitor (L-NAME) did not block the GH response to insulin-induced hypoglycemia, but did amplify the ACTH-secretory response to this stimulus (676). The converse experimental strategy of administering a nitric oxide donor, molsidomine, also failed to alter GH secretion (668). Thus, available clinical experiments fail to show that nitric oxide drives GH secretion in the human, although this conclusion cannot be viewed as definitive until highly specific nitric oxide probes become available and are tested with appropriate positive and negative controls.
|
VI. Role of Metabolic Substrates in the Regulation of the GH Axis |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
).
A. Blood glucose
1. Hypoglycemia. A mechanistically important observation is
that the clinically well known and marked GH-secretory response to
insulin-induced hypoglycemia is not mediated solely via GHRH, since
GHRH pretreatment abolishes the subsequent GH response to GHRH, but not
to hypoglycemia (681). This finding is compatible also with a true
pituitary desensitization phenomenon specific for GHRH’s stimulation
of somatostatin release, or some degree of pituitary GH depletion
(namely, loss of a specific GHRH-releasable pool). On the other hand,
combined GHRH stimulation and insulin-induced hypoglycemia exert
additive effects on GH release (681), which is consistent with proposed
somatostatin withdrawal during hypoglycemia. Alternatively, GHRH and
hypoglycemia may act on different releasable pools of GH within the
somatotrope population. In addition, hypoglycemia may inhibit
somatostatin release and/or action in man, although the former is not
observed by portal vein sampling in the sheep (51). Moreover, the rat
responds to the stress of hyoglycemia with decreased GH secretion, thus
highlighting major species disparities in the response of the GH axis
to glucose deprivation.
2. Hyperglycemia.
a. Normal physiology.
A rise in serum GH concentrations occurs
3–5 h after oral glucose administration in man (682, 683). This
delayed GH increment is preceded by initial suppression of plasma GH
levels for 1–3 h (682, 683). The clinically well known rapid
inhibitory effect of glucose on GH release may be due to a discharge of
somatostatin from the hypothalamus. In accordance with this proposed
mechanism, acute hyperglycemia blocks GH secretion stimulated by GHRH
(684, 685, 686, 687, 688, 689, 690). Conversely, central cholinergic receptor activation by
pyridostigmine, which is inferred to suppress hypothalamic somatostatin
release, counteracts the acute inhibitory action of glucose on GH
release (687, 689, 690). Oral glucose does not likely have a direct
pituitary effect, since glucose does not influence basal GH secretion
or GH responses to GHRH in vitro (691, 692). Thus, glucose
appears to modulate GH secretion through one or more hypothalamic
mechanisms, probably somatostatinergic, in the human.
In the human, we infer that oral glucose acts rapidly to suppress GH release by increasing hypothalamic somatostatin release, thus suppressing serum GH for 1–3 h. The reduced GH release results in an increase in the pituitary stores of GH. When somatostatin release declines, endogenous GHRH secretion is activated reciprocally, and available pituitary stores of GH are released, leading to the "rebound" increase in serum GH levels. Since the administration of pyridostigmine, which may inhibit somatostatin secretion and release GHRH, potentiates the late GH rise induced by the oral glucose, we infer that hypothalamic somatostatin release may not be shut off completely, and/or GHRH discharge may be increased partially, at the time of the 3–5 h delayed GH rise after oral glucose injection (693).
Studies carried out in vitro using perifused rat hypothalamic fragments reveal an inverse relationship between glucose concentration and somatostatin (694, 695) or GHRH (695) release. In the rat in vivo, however, basal or GHRH-stimulated GH release is not altered by acute hyperglycemia; therefore, the relevance of (metabolic) data obtained in the rodent to man is often limited. Moreover, the control mechanisms that confer these species differences are not understood at present.
b. Diabetes mellitus.
GH secretion is markedly altered in
diabetes mellitus in both rats and humans, albeit in opposite
directions. In principle, abnormal GH secretion in the diabetic context
could originate via several mechanisms: 1) hypothalamic
neurotransmitter alterations resulting in changes in GHRH and/or
somatostatin secretion; 2) pituitary variations in the affinity and/or
number of receptors for GHRH and/or somatostatin; 3) pituitary changes
in GH synthesis and release; and 4) altered feedback regulation of GH
secretion at the hypothalamic and/or pituitary levels (696).
The adverse effects of diabetes on the circulatory, visual, renal, and peripheral nervous systems are commonly recognized; neuroregulatory disruption is also known, as reviewed in part in 1994 (654). The impact of decreased insulin secretion or attenuated insulin action on the function of other endocrine glands is not so well documented. Both clinical and animal research demonstrate that diabetes mellitus is commonly associated with altered thyroid, adrenal, and gonadal function. Some of these changes are reversed by insulin replacement therapy, but endocrine function is not always normalized even with rigorous glycemic control. For example, patients with poorly controlled diabetes mellitus exhibit GH hypersecretion basally and when variously stimulated, while patients with good metabolic control still present with diurnal and exercise-induced GH hypersecretion (654). In evident contrast, diabetes suppresses GH secretion in the rat. Clinical and experimental evidence exists for diabetes-associated changes in GHRH and somatostatin release as well as in the pituitary response to these hypothalamic hormones (697).
Despite GH hypersecretion, basal levels of IGF-I (the principal tissue mediator of GH activity), are low in type I diabetes mellitus (698). Serum concentrations of IGFBPs, the carriers that transport IGFs in serum and modulate their tissue-specific bioactivity, are also altered. Specifically, blood concentrations of IGFBP-3, the major transport protein for IGF-I, are decreased in poorly controlled diabetics (698). In contrast, IGFBP-1, a carrier protein that also likely functions in the autocrine/paracrine modulation of IGF bioavailability, is increased in serum of type 1 diabetic patients, and IGFBP-1 levels are inversely correlated with glycemic control (698). How such disturbances in circulating IGFBPs influence GH and/or IGF-I autonegative feedback on the GH axis in diabetic individuals is not known.
Understanding of the mechanisms responsible for the abnormalities in the GH-IGF-I axis in type I diabetes mellitus is incomplete, although recent evidence strongly suggests that portal vein insulinopenia contributes to dysregulation. Down-regulation of hepatic GH receptor expression secondary to portal insulin deficiency could explain the apparent GH resistance observed in this disease. Consistent with this hypothesis, there is a decrease in circulating concentrations of GH binding protein, a putative index of GH receptor number, in patients with type I diabetes mellitus (698). The abnormalities in the GH/IGF-I axis have been implicated in the worsening of metabolic control that occurs in some diabetic patients, as well as in the development of microvascular complications, such as retinopathy and nephropathy (699).
Diabetes mellitus in the rat, whether induced by streptozotocin administration or occurring spontaneously, markedly suppresses pulsatile GH secretion and blunts the GH-secretory response to GHRH (700). Hypothalamic somatostatin appears to play a major role in this animal model (701, 702), since 1) there are increases in immunoreactive somatostatin levels in hepatic portal and peripheral blood in spontaneously diabetic rats (703); 2) the attenuated GH response to GHRH is normalized after administration of somatostatin antibodies (702), or deep pentobarbital anesthesia (563); and 3) pituitary GH content is unaltered in spontaneously diabetic rats (702). Indeed, increased hypothalamic somatostatin release with unaltered GHRH secretion might be expected to increase pituitary GH content in the diabetic rodent. Unexpectedly, direct hypophysial-portal blood sampling in the streptozotocin-induced diabetic male rat revealed decreased hypothalamic release of both GHRH and somatostatin (704). Unfortunately, similar data are not yet available in other species, such as the sheep or primate, thus rendering this issue indeterminate at present.
The male BB/Worcester rat with insulinopenic diabetes mellitus exhibits a markedly reduced mass of GH secreted per bust and the GH-secretory rate (705), but GH-secretory burst frequency, secretory burst half-duration, and serum GH half-life are unchanged. In vitro studies of acutely dispersed somatotropes obtained from rats with diabetes mellitus show increased sensitivity to GHRH, as quantitated by a greater mean hemolytic plaque area after exposure to an EC50 dose of the secretagogue and diminished sensitivity to somatostatin’s inhibition of GH release driven by an EC50 dose of GHRH. The numbers of pituitary cells and somatotropes were indistinguishable in diabetic and normoglycemic animals (705). Thus, somatotrope in vitro responsitivity in the diabetic rodent seems to manifest a "denervation-like" hypersensitivity, perhaps reflecting endogenous GHRH withdrawal (and somatostatin excess).
The GK Wistar rat is a new model of diabetes mellitus in nonobese animals with significant fasting hyperglycemia, hyperinsulinemia, and absent insulin response to iv glucose. The GH response to GHRH is reduced at 16 weeks of age compared with normal, age-matched Wistar rats, but no differences are observed at 6 weeks of age. Pretreatment of older rats (16 weeks) with somatostatin antibodies significantly increases the GH response to GHRH in both normal or young (6 weeks old) GK rats (706). These results support the thesis (above) that accentuated somatostatin release mediates the blunted GH response to GHRH in diabetic GK rats. Reduced hypothalamic cholinergic signaling to the somatostatinergic neuron might, in turn, permit excess somatostatin release. This view is supported by the results of both in vitro and in vivo studies. In vitro, cholinergic muscarinic blockade with pirenzepine causes dose-related stimulation of SS release from normal hypothalami, but does not affect GK rat hypothalami (706). In vivo, concentrations of immunoreactive somatostatin are higher in hepatic portal and peripheral blood in spontaneously diabetic rats (703). Finally, the attenuated GH response to GHRH normalizes after pentobarbital anesthesia, which presumably suppresses hypothalamic somatostatin (and other neuropeptide) release (563, 702). Therefore, increased hypothalamic somatostatin secretion may be the primary determinant of the GH suppression observed in the diabetic rodent. We emphasize that amplified somatostatin release would also be expected to quash GHRH output via intrahypothalamic reciprocal regulation (131, 265).
c. Type 1 diabetes mellitus patients.
Available clinical data
show elevated 24-h GH release in untreated and treated type I diabetic
patients (707, 708, 709, 710). Detectable serum GH peak frequency and interpeak
GH concentrations are higher in these patients (708), and exaggerated
GH release in response to GHRH occurs consistently (420, 711, 712).
Mechanistic studies with the acetylcholinesterase inhibitor,
pyridostigmine, which is an indirect cholinergic agonist hypothesized
to decrease somatostatin release (262), indicate that type I diabetic
patients with an exaggerated GH response to GHRH fail to respond with
further GH release to pyridostigmine treatment (Fig. 11). This finding suggests, but does not
prove, maximally decreased endogenous somatostatin release in (human)
type I diabetes mellitus (420, 558), and/or suggests maximal
somatotrope output of GH.
Presumed decreased somatostatin release in human type I diabetes in turn may arise from impaired GH and/or IGF-I autofeedback. Thus, although GH pretreatment reduces GHRH-stimulated GH secretion in both healthy man and rat, probably by increasing hypothalamic somatostatin secretion via cholinergic pathways (259), GH pretreatment (with or without pyridostigmine) does not limit the GH-secretory response to GHRH in type I diabetic patients (260, 713). Conversely, pirenzepine, a cholinergic muscarinic antagonist, acutely inhibits spontaneous, sleep-related, and stimulated GH release in patients with insulin-dependent diabetes mellitus (714, 715), suggesting the releasability of additional somatostatin. Pirenzepine truncates the amplitude of GH pulses, but has no effect on GH peak frequency. However, in more chronic studies, pirenzepine given in a daily dose of 75 or 150 mg for 4 days does not affect GH secretion (716, 717), indicating "escape" of somatostatin withdrawal, drug tachyphylaxis, etc. Autoimmunity also has been speculated to contribute to derangement in somatostatin tone, e.g., via GAD antibodies (718). However, the GH response to GHRH plus pyridostigmine was not significantly correlated with GAD antibodies, which (if anything) would be expected to decrease GH secretion via reduced GABA-ergic stimulation in the human (above). Thus, overall data favor the thesis that decreased somatostatinergic tone accounts primarily for elevated pulsatile GH secretion and exaggerated GH release in response to various pharmacological stimuli in type I diabetic individuals (718). Withdrawal of IGF-I’s negative feedback actions also likely augments GH secretion in these catabolic patients. However, serum IGF-I levels fail to correlate with lack of responsiveness to pyridostigmine, thus suggesting that additional pathophysiological mechanisms may also operate (255).
The pathophysiology of the presumptively altered somatostatin neuroregulation in type I diabetes mellitus has been studied traditionally in three venues: 1) role of metabolic control, 2) neuroendocrine pathways, and 3) altered IGF-I and BP feedback signaling. In patients with insulin-dependent diabetes mellitus, there is an increase in both the amplitude and frequency of pulsatile GH secretion compared with normal subjects (707, 708), which is not affected by maintenance of overnight normoglycemia. Type I diabetics with poor metabolic control have significant increases in GH peak maxima, incremental amplitudes, and pulse durations, when compared with a period of better control (654). Mean 24-h GH secretion decreased significantly during improved (insulin-treated) glycemic control, although the fraction of pulsatile GH secreted per 24 h did not change significantly (719). In contrast, intraportal delivery of insulin nearly normalizes the GH/IGF-I/IGFBP axis in recent-onset type I diabetes mellitus (720).
Chronic subcutaneous recombinant human (rh) IGF-I administration to children with insulin-dependent diabetes mellitus corrects many of the preexisting abnormalities in the GH/IGF/IGFBP axis. Specifically, 28 days of rhIGF-I therapy to overcome endogenous IGF-I deficiency suppresses both serum IGFBP-1 and GH concentrations (721). Studies by Cheetham et al. (722) further demonstrate that a single subcutaneous injection of rhIGF-I increases serum IGF-I levels, decreases overnight GH secretion, and reduces insulin requirements in adolescent type I diabetics (n = 9). Similarly, Bach et al. (723) found that 10-h continuous subcutaneous infusions of rhIGF-I given on three successive days to each of four diabetic adolescents increased serum IGF-I, reduced IGF-II, and suppressed GH concentrations (724). Thus, we infer that feedback responsiveness to IGF-I is preserved in type I diabetes mellitus with GH hypersecretion, and that IGF-I deficiency contributes to the GH hypersecretory state via feedback withdrawal. Practically, compared with insulin treatment in type I diabetic children, dual injections of insulin and rhIGF-I tended to lower GH and increase IGFBP-1 concentrations while improving glycemic control, more than insulin treatment alone during 1–4 weeks of therapy (698).
Finally, although not fully established, in principle, suppression of endogenous GH secretion with somatostatin analogs might ameliorate renal hyperfiltration, an early feature of diabetic nephropathy linked indirectly to GH hypersecretion in type I diabetes, since GH treatment itself tends to promote renal blood flow (725, 726, 727, 728, 729).
d. Type II diabetic patients.
Clinical investigations of GH
secretion in patients with type II diabetes have yielded conflicting
inferences. Spontaneous GH secretion may be increased, normal, or
decreased (730, 731, 732, 733), and secretagogue-stimulated GH release also may
be augmented (734), normal (735), or reduced (654, 655, 736, 737). Many
studies of GH secretion in type II diabetic patients are confounded
significantly by major covariates and codeterminants, such as age and
body composition, which profoundly influence GH secretion. GH secretion
decreases markedly with age and is strongly inversely related to the
degree of obesity (93, 119, 738, 739, 740) (e.g. see Fig. 12),
both of which factors are operative in diabetic (and control)
populations.
|
An informative recent study demonstrated that spontaneous GH secretion in type II diabetes mellitus is influenced by two independent factors acting in opposite directions: the marked inhibitory effect of obesity on GH-secretory burst amplitude or mass (119, 733, 740), and the apparent stimulatory effect of the diabetic state on GH burst frequency (693, 708, 743). The observed suppression of GH burst mass in proportion to the degree of obesity in diabetic patients illustrates the importance of body composition in the regulation of GH secretion in type II diabetes mellitus. In contrast to the reduced GH-secretory burst mass (733), the apparent increase in GH-secretory burst frequency in the type II diabetic patients is not explained by obesity. Augmented GH pulse frequency also occurs in adult and adolescent patients with type I diabetes mellitus, where it leads to an increased overall GH production rate (707, 708). Thus, a disruption of the hypothalamo-pituitary interplay of GHRH and somatostatin that regulates pulsatile GH release may be associated with the diabetic state, whether type I or type II. In short, chronic hyperglycemia may elicit variable alterations in pituitary GH secretion in type II diabetics depending on the degree of obesity (743), viz., on the one extreme, the GH-hypersecretory pattern of type I diabetics (lean young subjects with low circulating insulin levels, and frankly elevated blood glucose) and, on the other extreme, the GH hyposecretory state of pure obesity (older patients with hyperinsulinemia and milder hyperglycemia) (693, 708, 733). This concept may explicate many disparities in reported GH axis activity in type II diabetic patients.
B. Leptin and FFA
In the hyperphagic leptin-deficient (ob/ob) mouse, either central
or peripheral infusion of leptin induces weight loss and satiety by
reducing hypothalamic NPY levels (745, 746, 747, 748). Thus, in the rodent, NPY
is one likely link among obesity, eating, and leptin feedback.
The exact metabolic mechanisms subserving reduced GH secretion in
relative adiposity or absolute obesity are not known (312, 742).
Relative hyposomatotropinemia appears to correlate with total and
especially visceral (intraabdominal) fat (121, 749) (Fig. 11). Indeed,
visceral fat mass as estimated by computed tomography scanning is the
primary (negative) statistical determinant of GH secretion in
middle-aged men and women and accounts for both the gender and age
differences in GH secretion in this context (750). The precise
neuroregulatory basis for the inverse relationship between body fat
mass and GH secretion is not fully understood, but the recent discovery
of a polypeptide, leptin, secreted by fat cells allows several
plausible hypotheses (751). GH secretion in postmenopausal women (fed)
as well as young midluteal phase women (fed and fasted) is inversely
correlated with serum leptin concentrations (150, 524) (see Fig. 10).
Although direct hypothalamic actions of leptin on somatostatin and/or
GHRH release are imperfectly defined, leptin stimulates GH secretion in
the fasted adult male rat (522). Akin to other metabolic antitheses in
the rat and human, an opposite action of leptin would be expected in
the human, viz., to inhibit GH release. Alternatively, serum
leptin may be a marker of acute nutrient intake (albeit not
macronutrient composition) and longer-term fat depots in the human, and
not a direct regulator of the GH axis (150, 523, 524, 752, 753). Thus,
rh GH treatment decreases serum leptin concentrations in proportion to
fat loss per se in GH-deficient adults (753).
Full-length leptin receptors are present in rat hypothalamic neurons, and leptin administration in the rodent inhibits the expression of the potent orexigenic and GH-suppressing peptide, NPY (748, 754, 755). In addition, leptin advances puberty in mice by stimulating the hypothalamic neuronal secretion of GnRH and overcomes the suppressive effects of fasting on the reproductive axis in the same manner (756, 757). Leptin also acts on the thyroidal axis by stimulating TRH-secreting neurons in the paraventricular nucleus (758). Accordingly, leptin, which can enter brain via a saturable system (759), can act on multiple hypothalamic neurons (754, 760, 761), and may thus directly or indirectly modify somatostatin and/or GHRH release, thereby mediating or modulating the apparent negative-feedback effects of increased body fat on the GH axis in the rodent. However, other (nonleptin) hypothalamically targeted regulators might also be released by adipose or other tissues and contribute relevant long-term feedback information concerning body composition. Indeed, the decrease in IGFBP-1 levels in obesity, with possible relative increases in free IGF-I, has been postulated to suppress GH in the human (762), and possibly exacerbate obesity, whereas IGF-I (or GH) treatment will reduce body fat content (763).
The clinical pathophysiology of leptin secretion and the nature of leptin actions in the human are being actively explored. From a neuroregulatory perspective, leptin is secreted via both circadian and pulsatile rhythms and achieves higher serum concentrations in women than men (764, 765). Leptin is likely to interface with the GH axis via GHRH-somatostatin neurons and other putative hypothalamic modulators, e.g., NPY-expressing neurons, GnRH neurons, and possibly TRH neurons (748, 754, 756, 760, 761). Central actions of leptin are plausible in view of the saturable transport of this peptide into the brain (759). Since the leptin level is a strongly positive correlate of fat cell mass, and full-length leptin receptors are expressed in the hypothalamus (above), the observation is not surprising that treatment for 1 yr (but not 1 month) of adult GH-deficient adults with rh GH reduces serum leptin concentrations significantly (from 16 ± 3 to 11 ± 2 µg/liter) in direct proportion to the reduction in total percentage body fat (r = 0.763, P = 0.002) (753). In the presence of concomitant glucocorticoid treatment, GH treatment for 1 week appears to increase serum leptin concentrations further (above that stimulated by methylprednisolone alone) (766). These preliminary sallies into leptin pathophysiology indicate the need for further clinical studies to clarify specific mechanistic interactions among GH, leptin, and various modulators of leptin production.
FFA may well participate in the regulation of pituitary GH secretion in the human and sheep, since GH secretion is stimulated when plasma FFA levels are reduced, whereas conversely increased FFA levels block GH secretion provoked by virtually all stimuli (767, 768, 769). For example, the drug acipimox lowers plasma concentrations of FFA by blocking their release from adipose tissue (770) and concomitantly enhances basal GH secretion (771) and the GH-secretory response to GHRH in normal or obese subjects (772). Acipimox also potentiates the GH response to GHRH after pyridostigmine pretreatment (773). The synergistic action of acipimox and pyridostigmine can be explained by hypothesizing a hypothalamic site of action of the cholinergic agonist (262), and a direct pituitary-inhibitory effect of FFA (774). This concept is supported by in vitro studies showing that fatty acids inhibit basal (775) and GHRH-stimulated (776) GH secretion from cultured rat somatotropes. In obesity, acipimox treatment virtually normalizes GH secretion, pointing to an important role of relatively increased FFA flux in this relatively hyposomatotropic state (771). In addition, reportedly elevated free IGF-I levels (that accompany insulin-driven IGFBP-1 reductions) in human obesity may suppress pituitary GH secretion directly (777), as suggested in the rat in vitro, and the human and sheep in vivo (62, 63, 64).
C. Amino acids
L-Arginine is an essential amino acid and a well known
potent GH secretagogue in man (778). In the rat, amino acids do not (or
less overtly) stimulate GH secretion (563). The most plausible
mechanism through which L-arginine stimulates GH secretion
is somatostatin withdrawal. Arginine does not influence either basal or
GHRH-induced GH secretion from rat anterior pituitary cells in
vitro (779), which speaks against a direct effect of
L-arginine at the somatotrope level. This interpretation is
also concordant with the observation that patients with idiopathic GH
deficiency or pituitary dwarfism, which is often due to a lack of
endogenous GHRH (780), have no GH rise after L-arginine
infusion. In man, several investigators have shown that
L-arginine enhances the GH response to GHRH (779, 781).
Since these experiments used a maximal dose of GHRH combined with
L-arginine, one can infer that this amino acid stimulates
GH secretion by inhibiting endogenous somatostatin, rather than by
promoting endogenous GHRH release. This inference is in agreement with
data showing that L-arginine infusion after an earlier
(desensitizing) dose of GHRH will still induce a pronounced GH rise
(171). Moreover, L-arginine fails to augment the GH
response to pyridostigmine (781), which is believed to act in part via
somatostatin withdrawal (above).
The hypothesis that L-arginine suppresses endogenous somatostatin release is consonant with the ability of L-arginine, as well as pyridostigmine (782), to amplify the TSH response to TRH (which is at least partially under inhibitory somatostatin control) in normal humans (779). Furthermore, administration of L-arginine normalizes the GH response to GHRH in patients with nonendocrine diseases receiving chronic immunosuppressive glucocorticoid therapy (783). Indeed, the L-arginine-induced enhancement of the GH response to GHRH is greater in glucocorticoid-treated than normal subjects. The ability of L-arginine to reverse glucocorticoid inhibition (see below) strengthens the hypothesis that this amino acid decreases somatostatin release by the hypothalamus. Whether L-arginine, a precursor in the nitric oxide- signaling pathway, works via nitric oxide to stimulate GH release in the human is unknown, but seems unlikely based on a recent clinical study (668).
|
VII. Other Hormonal Regulators of the GH Axis |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
A. Glucocorticoids
1. In vitro studies. Combined immunocytochemical (below) and
mRNA data demonstrate glucocorticoid suppression of GHRH message and
peptide synthesis, which could account in part for glucocorticoid’s
inhibition of GH secretion (784). As reviewed in part earlier (171, 550), studies from at least three laboratories indicate that
hypothalamic GHRH mRNA levels are decreased by chronic glucocorticoid
treatment in the rat (785, 786, 787). Glucocorticoids likely significantly
inhibit GHRH gene transcription in the arcuate nucleus, plausibly by
acting via a known glucocorticoid-responsive element in the 5'-upstream
promoter of the GHRH gene (785). This region of the promoter may
mediate repression of GHRH gene transcription by an activated
glucocorticoid receptor. Inhibition of GHRH gene expression is
dependent on the dose and duration of glucocorticoid treatment and
occurs even after pituitary removal (787), which indicates that GH
feedback per se is not required for these brain actions of
glucocorticoid.
Immunologically based studies have shown that total hypothalamic GHRH peptide content falls in glucocorti-coid-treated rats compared with controls (788). Similarly, Fernandez-Vasquez et al. (789) reported that treating rat hypothalamic cells in vitro with high doses of corticosterone decreased neuronal GHRH release. Other recent immunocytochemical data define a reduction in optical density and percentage area of immunostaining for GHRH only in the rostral region of the median eminence of the hypothalamus in glucocorticoid-exposed rats (784). In vitro, corticosterone has dual effects on hypothalamic neuronal GHRH peptide expression, i.e., concentrations in the range of the glucocorticoid receptor Kd (3 nM) increase GHRH content, whereas higher concentrations (30 and 300 nM) decrease GHRH content in GHRH-secreting cells (789). Biphasic regulation (low-dose stimulation, and higher-dose inhibition) of GH secretion by synthetic glucocorticoids is also recognized in the human in vivo (below).
Available investigations of glucocorticoid effects on hypothalamic somatostatin mRNA accumulation in the rat are conflicting. Nakagawa et al. (790) observed that high-dose glucocorticoid administration (330 µg/day) for 3 days raised hypothalamic somatostatin mRNA content in adult female rats. Conversely, Lam et al. (791) found that more prolonged (10 days) administration of dexamethasone in the drinking water at a concentration of 5 µg/ml slightly reduced somatostatin mRNA content in prepubertal male rats. The latter results most likely reflect the lower dose, different route of delivery, and/or animal age and sex used in that study. A bipotential effect of glucocorticoids on hypothalamic somatostatin mRNA levels was observed recently, i.e., a significant increase in specific somatostatin transcripts in the periventricular nucleus after 3 days of dexamethasone treatment (1 mg/kg/day), but a reduction after 8 days for all glucocorticoid doses tested (786).
The effects of glucocorticoid treatment on the somatostatinergic system are tissue specific and apply to both somatostatin peptide and mRNA content (792). The rise in hypothalamic somatostatin content due to glucocorticoids seems to reflect an increase in transcription of the somatostatin gene (784). RNAse protection assay also revealed a second lower molecular weight somatostatin gene-transcription product in glucocorticoid-treated rats, suggesting possible control of somatostatin gene expression both quantitatively and qualitatively.
In vitro incubation of fetal hypothalamic cells with corticosterone increases the content and release of immunoreactive somatostatin (789). Using immunohistochemical techniques, others reported in vivo increases in somatostatin content in the rostral, middle, and caudal regions of the median eminence of the rat hypothalamus after glucocorticoid treatment (784). In recent experiments, freely moving animals received a stereotaxic implant of a push-pull cannula into the median eminence for 10 days, which was then perfused with artificial cerebrospinal fluid for 120–150 min. An intraperitoneal injection of dexamethasone (200 or 300 µg/100 g) induced a mean increase in hypothalamic somatostatin output of 63 ± 6.2% above basal 15–30 min later (793). Thus, glucocorticoids seem to stimulate hypothalamic somatostatin production, storage, and release. Conversely, glucocorticoids appear to repress somatostatin receptor expression, at least by GH-secreting tumoral cell lines, e.g., dexamethasone inhibits mRNA expression of several somatostatin receptor subtypes in rat GH4C1 cells (794), whereas sex steroids oppose this action (below). Thus, multiple actions of glucocorticoids are likely in the hypothalamus as well as in the pituitary gland.
Chronic glucocorticoid treatment increases pituitary GH mRNA content in the adult male rat, as recently assessed by RNAse protection assays. This stimulation may reflect direct pituitary actions of glucocorticoids to increase GH gene transcription (795). Such an interpretation is consistent with the report of Slater et al. (796), showing the existence of a glucocorticoid-regulatory element in the human GH gene promoter.
Glucocorticoids can stimulate GH synthesis and secretion in vitro, e.g., by somatotropes obtained from human and normal monkey pituitaries (797). Similar effects are observed in studies of normal rat pituitaries (798) and pituitary clonal cell lines (799). Enhanced GH secretion occurs after several days of incubation and is magnified when the cells are first cultured for long periods without glucocorticoids. The glucocorticoid-induced increase in pituitary GH secretion is abolished by cycloheximide or actinomycin D pretreatment, suggesting requirements for both protein and RNA synthesis in this GH-stimulating effect of glucocorticoid (800).
Glucocorticoids can also increase the GH-secretory response to GHRH stimulation in vitro. Pretreatment of rat pituitary cells for 24 h with the potent synthetic glucocorticoid, dexamethasone, increases the GH response to GHRH more than 5-fold (798). Facilitation of GHRH action by glucocorticoids is time dependent; e.g., when pituitary cells are exposed to dexamethasone for less than 4 h, GHRH and forskolin-induced GH release declines, but after 18 h, basal and GHRH-elicited GH secretion rises.
A recent study examined the effects of glucocorticoid on GHRH receptor gene expression in adult male rat anterior pituitary cells in primary cell culture using a highly sensitive and quantitative reverse-transcribed PCR methodology (801). Dexamethasone (5 nM) significantly increased GHRH receptor mRNA levels after 6- and 24 h-incubations, with a maximal effect at 25 nM. The glucocorticoid receptor-specific antagonist, RU 486, completely eliminated dexamethasone’s enhancement of GHRH receptor mRNA expression (801). Hydrocortisone also stimulated a marked dose-dependent increase in GHRH-receptor mRNA (802).
Dexamethasone increases GH receptor mRNA expression in the growth plate of long bones and in the liver, target tissues important for body growth, but not in kidney. Glucocorticoid also induces liver GH-binding protein expression, but with a biphasic dose-response (803). In other tissues, dexamethasone directly antagonizes GH action at the cellular level; e.g., glucocorticoid inhibits the ability of GH to elicit several early events in GH signaling in 3T3-F442A fibroblasts (804).
In conclusion, a plethora of competing actions of glucocorticoids are exerted in the hypothalamus, the pituitary gland, and in different peripheral tissues. The complex interplay among these multiple opposing actions of glucocorticoid determines the net in vivo effect of the corticosteroid in the intact individual. Final responses are complicated by the bipotential dose and time dependencies of glucocorticoid actions on the GH axis. Our overall unifying working hypothesis is that glucocorticoids are both inhibitors and stimulators of regulated GH secretion and action.
2. In vivo studies. Adrenalectomy significantly decreases the GH-secretory response to a submaximal dose of GHRH in the rat (805). Since this attenuation is reversed by dexamethasone replacement, glucocorticoids likely maintain pituitary sensitivity to GHRH and/or the releasable pool of GH. In favor of the former, adrenalectomy results in a significant decrease in pituitary GHRH-binding capacity (239). On the other hand, and typical of bipotential actions of glucocorticoids, the GH response to a physiological dose of GHRH is significantly lower in animals treated with glucocorticoids either subacutely or chronically. In contrast, cortisol infusion into castrated male Suffolk sheep (1 to 1.5 yr old) acutely attenuates GHRH-stimulated GH release in this species (806).
Nakagawa et al. (788) reported that chronic (several-day) dexamethasone treatment suppresses serum GH concentrations, increases hypothalamic somatostatin concentrations, and decreases hypothalamic GHRH concentrations in conscious female rats. In male rats, glucocorticoids reduce and prolong trough GH values (807). When GH pulses occur, however, they are significantly higher than those seen in normal animals. This resembles a more masculine-like GH release pattern. According to the model of reciprocally coupled GHRH and somatostatin neuronal systems in the rodent (265), and the ability of somatostatin withdrawal to evoke rebound GH release (54, 157, 160), the foregoing dexamethasone-induced patterns of pulsatile GH release (807, 808) would suggest increased somatostatin tone induced by dexamethasone, with or without augmented endogenous GHRH release (and/or pituitary actions) during somatostatin withdrawal episodes.
Given that deep pentobarbital anesthesia can suppress the release of
several hypothalamic hormones, e.g., somatostatin (563, 809), the dichotomous effects of glucocorticoids in anesthetized
vs. conscious rats suggest multisite actions of
glucocorticoids, including at hypothalamic loci. A unifying
interpretation is that at the hypothalamic level glucocorticoids reduce
GHRH and increase somatostatin tone, thus promoting inhibition of GH
secretion; in contrast, at the pituitary level, corticosteroids enhance
GH synthesis and the GH response to GHRH (e.g., via
heightened GHRH receptor expression), thus favoring a stimulatory
response (Fig. 13). In support of this
proposed thesis, experiments in conscious, freely moving rats receiving
chronic (4-day) dexamethasone treatment show a diminished GH-secretory
response to a near-physiological dose of GHRH compared with
saline-treated rats. However, passive immunization with somatostatin
antiserum completely alleviated the attenuated GH response;
indeed, somatostatin-immunoneutralized glucocorticoid-treated rats had
a significantly higher GH response to GHRH than control animals (810).
These experiments further corroborate the notion that glucocorticoids
increase somatostatin’s inhibitory activity. In this regard,
somatostatin antiserum partially counteracts glucocorticoid’s
inhibition of body growth in the rat (811).
|
In recent experiments, three doses of GHRP (1 µg, 4 µg, and 25
µg/kg) and one dose of GHRH (500 ng/kg) were evaluated alone and
together in the glucocorticoid-treated rat. These studies revealed that
GHRP-6 alone increases plasma GH concentrations in a dose-dependent
fashion both in vehicle- and dexamethasone-treated rats, whereas the GH
responses to GHRP-6 and GHRH are significantly reduced selectively in
dexamethasone-pretreated animals. Thus, GHRP-6 is able to stimulate GH
secretion in glucocorticoid-treated rats, but only partially overcomes
glucocorticoid-induced inhibition of GHRH-stimulated GH secretion (812)
(see Fig. 13). This observation may indicate that GHRP only partially
reverses somatostatin excess induced by glucocorticoids and/or that
glucocorticoids inhibit via mechanisms that do not interface with the
GHRP stimulus pathway.
Galanin increases serum GH similarly in both vehicle- and dexamethasone-treated rats. The response to galanin plus GHRH is similar to that of galanin plus saline in vehicle-treated rats, but is significantly enhanced in rats chronically treated with dexamethasone. We infer that galanin-mediated GH release in rats may involve both increased GHRH secretion and possibly reciprocally reduced activity of somatostatinergic pathways and thereby oppose the inhibitory effect of glucocorticoids (813).
Overall, available data are consonant with the hypothesis that in vivo inhibition of GH secretion by glucocorticoids occurs via increased somatostatin release in the rodent, with or without a concurrent (e.g., somatostatin-mediated) decline in GHRH release.
3. Human studies. The impact of untreated primary adrenal failure (Addison’s disease) on GH secretion in humans is inadequately studied owing to the clinical severity of this disease and the urgent need to start glucocorticoid replacement promptly. On the other hand, there are more reports concerning pituitary function in patients with isolated glucocorticoid deficiency secondary to idiopathic ACTH deficiency. The first female patient with this rare autoimmune disease was described in 1988 and exhibited impaired GH-secretory responses to several different pharmacological stimuli (L-arginine and insulin). Stimulated GH release was restored fully during glucocorticoid replacement therapy (814). Glucocorticoid-remediable reduction in GH secretory-reserve was confirmed in several other (usually female) patients with ACTH deficiency (815) and also in some men (816). In addition, when glucocorticoid treatment is suspended in Addisonian patients, there is a progressive loss of the GH-secretory response to a GHRH stimulus (817), which is reversed in a dose-dependent manner by cortisol infusions (818). On the other hand, acute withdrawal of replacement glucocorticoid with resultant prompt biochemical hypocortisolemia (cortisol <55 nmol/liter) does not impair the GH-secretory response to GHRH compared with that observed during eucortisolemia (serum cortisol, 280–420 nmol/liter) in the same patients. On the basis of these clinical studies, we hypothesize that the impaired GH response to GHRH in the human with subacute or chronic adrenal insufficiency is a consequence of sustained low circulating cortisol concentrations. Therefore, physiological amounts of glucocorticoids probably play an important long-term role in maintaining basal somatotrope secretory reserve in humans, as well as the rat. Accordingly, small amounts of exogenous synthetic glucocorticoids (approximately twice the basal daily cortisol secretion rate) stimulate pulsatile GH secretion over several days in the healthy adult (819), and the serum cortisol concentrations correlate positively with GH release in fasting women (524).
In contrast to the foregoing circumstances of primary or secondary Addison’s disease, the immediate and short-term effects of exogenous glucocorticoids in normal humans are suppression of GH secretion. Hydrocortisone inhibits the GH-secretory response to GHRH within 1 h of its administration (462), and the GH response is still inhibited 8 h after treatment (820). The rapid (1-h) inhibitory effect is similar to that observed in animals (806) and may be somatostatin dependent, since pretreatment of the subjects with the cholinergic agonist, pyridostigmine, an agent thought in part to decrease hypothalamic somatostatin release, relieves glucocorticoid’s inhibition (464). In contrast to the effect of cortisol per se, treatment with the potent synthetic glucocorticoid, dexamethasone, for 3 or 4 h enhances both basal and GHRH-stimulated GH secretion (821, 822). Although the mechanisms involved in this transient stimulatory response are not well understood, one possibility is that glucocorticoid treatment decreases serum IGF bioactivity, perhaps through pathways involving the IGFBPs, e.g., BP-1, which is rapidly insulin-sensitive (823). This would result in decreased free IGF-I-negative feedback of GH secretion. Alternatively, one can hypothesize that temporally biphasic somatostatin release may occur after acute dexamethasone administration, and therefore a reciprocally biphasic pattern of GH release. Curiously, 8 h after the administration of dexamethasone, the GH response to GHRH is inhibited, apparently also at least in part, via a somatostatinergic mechanism (820). Inhibition of the GH response to GHRH continues after 4 days of administration of prednisone in normal subjects (824). Thus, acute administration of natural glucocorticoids may stimulate a rapid burst of somatostatin release. Synthetic glucocorticoid may promote rapid rebound GHRH (and GH release) after spontaneous somatostatin withdrawal or act directly to facilitate pituitary GH gene transcription. A later tonic hypersomatostatinergic state may follow in the case of long-acting glucocorticoids, such as dexamethasone and prednisone. Overall, available clinical data thus indicate that glucocorticoid administration can stimulate or inhibit GH release in a dose- and time-dependent manner, and in an agent-specific fashion, presumably through alterations both in somatotrope responsiveness to GHRH and in hypothalamic somatostatin release.
The chronic effects of glucocorticoids in humans are recognized in patients with spontaneous hypercortisolism (Cushing’s syndrome) and in patients undergoing high-dose long-term immunosuppressive glucocorticoid treatment. The increases in serum GH concentrations in response to commonly used pharmacological stimuli, such as insulin-induced hypoglycemia, L-arginine, lysine-vasopressin, and GHRH, as well as spontaneous sleep-associated GH secretion, all are typically inhibited in patients with significant endogenous hypercortisolism (825, 826, 827). In 14 patients with Cushing’s disease, blood sampling was performed every 20 min over 24 h for later determination of serum GH profiles before and 10–11 days and 3, 6, and 12 months after neurosurgical cure. Before intervention, Cushing’s disease patients manifested markedly decreased mean 24-h serum GH concentrations, GH peak heights, and GH peak areas, but GH pulse frequency was similar to values in controls (828). Several clinical studies have linked this suppression of GH release to hypercortisolemia, since GH secretion tends to normalize after the removal of the source of cortisol excess (829). However, surprisingly, a pattern of GH suppression similar to that observed in patients with acute hypercortisolism was also evident in Cushing’s disease patients who were studied 10–11 days and 3, 6, and 12 months after cure. The basis for long-term suppression of the GH axis is not known. In addition, more recent clinical studies have unmasked a more disorderly pattern of 24-h GH release in Cushing’s disease (830), which on theoretical grounds points to a defect in feedback control of the GH axis (144).
Recent clinical experiments have focused further on the neuroendocrine mechanisms involved in GH axis suppression in Cushing’s disease. Pyridostigmine, used in an attempt to decrease hypothalamic somatostatin release, augments, but does not restore fully, GH release in response to GHRH stimulation in patients with Cushing’s disease (831, 832). Thus, either this drug only partially opposes the somatostatin excess presumed to operate in Cushing’s disease, or the inferred glucocorticoid-mediated increase in hypothalamic somatostatin tone may not be the sole factor inhibiting GH secretion in these patients. Indeed, increased IGF-I feedback inhibition of the GH axis cannot be excluded, since in patients with Cushing’s syndrome the marked suppression of endogenous GH secretion is accompanied by normal GHBP, normal to low IGF-I, and low BP-1 and BP-3. Such biochemical findings may suggest enhanced tissue sensitivity to GH (high GHBP despite low GH secretion), and normal or increased free IGF-I bioavailability (833). Other relevant contributing factors in Cushing’s disease could include elevated blood glucose and FFA concentrations, and/or obesity (see above), all of which are known to suppress GH secretion in the human. Indeed, obesity also may be marked by increased (measurable) free IGF-I concentrations (777) with resultant feedback suppression of pituitary GH secretion directly [see Section VI.B above].
The type of glucocorticoid used also may influence the response of the GH axis. For example, prednisone blunted GH release more than deflazacort during an insulin tolerance test in two matched groups of subjects given nominally equivalent doses of the two glucocorticoids (834). In other experiments, propanolol (40 mg), clonidine (0.3 mg), or pyridostigmine (120 mg) was administered orally followed by intravenous GHRH infusion 60 min later with or without a prior nocturnal dose of dexamethasone (8 mg orally). Both propanolol and pyridostigmine were able to reverse (partially) the inhibitory effect of dexamethasone on GHRH-elicited GH release. These data suggest that the inhibitory effect of glucocorticoid excess on GH release is due to increased hypothalamic somatostatin secretion, which appears to be dependent on dexamethasone-enhanced ß2-adrenergic activity or glucocorticoid-suppressed cholinergic tone (835). However, direct pituitary effects are suggested by other studies showing an inhibitory effect of an acute intravenous infusion of hydrocortisone (cortisol) on baseline GH release in acromegaly (836). Moreover, bolus intravenous injection of hydrocortisone hemisuccinate, 100 mg, followed by intravenous continuous infusion of 250 mg over 120 min, in acromegalic patients, blunted the GH-secretory response to the physiological stimulus GHRH (837) and/or reduced the paradoxical GH response to TRH (484). In counterpoint, when acromegalic patients are pretreated with the functional somatostatin antagonist, L-arginine, acute hydrocortisone infusion does not suppress baseline serum GH concentrations, which favors an hypothesis of a cortisol-stimulated somatostatin-mediated inhibition of baseline GH release (838). Other confounding factors in glucocorticoid-treated patients, such as residual organ system disease in posttransplantation use of glucocorticoids for immunosuppression (839), must be considered in interpreting clinical studies.
Overall, available clinical data thus indicate that acute or short-term glucocorticoid exposure may either inhibit or stimulate GH release in normal subjects or acromegalic patients. In our view, such effects likely occur via several mechanisms, i.e., namely, changes in somatostatin release, which may in turn reflect direct hypothalamic actions of glucocorticoids (see above), changes in hypothalamic neurotransmission, and/or altered peripheral feedback signals. Furthermore, the magnitude and direction of glucocorticoid effects are influenced by the nature, potency, and half-life of the particular glucocorticoid administered, its route of administration, and the time of observation.
Numerous studies have examined GH secretion in prepubertal children undergoing long-term immunosuppressive glucocorticoid treatment (552, 840). In these children, GH responses to various pharmacological stimuli are reduced, as expected from the adult paradigm. Similarly, spontaneous GH secretion is decreased (552). Pyridostigmine significantly enhances both GHRH and sleep-induced GH release in children, but such actions are reduced during long-term glucocorticoid treatment. We note that acetylcholinesterase inhibitors may also act via stimulating GHRH secretion, the action of which is impaired during chronic glucocorticoid treatment (552). Thus, attenuated pyridostigmine actions are consistent with, but not proof of, somatostatin’s involvement in glucocorticoid inhibition of GH secretion in children. Indeed, since neostigmine also releases GHRH into the sheep hypothalamo-pituitary portal circulation (29), and a GHRH antagonist blocks pyridostigmine’s action in the human (269), pyridostigmine’s limited efficacy in chronic glucocorticoid therapy could also reflect restricted GHRH release and hence reduced stimulation of GH secretion by pyridostigmine.
GH secretion in adults receiving chronic immunosuppressive therapy with glucocorticoids is also impoverished. Pharmacological agents such as clonidine and galanin, which are presumed to effect GH release in part via increased hypothalamic GHRH release, are less effective in enhancing baseline GH concentrations and facilitating GHRH-induced GH release in glucocorticoid-suppressed adults compared with controls (441). On the other hand, L-arginine and hexarelin or GHRP-6, which are thought to serve as GH secretagogues by acting in part as so-called functional somatostatin antagonists, virtually normalize the GH-secretory response to GHRH and galanin in adults treated with glucocorticoids (783).
Although the natural ligand for the GHRP receptor has not been
identified, patients with Cushing’s disease show blunted GH release to
the otherwise synergistic stimuli of GHRH and GHRP (334), as reported
in the glucocorticoid-treated rat (above). Whereas endogenous
hypercortisolism blocks both GHRH and GHRP-6 pathways, exogenous
glucocorticoids do not always impair GHRP-6- releasing mechanisms
(841). The latter inference was confirmed in other patients undergoing
long-term (no fewer than 6 months) immunosuppressive glucocorticoid
treatment for nonendocrine disease, who exhibited unimpeded GH release
after an intravenous injection of GHRP or hexareliln (without or with
GHRH). The suppressive influence on GH secretion of prednisolone is
reversed by a high dose of L-692,429 (MK0677), a nonpeptidyl mimic of
GHRP-6 (842). Thus, the balance of evidence in the human favors the
interpretation that high-dose and/or chronic glucocorticoid treatment
inhibits basal GH secretion and that driven by GHRH, but to a lesser
extent that driven by GHRP. In contradistinction, physiological
amounts of glucocorticoids are required to maintain normal basal and
GHRH-stimulated pituitary GH reserve (Fig. 13).
The clinical significance of the suppressed GH/IGF-I axis in adults treated with glucocorticoids is reinforced by short-term intervention trials. GH administration elicits a significant increase in nitrogen balance, serum osteocalcin, the carboxy-terminal propeptide of type I procollagen, and carboxy-terminal telopeptide of type I collagen. GH treatment also improves protein synthesis without altering protein breakdown in patients receiving glucocorticoid treatment (843). Indeed, glucocorticoid-induced protein catabolism is reversed during coadministration of GH (844), whereas cotreatment with IGF-I and GH elicits net anabolism (845). In patients undergoing long-term glucocorticoid therapy for nonendocrine diseases, GH coadministration remains able to significantly lower high-density lipoprotein and low-density lipoprotein cholesterol, but increases serum triglyceride levels (678).
B. Gonadal sex hormones
1. In vitro and animal studies. To establish the necessary
foundation for interpreting more recent data on gonadal sex hormone
regulation of the GH axis, we will first highlight issues introduced in
an earlier review in the Journal in 1992 (1). This
background will be reappraised in view of new insights and technologies
developed within the last half-decade (76). Moreover, as highlighted in
Fig. 14, wherever possible the actions
of testosterone will be distinguished as via its aromatized product
estradiol, or by way of its 5- reduced potent androgen metabolite,
dihydrotestosterone (143, 846, 847, 848).
|
Whereas there is sexual dimorphism of GHRH mRNA expression, the specific role of gonadal steroids in regulating this difference is less clear. For example, hypothalamic GHRH gene expression in adult and aged male rats in one study was not influenced by castration or chronic testosterone replacement (853). In females, GHRH mRNA levels remain constant throughout the estrous cycle (849) and pregnancy (850). The latter point is particularly noteworthy, since placentally derived GH (and hence presumably GH negative feedback) rises during pregnancy (854). In contrast, other authors report that brain GHRH mRNA content in the rat is reduced by castration, and this drop is reversed by testosterone or dihydrotestosterone (but not estradiol) replacement (855). Such experiments suggest that in the rodent testosterone exerts a stimulatory effect on GHRH gene expression via androgen receptors rather than after its aromatization (856).
In relation to GHRH action, one study disclosed that in vivo treatment with testosterone increased the subsequent stimulation by GHRH of pituitary GH release in vitro (857). In another in vivo investigation, gonadectomized male rats treated with testosterone exhibited enhanced responsiveness to GHRH (858). Such observations suggest that testosterone may also modify GHRH actions on the pituitary gland, via one or more mechanisms activated in vivo and persisting in vitro (see below).
The role of estrogens in modulating GHRH expression or action in the rat seems to be minor compared with that inferred for androgens. For example, treating adult ovariectomized rats with estradiol does not alter GHRH-stimulated GH release (858). In a study of male rats, estrogen administration reduced hypothalamic GHRH mRNA content (859). Although subpopulations of immunoreactive GHRH-containing neurons take up [3H]estradiol in the rat (860), how and the extent to which arcuate nucleus GHRH synthesis and release are regulated by estrogens, and whether such putative regulation occurs directly or via changes in somatostatin or other neurotransmitter input to or actions on GHRH neurons (861), remain largely unknown.
The impact of gender on basal as well as secretagogue-stimulated release of hypothalamic GHRH has been studied in vitro in rats ranging from 10 days to 14 months of age. Ge et al. (862) showed that, although there is no difference in baseline hypothalamic GHRH secretion in vitro, pharmacologically stimulated GHRH release is higher in male than female rats. This is consistent with, but does not necessarily account for, higher serum GH pulses in male rats (863). Indeed, a sensitizing action of somatostatin withdrawal on GHRH release and action, and inferred intrahypothalamic interactions between GHRH and somatostatin neurons, are likely to contribute to the biological rhythmicity of GH release in the male rat (140, 157, 265, 555, 864, 865).
Based on experiments employing passive immunization against, and
controlled infusions of, GHRH and somatostatin in the rat, a current
thesis is that putative sex-specific differences exist in reciprocal
GHRH and somatostatin release patterns in male and female animals,
which dictate the prominent sex-specific features of pulsatile GH
secretion that typify male and female rodents (863, 866). To date, to
our knowledge, no direct evidence documents disparate modes of
hypothalamo-pituitary portal vein GHRH and/or somatostatin release in
male and female rats. However, indirect studies conducted in monosodium
glutamate-treated animals to lesion GHRH-secreting neurons, and in
GHRH-passively immunized rats, suggest that baseline GHRH secretion may
be elevated with lower amplitude (and possibly more frequent) GHRH
pulses in females, and conversely that baseline GHRH release is lower
with higher-amplitude (and possibly less frequent) GHRH secretory
bursts in males (863). In the male, more regularly recurrent intervals
of significant somatostatin withdrawal are also inferred to contribute
mechanistically to the 3- to 3.5-hourly periods of volley-like GH
release (Fig. 2). Further experiments will be necessary to establish
whether such proposed gender differences exist with respect to in
vivo hypothalamic GHRH and somatostatin release, and if so how
they are endowed specifically by sex steroids (143, 867). It is
pertinent that a recent study indicated that ovariectomy in rats
increases hypothalamic GHRH receptor expression, which is diminished by
subsequent estradiol supplementation (255). Thus, sex differences in GH
release in the rat may reflect not only unequal GHRH and somatostatin
release per se, but also differences in GHRH reception at
one or more neural levels.
3. Somatostatin. The extent to which somatostatin mRNA content
in the periventricular nucleus differs in male and female rats remains
controversial. Werner and colleagues (868) and others (869) failed to
observe any sex difference in hypothalamic somatostatin mRNA
concentrations in intact animals. However, when evaluating
pre-pro-somatostatin mRNA, Zorrilla et al. (869) reported
that periventricular in situ expression falls after male or
female gonadectomy, and is restored by 5-dihydrotestosterone (DHT),
estradiol, or a dopamine agonist. Others describe a markedly lower
concentration of periventricular nuclear somatostatin mRNA in
proestrous females compared with males (851). In addition, a sex
difference in somatostatin gene transcript levels was evident on the
10th day of life (851), the earliest time studied. Developmentally,
male rats exhibited significant increases in somatostatin mRNA between
days 10–25 and females between 10–35 days of age. However, such
differences in somatostatin mRNA content in relatively young males and
females are at variance with the lack of sexual dimorphism in the
GH-secretory pattern, somatostatin peptide concentrations, or growth in
animals before 33 days of age (870). Thus, differential mRNA expression
taken alone may not accurately reflect peptide expression and secretory
output of the GH axis.
There is less ambiguity concerning the effects of gonadal steroids on hypothalamic somatostatin mRNA concentrations in the rat. Castration of adult male rats significantly lowers concentrations of somatostatin mRNA in the periventricular nucleus (868, 871). Testosterone or DHT replacement therapy reverses this effect of castration. With the exception of one study, estrogens do not appear to be so effective in altering somatostatin gene expression in males (851, 868). Similarly, in females, ovariectomy significantly decreases hypothalamic somatostatin mRNA content, whereas treatment with estradiol, or in one study DHT (872), reverses these decrements. Thus, sex steroids likely facilitate somatostatin gene expression, with a reduction in gonadal sex steroids precipitating similar decreases in somatostatin gene transcripts in males and females.
Since present data suggest that no estrogen-receptor binding consensus sequence exists in the somatostatin gene promoter (873), indirect mechanisms or sites of sex steroid action may be relevant. Of interest, 40–70% of somatostatin-secreting neurons, including these in the periventricular area, express nuclear androgen, but not estrogen, receptors, which makes these neurons a plausible (albeit not established) target of direct androgen actions (874, 875, 876). The male rat shows greater androgen receptor expression in somatostatin neurons than the female (875).
Two of three reports disclosed no differences in hypothalamic somatostatin peptide content in the male vs. female rat (877, 878, 879). Baseline and stimulated somatostatin secretion from hypothalamic fragments in vitro also was similar in the two sexes (862, 878). Thus, somatostatin peptide, unlike its mRNA, shows little or no evident sexual dimorphism in the rodent. This disparity attests further to the need to establish gender-specific patterns of somatostatin release and action by more direct means, e.g., based on hypothalamo-pituitary portal vein sampling and somatostatin-receptor antisense knockout.
Other indirect studies also offer insights into plausible gender-based differences in GHRH/somatostatin secretion and/or action. Wehrenberg et al. (880) inferred that discrepant responses to exogenous GHRH stimuli in conscious freely moving male vs. female rats reflect unequal endogenous somatostatin secretion and action. Other experiments by Clark and Robinson (881) show that female rats respond consistently with unabated GH secretion to repeated GHRH administration, whereas male rats respond only briefly and intermittently, further suggesting that patterns of hypothalamic somatostatin are different in male and female animals. Somatostatin immunoneutralization also evokes disparate GH release profiles in male and female rats (863, 882). Thus, a plausible inference in the rodent is that in males somatostatin is expressed (129) and released (863) in more regular (3- to 3.5-hourly) and higher-amplitude discharges than in females (870). Conversely, female rats may secrete somatostatin approximately continuously, more nearly randomly, and in smaller amounts (53, 883).
Recent approximate entropy analyses of both human and rat GH release
profiles in vivo can quantify a notion of decreased
orderliness of GH release in the female (139). This contrast is shown
in the human in Fig. 15, wherein serum
GH concentrations in blood collected every 10 min for 24 h were
measured by high-sensitivity immunofluorometric assay (118, 124, 139).
This entropy difference also is consistent with a postulated reduced
regularity of cyclic somatostatin withdrawal in the human and rodent
female. An antiestrogen attenuates the rat’s male-like pattern of GH
pulsatility, possibly via reducing somatostatin release (884),
suggesting a role for estrogen in directing the male rodent’s GH
pattern. However, in the human, estrogen (or testosterone, but not DHT)
increases the entropy of GH release, thereby inducing a more female
pattern (139, 142) (Fig. 15). We speculate that
GHRH-somatostatin-interactions within the hypothalamus further modulate
such apparent gender differences in GH neuroregulation, but how sex
steroids in particular govern these interactive mechanisms remains
unclear (183, 861, 863).
|
In relation to sex differences in pituitary GH gene expression, the concentration of GH mRNA is approximately 2 to 6 times greater in adult male than female rats (885, 886). Whether this distinction arises from direct actions of gonadal steroids on pituitary GH gene transcription or mRNA stability, or via indirect sex-steroid effects on hypothalamic inhibitory and releasing hormones, or their receptors, is not established. The former indirect mechanisms would include higher GHRH pulses in males, which in turn induce GH gene transcription (887). Since somatostatin inhibits GH release, but not GH biosynthesis, presumptively greater somatostatin release in the male rat would not directly limit GH mRNA accumulation, but would allow greater GH storage in somatotropes awaiting the next GHRH pulse (888).
Few studies have evaluated direct actions of gonadal steroids on GH gene expression. Estradiol does not directly influence GH mRNA accumulation in the bovine pituitary gland (889) or alter in vitro GH secretion by macaque somatotropes (890). On the other hand, estrogen stimulates GH mRNA gene expression in MtT/F4 transplantable tumors in 40-day-old female rats (891). The latter finding diverges from the general theme that estrogen inhibits pituitary responses to GHRH in the normal rat (see below), and that serum GH concentrations are lower in the female of this species. A plausible explanation is that the MtT/F4 tumor is a unique estrogen-responsive neoplasm and therefore may not be an appropriate model for evaluating normal physiology. On the other hand, in one study, 80% of normal rat somatotrope cells expressed [3H]estradiol binding sites (860), thus allowing possible direct actions of estrogen on GH-secreting cells. In this regard, estrogen treatment of male rats in vivo reduces in situ GH mRNA expression in single somatotropes in vitro (892). Such in vitro observations differ from other in vivo studies in the ovariectomized female rat, in which androgens, but not estrogen, can augment pituitary GH content (893). Consequently, far more study will be needed to clarify the nature of sex steroid actions (if any) directly on somatotrope cells.
The GH-secretory response to GHRH stimulation in the rat tends to be sexually dimorphic. In vitro perifusion studies reveal that pituitaries from male rats release more GH in response to GHRH than those of females. Furthermore, either neonatal or prepubertal orchidectomy reduces baseline and GHRH-stimulated secretion of GH in vitro (857). Male rats gonadectomized on day 22 of age show a decreased in vitro response to GHRH, which is reversed by in vivo testosterone replacement (894). On the other hand, in female rats prepubertal ovariectomy does not affect GH secretion in vitro, and estradiol replacement in adult castrate animals lowers baseline and GHRH-stimulated GH secretion in vitro (857).
A gender-appropriate sex-steroid environment is required throughout adult life to maintain physiological GH secretion in the rat in vivo (883). Basal or GHRH-stimulated secretion of GH in vitro is significantly higher by pituitary tissue obtained from adult male compared with female animals (878, 895). Orchidectomy followed by various sex hormone replacement therapies suggests that this gender distinction is mediated primarily by testosterone (896). Androgen’s enhancement of GH release reflects, in part, increased sensitivity to GHRH and greater GH secretory capacity of individual somatotropes, as studied in vitro in the reverse hemolytic plaque assay (897, 898). Greater somatotrope GH content also has been inferred on a histological basis (877). However, other (non-GHRH) modulatory inputs to somatotropes may be influenced by androgens, such as somatostatin, and possibly the putative endogenous GHRP pathway, galanin, NPY, IGF-I feedback, and GH autofeedback. For example, the exogenous GHRP stimulus is enhanced by both testosterone and estrogen in the pituitary-transplanted rat (899).
The effects of estrogens on pituitary responsiveness to GHRH in the rodent are controversial. Simard et al. (900) reported that estradiol exposure stimulates in vitro basal and GHRH-stimulated GH secretion by pituitaries obtained from female rats, whereas Fukata and Martin (901) reported no effect of estrogen on these measures. Other experiments indicated that estradiol treatment of castrated male rats leads to reduced in vitro GH secretion similar to that reported in female animals (896). Thus, whereas androgens consistently facilitate the pituitary GH response to GHRH in the rat, estrogen effects are inconsistent or inhibitory.
Understanding the mechanistic basis of the distinctive patterns of GH secretion in the two sexes, i.e., higher and more regular GH volleys in males, and lower-amplitude irregular peaks in females, has been aided by studies in androgen receptor-deficient rodents (e.g., testicular feminized rat). The latter genetically male, but androgen receptor-deficient, rodent shows intermediate (to intact male and female, with intact female having the greatest) single somatotrope cell sensitivity to somatostatin’s inhibition in the reverse hemolytic plaque assay (902). The testicular feminized rat also exhibits an intermediate in vitro GH secretion rate in response to a maximal GHRH stimulus, i.e., less than of an intact male and more than that of an adult female (895). In further single somatotrope studies, Martha et al. (903) identified more functional somatotropes in the male testicular-feminized rat than the normal female, but each somatotrope cell secreted less GH basally than the wild-type male. These differences could explain the more nearly "female" GH release pattern in vivo of the testosterone-producing but androgen receptor-deficient testicular-feminized animal (856) and is consistent with a notion of greater somatotrope-secretory activity in the male. How such differences arise mechanistically is not yet established.
In nonhuman and nonrodent species, estrogens typically modify the GH axis in a stimulatory manner. For example, studies with single bovine pituitary somatotrope cells using the reverse hemolytic plaque assay document increased mammosomatotropes (PRL- and GH-secreting cells) in the luteal phase (904). On the other hand, in this species, bulls have higher mean serum GH concentrations, GH pulse frequencies, amplitudes, and GHRH-stimulated GH release than heifers (905). In sheep, increased pituitary GH mRNA content and serum GH concentrations develop in the late follicular phase, but the estrogen-induced LH surge 22 h later occurs without a further change in GH mRNA concentrations (906). In the baboon, estradiol in either physiological or pharmacological amounts delivered via silastic subcutaneous capsules increases serum IGF-I concentrations and stimulates GH release in long-term ovariectomized animals (907). In this primate, basal serum GH concentrations also are higher in females than in males. These gender differences are likely mediated via the hypothalamus, since estrogen stimulates in vivo GH secretion but has little effect in vitro on GH release by primate (macaque) pituitary cells (890).
4. GH receptor and BPs. Sex steroids modify GH receptor expression in the liver in a sex hormone- and species-specific manner. For example, in the rabbit (908), estradiol inhibits GH receptor gene expression. On the other hand, in the rat, estrogen consistently increases liver GH receptor and GHBP expression (909), as discussed further below. In the human, oral estrogen treatment also augments plasma GHBP levels, which is believed to reflect hepatic GH receptor (extracellular domain) expression (910). Contrastingly, whether in the presence or absence of endogenous GH, testosterone treatment in men lowers serum GHBP concentrations (911).
Since the demonstration that the GHBP in rat plasma is an alternatively spliced transcript of the GH receptor gene (912), the tissue distribution and regulation of GHBP mRNA and protein have been studied extensively (912, 913, 914, 915, 916, 917, 918). Molecular experiments document species variations in the spectrum of mRNA transcripts that encode the GH receptor and GHBP, e.g., in the pig, rat, chicken, and human (917), as well as in their gestational, developmental, and tissue-specific regulation in the rodent (916). There is gender-specific (sexually dimorphic) liver GHBP production in the rat, as reflected by sex steroid and pregnancy’s regulation of GHBP transcript concentrations. GHBP gene and protein expression patterns are also controlled by continuous (female-like) vs. pulsatile (male-like) GH exposure (915). For example, the so-called GHR-1 is an alternatively spliced mRNA class encoding the GHBP, which is GH dependent, signal specific (responsive to continuous rather than pulsatile GH), induced by gonadal steroids, and eliminated by hypophysectomy (915). Such in vivo evidence of steroidal regulation of specific GHBP expression has been corroborated more directly recently in an in vitro model of primary adult rat hepatocytes (914).
Other hepatic genes that are responsive to a sex-specific pattern of GH
secretion, and are subject to androgen imprinting, include the hepatic
cytochrome P450 2C11, 3A2, and 5-reductase enzymes (919) (see Table 3). Gender differences in GH tissue actions are attributed to pulsatile
(male-like) vs. more nearly continuous (female) patterns of
GH secretion (920). Indeed, specific gene induction in various target
tissues of GH action (e.g., liver, muscle, etc.) is
differentially responsive to the GH pulse pattern (921, 922).
Second-messenger intracellular signaling per se is activated
differentially via a pulsatile vs. continuous GH stimulus,
with the former (but not the latter) activating the STAT5b-signaling
cascade (866). Demonstrating the critical nature of the STAT5b
signaling pathway, STAT5b knockout mice fail to respond with male
growth rates or gene induction despite "masculine" GH pulse
patterns in the blood (32).
5. Human studies.
a. Spontaneous GH secretion.
Gender differences in human GH
secretion were reviewed in part by Kerrigan and Rogol in the
Journal in 1992 (2), and elsewhere (76). Here, more recent
findings are evaluated, and earlier observations reinterpreted in the
context of new hypotheses and investigative strategies (76). For
example, although androgens clearly act on the GH-IGF-I axis, recent
experiments also emphasize their capacity to alter the orderliness of
GH release and for independent trophic effects without evident
stimulation of the GH axis, e.g., to exert anabolic tissue
effects (142, 923).
Androgens in the absence of GH are insufficient in the human to drive the fully normal tempo of clinical pubertal maturation, since hypopituitary boys replaced with testosterone but not GH exhibit a protracted 7- to 8-yr (rather than 3- to 4-yr) long pubertal growth period (924). In contrast, a marked androgen-GH synergy unfolds in normal puberty (848), as well as in pubertally delayed boys treated with combined GH-testosterone replacement therapy. Serum testosterone concentrations in the course of normal puberty span a wide spectrum, viz., from prepubertal values (<1 nmol/liter) to adult concentrations (20–25 nmol/liter). Cross-sectional and longitudinal studies of healthy pubertal boys have documented the physiological association between rising blood androgen concentrations and augmented neuroendocrine activity of the somatotropic axis. Initial studies reported GH pulse activity using discrete peak detection methodologies and demonstrated that the mean serum GH concentration peak amplitude in boys increases 2- to 3-fold in the mid- to later stages of puberty, with no change in the number of detected GH peaks by cluster analysis (925). Later, a deconvolution technique for estimating endogenous hormone secretion rates and/or half-life has been implemented to quantitate 24-h serum GH profiles in males whose pubertal development spanned the range of Tanner stages I (prepubertal) through V (fully mature adult). In these cross-sectional analyses, boys in mid-to-late puberty exhibited higher mean and pulsatile serum GH concentrations, mechanistically explained by amplification of GH-secretory burst mass and amplitude due to a rise in the maximal rate of GH secretion (amplitude) attained within each secretory pulse (926). Linear correlation analysis demonstrated that serum total testosterone concentrations in pubertal boys are strongly positively correlated with plasma total IGF-I levels, the calculated daily pulsatile GH secretion rate, and GH- secretory burst mass (926). In a novel clinical model of leuprolide-induced gonadotropin down-regulation in normal young men (923), Fryburg et al. showed dose-dependent actions of testosterone repletion on each of these measures of the GH axis. A nonaromatizable synthetic androgen did not activate the GH axis (or increase serum IGF-I levels), but promoted nitrogen retention, thus emphasizing the non-GH-dependent anabolic effects of androgens (923).
Other clinical intervention investigations in boys with
constitutionally delayed puberty revealed that testosterone
administration significantly increases the amplitude (or mass) of
GH-secretory peaks in a selective manner without altering the GH
half-life, pulse frequency, or pulse duration (927). A qualitatively
similar response was observed in four of five boys after administration
of oxandrolone, a synthetic nonaromatizable androgen, but this substudy
did not include controls for possible placebo or order effects of
treatment intervention. More recently and informatively, a
longitudinal/prospective study of nocturnal pulsatile GH secretion was
performed in six previously untreated boys with isolated
hypogonadotropic hypogonadism. Deconvolution analysis showed that the
amount of GH secreted in pulses was significantly increased by as
little as 25 mg of injected testosterone (928). Both the number of
spontaneous GH peaks and their mean amplitude rose significantly after
doses of 50 and 100 mg testosterone. This suggests that in young human
males even minimal increments in serum testosterone concentrations
occurring during the early stages of puberty can amplify both the
amount and frequency of spontaneous GH-secretory bursts (928).
Moreover, the orderliness or regularity of GH release, as quantified by
approximate entropy (132), decreased rapidly during testosterone
treatment. The more irregular GH release patterns induced by
testosterone injections in these boys closely mirrors the more
disorderly mode (reduced pattern reproducibility) of GH secretion that
emerges before and during maximal linear growth in normal pubertal boys
(142) (see Fig. 16). Thus, androgens
control not only the quantity but also the quality (orderliness or
regularity) of the GH release process, presumably indicate network
effects of sex steroids on the integration of
somatostatin-GHRH-GH-IGF-I feedback/feedforward interactions.
|
Testosterone treatment of primary hypogonadal patients increases both serum GH and IGF-I concentrations (142, 923, 930), whereas an antiestrogen antagonizes this response (930). Administration of a nonsteroidal antiandrogen to healthy young men significantly increases pulsatile GH secretion (931), and conversely, an antiestrogen significantly reduces GH production (932). Thus, the inferred stimulatory action of testosterone on pulsatile GH release in the human (unlike the rat) is likely mediated via the estrogen, rather than androgen, receptor. Indeed, administration of a nonaromatizable androgen, DHT, to pubertally delayed boys fails to stimulate GH secretion or alter its orderliness of release (approximate entropy), unlike the actions of testosterone or ethinyl estradiol (142). Thus, in man, androgen action after its aromatization to estrogen dictates both quantitative and qualitative features of GH secretion (848, 932). The ability of incremental doses of testosterone replacement in hypogonadal boys to progressively increase GH secretion basally, as well as after GHRH coadministered with the functional somatostatin antagonist, L-arginine, further suggests that aromatizable androgens augment maximal pituitary GH secretory capacity (928).
Serum GH concentrations rise throughout puberty in both sexes. In normal girls, pubertal GH elevations are proportionate to the rise in serum estradiol levels. Mean serum GH peak amplitude increases 2- to 3-fold from prepuberty to menarche, which increment (especially in the daytime) is correlated to serum estradiol concentrations (933). This relationship supports the notion that in puberty in girls endogenous estradiol, or a pertinent covariate, drives amplified GH secretion. Similar GH axis activation by estrogen is demonstrable in the castrate (female) baboon (907) and macaque (890). Moreover, in prepubertal girls with previously untreated Turner’s syndrome, oral administration of a small dose of ethinyl estradiol (100 ng/kg/day) amplifies the GH secretion rate 2- to 3-fold and heightens the disorderliness of GH release, the latter as quantified by the approximate entropy statistic (934). Similar responses are achieved in prepubertal boys treated with testosterone, but not DHT. This points to an augmentative role of estrogen (or aromatized androgen) in GH axis neuroregulation in both boys and girls (142).
Gender differences in GH release in adults were recognized in 1965 by Frantz and Rabkin (935), who described similar fasting serum GH concentrations in men and women in a cohort of 79 individuals (ages 20 to 80 yr) (935), but a marked (6-fold) increase in serum GH concentrations after ambulation in women only. Administration of a potent oral nonsteroidal estrogen, diethylstibestrol, 2.5 mg twice daily for 5 days in 4 men evoked a "female pattern" of GH release after exercise. These authors further identified a 2-fold rise in serum GH concentrations across the menstrual cycle and inferred that gender contributes to the regulation of GH release and probably does so via estrogen (935). Later studies by Unger et al. (936) described a sex difference in basal serum GH concentrations. Thompson et al. (92) extended the inferred role of estrogen in stimulating higher GH secretion rates in premenopausal than in postmenopausal women. More directly, Dawson-Hughes et al. (937) demonstrated that 20 µg ethinyl estradiol administered orally for 15 days elicits an approximately 45% increase in the 24-h mean serum GH concentration in older women. Similarly, Dursma et al. (938) showed that 3 weeks of this estrogen regimen will stimulate a rise in serum GH concentrations, while reducing plasma somatomedin C (IGF-I) levels.
By way of possible physiological regulation, serum GH concentrations rise 2-fold in the late follicular phase of the normal menstrual cycle, based on specific GH IRMA in young women sampled every 10 min for 24 h (939). More dramatic increases in endogenous estradiol secretion induced by superovulation treatment of infertile women can stimulate serum GH concentrations by up to 4-fold (940). Conversely, suppressing ovarian estrogen secretion by GnRH agonist down-regulation of the gonadotropic axis in premenopausal women or girls with precocious puberty reduces basal and GHRH-stimulated GH release (941, 942), as well as plasma IGF-I concentrations (943). A recent analysis using longitudinal (within-subject) comparisons in the same menstrual cycle further disclosed that GH-secretory pulse frequency and amplitude, and plasma IGF-I concentrations, rise concomitantly in the preovulatory phase with the serum estradiol zenith (944). This physiological reactivity of the GH-IGF-I axis with combined augmentation of GH and IGF-I concentrations is not fully recapitulated by oral or transdermal estrogen delivery in postmenopausal women (below), but is achieved in the long-term castrate (female) baboon by crystalline estradiol delivery via subcutaneous silastic capsules (907) or parenteral estrogen in the macaque monkey (890).
Weissberger et al. (945) suggested that oral (but not transdermal) estrogen delivery augments GH secretion in older women and does so by reducing plasma IGF-I concentrations. This would attenuate IGF-I’s negative feedback on the hypothalamic-somatotropic axis. Three different oral estrogens exerted this effect and concurrently increased serum GHBP concentrations (24, 946). Other studies show that oral estrogen treatment may not always lower plasma IGF-I levels (947), and that higher doses of transdermal estrogen can indeed stimulate GH secretion (204). In contrast to these paradigms, normal female puberty and the preovulatory phase of the menstrual cycle are marked by conjoint elevations in circulating estrogen, IGF-I, and GH concentrations (140, 933). Thus, whether during puberty or just before ovulation, reduced plasma (total) IGF-I feedback cannot provide the basis for amplified GH release. In addition to physiological follicular phase estrogen secretion, parenteral estradiol treatment of castrate or intact baboons increases serum GH and IGF-I levels concurrently (907). This experimental paradigm mimics that observed in puberty and in the normal preovulatory phase of the menstrual cycle and suggests that estradiol can enhance the net central drive to GH release and thereby concurrently stimulate IGF-I secretion. Thus, the route and amount of estrogen, as well as the pubertal or menstrual context, likely govern the nature of estrogen’s effects on the GH-IGF-I axis (204, 910).
The relationship between estrogen and IGF-I may be complex, and tissue specific, e.g., different in liver vs. uterus. For example, estrogen inhibits liver but stimulates uterine IGF-I accumulation in the hypophysectomized rat (948). This could explain the ability of low doses of oral estrogen to stimulate linear growth in patients with Turner’s syndrome, whereas the same or higher doses of estrogen reduce hepatic IGF-I production (934). Other indirect evidence suggests that estrogen administration may attenuate plasma IGF-I responses to fixed doses of exogenous (human recombinant) GH in postmenopausal women (949); i.e., there is a possible gender difference in tissue sensitivity to GH actions, with reduced end organ (at least liver) responsiveness to GH in women treated with estrogen (950). This might reflect the ability, in some species, of estrogen and GH itself to down-regulate hepatic GH receptors (above), or impede GH tissue responses. In contrast, multiple reports demonstrate that parenteral testosterone stimulates increases in both GH secretion and plasma IGF-I concentrations (847, 910, 923, 930). This IGF-I response to aromatizable androgen mechanistically distinguishes it from oral estrogen action in the human.
Analyses of the responsiveness of the anterior pituitary gland to GHRH injections in women compared with men and in women throughout the menstrual cycle are conflicting. For example, the evaluation by Gelato et al. (205) of the GHRH dose-response curve (0.01 to 10 µg/kg) throughout the normal menstrual cycle in young women detected no systematic variations in the magnitude of GH responses. Although maximal GH release was similar in men and women, the half-maximally effective dose of GHRH was reduced in women at 0.2 µg/kg compared with 0.4 µg/kg in men. Evans et al. (210), utilizing a single maximal dose of 3.3 µg/kg GHRH, also found no GH-secretory differences among the early follicular, late follicular, and midluteal phases. In a later study, there were similar maximal GHRH effects in men and women basally and after GnRH agonist-induced down-regulation of the gonadal axis for 40 days (951). However, to our knowledge, studies of GHRH dose-dependent actions in the same woman at different stages of the menstrual cycle with vs. without somatostatin withdrawal are not yet available.
Unlike the above findings, other investigators identify larger effects of GHRH in women than men (952) and greater maximal GHRH-stimulated (incremental) GH release in the late follicular phases of the menstrual cycle. The latter rises correlated with higher serum estradiol concentrations (952). In a large study of 116 men and women ages 18–95 yr, Lang et al. (953) unmasked higher incremental release to GHRH stimulation in premenopausal women vs. age-matched men. Serum estradiol concentrations correlated with the magnitude of maximal GHRH-stimulated GH release. GH increments were equal in postmenopausal women and older men, who have similar estrogen levels.
Greater GH release in women than men is also reported for L-arginine and insulin-stimulated hypoglycemia (954), stimuli acting at the hypothalamic level (681, 783). Treatment with estrogen (but not testosterone’s methylated derivative) augmented L-arginine and insulin-induced GH release in men, which suggests that these gender differences reflect unequal estrogen exposure. Another peptide also thought to stimulate GH secretion via hypothalamic pathways, i.e., galanin (435), elicits more GH secretion in young women than men (432, 445, 955). Whereas the non-GHRH hexapeptide, GHRP-6, had no gender bias in its maximal GH-releasing properties (efficacy), this study did not evaluate the half-maximally effective GHRH-6 doses (potency of, or GH-axis sensitivity to, this potent secretagogue) (291, 416).
Although the mechanism of glucagon’s stimulation of GH release is not
clear, no sex difference is evident (956). Conversely, clonidine, an
2-agonist, stimulates more GH release in men than
premenopausal women (597). Most recently, whereas young early
follicular phase women were more responsive to L-arginine
than men, young men exhibited greater GH release after GHRP-2
stimulation than comparably aged women (957). Combining these two
secretagogues with an exercise stimulus yielded maximal GH secretion,
which was equivalent in the two sexes (957).
We infer that the ability of pharmacological agonists to stimulate GH secretion varies as a function of gender and depends on the specific secretagogue and neurotransmitter pathway employed, the age and the gonadal status of the individual, the stage of the menstrual cycle, the intensity of the stimulus, and possibly the estrogenicity of the endocrine milieu and route of estrogen delivery.
6. Influence of sex steroids on GH clearance. Steady-state GH infusions in the rat indicate that the gonadal sex steroid environment may influence the MCR of GH. For example, the orchidectomized rat shows a decrease in GH MCR during estrogen add-back, and an increase after testosterone add-back (958). Since the half-life of GH disappearance from steady state was similar in male and female animals, this study suggests that sex steroids might alter the volume of distribution of GH, which was not measured here. An infusion study in the human suggested as much as a 30% mean decrease in the equilibrium GH MCR in women compared with men, although the distribution volume and half-life of GH were not estimated (959). On the other hand, another recent clinical investigation of the half-life of bolus-injected human recombinant GH in young men revealed an inverse correlation of half-life and serum estradiol concentrations, which would suggest a direct (rather than negative) relationship between estradiol and GH MCR, at least in men (960). In young women, stage of the menstrual cycle had no influence on exogenous (bolus) biosynthetic human GH metabolic clearance, half-life, or distribution volume (961). Men and women had similar GH half-lives and distribution volumes (adjusted for body surface area) in this study. Equivalent half-lives of endogenous GH were also estimated recently by deconvolution analyses in men and women (118). Thus, kinetic differences based on gender are not settled in the human. However, quantitatively, unequal GH removal rates, if confirmed in women and men, at a difference of only 20–30% would account marginally for the 200–300% differences in mean serum GH concentrations between the sexes in young adults and across the menstrual cycle (939, 944, 962).
C. Thyroid hormones
1. In vitro and animal studies. Brain GHRH expression in the
rodent is affected by alterations in circulating levels of thyroid
hormones (963). Propylthiouracil-induced hypothyroidism in adult rats
increases hypothalamic GHRH mRNA expression (964), while
hyperthyroidism acts conversely to decrease GHRH mRNA content (965).
The hypothalamic content of GHRH peptide is also affected by
hypothyroidism and hyperthyroidism in an age-related manner. In rats up
to 21 days of age, thyroid status does not affect the hypothalamic
content of GHRH or GHRH positive-staining structures (966). After that
age, hypothyroidism decreases GHRH content (967). These experiments are
consistent with, but not proof of, the hypothesis that there is
increased GHRH release in a homeostatic response by the animal to the
low serum GH concentrations that prevail in hypothyroidism.
Hyperthyroidism reportedly also decreases GHRH content, which would
concur with the fall in GHRH mRNA under these conditions (965).
The hypothyroid state enhances both basal and K+-stimulated GHRH secretion from in vitro incubated hypothalamic fragments (968), which might represent a compensatory mechanism to the low GH levels. Conversely, excess T4 inhibits hypothalamic GHRH secretion up to 50% (compared with control values) in this model. Using in vitro pulse-labeling techniques, De los Frailes et al. (969) observed that GHRH synthesis in fetal cortical cells is stimulated by a low concentration of T3 (10-7 mol/liter), but inhibited by higher amounts (10-3 mol/liter). Baseline culture medium contained no T3, and thus may have represented a biochemical "hypothyroid" state, while the low and high concentrations of T3 may resemble euthyroid and hyperthyroid states, respectively. Thus, these experiments suggest that hypothyroidism, as well as hyperthyroidism, may decrease GHRH synthesis at least in cortical neurons.
In vivo studies indicate that neural somatostatin synthesis is not affected by thyroid hormone during early neonatal rodent development (963). In contrast, in adult rats pharmacologically induced hypothyroidism decreased hypothalamic somatostatin mRNA concentrations (964). The hypothalamic content of somatostatin peptide also varies under hypo- and hyperthyroid conditions. In vitro cultures of fetal cortical cells obtained on day 18 of gestation, and treated with T3 or T4, become impoverished in somatostatin (970). However, the neonatal and prepubertal hypothalamic content and synthesis of somatostatin are not affected by thyroid hormones (964). In adult male rats, the somatostatin content in the median eminence falls with thyroidectomy compared with thyroid-intact animals and returns to normal with T4 replacement therapy (971). Hyperthyroidism, induced by a subcutaneous implant of T4, does not seem to alter the level of hypothalamic somatostatin peptide (972). However, peptide concentrations in neuronal loci must not be equated faciley with their de facto release rates.
Although the secretion of somatostatin from the hypothalamus is difficult to assess in vivo, in vitro release of somatostatin from fetal hypothalamic tissue is stimulated by a low concentration of T3 (10-11 mol/liter), but inhibited by a high dose (10-7 mol/liter) (969). As noted earlier, the baseline medium was devoid of T3 and therefore could reflect a "hypothyroid" condition, while the low and high dose of T3 might approach a normal and hyperthyroid state, respectively. This argument conforms with the observation that in vitro basal and stimulated somatostatin secretion from hypothalamic tissue of hypothyroid rats is decreased compared with that in euthyroid animals (971). Furthermore, chronic exposure to T3 or T4 suppresses neural somatostatin release after K+-induced depolarization (970). However, one publication suggests that thyroid hormone has no effect on somatostatin content or release despite altering pulsatile GH secretion (973).
Direct in vivo quantification of hypothalamic somatostatin secretion has been reported in pentobarbital-anesthetized rats. In this model, neither hypothyroidism nor hyperthyroidism affected hypothalamic-hypophyseal portal blood concentrations of somatostatin (974). Although such results would suggest that the thyroid environment does not affect somatostatin secretion, the conclusions must be tempered by the recognition that anesthesia has inhibitory and possibly other confounding effects on hypothalamic somatostatin (and/or other neuropeptide) secretion (809).
In somatotrope tumor cell lines, thyroid hormones stimulate GH mRNA accumulation (975, 976) by increasing the rate of GH gene transcription (977, 978). This effect of T4 also occurs in cultured normal rat anterior pituitary cells (979). Of note, thyroid hormone’s effects on GH gene expression and GH synthesis vary significantly among mammalian species; e.g., in bovine anterior pituitary cells, T3 fails to augment basal and GHRH-stimulated GH secretion (889), and cultured human fetal pituitary cells supplemented with T3 show reduced basal GH release and attenuated responses to GHRH (980). Also, in cultured human pituitary tumor cells obtained at neurosurgery from patients with acromegaly, T3 has no major impact on GH secretion (981). Furthermore, after transfection into rat pituitary cells, human GH gene expression is reduced by T3 administration (982), although expression of the rat and bovine GH promoters are induced by T3 (983). Thus, although the most prominent action in the rat of thyroid hormones on GH secretion at the somatotrope level is to promote the transcription of the GH gene and hence GH synthesis and accumulation, there are evident species differences.
Thyroid hormones also have important effects on pituitary GH protein content. In studies conducted in 10- to 21-day-old rat pups, hypothyroidism decreased pituitary GH content (966). Similarly, thyroidectomy in adult rats profoundly reduces pituitary GH concentrations with values dropping to less than 1% of those in normal animals (967, 973). T4 replacement therapy partially restores pituitary GH content.
Another mechanism through which thyroid hormones may act is regulation of pituitary sensitivity to GHRH. In fact, the GH response to GHRH is significantly lower in normal pituitary cells maintained in vitro without T3 compared with T3-exposed cells (798). The blunted action of GHRH reflects primarily a reduced maximal GH response to GHRH as distinct from a change in the ED50 of GHRH action. Cells obtained from hypothyroid animals likewise show a decreased response to GHRH in vitro. Most recently, decreases in not only GH and GHRH gene expression, but also in somatostatin and GHRH receptor mRNA, have been described in hypothyroid animals, thus highlighting an array of GH axis disturbances in this metabolic state (984).
2. Human studies.
a. Hypothyroidism.
In humans, as in other species,
hypothyroidism severely impairs postnatal growth (963). Moreover,
spontaneous nocturnal GH secretion, which is one indicator of
physiological GH production, is consistently low in the hypothyroid
state and is correlated with reduced circulating IGF-I levels (985).
Chernausek and Turner (985) reported that mean nocturnal serum GH
concentrations in hypothyroid individuals are reduced to 58% of values
obtained during T4 replacement therapy. Concurrently with
reduced spontaneous GH secretion, plasma IGF-I concentrations are lower
in the same patients before, compared with during, T4
replacement therapy. Replacement T4 therapy returns GH and
IGF-I levels to normal (986).
Consistent with the data derived from studies of spontaneous GH secretion, acute GH-secretory responses to the most common pharmacological stimuli, such as insulin-induced hypoglycemia, L-arginine and GHRH, are all blunted in hypothyroid patients (494, 987, 988). When hypothyroid patients are replaced with T4, plasma IGF-I concentrations rise toward normal values. Administration of T4 to hypothyroid patients also enhances the GH response to GHRH. In fact, in hypothyroid humans, the GH response to GHRH is normalized 2 weeks after T4 replacement therapy is begun (494, 989).
The possible hypothalamic origin of hyposomatotropism and pituitary hyporesponsiveness to GHRH in clinical hypothyroidism is supported by investigations using the acetylcholinesterase inhibitor, pyridostigmine, and the amino acid, L-arginine, both of which are hypothesized to decrease hypothalamic somatostatin release. Indeed, hypothyroid patients manifest reduced GH-secretory responses to putative somatostatin withdrawal via either pyridostigmine or L-arginine (989). Such observations speak against heightened somatostatin tone as the sole cause of GH deficiency and suggest that either endogenous GHRH action and/or GH synthesis in the pituitary gland may also be impaired during the chronic hypothyroid state in the human. These mechanisms cannot be distinguished based on available data, given the reduced GH response to GHRH demonstrated in hypothyroid subjects. Indeed, decreased pituitary GHRH-releasable pools of GH could result secondarily from reduced GHRH release and/or action, which would account for reduced 24-h GH secretion, as well as impaired acute GH-secretory responses to other pharmacological stimuli.
b. Hyperthyroidism.
The 24-h GH-secretory rate and the amount
of GH released during each pulse are reportedly reduced in adolescents
with untreated thyrotoxicosis compared with normal controls (990).
During antithyroid treatment, GH secretion rises in all previously
hyperthyroid subjects, but normalizes only in the subjects in whom
euthyroidism was achieved. Mean and peak sleep-related GH release also
are low in hyperthyroid adult patients (991). However, 24-h GH
secretion was normal or increased in another 24-h sampling study in men
studied in a more sensitive GH IRMA (992). In thyrotoxicosis, serum
total immunoreactive IGF-I levels are either elevated (993) or normal
(994). However, IGF-I bioactivity is low and rises after treatment
(995), which speculatively may be due to concomitant alterations in the
IGF-I BPs. Thyroid status also may influence IGF-I expression in
tissues other than the liver. In fact, IGF-I production is increased in
the hypothalamus after T3 administration (995) and
plausibly may participate in negative feedback by triggering either
increased somatostatin tone and/or decreased GHRH production. Such
postulated neuromodulatory responses might account for the
low-amplitude (but high-frequency) GH-secretory bursts reported
in hyperthyroid men (992).
Thyroid hormone excess consistently impairs GH release as stimulated by
several secretagogues in man (963). For example, the GH-secretory
response to insulin-induced hypoglycemia is markedly impeded in
hyperthyroid patients (996), and GHRH’s effects are reduced to less
than 50% of normal (Fig. 17).
Moreover, the timing of GH peaks is delayed after GHRH injection in
hyperthyroidism (997). After 1 month of antithyroid drug (methimazole)
therapy, hyperthyroid volunteers achieved serum thyroid hormone levels
within the normal range, but the GH-secretory response to GHRH was
still reduced in amplitude and delayed in time. Although not yet
explained, this persisting defect could reflect chronically reduced
pituitary GH stores, possibly due to prolonged GHRH deficiency. The
latter hypothesis would also explain why clonidine, a drug that is
believed to release GH through the stimulation of endogenous GHRH,
evokes less than the expected amount of GH secretion in hyperthyroid
subjects (594). This putative GHRH-GH deficiency state resolves when
euthyroidism is maintained for at least 3 months (997) (Fig. 17).
|
Other mechanistic investigations in thyrotoxicosis have exploited agents presumed to decrease hypothalamic somatostatin release. For example, the lack of effect of pyridostigmine in hyperthyroidism suggests that the diminution in GH secretion is not due solely to reduction in the activity of cholinergic pathways (999). On the other hand, L-arginine is able to enhance, but not normalize, the GH response to GHRH in patients with hyperthyroidism (1000). Thus, any inferred excess in hypothalamic somatostatin release in hyperthyroidism is not likely to represent the exclusive cause of suppressed GH secretion. Alternatively, we hypothesize that hyperthyroidism may suppress GH secretion directly at the pituitary level by impairing GHRH receptor activity and/or inhibiting GH biosynthesis or storage. The latter conjecture may explain both the blunting of the serum GH response to GHRH in hyperthyroid patients and the slow recovery of pituitary responsivity to GHRH after correction of hyperthyroidism (997).
Hyperthyroidism decreases GH release after GHRH stimulation alone or combined with GHRP-6, but not after GHRP-6 alone (1001). To our knowledge, altered expression of the recently cloned GHRP receptor in the hypothalamus and/or pituitary gland in hyperthyroidism remains unstudied.
The foregoing studies taken together suggest, but do not establish, that reduced GH release stimulated by GHRH in hyperthyroidism is explained, at least in part, by an increase in hypothalamic somatostatin tone with concurrent GHRH deficiency and suppressed pituitary GH production. Whether these putative alterations are caused directly by excess thyroid hormone or by the elevated circulating IGF-I levels remains to be elucidated. Alternatively, deficient GH responses to acute pharmacological challenges may be due to either sustained GHRH deficiency leading to depletion of somatotrope GH stores or to direct inhibitory effects of thyroid hormone excess on somatotropes.
|
VIII. Regulation of the GH Axis Throughout the Human Lifetime |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
A. Birth and infancy
Within the first hours of postnatal life in the human,
markedly amplified GH-secretory bursts emerge throughout the day and
night (1002) (Fig. 18). The calculated
half-life of GH released in the neonate is similar to that inferred in
the child or adult, but the amount of GH secreted per unit time and
surface area is amplified many times. Infants born prematurely also
exhibit exuberant GH secretion (1003) and reduced plasma IGF-I
concentrations (738). Such amplified GH secretion in the neonate may be
due to withdrawal of negative feedback signaling by IGF-I. Although
dopamine infusion suppresses GH and PRL hypersecretion in the newborn
(1004), the exact hypothalamo-pituitary mechanisms that drive increased
GH secretion in this context are not yet known. The dynamic neonatal GH
hypersecretory pattern also is typical of mild GH-resistance states,
e.g., as inferred in starvation, chronic hepatic or renal
failure, type I diabetes mellitus, or end-stage congestive heart
failure (108, 110, 113, 654, 1005, 1006, 1007).
|
C. Puberty
With the onset of biochemical and clinical manifestations of
increased sex steroid hormone action in pubertal boys and girls, a
remarkable 1.5- to 3- fold amplification of pulsatile GH secretion
occurs (848, 926). The dispersion or variability in GH secretion rates
is large interindividually (1009). Similarly high GH production rates
expressed per unit of surface area are observed in the first days of
life (1002), during fasting (1007), and in a large range of disease
states [end-stage renal failure (1005), liver disease (108),
Laron-type dwarfism (33, 35), type I diabetes mellitus (708), metabolic
acidosis without renal failure (1010), and mild acromegaly (134)].
However, the increase in pubertal GH secretion is virtually unique
mechanistically, because plasma IGF-I concentrations rise
commensurately by several fold (1011). Other than in acromegaly and
after testosterone administration (see below), plasma GH and IGF-I
concentrations rarely, if ever, increase together to these levels
within the normal human lifetime. Rather, the disease states above are
marked by relative GH resistance, since GH levels are increased but
IGF-I concentrations are reduced or low-normal (740, 1012). In
experimental animals and humans, diminished hepatic GH receptor
expression or other features of GH resistance are often evident in
these contexts (1013). The striking pubertal stimulation of the GH axis
creates a critical window of physiological hypersomatotropism that is
self-limited despite continuing adult sex-steroid hormone
concentrations (1014, 1015). Why, beginning at age 18–25 yr, GH
secretion falls toward and eventually below prepubertal levels in the
postpubertal adult is not known (92, 97, 119, 926, 1008, 1014, 1015).
In addition, recent studies using more modern techniques confirm
earlier clinical inferences that some (whole-body) anabolic actions of
testosterone require little if any measurable changes in the GH axis
(923, 1016).
The augmentation of GH secretion in puberty has been shown via 20-min sampling studies to originate neuromechanistically by way of increased GH secretory burst amplitude and mass, with no evident acceleration of detectable GH secretory burst frequency or major prolongation of the endogenous GH half-life (926). Although not proven experimentally, we postulate that the combined increases in GH secretion and plasma IGF-I concentrations in boys and girls in puberty reflect decreased sensitivity of the hypothalamo-pituitary unit to effective feedback inhibition by GH and/or IGF-I during the active growth phase. This could be due to altered IGFBP, IGF-I clearance, or (more likely) central hypothalamo-pituitary changes in GH drive and in autofeedback sensitivity.
The unique state of physiological GH hypersecretion in normal puberty is driven by sex-steroid hormones and can be recapitulated in part by administering small doses of estrogen to girls with ovarian dysgenesis (847, 934) or parenteral testosterone to boys with constitutionally delayed puberty (925, 927) or hypogonadotropic hypogonadism (928). During oral estrogen treatment, plasma IGF-I concentrations often decrease, whereas in response to parenteral testosterone administration, plasma IGF-I levels typically increase concurrently with GH hypersecretion. The latter more closely resembles normal puberty in boys. Notably, either estrogen or testosterone treatment selectively augments the mass of GH released per secretory burst, which is likely to be the dominant neuroendocrine mechanism subserving the physiological relative hypersomatotropism of puberty. Although sex steroids exert many direct effects on the liver, the testosterone-stimulated rise in plasma IGF-I concentrations is believed to require pituitary GH release, since hypopituitary boys who are markedly GH-deficient fail to show a robust rise in plasma IGF-I levels during testosterone repletion (1017, 1018).
Serum total and free testosterone concentrations strongly predict the total daily GH secretion rate in pubertal boys and men (926, 928). Serum GH concentrations in girls at different phases of puberty and in menstruating women at different stages of the cycle are positively correlated with serum estradiol levels (939, 1019). In a cross-sectional study in an adult population of both sexes spanning a wide age range, a strongly positive correlation between total or free estradiol and peak GH concentrations also pertained (1020). Longitudinal studies in boys and girls as they enter and traverse puberty will be required to clarify further the temporal linkages between endogenous androgen and estrogen concentrations and increased GH secretion.
D. Adulthood
Although incompletely explained, serum GH concentrations and
pulsatile GH secretion rates fall in adulthood to less than one-quarter
or one-half of the maximal values achieved in mid- to late puberty
(929, 1021, 1022, 1023) (see Fig. 19).
Indeed, there is an exponential decline (t1/2 of 7 yr) in
the 24-h mean serum GH concentration in both men and women beginning at
the late pubertal-young adult transition (ages 18–25) and continuing
into the middle years of life. This decline occurs despite adult levels
of sex steroid hormones and is accompanied by a gradual fall in plasma
IGF-I concentrations. The neurotransmitter mechanism(s) subserving this
progressive decline or "reset" of the hypothalamo-somatotrope-IGF-I
axis in healthy aging is (are) not known. The term "somatopause"
has been considered to describe this state.
|
Recent studies indicate that gender itself is a strong modifier of the
impact of age, obesity, and physical fitness on GH secretion (749). For
example, as shown in Fig. 20, compared
with men, healthy women of premenopausal age (and thus with normal
premenopausal concentrations of estrogen) are significantly protected,
by approximately 2-fold, from the negative impact of increasing age,
relative adiposity, and diminishing physical fitness on daily GH
secretion (assuming nominally similar GH half-lives in men and women,
and similar weight-adjusted GH distribution volumes). For example, mean
24-h serum GH concentrations decline twice as rapidly with increasing
age or increasing percentage body fat in men as in women between the
ages of 21 and 48 (749). Based on earlier studies, this gender
difference vanishes in the hormone-unreplaced postmenopausal age group
(1024). Increasing total adiposity (as quantitated by underwater
weighing to estimate total percentage body fat) also exerts an
approximately 2-fold greater negative impact on mean serum GH
concentrations in men than premenopausal women. Indeed, in a recent
study, the linear relationship between percentage body fat and GH
concentrations was significantly negative in men volunteers (n =
12), but not in the premenopausal women subjects (n = 32) (749).
Furthermore, increased physical fitness, as defined objectively by a
higher maximal rate of oxygen consumption, is a positive predictor of
serum GH concentrations in men, but the strength of this effect
(reflected in the standardized regression coefficient) is reduced by
approximately 50% in premenopausal women (749). Consequently, gender
has a major impact on the expression of several critical determinants
of GH secretion, including age, body composition, and physical fitness.
|
Mechanistic clinical studies indicate that testosterone and estradiol
both stimulate GH secretion and do so via a common neuroendocrine
mechanism of increased GH secretory burst amplitude or mass (928, 1025). This steroidal effect is likely to be mediated by the estrogen
receptor (142, 848, 930) and could in principle be brought about by
withdrawal of somatostatin’s inhibitory tone, amplification of
endogenous GHRH release or its pituitary actions, and/or recruitment of
other mechanisms that stimulate GH release (e.g., via the
theoretical endogenous ligand for the GHRP receptor). Studies do not
yet clarify whether sex steroid hormones in the adult might also reduce
the negative feedback efficacy of IGF-I and/or GH in suppressing GH
release. However, marked changes in the so-called approximate entropy
or orderliness of the GH release process at puberty (Fig. 16) strongly
suggest that feedback changes occur within the GH-IGF-I axis at this
time (142).
In addition to alterations in the pulsatile mode of GH release throughout the human lifetime, a recent observation is that a novel regularity statistic, approximate entropy (ApEn), can provide a measure of the regularity, orderliness, or subpattern reproducibility in hormone release profiles. Orderliness of GH release declines significantly in acromegaly and at puberty, is lower in women compared with men, as well as with greater relative or absolute adiposity, with aging, and during estrogen or testosterone (but not DHT) treatment (81, 121, 134, 136, 139, 142, 261). The exact neuroendocrine mechanisms that govern the moment-to-moment regularity of the GH release process are not known, but probably involve feedback control, and represent an important focus for further study.
Investigations of spontaneous GH secretion in the human (and rat) must be understood critically with respect to prevailing technical limitations. First, with contemporary methodologies, GH secretion rate estimates are stable on successive control days in healthy men (81). Second, the majority of prior RIA and IRMA estimates of serum GH concentrations had an absolute sensitivity of 0.1–0.5 µg/liter. Thus, in middle-aged and older, obese, hypothyroid, or fed healthy humans, daytime serum GH concentrations can fall below such assay thresholds in 20–97% of blood samples (118, 119). Consequently, the recent emergence of immunofluorometric and chemiluminescence GH assays with 10- to 30-fold increased sensitivity have provided new insights into the regulation of GH release during daytime hours and in selected pathophysiologies (118, 119). For example, an ultrasensitive chemiluminescence assay disclosed the existence of a 140-fold range in mean serum GH concentrations in healthy men. In addition, the assay suggested that a low rate of basal GH secretion exists between distinct GH-secretory bursts. Similar results were obtained in three independent ELISA-based studies (122, 123, 124). Notably, low basal rates of GH release were correlated to the serum concentration of estrogen as well as influenced by age and body composition (120). In the high-sensitivity chemiluminescence assay, all blood samples in more than 100 individuals spanning a range of ages and body compositions and pathophysiologies have contained detectable GH concentrations (sensitivity limits 0.002 µg/liter at 2 SDs and 0.005 µg/liter at 4 SDs above blank) (119). This assay further revealed that serum GH concentrations are suppressed by glucose to less than 0.7 µg/liter in healthy women and less than 0.07 µg/liter in healthy men, evincing a novel sex distinction in GH axis function (120). Other studies with a high-sensitivity (0.013 µg/liter threshold) immunofluorometric assay in middle-aged and perimenopausal women and age-matched men show that the quantitative basis for the gender difference in pulsatile GH release arises from higher GH-secretory burst mass (pulse amplitude and duration) in women compared with men, with no evident differences in the GH half-life or GH-secretory burst frequency (118).
Postpartum relative hyposomatotropism in the mother is also recognized in the human (1026). Although the neuroendocrine mechanisms are not established, suppression of pituitary GH secretion by the placental GH variant (V) synthesized during pregnancy is possible (1027, 1028, 1029). Interestingly, neither GHRH nor GHRP-1 is an effective GH secretagogue in this context, until PRL levels decline peripartally (1028). If GH autonegative feedback is relevant in postpartum maternal GH deficiency, the reasons for such sustained inhibition of the GH axis in lactating women remains enigmatic.
E. Aging
The manifold changes in the GHRH-somatostatin-GH-IGF-I axis with
healthy aging were reviewed (83, 84, 1021, 1030, 1031, 1032, 1033). Additional
insights have emerged into the mechanisms and/or treatment issues
concerning relative GH deficiency in healthy older individuals, as
discussed briefly below. The reduction in GH secretion with healthy
aging begins after the pubertal maximum and is correlated with
increased percentage total body and visceral fat, decreased physical
fitness, and decreased testosterone concentrations in men, or the
menopause in women (84, 121). The age effect in younger adults is more
evident in men than premenopausal women (above) (749).
The response of pituitary GH release as stimulated by multiple
individual secretagogues is decreased in aging, e.g., sleep,
exercise, L-dopa, insulin, L-arginine, GHRH,
GHRP, and clonidine (1032, 1034, 1035). However, intravenous (or oral)
L-arginine, which is presumed to withdraw somatostatin,
enhances GHRH’s stimulatory effect on GH release in the elderly by
4-fold, and these combined secretagogues nearly evoke normal GH
secretion acutely (1036). Similarly, L-arginine in
combination with GHRP reinstates a virtually young adult level of
acutely stimulated GH release in older humans (1037). In the dog,
pharmacologically induced somatostatin withdrawal elevates GH release
in older animals (574, 1038), thus also suggesting a partial role for
somatostatin excess in the hyposomatotropism of aging. A recent
clinical study in healthy men of varying ages receiving 2 days of oral
pyridostigmine treatment (putatively to withdraw somatostatin and
secondarily stimulate GHRH release) support an inference of combined
somatostatin excess and GHRH deficiency with aging (261). Such
treatment nearly normalized mean 24-h serum GH and IGF-I concentrations
in older men, compared with levels attained in younger men receiving
the same intervention, but failed to overcome the hyposomatotropism
associated with increased total percentage body fat (see Fig. 21).
|
). An independent earlier
investigation showed that 15 days of GHRH treatment alone once daily
and/or combined with L-arginine failed to fully normalize
GH secretion in older volunteers (1039). Another study examined the
effects of healthy aging on the GH response to acute GHRH stimulation
alone or in combination with either the neuropeptide, galanin, or the
acetylcholinesterase inhibitor, pyridostigmine. Pretreatment with
either galanin or pyridostigmine enhanced, but did not normalize, the
GH response to GHRH in aged subjects (443), which is consistent (as
above) with combined GHRH deficiency and somatostatin excess in healthy
aging humans (228, 261, 1036, 1037). This notion is supported by recent
studies showing diminished somatostatin-withdrawal-induced
"rebound" GH secretion (thought to be driven by endogenous GHRH
activation) in older men and women (1040).
|
Of analogous interest to the aging phenomenon, prolonged critical illness is associated with deficiencies of various neuroendocrine axes, e.g., gonadotropic, somatotropic, and thyrotropic (1043). Recent studies suggest that a hypothalamic component of relative deficiency of releasing factors plays a role (319, 320). Indeed, combined GHRP and GHRH infusions in extended critical illness activate GH, IGF-I, and BP-3 production significantly (1044).
F. GH treatment in older humans
Since this topic has been reviewed earlier, the reader is referred
to relevant background (315, 1032, 1045, 1046, 1047), and only more recent
issues are alluded to here. Our comments are restricted to GH treatment
in healthy volunteers rather than obese (763) or ill patients
[e.g., congestive heart failure (1012)].
Twice daily injections of GHRH in healthy older men can increase serum GH and IGF-I concentrations (1045). Short-term GHRH treatment can enhance muscle strength (1048), and long-term treatment can favorably modified body composition in elderly people (1049). Single nightly doses of GHRH for 6 weeks in older men are less effective than multiple daily doses (1048). The role of GHRH therapy remains largely unexplored otherwise in the biochemical and clinical hyposomatotropism of aging.
GH replacement in healthy older individuals improves body composition (see below) and enhances some measures of functional capability. For example, in a recent study of 52 men aged 70 yr or older, 30 µg/kg of human GH given 3 times weekly for 6 months augmented lean body mass by 4.3%, reduced body fat mass by 13%, but did not improve strength or physical endurance, or enhance mentation. Approximately 50% of the subjects experienced adverse incidents on this dose, and the GH dose was reduced in a quarter of the patients (1050). Treatment over 4 weeks with either recombinant human GH or IGF-I also decreased total fat mass, and increased lean body mass and nitrogen retention in elderly women (GH dose 25 µg/kg and effective IGF-I dose 60 µg/kg twice daily) (1051). However, with both of these treatments, patients often had side effects, e.g., headaches, edema, lethargy, joint swelling and pain, and a bloated sensation. In addition to the difficulties inherent in cost and repeated injections, parenteral GH treatment can elicit the above side effects, as well as carpal tunnel syndrome and gynecomastia, and cause mildly decreased insulin sensitivity (1046).
GH treatment objectively improved muscle mass and strength, but did not rejuvenate myofibrillar protein synthesis in men over the age of 60 yr, who received 30 µg/kg three times weekly for 3 months (1052). In another treatment study, fat mass decreased and lean body mass increased, but no systematic increases in muscle strength occurred in an exercise trial in men over the age of 65 yr (GH treatment 20 µg/kg/day for 10 weeks, after a 14-week training interval) (1053).
Older men release GH in response to exercise, albeit at a reduced level, and can benefit from exercise training with increased strength, increased muscle mass, and decreased total body fat over 4 months of a resistive exercise program (1054). Nonetheless, resistive exercise in older men and women is less effective in stimulating GH release than in younger individuals (1055). To our knowledge, exactly how GH, testosterone, and/or exercise interact at the muscle level is not yet known; e.g., the local IGF-I-BP system is likely involved, and testosterone itself stimulates human muscle IGF-I expression, while reciprocally reducing BP-4 levels in muscle (1056, 1057). However, less is understood concerning other muscle paracrine factors, which also might be regulated by GH and/or androgen, such as the recently cloned myostatin gene/protein, deletion of which markedly augments muscle growth in mice and cows (Belgian Blue and Piedmontese double-muscled cattle) (1058, 1059, 1060, 1061).
Adult-onset organic GH deficiency, which should be distinguished from healthy aging per se, provides a context in which GH replacement has demonstrably favorable effects on body composition; e.g., in one recent study of 32 men with adult-onset GH-deficiency, lean body mass increased by 2.5 kg, percentage body fat decreased by 4%, and bone mineral density rose by 2.4–5.1% over 1.5-yr treatment interval. The bone mineral density improvement was 5.1% in the lumbar spine and 2.4% in the femoral neck (1062). How and whether these changes translate into enhanced functional quality of life with consequent employment benefits, reduced bone fracture rates, etc., is not yet established (3, 4, 1033, 1063, 1064, 1065, 1066, 1067, 1068, 1069).
Although GH treatment in variably GH-deficient humans is capable of increasing plasma IGF-I concentrations, a gender difference exists in that serum IGF-I levels in hypopituitary women are almost 2-fold less responsive to any given weight-adjusted dose of GH (1033), especially when the patients are receiving estrogen therapy (949). A similar inference was made recently in a large 6-month multicenter (Australian) GH treatment trial in GH-deficient adults (3), in analogous patients in Holland (1070), and in a comparable 2-year study in Sweden (1063). After a single intravenous pulse of rhGH, greater acute lipolysis occurs in young males, whereas greater suppression of IGFBP-I occurs in older females (1071), thus indicating that both age and gender differences exist in tissue actions of GH. Moreover, sex differences further depend on the specific endpoint of GH action studied.
Given the costs, inconvenience, and side effects of GH replacement via
conventional once-daily injections, long-acting novel secretagogues of
endogenous GH that maintain pulsatile GH release and presumptively
remain susceptible to negative feedback control (thus, possibly not
eliciting pharmacologically increased IGF-I secretion) are therefore of
major interest, as summarized recently elsewhere (289, 291, 295, 315).
For example, a nonpeptidyl mimetic of GHRP, MK0677, administered once
daily orally for days or weeks to young, older, or obese humans can
stimulate sustained pulsatile GH secretion and increase (double) plasma
IGF-I concentrations commensurately (290) (see Fig. 8). Substantial
further study and discussion will be required to clarify the clinical
roles of GHRP-mimetics in boosting endogenous (pulsatile) GH secretion
in elderly healthy or frail individuals.
Intranasal peptide delivery also requires further exploration (see Sections III.C and IV.A). Recent studies suggest that intranasal peptide administration might evoke some direct (nonsystemically mediated) brain effects. For example, a "nose-brain pathway" has been inferred in recent clinical investigations, which show that vasopressin and cholecystokinin elicit rapid CNS electrical changes when given intranasally in the human (1072, 1073). CRH may also exert CNS effects (1074), as discussed above for GHRH and GHRP (e.g., Refs. 338, 384, 385, 386, 388).
|
IX. Exercise-Induced Modulation of the GH Axis |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
A more comprehensive study in the rat compared GH release in
aerobically exercised vs. sedentary male and female animals.
Nonexercising male rats exhibited typical pulses of serum GH
concentrations with values of more than 100 µg/liter every 3–4 h
with intervening trough periods of less than 10 µg/liter (1076). The
acutely and chronically exercised males manifested a decline in serum
GH concentrations during the actual exercise bout. In principle, such
suppression of GH release could be attributed to a decrease in
hypothalamic GHRH and/or an increase in somatostatin secretion.
Analysis revealed that GH, GHRH, and somatostatin mRNAs all fell in
acutely exercised male rats, suggesting that both hypothalamic and
pituitary biosynthetic capacities decline when untrained animals are
challenged with the energy requirements of acute exercise. In trained
male rats, more prolonged decreases in plasma GH concentrations and GH
and somatostatin mRNA levels occurred. A concomitant increase in GHRH
mRNA occurred, which may reflect the hypothalamic response to
somatostatin withdrawal (see Fig. 1).
The sedentary female rat (control) exhibited typical "feminized" GH secretory patterns, wherein peak serum GH concentrations before exercise did not exceed 100 ng/ml, and trough concentrations averaged 5–15 ng/ml. As in the male animals, the acute and chronically exercised female exhibited a decline in GH secretion during the 30-min exercise bout. At the molecular level, the acutely exercised female showed a 5-fold increase in pituitary GH mRNA concentrations, whereas chronically exercised females exhibited a 2-fold decrease. Messages for GHRH (increased) and somatostatin (decreased) also changed in chronically exercised females, as in their male counterparts. Thus, in contrast to the human (below), GH secretion in the rat is inhibited in response to acute exercise in both untrained and trained animals of either sex. However, compelling interspecies comparisons are limited, and the foregoing experiment focused only on low-intensity (aerobic) exercise. However, this work provides the first evidence that trained female rats show more rapid GH adaptive responses to exercise than trained males (1076).
B. Humans
Exercise of appropriate intensity and duration serves as a potent
physiological stimulus to GH secretion in humans (1077, 1078). In one
study, an increase in core body temperature was required to elicit GH
release with exercise (1079). However, the neuroendocrine mechanisms of
exercise-induced GH release remain incompletely elucidated,
e.g., whether the final pathway involves GHRH stimulation
and/or decreased somatostatin release, and/or possibly non-GHRH and
nonsomatostatin modulators, such as an endogenous GHRP ligand, galanin,
etc. Exercise-induced GH release is delayed until approximately 15 min
into exercise (1080, 1081) and peaks by the end of short-term exertion
(1080, 1081) or shortly thereafter (1082). Felsing et al.
(1077) concluded that high-intensity exercise must exceed 10 min to
reliably stimulate GH release.
Intersubject variability in peak exercise-induced GH concentrations is large (1081), some of which variation is attributable to differences in the age, sex, and body composition of subjects participating in different studies. In addition, exercise-induced GH release may be intensity dependent (1083, 1084). Vanhelder et al. (1083) demonstrated that high-intensity anaerobic work evoked greater GH release than lower intensity aerobic work, despite similar duration and total work effort in the two sessions. In addition, exceeding a threshold intensity of exercise is required for the release of GH (1083).
Hartley et al. (1084) reported that extremely high-intensity exercise paradoxically attenuates GH release compared with moderate intensity exercise. However, in this study the high-intensity exercise bout was completed on the same day as the moderate- and low-intensity bouts and was always performed last. Others have suggested that multiple bouts of exercise in a single day decrease exercise-induced GH release after several sessions (1081). The inferred ad seriatim decrease in exercise-induced GH release was alleviated when longer periods of rest (2–3 h) were introduced between exercise sessions (1081). Conversely, Felsing et al. (1077) showed that three bouts of exercise each separated by 1 h evoked escalating amounts of GH release. However, the exercise bouts also increased in duration in the latter study. Recently, Kanaley et al. (1086) reported that, independently of the length of recovery, three bouts of exercise in a single day progressively augmented GH secretion. These clinical data suggest to us the hypothesis that high-intensity aerobic exercise can overcome GH autonegative feedback, for which thesis preliminary evidence has been obtained recently (1087).
Differences in GH release during exercise between genders are incompletely defined, e.g., whether the mechanism and amount of exercise-induced GH release are similar in both sexes. Some authors have noted that women tend to have higher serum GH levels before exercise that do not return to baseline within 1 h, but the response pattern to exercise is similar to that of men (1080, 1086). Exercise-induced GH release was similar during all stages of the menstrual cycle in another analysis (1086).
Hartley et al. (1084) reported that 7 weeks of physical training decreased exercise-induced GH release in association with mild, moderate, and high intensity work. This study measured only the GH response to acute exercise and evaluated a small number of "recovery" blood samples. As expected, there was large variability in resting and exercise-induced GH concentrations both before and after training (1081). Long-term exercise training in women approximately doubles integrated 24-h serum GH concentrations on the nonexercising day, and GH pulse amplitude specifically, if the training is performed at an intensity greater than the individual’s lactate threshold (1088).
C. Neural control of GH release during exercise
Epinephrine, norepinephrine, acetylcholine, and opioids may
modulate exercise-induced GH release. Many clinical studies have
attempted to discern the individual contributions of these general
neurotransmitters to the exercise-stimulated release of GH, but no one
mechanism is agreed to be primary. Possibly several mechanisms are
operative or interactive during exercise-induced GH release. Regardless
of the neurotransmitters implicated, the final common pathway likely
involves either increases in GHRH and/or decreases in somatostatin
release and/or action. In addition, a role for putative endogenous
GHRP-receptor ligands and other neuromodulators cannot be excluded.
1. Adrenergic. Both epinephrine and norepinephrine
concentrations rise significantly in the peripheral blood after
high-intensity exercise (1089). However, the temporal patterns of
circulating GH and catecholamine responses to exercise differ. Although
blood catecholamines do not faithfully reflect adrenergic inputs to the
hypothalamus, they might stimulate human pituitary GH secretion
directly, at least as they do in the rat in vitro (see
Section V). More likely, the hypothalamic adrenergic system
modulates the exercise-driven release of GH, since propranolol (a
nonselective ß-blocker) amplifies exercise-induced GH secretion
(1090). Uusitupa et al. (1090) reported that both a
cardioselective ß-blocking agent (metoprolol) and a nonselective
ß-blocker (propranolol) increased the GH response to exercise.
Conversely, salbutamol and broxaterol (ß2-agonists)
blocked the exercise-induced release of GH (608). The consequences of
-receptor blockade are less uniform. Hansen (1091) observed that
phentolamine (an -receptor antagonist) nearly totally suppresses GH
release during exercise, whereas Sutton and Lazarus (1092) reported
that this drug has little effect. The experimental conditions and the
exercise bouts in the two studies were similar, making the disparity in
results difficult to reconcile. However, Hansen (1091) studied only 4
subjects while Sutton and Lazarus (1092) studied 10. We infer from
these data that any role of -adrenergic pathways in
exercise-stimulated GH release is not yet clear, whereas imposed
ß2-adrenergic neurotransmission remains capable of inhibiting the
exercise-triggered signal(s) directing GH secretion. Thus,
ß2-receptor input may occur either downstream of the
primary exercise effector pathway or be capable of suppressing it.
Exercise-mediated increases in serum GH concentrations may also be modulated by the metabolic changes that occur during maximal exercise. In one study, there were no significant changes in blood glucose during the four exercise bouts evaluated (608). Although the inhibitory effect of ß2-agonists on GH release occurred concomitantly with a slight increase in preexercise blood glucose levels, the lack of correlation between these two variables, as well as the minimal hyperglycemia attained, suggests that a transient glucose elevation after ß2-agonists is not the primary mechanism involved in GH suppression. The increase in serum lactate, which normally occurs during maximal exercise (1089), is also not likely to modify GH secretion, because infusion of lactate in man does not alter GH release (1093). Finally, since the ß2-agonist-mediated decrease in the GH response to exercise occurs despite an increase in plasma lactate levels during the exercise (608), ß2-agonists do not inhibit GH release by suppressing lactate production.
2. Cholinergic. Cappa et al. (1094) reported that pyridostigmine (an indirect cholinergic agonist) administered concomitantly with an exercise stimulus increased GH release additively. In the rat, pyridostigmine may decrease hypothalamic somatostatin release, whereas in the sheep it increases GHRH secretion (above). Since the combination of pyridostigmine plus exercise is additive, one could infer that exercise modifies release of yet an additional (non-GHRH, nonsomatostatin) effector and/or acts more effectively than a submaximally effective dose of pyridostigmine. Thompson et al. (1095) corroborated that oral pyridostigmine, alone or in combination with the opiate receptor antagonist, naltrexone, potentiates exercise-induced GH release. Last, atropine (a muscarinic blocker) inhibits the GH response to exercise (535). In fact, putatively enhanced somatostatinergic tone due to this cholinergic antagonist impairs GH responsiveness to several stimuli, including GHRH. Thus, exercise cannot overcome pharmacological antagonism of the cholinergic (muscarinic) pathway or activation of ß2-receptor neurotransmission (above). Observations with atropine suggest, but do not prove, that the exercise signal converges on the GH release pathway at, or proximal to, somatostatinergic systems normally suppressed by cholinergic inputs.
3. Opioids. The role of opioids in the control of exercise-stimulated GH release is controversial. Experiments investigating the effects of naloxone (an opiate antagonist) on exercise-induced GH release have yielded equivocal results. Moretti et al. (482), using high doses of naloxone and 20 min of high-intensity exercise recorded that naloxone completely blocked exercise-induced GH release in well trained competitive athletes. The ability of naloxone to impede exercise-stimulated GH release in this context may have reflected a propensity of these competitors to perceive greater stress, possibly because of the achievement-oriented task. In contrast, Coiro et al. (1096) noted that naloxone did not inhibit exercise-induced GH release in subjects who were not trained athletes. This discrepancy might be explained if chronic exercise training induces the activity of opioid pathways regulating GH release during acute exercise. The study by Coiro et al. (1096) further revealed that the GABA-ergic agonist, sodium valproate, significantly attenuated exercise-elicited GH secretion, and that the combination of sodium valproate and naloxone completely abolished GH release. Thus, GABA-ergic mechanism may be able to inhibit the GH response to exercise in humans, and an opioid pathway may be able to relieve the inhibitory effects of GABA.
The interaction of opioid and cholinergic pathways to prompt GH release during exercise was investigated by Thompson et al. (1095). Naltrexone (an opioid antagonist) did not alter resting serum GH concentrations or exercise-stimulated GH release. Naltrexone also did not influence the rise in mean and integrated serum GH concentrations stimulated by pyridostigmine, thus disclosing no evident interaction between opiatergic and cholinergic pathways in the effect of exercise. We conclude provisionally that endogenous opiatergic activation is not crucial for acute exercise-induced GH release in untrained individuals, but may be relevant in competitively trained athletes.
D. Kinetics of exercise-induced GH-release
Lassarre et al. (1080) and Thompson et al.
(1095) reported that the apparent half-life of endogenous GH during
exercise was lower than that computed at rest (namely 19 min at rest
vs. 16 min after exercise). Since a shorter GH half-life
would tend to disproportionately reduce the serum GH concentration, the
marked rise in blood GH after exercise (assuming no major changes in
distribution volume) clearly reflects amplified GH secretion, rather
than decreased GH clearance (1088, 1097). As revealed by deconvolution
analyses, the dominant neuroendocrine mechanism of exercise-induced GH
secretion specifically entails a markedly augmented mass of GH secreted
per burst, which would be consistent with somatostatin withdrawal
and/or amplified GHRH release, or release of one or more
cosecretagogues, such as a putative endogenous GHRP receptor
ligand, yet to be identified.
|
X. Summary |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
To the above end, we review here the pathophysiology of the GHRH somatostatin-GH-IGF-I feedback axis consisting of corresponding key neurotransmitters, neuromodulators, and metabolic effectors, and their cloned receptors and signaling pathways. We propose that this system is best viewed as a multivalent feedback network that is exquisitely sensitive to an array of neuroregulators and environmental stressors and genetic restraints. Feedback and feedforward mechanisms acting within the intact somatotropic axis mediate homeostatic control throughout the human lifetime and are disrupted in disease. Novel effectors of the GH axis, such as GHRPs, also offer promise as investigative probes and possible therapeutic agents.
Further understanding of the mechanisms of GH neuroregulation will likely allow development of progressively more specific molecular and clinical tools for the diagnosis and treatment of various conditions in which GH secretion is regulated abnormally. Thus, we predict that unexpected and enriching insights in the domain of the neuroendocrine pathophysiology of the GH axis are likely be achieved in the succeeding decades of basic and clinical research.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 This work was partially supported by the Centro Studi e Ricerche di
Neuroendocrinologia (Brescia, Italy), by MURST, and by Regione
Lombardia (Project microalbuminuria (A.G.), NIH Grant RR-00847 to the
General Clinical Research Center, the National Science Foundation
Center for Biological Timing, and NIH Grants NIA RO1 AG 147991 and RO3
AG14873 (to J.D.V.). 
|
References |
|---|
|
TopAbstractI. IntroductionII. Contemporary Tools for...III. Neuropeptide Regulation of...IV. Other GH-Regulating...V. Neurotransmitter Regulation...VI. Role of Metabolic...VII. Other Hormonal Regulators...VIII. Regulation of the...IX. Exercise-Induced Modulation...X. SummaryReferences |
|---|
, Acromegalic males;
, acromegalic females; 
, normal fed males; 
, normal fasted
males; •, normal fed females; 
, normal fasted females; 
,
acromegalic males in remission; 
, acromegalic females in remission.
[Redrawn with permission from M. L. Hartman et
al.: J Clin Invest 94:1277–1288, 1994
(
