Endocrine Reviews 20 (4): 460-500
Copyright © 1999 by The Endocrine Society
The Corticotropin-Release Inhibitory Factor Hypothesis: A Review of the Evidence for the Existence of Inhibitory as Well as Stimulatory Hypophysiotropic Regulation of Adrenocorticotropin Secretion and Biosynthesis1
Dennis Engler,
Eva Redei and
Ismail Kola
Laboratory of Molecular Genetics and Development (D.E., I.K),
Institute of Reproduction and Development, Monash Medical Centre,
Clayton, Victoria, Australia 3168; and The Asher Center (E.R),
Department of Psychiatry and Behavioral Sciences, Northwestern
University Medical School, Chicago, Illinois 60611
 |
Abstract
|
|---|
- I. Introduction
- II. Hypothalamic Stimulation of ACTH Release and Biosynthesis
- A. Corticotropin-releasing factor (CRF)
- B. Arginine vasopressin (AVP)
- C. The hypothalamic CRF and AVP neurons and their connections
- D. Studies of the secretion of CRF and AVP into the hypophysial-portal
circulation
- E. Studies of the neural regulation of CRF and AVP secretion and
biosynthesis
- III. Hypothalamic Inhibition of ACTH Release and Biosynthesis
- A. Historic studies
- B. Effect of hypothalamo-pituitary disconnection in adult and fetal sheep
- C. Effects of the opiate alkaloids and opioid peptides on the HPA axis
- D. The role of the posterior pituitary in the regulation of corticotropic
function
- E. Hypothalamic ACTH release-inhibitory activity
- F. Definition of corticotropin release inhibitory factor (CRIF)
- IV. CRIF: A Consideration of Possible Candidates
- A. Somatostatin (SST)
- B. Dopamine
- C. Atrial natriuretic peptide (ANP)
- D. Prepro-TRH-(178199)
- E. Other substances
- V. Future Directions
 |
I. Introduction
|
|---|
IT HAS been traditionally thought that the hypothalamus
only exerts a stimulatory influence upon the secretion and synthesis of
ACTH by the anterior pituitary gland and that this is mediated by
neuropeptides such as corticotropin-releasing factor (CRF), arginine
vasopressin (AVP), and oxytocin (OT), which are secreted into the
hypophysial-portal circulation (Refs. 1, 2, 3, 4, 5, 6 and Fig. 1
). However, an analysis of the effects
on the hypothalamic-pituitary-adrenal (HPA) axis of surgically
isolating the anterior pituitary from the hypothalamus suggests that
the hypothalamus may exert both inhibitory and stimulatory regulation
over ACTH secretion and POMC biosynthesis and that the inhibitory
regulation might be mediated by a currently unidentified substance that
has been named corticotropin release-inhibitory factor (CRIF) (7, 8).
In this review, we will initially summarize some of the recent
knowledge of the mechanisms by which the hypothalamus stimulates
corticotropic function and describe some of the central neural pathways
regulating the HPA axis. We will then review the historical, and more
current, literature describing the effects on the HPA axis of
surgically isolating the pituitary from the hypothalamus. We outline
several postulates that may need to be fulfilled for a substance to
qualify as a CRIF, describe studies performed to date to determine the
ACTH release-inhibitory activity of a number of possible candidates,
and finally point to possible future directions in this emerging area.

View larger version (35K):
[in this window]
[in a new window]
|
Figure 1. A schematic representation of the current model by
which the hypothalamus is thought to regulate ACTH secretion. This
model proposes that the hypothalamus only stimulates ACTH secretion by
secreting neuropeptides such as CRF and AVP into the hypophysial-portal
circulation. ACTH then stimulates the adrenocortical production of
cortisol, which then restrains the secretion of ACTH by exerting
negative feedback effects on the anterior pituitary, hypothalamus, and
various extrahypothalamic brain sites.
|
|
 |
II. Hypothalamic Stimulation of ACTH Release and Biosynthesis
|
|---|
A. CRF
CRF is a 41-residue peptide that was first isolated in 1981 by
Vale et al. (9) on the basis of its ability to markedly
augment the secretion of ACTH and ß-endorphin from cultured rat
anterior pituitary cells (Fig. 2
). CRF is
the most potent ACTH secretagog in the rat and in man although its
ability to stimulate ACTH secretion is potentiated severalfold by
agonists such as AVP, OT, angiotensin II, norepinephrine (NE), and
epinephrine (EPI) (10, 11, 12). However, CRF may not be the most potent
ACTH secretagog in all species as it appears to be equipotent with AVP
in the bovine species (13), and in the ovine species, AVP appears to be
even more potent than CRF (Refs. 14, 15, 16 and Fig. 3
). Moreover, in the rat, CRF is the only
hypothalamic neuropeptide known to increase POMC biosynthesis, and none
of the aforementioned ACTH secretagogs are able to potentiate this
effect of the peptide (17, 18, 19, 20, 21, 22, 23).
CRF exerts its effects on the anterior pituitary by binding to a
specific receptor that is linked to the adenylate cyclase complex. The
binding of the hormone to the corticotrope cell membrane results in
increased intracellular concentrations of cAMP, an increased influx of
extracellular Ca2+, an activation of the cAMP-dependent
protein kinase [protein kinase A (PKA)], and the phosphorylation of a
number of intracellular proteins (24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40). The brain CRF receptor is
also linked to the adenylate cyclase complex, and it is therefore
likely that a similar sequence of second messenger responses underlies
the actions of the peptide in the brain (41). The characteristics of
the CRF receptor were determined in the 1980s by studies that examined
the binding of radiolabeled CRF ligands to pituitary and brain
membranes (42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56). The anterior pituitary CRF receptor has high
affinity, its activity is modulated by divalent cations and guanyl
nucleotides, and receptor down-regulation occurs in response to
adrenalectomy, corticosterone (CS) treatment, and chronic stress. In
contrast, none of these interventions affect CRF receptor number in the
brain or intermediate pituitary. As determined by autoradiography, the
highest CRF receptor density in the primate brain is found in the pars
tuberalis of the pituitary and throughout the cerebral cortex (Ref. 51
and Fig. 4
). A very high binding density
is also found in the hippocampus and arcuate nucleus, high levels are
seen in the amygdala, and lower levels of receptor binding are also
found in a variety of other brain areas.

View larger version (160K):
[in this window]
[in a new window]
|
Figure 4. Color-coded image of
[[125I]Tyr]oCRF autoradiographs of representative
coronal sections from rostral to caudal of monkey brain. Areas of very
high density are shown in red, and densities decrease
through yellow, green, and light
blue. Dark blue and black
correspond to nonspecific background. CC, Cingulate cortex; P, putamen;
OC, orbital cortex; C, caudate; OT, olfactory tubercle; Cl, claustrum;
TC, temporal cortex; NST, nucleus of stria terminalis; IC, insular
cortex; POA, preoptic area; AA, anterior amygdala; NPT, paraventricular
nucleus of thalamus; A, amygdala; MDT, medial dorsal thalamus; MN,
mamillary nucleus; Hi(gd), hippocampus (gyrus dentatus); SC, superior
colliculus; LPTN, lateral posterior thalamic nucleus; LGB, lateral
geniculate body; Hi, hippocampus; Cb, cerebellum; ICo, inferior
colliculus; LC, locus coeruleus; DPN, dorsal parabrachial nucleus.
[Reproduced with permission from M. A. Millan et
al.: Proc Natl Acad Sci USA 83:19211925, 1986
(51 ).]
|
|
Two CRF receptor subtypes are encoded by distinct genes and they are
members of the calcitonin/vasoactive intestinal peptide/GHRF
subfamily of G-protein coupled receptors. The CRF-R1
receptor was the first to be cloned and characterized from a human
corticotropic adenoma, from mouse AtT20 pituitary tumor cells, and from
human and rat brain (57, 58, 59, 60). It is a 415-amino acid protein comprised
of seven membrane-spanning domains that binds rat/human CRF with high
affinity. In addition, the nonmammalian CRF-related peptides, sauvagine
and urotensin I, also bind to CRF-R1 and are nearly
equipotent with CRF in stimulating cAMP accumulation in cells
containing the expressed receptor. CRF-R1 mRNA is expressed
throughout the intermediate pituitary and in the corticotropes of the
anterior lobe (61, 62), and within the brain, its distribution conforms
well with the results of the autoradiographic studies. However, the
occurrence of sparse CRF-R1 mRNA expression in several
brain areas known to bind CRF has been observed, and this
"mismatch" may be partly due to the existence of the
CRF-R2 receptor in these sites.
The second CRF receptor to be isolated (CRF-R2) was cloned
from mouse heart and rat brain cDNA libraries (63, 64, 65, 66) and consists of
two splice variants, CRF-R2
and CRF-R2ß,
which differ at their N-terminal domain. CRF-R2
is a
411-amino acid protein cloned from the rat brain (63), whereas
CRF-R2ß is a 431-amino acid protein that was isolated
from mouse heart (64, 65, 66). CRF and the nonmammalian peptides, sauvagine
and urotensin I, all bind to CRF-R2 and stimulate adenylate
cyclase activity, and the rank order of potency of their interaction
(sauvagine > urotensin> rat/human CRF > ovine CRF) differs
markedly from their interaction with CRF-R1. The
distribution of CRF-R2 mRNA in rat brain differs from that
of CRF-R1 mRNA (67) since CRF-R1 mRNA
expression is highest in neocortical, cerebellar, and sensory relay
structures, whereas CRF-R2 mRNA expression is mainly
confined to subcortical structures and is most abundant in the lateral
septal nucleus, the ventromedial hypothalamus, and choroid plexus (Fig. 5
). In further contrast with the
CRF-R1 mRNA, the CRF-R2 mRNA is in very low
abundance within the anterior pituitary.

View larger version (123K):
[in this window]
[in a new window]
|
Figure 5. Color-coded digitized images of CRF1
and CRF2 receptor mRNA expression in adjacent horizontal
brain sections. Regions exhibiting high levels of mRNA expression are
coded in red and orange, while the lowest
levels of expression are coded in blue. [Reproduced
with permission from D. T. Chalmers et al.:
J Neurosci 15:63406350, 1995 (67 ).]
|
|
As noted, both CRF-R1 and CRF-R2 bind the
nonmammalian CRF-like peptides, sauvagine and urotensin-l, albeit with
different affinity. It was originally presumed that sauvagine and
urotensin-1 were CRF homologs involved in stimulating ACTH release in
amphibia and fish, respectively. However, in 1988, Okamura et
al. (68) isolated a cDNA in the teleost fish, Catostomus
commersoni, that encoded a peptide possessing 95% homology with
rat CRF and in 1992, Stenzel-Poore et al. (69) characterized
a cDNA in the frog, Xenopus laevis, which encoded a peptide
sharing 93% homology with mammalian CRFs and only 50% with sauvagine.
These findings demonstrated that sauvagine and urotensin-1 were not the
primary ACTH secretagogs in amphibia and fish and showed that at least
two members of the CRF superfamily could coexist in a single species.
Furthermore, the findings raised the possibility that a nonmammalian
CRF-like peptide might also occur in mammals. This prediction was
subsequently confirmed by the isolation of urocortin (70). In the
initial studies, urotensin-like immunoreactivity was located in the
Edinger-Westphal nucleus of the rat midbrain, and a unique cDNA was
cloned from a library encompassing this nucleus, which encoded a
precursor and a peptide that was named urocortin. The human and rat
peptides share 95% identity within the mature peptide region, and the
gene encoding the human peptide has been localized to chromosome 2
(71). In the rat brain, the most abundant immunoreactive (ir)-urocortin
perikarya are located in the supraoptic, paraventricular, and
periventricular hypothalamic nuclei, the Edinger-Westphal nucleus, the
dorsal raphe, and the dorsal and laterodorsal tegmental nuclei.
However, despite its hypothalamic distribution, no ir-urocortin has
been found in the median eminence or posterior pituitary, suggesting
that it is unlikely that the peptide is secreted into the
hypophysial-portal circulation. The distribution of ir-urocortin
overlaps with the CRF-R2 mRNA in several brain areas,
suggesting that it may mediate the functions of CRF in these sites
(72).
B. AVP
The nonapeptide AVP is also secreted into the hypophysial-portal
circulation and acts on the anterior pituitary to stimulate ACTH
release. AVP is a weak ACTH secretagog in the rat and in man, although
it appears to be as potent as CRF in the bovine species and even more
potent than CRF in the ovine species. Although CRF and AVP both
increase ACTH secretion and total ACTH content in ovine anterior
pituitary cells, AVP is clearly the more potent agonist, and these
observations gave rise to the hypothesis that both secretagogs might
increase POMC gene expression in ovine corticotropes (15). This
suggestion initially received little experimental support (73), but a
recent study by van de Pavert et al. (74) has demonstrated
that AVP, but not CRF, does increase steady-state levels of POMC mRNA
in ovine corticotropes. This finding contrasts sharply with the
situation in the rat where CRF is the only neuropeptide known to
stimulate POMC gene expression and points toward species-specific
differences in regulation of the POMC gene. The molecular basis for
these differences requires further investigation.
The pituitary AVP receptor (V1b) is clearly different from those AVP
receptors found in the liver and on vascular smooth muscle cells (V1)
or in the kidney (V2) (75, 76, 77, 78, 79, 80, 81, 82). Adrenalectomy markedly decreases AVP
binding in the pituitary, an effect that is prevented by glucocorticoid
replacement (78, 79). Moreover, Vlb receptor content is increased or
decreased by manipulations that respectively augment or reduce HPA axis
activity (83). When taken together, these findings suggest that V1b
receptor expression is positively regulated by glucocorticoids. The
nucleotide and amino acid sequences of the Vlb receptor were first
derived from a human pituitary cDNA library (84). The deduced protein
was found to contain 424 amino acid residues and 7 transmembrane
domains, and Northern blot analysis showed a single mRNA transcript
that was detectable only in the anterior pituitary. However, a
subsequent study has predicted a protein of 421 amino acids and shown
expression of 2 mRNA species in the anterior pituitary, in several
brain regions, and in a number of extrapituitary tissues (85). The
precise basis for these different findings awaits further elucidation.
The anterior pituitary AVP receptor is coupled to the
phosphatidylinositol (PI) pathway (86, 87, 88), and hormone binding
therefore stimulates PI hydrolysis and increases the production of
inositol 1,4,5-trisphosphate (IP3) and diacylglycerol
(Refs. 89, 90, 91 and Fig. 6
). The
diacylglycerol is required for the activation of protein kinase C,
which then causes the phosphorylation of a number of cellular
substrates such as the myristoylated alanine-rich C kinase substrate
(MARCKS) protein (15, 37, 92, 93, 94, 95, 96). The IP3 causes the
liberation of Ca2+ from intracellular stores and, together
with the influx of extracellular Ca2+, causes the rise in
intracellular Ca2+ that is required to mediate ACTH release
(33, 40). Since the pituitary AVP receptor is not linked to the
adenylate cyclase, it does not increase cAMP levels in this site (36, 97, 98, 99). However, it does potentiate the cAMP response to CRF, an
effect that is due to the activation of protein kinase C since it is
also mimicked by phorbol esters (36, 98, 99).

View larger version (48K):
[in this window]
[in a new window]
|
Figure 6. A schematic representation of the subcellular
mechanisms activated by the binding of AVP to the anterior pituitary
V1b receptor. A, A unified hypothesis of the spatiotemporal aspects of
calcium signaling, based on ideas concerning calcium oscillations and
the propagation of calcium waves. A typical response begins with an
agonist (such as AVP) generating Ins(1 4 5 )P3, which then
mobilizes calcium from an Ins(1 4 5 )P3-sensitive calcium
store and also promotes the entry of external calcium to give an
initial calcium signal. In some cells, this
Ins(1 4 5 )P3-induced calcium signal can act as a
"primer" to drive a process of calcium-induced calcium release from
the Ins(1 4 5 )P3-insensitive pools to produce a spike that
might be organized in the form of a wave, thereby spreading the
signal throughout the cell. [Reproduced with permission from M.
J. Berridge and R. F. Irvine: Nature 341:197205,
1989 (86 ). © Macmillan Magazines, Ltd.]
B, Schematic representation of proposed mechanisms of protein
kinase C (PKC) activation. There are two pathways leading to the
activation of intracellular PKC from the extracellular environment: (a)
by the receptor-mediated production of diacylglycerol (DAG) via
a G protein-coupled phospholipase C; and (b) by direct pharmacological
activation with phorbol esters. Both DAG and phorbol esters [such as
phorbol 12-myristate 13-acetate (PMA)] bind to the C1 region of the
PKC-regulatory subunit that causes physical association of PKC with the
plasma membrane (translocation). DAG promotes the formation of a
PKC-phosphatidylserine (PS)-Ca2+ complex, whereas phorbol
esters cause plasma membrane insertion of the kinase and its subsequent
down-regulation. Both DAG and phorbol esters induce a conformational
change in the structure of PKC so that the inhibitory pseudosubstrate
sequence moves away from the substrate binding site. This then permits
PKC substrates to bind to the kinase and to be phosphorylated. (1 )
Receptor activation; (2 ) translocation; (3 ) down-regulation; (4 )
membrane insertion; and (5 ) substrate phosphorylation. [Reproduced
with permission from J. P. Liu: Mol Cell Endocrinol
116:129, 1996 (88 ). © Elsevier Science.]
|
|
C. The hypothalamic CRF and AVP neurons and their connections
The CRF and AVP neurons that project to the external zone of the
median eminence are located on either side of the midline in the
paraventricular nucleus of the hypothalamus (PVH) (100-120). The rat
PVH is a complex structure that contains three dense clusters of
magnocellular neurons (anterior, medial, posterior) that project to the
posterior pituitary surrounded by a large shell of parvocellular
neurons. The parvocellular component can, in turn, be subdivided into
five parts (anterior, medial, dorsal, lateral, and periventricular),
which are interrelated with autonomic cell groups in the brainstem and
spinal cord via bidirectional pathways and with the external zone of
the median eminence (Fig. 7
).

View larger version (18K):
[in this window]
[in a new window]
|
Figure 7. A schematic diagram to show the major cell groups
of the paraventricular nucleus of the hypothalamus in the rat, as
viewed from above. The three parts of the magnocellular division are
shown in stipple,and are embedded in the parvocellular
division, which consists of five parts. The abbreviations are as
follows: am, anterior magnocellular; ap, anterior parvocellular; dp,
dorsal parvocellular; lp, lateral parvocellular; mm, medial
magnocellular; mp, medial parvocellular; pm, posterior magnocellular;
pv, periventricular. [Reproduced with permission from L. W.
Swanson et al.: J Comp Neurol
196:271285, 1981 (129 ). © Wiley-Liss, Inc., a division of John Wiley
& Sons, Inc.]
|
|
The rat brain contains about 2,000 CRF-stained perikarya distributed
throughout all 8 parts of the PVH, and most of these cells are found in
the parvocellular division (111). The medial, periventricular, and
medial lateral parts of the parvocellular division contain about half
of the total number of CRF-stained neurons, and these areas are known
to send massive projections to the external zone of the median
eminence. In addition, about 15% of the total CRF-stained population
is found in those areas of the magnocellular division that
predominantly contain oxytocinergic cells. In addition to the PVH,
CRF-stained cells are also found in the basal hypothalamus,
telencephalon, and brainstem, and these areas are involved in the
functioning of the autonomic nervous system. Finally, CRF-stained cells
are found scattered throughout the cerebral cortex and, in this region,
they are concentrated in layers II and III (Fig. 8
).

View larger version (17K):
[in this window]
[in a new window]
|
Figure 8. A schematic illustration of the major CRF-stained
cell groups (black dots) and fiber systems in the rat
brain. Most of the immunoreactive cells and fibers appear to be
associated with systems that regulate the output of the pituitary and
the autonomic nervous system, and with cortical interneurons. Most of
the longer central fibers course either ventrally through the medial
forebrain bundle and its caudal extension in the reticular formation,
or dorsally through a periventricular system in the thalamus and
brainstem central gray. The direction of fibers in these systems is
unclear because they appear to interconnect regions that contain
CRF-stained cell bodies. Thus, for example, three adjacent CRF-stained
cell groups, the laterodorsal tegmental nucleus (LDT), locus ceruleus
(LC), and parabrachial nucleus (PB), lie in the dorsal pons. However,
it is uncertain which of these cell groups contributes to each of the
pathways shown, and which of them receives inputs from the same
pathways. Abbreviations: HIP, hippocampus; SEPT, septal region; BST,
bed nucleus of the stria terminalis; SI, substantia innominata; CeA,
central nucleus of the amygdala; PVH, paraventricular hypothalamic
nucleus; MPO, medial preoptic area; ME, median eminence; PP, posterior
pituitary; LHA, lateral hypothalamic area; MID THAL, midline thalamic
nuclei; POR, perioculomotor nucleus; CG, central gray; DR, dorsal
raphe; MR, median raphe; LDT, laterodorsal tegmental nucleus; LC, locus
ceruleus; PB, parabrachial nucleus; MVN, medial vestibular nucleus;
DVC, dorsal vagal complex; A1, A5, noradrenergic cell groups; ac,
anterior commissure; cc, corpus callosum; st, stria terminalis; mfb,
medial forebrain bundle. [Reproduced with permission from L. W.
Swanson et al.: Neuroendocrinology
36:165186, 1983 (111 ). © Karger, Basel.]
|
|
The CRF neurons in the parvocellular division of the PVH may be further
subdivided into two populations that are distinguished by the
colocalization of the AVP precursor (pro-AVP)-derived peptides AVP, the
vasopressin-neurophysin (NP), or the pro-AVP C-terminal
glycopeptide (121, 122). The CRF+/AVP+
and CRF+/AVP- subpopulations are similarly
distributed in the rostral and caudal regions of the PVH but
differently distributed in the region midway between the rostral and
caudal ends of the nucleus. In the normal rat, almost half of the
CRF-positive axon terminals in the external zone of the median eminence
stain intensely for AVP and NP, and both CRF and the pro-AVP-derived
peptides are found within the same neurosecretory vesicles. However,
the majority (>90%) of the pro-AVP-deficient neurons and virtually
all the CRF-positive structures in the median eminence stain intensely
for AVP, NP, or glycopeptide in the adrenalectomized rat. These
findings indicate that the normal circulating level of glucocorticoid
is sufficient to inhibit pro-AVP gene expression in the
CRF+/AVP- neurons but not in the
CRF+/AVP+ cells and are consistent with the
hypothesis that modulation of the CRF:AVP molar ratio in the
hypophysial-portal circulation could be partly mediated by differential
regulation of the two CRF neuronal subpopulations in the PVH.
The PVH is also connected with a number of brain regions, and prominent
among these are the forebrain, the limbic system, and the brainstem.
For example, the parvocellular part of the PVH receives moderately
dense projections arising from all areas of the hypothalamus (except
the supraoptic nucleus, the medial and lateral mamillary nuclei, and
the magnocellular preoptic nucleus), from the subfornical organ, and
the bed nucleus of the stria terminalis, but the magnocellular division
of the PVH receives relatively few inputs from these structures (123).
Although a large amount of physiological and functional evidence
indicates that several parts of the limbic system (septum, amygdala,
and hippocampal formation) may regulate the activity of the PVH,
detailed autoradiographic studies have failed to detect direct
projections from any of these areas to the PVH or the supraoptic
nucleus (124, 125, 126). Rather, these structures may give rise to
pathways that end in cell groups lying immediately adjacent to the PVH.
The only exception to this statement is the projection from the ventral
subiculum, which may interact directly with the PVH by means of a relay
in the bed nucleus of the stria terminalis (127). As noted, the PVH is
densely innervated by aminergic and peptidergic axon terminals that
arise from cell bodies located in brainstem nuclei. The aminergic
terminals contain NE, EPI, dopamine (DA), and serotonin and, to
date, the noradrenergic and adrenergic projections have been subjected
to the most detailed analysis (128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139).
The noradrenergic input to the PVH arises almost exclusively from three
interrelated cell groups in the brainstem, namely the A2 region in the
nucleus of the tractus solitarius (NTS), the A1 region in the
ventrolateral medulla, and the A6 area in the locus ceruleus (Fig. 9
). The projections from the A1 region
are almost entirely directed toward the magnocellular division and
terminate preferentially on vasopressinergic cell bodies. The fibers
originating from the A2 region are primarily distributed throughout the
parvocellular division and are most dense in the dorsal medial part, a
region known to contain a large population of CRF-immunoreactive
neurons. The projections arising from the A6 area are almost entirely
distributed to the parvocellular PVH, and their most prominent input is
localized to the periventricular zone, an area known to contain DA-,
somatostatin (SST)-, and TRH-staining neurons (138). When taken
together, these findings suggest that the A1, A2, and A6 noradrenergic
cell groups may each regulate the activity of anatomically and
functionally distinct groups of neurosecretory neurons.

View larger version (14K):
[in this window]
[in a new window]
|
Figure 9. Schematic drawing of a sagittal section through
the rat brain to indicate the dominant biochemical makeup and
distribution of catecholaminergic and NPY-immunoreactive inputs from
the brainstem to the PVH. Adrenergic (E) projections arise from the C1,
C2, and C3 regions, are distributed overwhelmingly to the parvicellular
(pc) division of the nucleus, and generally stain positive for NPY
immunoreactivity. Noradrenergic (NE) projections from the locus
coeruleus and A2 cell groups are also distributed primarily to the
parvicellular division, but are, for the most part, NPY negative. A
heterogeneous input arises from the A1 region and is distributed to
both the parvicellular division and preferentially to those parts of
the magnocellular division in which vasopressinergic neurons (V)
predominate over oxytocinergic ones (o). One component appears also to
contain NPY immunoreactivity, whereas a second one does not.
[Reproduced with permission from P. E. Sawchenko et
al.: J Comp Neurol 24:138153, 1985 (149 ).
© Wiley-Liss, Inc., a division of John Wiley & Sons, Inc.]
|
|
The ascending adrenergic projections to the PVH are also derived from
three discrete brainstem cell groups, namely the C1 group (in the
rostral ventrolateral medulla), the C2 group (in the rostromedial part
of the nucleus of the tractus solitarius), and the C3 group (in the
medial longitudinal fasciculus and nucleus prepositus hypoglossi).
However, in contrast to the highly differentiated noradrenergic
projection to the PVH, the projections from each of the adrenergic cell
groups are very similarly distributed within the PVH and, in each case,
the most dense innervation is seen in the dorsal medial parvocellular
part, a region that is rich in CRF-staining neurons. These findings
suggest that the C1, C2, and C3 adrenergic cell groups may each
influence similar groups of parvicellular neurosecretory and/or
autonomic neurons in the PVH (139).
As stated above, the PVH is also innervated by peptidergic axon
terminals, and those that stain for neuropeptide Y (NPY) appear
particularly prominent (140, 141). NPY-stained perikarya and axon
terminals are widely distributed within the brain, and the PVH and
hypothalamic arcuate nucleus, respectively, contain the highest density
of NPY-stained axon terminals and perikarya within the brain
(142, 143, 144, 145, 146). NPY is extensively colocalized within brainstem adrenergic
neurons that project to the PVH, while its expression in noradrenergic
neurons appears limited to a subpopulation of cells in the A1 group
(147, 148, 149, 150, 151). NPY-stained projections are most dense in the anterior and
medial parvocellular parts of the PVH, and these areas are known to
contain CRF- and TRH-stained neurons (Fig. 10
). Furthermore, the ultrastructural
demonstration of direct synaptic contacts between NPY-stained axon
terminals and CRF-stained perikarya in the PVH provides the anatomical
basis for considering NPY a potentially important regulator of the HPA
axis (152).

View larger version (91K):
[in this window]
[in a new window]
|
Figure 10. Darkfield photomicrographs of avidin-biotin
immunoperoxidase preparations to show the distribution of
fibers and varicosities stained for
phenylethanolamine-N-methyltransferase (PNMT),
dopamine-ß-hydroxylase (DBH), and neuropeptide Y (NPY)
immunoreactivity at a similar level through the paraventricular nucleus
(PVH; the third ventricle is at the extreme left of each micrograph).
At this midcaudal level, basic similarities and differences in the
distribution and density of each input may be appreciated, although in
these thicker (3035 µM) sections, details of the
distributions cannot necessarily be inferred. Note that the
distribution of PNMT-stained elements is largely limited to a discrete
part of the parvicellular division of the nucleus; few are seen in the
magnocellular division. The DBH-stained projection encompasses and
exceeds that localized with anti-PNMT, providing a prominent input to
the magnocellular division, which is located at the right-hand margin
of the nucleus at this level. The NPY-stained input encompasses and
exceeds the distribution and density provided by DBH-immunoreactive
inputs, providing perhaps the most prominent chemically specified input
to the PVH yet described. [Reproduced from P. E. Sawchenko
et al.: J Comp Neurol
241:138153,1985 (149 ). © Wiley-Liss, Inc., a division of John Wiley
& Sons, Inc.]
|
|
D. Studies of the secretion of CRF and AVP into the
hypophysial-portal circulation
The studies of Porter and Jones (153) represent one of the
earliest efforts to detect the presence of ACTH-releasing activity in
hypophysial-portal plasma. These authors demonstrated that the
intravenous injection of canine hypophysial-portal vessel plasma into
the hydrocortisone-inhibited, intact rat caused ACTH release, as judged
by the adrenal ascorbic acid depletion assay. The effect was not
observed in animals injected with canine carotid artery plasma,
suggesting "that venous blood from the hypothalamico-hypophyseal
portal vessels contains a substance(s) which accelerates the release of
ACTH and that this substance(s) is not present or is present in much
less concentration in the blood from the carotid artery." The
response was also abolished by hypophysectomy, indicating that the
anterior pituitary is required for the effect to be seen.
In the 1980s and l990s, in vivo studies were performed in
the rat, the sheep, and the horse, which demonstrated that CRF, AVP,
and OT were secreted into the hypophysial-portal circulation
(154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186). The first studies to appear were performed in the
anesthetized rat (154, 155, 156, 157, 158), and when they are compared with those
subsequently performed in the conscious sheep and horse, it is apparent
that the stress of the anesthesia, and perhaps the surgical procedure
itself, elevated the baseline levels of these peptides. Nevertheless,
the studies in the rat demonstrated that maneuvers such as chemical
adrenalectomy, volume depletion, nitroprusside-induced hypotension,
insulin-induced hypoglycemia, the intracerebroventricular
administration of NE, and transection of the fornix all increased CRF
concentrations in portal blood (157, 158, 161, 162, 163, 166, 173, 174).
Furthermore, manipulations such as exposure to glucocorticoid-negative
feedback, the surgical or chemical destruction of the hypothalamic
noradrenergic input, and surgical lesions of the PVN all decreased the
basal or stress-induced increments in CRF concentrations in portal
plasma (155, 160, 168, 176, 177, 178).
These experimental manipulations were also found to exert effects on
the levels of AVP and OT in portal plasma. Volume depletion,
hypoglycemia, and fornix transection increased portal AVP and OT
concentrations (157, 158, 174), nitroprusside-induced hypotension did
not alter either of their levels (177), and chemical adrenalectomy had
no discernible effect on OT (161). In contrast, the exposure to
glucocorticoids also decreased portal levels of AVP and OT (176, 177),
but destruction of the hypothalamic noradrenergic input or the PVN
itself had no effect on portal AVP concentrations (168, 176). These
findings have given rise to the concept that a substantial proportion
of the AVP in portal plasma may be of magnocellular origin (176).
Finally, recent studies in the rat employing the technique of
microdialysis have shown that emotional stress may also increase the
release of AVP into the PVH (187). However, it appears that AVP, when
released into the PVH, exerts an inhibitory, rather than a stimulatory,
effect on the HPA axis. Although the mechanisms by which this
inhibitory effect is mediated were not elucidated, the findings
indicate that the same neuropeptide may exert opposite effects on the
HPA axis at the level of the PVH and anterior pituitary gland.
In the late 1980s, several studies appeared in which portal CRF and AVP
concentrations were measured in conscious, unanesthetized animals (159, 167, 169, 170, 171, 175, 185). In our studies, ovine pituitary portal blood
was collected by the technique developed by Clarke and Cummins (188).
This method allowed portal blood to be collected in the conscious
animal at 5- to 10-min intervals and thereby permitted the first
evaluation of the normal ultradian secretion of the hypothalamic
releasing factors and their responses to stressful stimuli. These
studies, and those of Caraty et al. (169, 175) performed in
conscious rams, demonstrated that CRF and AVP were secreted in a
pulsatile manner, that the two peptides were secreted synchronously or
asynchronously, that their basal concentrations were generally less
than 100 pg/ml, and that their basal secretory patterns differed in
each individual animal (Fig. 11
). They
also showed that volume depletion, a fear-associated audiovisual
stimulus, or insulin-induced hypoglycemia all increased CRF and AVP
levels in portal plasma and generally increased the AVP:CRF molar
ratio. Since fear-associated audiovisual stimuli initially activate the
auditory and visual cortices, respectively, and subsequently activate
the limbic system whereas hypoglycemia primarily activates several
subcortical brain areas, these findings showed that the CRF and AVP
neurons within the PVN may be activated by different neural inputs.
When sheep were treated with dexamethasone, the basal levels of CRF,
ACTH, and cortisol were decreased, the pituitary-adrenal response to an
audiovisual stimulus was inhibited, and the HPA response to
hypoglycemia was attenuated (171).
The studies in the horse measured CRF and AVP in the pituitary venous
effluent via a catheter that was introduced through the facial vein
(189), but since the tip of the catheter lay below the anterior
pituitary, the concentrations of CRF and AVP measured may be up to
10-fold lower than those actually present in the portal blood (159, 167, 172, 181, 183, 184). However, a high degree of concordance was
found between AVP and ACTH pulses, and the overall conclusions of these
studies were remarkably in accord with those derived from the studies
in the rat and sheep.
E. Studies of the neural regulation of CRF and AVP secretion and
biosynthesis
1. The central noradrenergic and adrenergic pathways. The
early studies of Krieger and Krieger (190) that were performed in the
conscious cat suggested that centrally administered catecholamines
activated the pituitary-adrenal axis. However, throughout the 1970s and
early 1980s, it was generally accepted that central NE inhibited ACTH
release, and this view was based on experiments by Ganong and
co-workers (191, 192) that were performed in the anesthetized dog .
However, in the late 1980s and early 1990s, studies in conscious
animals (rat, sheep) and in man appeared that strongly supported the
view that central catecholamines stimulate the pituitary-adrenal axis
(193, 194, 195, 196, 197, 198, 199, 200, 201, 202). The diametrically opposed findings of Ganong et
al. (191, 192) may be due to the use of anesthesia and are
reconciled with all these studies when one considers that anesthesia
may cause a single, centrally administered agonist to exert opposite
effects on brainstem NE release and/or biosynthesis (203).
Our studies in the conscious sheep were initiated to determine how
insulin-induced hypoglycemia might stimulate CRF and AVP secretion into
the hypophysial-portal circulation. The PVH does not appear to contain
glucose-sensitive neurons, but the NTS, the lateral hypothalamic area,
and the ventromedial hypothalamus do contain neurons that respond to
changes in glucose concentration by altering their firing rates
(204, 205, 206). Therefore, the effect of hypoglycemia on CRF and AVP
release is likely to be secondary to a primary activation of one or
more of these glucose-sensitive sites. The experiments were based on
the finding that hypoglycemia increases the hypothalamic turnover of NE
(207), and the anatomic relationships that exist between the PVH and
the brainstem catecholaminergic and peptidergic cell groups. It was
hypothesized that the effect of hypoglycemia on neuronal activity in
the NTS might increase NE synthesis within the nucleus or within the Al
and A6 areas, since all three regions are interconnected (131). This
conceptual framework would explain the increased hypothalamic turnover
of NE during hypoglycemia and would also predict that NE, and possibly
EPI, might stimulate the release of CRF and/or AVP.
When NE or EPI was injected into the lateral cerebral ventricle of the
conscious sheep, an acute and sustained increase in ACTH and cortisol
secretion was observed, and NE appeared to be the more potent agonist
(202). The finding that NE and EPI released only very modest amounts of
ACTH from cultured anterior pituitary cells suggested that both amines
were mainly acting on suprahypophysial brain sites to increase CRF and
AVP release. This prediction was tested by determining the effects of
intracerebroventricular (icv) NE on the secretion of CRF and AVP into
the hypophysial-portal circulation of the conscious sheep (Ref. 185
and Fig. 12
). The icv injection of 50
µg NE acutely increased CRF and AVP concentrations in portal plasma
and activated the pituitary-adrenal axis. In addition, NE caused
sustained hypersecretion of CRF and AVP even though plasma cortisol
levels were greatly elevated. This observation suggested that
activation of the central noradrenergic system might override the
normal glucocorticoid negative feedback on those areas of the brain
concerned with regulation of the HPA axis. This concept of "acquired
glucocorticoid resistance" has been recently addressed by De Kloet
et al. (208), and the following subcellular mechanisms may
underly this phenomenon.

View larger version (24K):
[in this window]
[in a new window]
|
Figure 12. The effect of an icv injection of 50 µg
norepinephrine on plasma CRF, AVP, ACTH, and cortisol in three ewes.
The arrow ( ) depicts the time of injection.
[Reproduced with permission from J.-P. Liu et al.:
J Clin Invest 93:14391450, 1994 (185 ). © The
American Society for Clinical Investigation.]
|
|
First, activation of hypothalamic adrenergic receptors by icv NE might
stimulate CRF (and AVP) biosynthesis to such a degree that the
glucocorticoid-inhibitory effect on CRF (and AVP) gene expression is
overridden. The suggestion that the central noradrenergic pathways
might stimulate CRF biosynthesis in vivo is supported by the
observations that microinjection of NE into the rat PVH increases CRF
mRNA expression in the nucleus (209) and that unilateral hemisection of
the brainstem between the locus ceruleus and NTS reduces both the basal
and stress-induced rise in CRF mRNA expression in the ipsilateral PVH
(210, 211). Second, chronic activation of the central catecholaminergic
pathways by repeated immobilization of the rat also attenuates
glucocorticoid-negative feedback by decreasing type 2 glucocorticoid
receptor (GR) mRNA in the PVH and hippocampus (212). The immobilization
causes hypercortisolemia, which may in turn mediate the effect on GR
gene expression since it is not observed in adrenalectomized animals
bearing implanted cortisol pellets.
The activation of the cAMP/protein kinase A pathway and the
cAMP-regulatory element binding protein (CREB) may constitute one of
the intracellular mechanisms that mediates the effect of NE on CRF
biosynthesis. The cAMP-regulatory element (CRE), generally represented
by the palindromic motif 5'-TGACGTCA-3', binds a number of
transcription factors including CREB, an intranuclear PKA substrate
protein that is phosphorylated upon activation of the cAMP-PKA pathway
(213, 214, 215, 216, 217). The CRE is also present in the rat, ovine, and human CRF
genes (218, 219, 223, 224) and is likely to be physiologically
important since activation of the cAMP-PKA pathway increases CRF mRNA
levels in vitro (220, 221, 222), and mutation of the CRE site
completely abolishes this effect (224). Moreover, the cAMP-PKA pathway
is also important in stimulating CRF gene expression in vivo
since 8-bromo-cAMP, when microinjected into the PVH, increases CRF
mRNA in the hypothalamus, increases POMC mRNA in the anterior
pituitary, and increases plasma ACTH concentrations (225). Furthermore,
the increase in CRF mRNA that occurs as a consequence of
insulin-induced hypoglycemia can be attenuated by the prior injection
into the PVN of a CREB antisense oligonucleotide. Since insulin-induced
hypoglycemia also increases NE turnover in the hypothalamus, these
findings point toward a major role for CREB in mediating the
stimulatory effect of the ascending catecholaminergic pathways on CRF
gene expression.
2. Neuropeptide Y. A number of studies have assessed the role
of Neuropeptide Y in the regulation of the HPA axis (185, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235).
This 36-residue peptide is widely distributed within the brain, and it
exerts its effects by acting on five receptor subtypes (designated
Y1-Y5) (236, 237). The cDNAs encoding all except the Y3 receptor have
been isolated (238, 239, 240, 241, 242, 243, 244), each one is expressed in the brain, and the
Y2 and Y5 receptor mRNAs seem restricted to this site (240, 244). NPY
mimics the action of icv NE on the pituitary-adrenal axis when
injected into the cerebral ventricles of the rat, dog, and sheep, and
the studies in the conscious sheep indicate that the effect is mediated
at one or more suprahypophysial sites since NPY increases the release
of CRF and AVP into the hypophysial-portal circulation, but does not
increase the release of ACTH from cultured ovine anterior pituitary
cells (185). Like NE, NPY also stimulates CRF biosynthesis since icv
NPY increases CRF mRNA in the rat hypothalamus (233). Although these
effects of NPY on CRF are likely to be mediated by binding of the
peptide to NPY receptors in the PVH, the exact NPY receptor subtype
responsible for this action is currently unclear. Moreover, the finding
that a broad range of NPY-related peptides increase plasma ACTH when
administered icv raises the possibility that multiple NPY receptor
subtypes are involved in mediating this effect (245).
3. CRF. Recent studies suggest that CRF may also exert an
ultrashort positive feedback regulation on its own biosynthesis within
the PVN (246). When injected icv in the rat, CRF initially increases
the PVN expression of the immediate-early genes c-fos and
NGFI-B and subsequently increases CRF mRNA in the nucleus. Since
these effects are markedly attenuated by the prior icv injection of the
CRF antagonist,
-helical CRF(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), they are likely to be mediated
by binding of the peptide to specific CRF receptors located either on
CRF neurons within the PVN or in other brain areas. The brainstem
catecholaminergic neurons could represent an alternative site of CRF
binding since the peptide could be transported to these neurons by the
cerebrospinal fluid after icv injection and since icv CRF has been
shown to increase c-fos expression in the A1 and C1 cell
groups. The activated brainstem catecholaminergic cell groups could
then signal to the PVH and other brain areas by means of the ascending
catecholaminergic pathways to increase neuronal c-fos
expression in these sites.
 |
III. Hypothalamic Inhibition of ACTH Release and Biosynthesis
|
|---|
A. Historic studies
Although the vast bulk of the studies related to the HPA axis can
be adequately explained by assuming the hypothalamus only exerts a
stimulatory influence upon ACTH release and biosynthesis, an analysis
of the early studies employing pituitary stalk section and other
ablative surgical procedures suggests that hypothalamic stimulation of
ACTH secretion may not completely explain all of the experimental
evidence. The earliest studies showed that pituitary stalk section in
the rat and the dog failed to prevent the adrenal hypertrophy that
occurs as a consequence of cold exposure (247, 248, 250). The surgery
also failed to prevent the adrenal depletion of ascorbic acid in
response to unilateral adrenalectomy in the rat (249), failed to
prevent cold-induced adrenal cholesterol depletion in the guinea pig
(251), and reduced, or abolished, the rabbits lymphopenic response to
restraint and cold exposure while exerting little effect on this
animals response to subcutaneous EPI or laparotomy (252). In 1954,
Keller et al. (253) sectioned the pituitary stalk and
destroyed the ventral hypothalamus in the dog and noted that the
procedure resulted in a chronic decrease in insulin tolerance, but did
not prevent the animals from withstanding further surgery such as a
unilateral adrenalectomy or pancreatectomy. Furthermore, the adrenal
gland weights (mg/kg body weight) and adrenal cholesterol contents
(mg/kg body weight) were also increased by the hypothalamectomy (Fig. 13
). In most instances, the
hypothalamectomy also caused an acute eosinopenia, and a drastic
eosinopenia was elicited by further surgery. The authors concluded that
their experiments "add crucial support to the conclusion drawn by
others that excitation of the adenohypophysis is not dependent upon
direct hypothalamic humoral or neurogenic influence," but the
findings could also be interpreted as providing support for the idea
that pituitary stalk section removed a dominant inhibitory influence
over ACTH secretion.

View larger version (38K):
[in this window]
[in a new window]
|
Figure 13. Adrenal data for four groups of animals: a) 10
dogs having hypophysectomies of varying magnitude (insulin dosages they
were able to tolerate ranged from 1/40 to 1/10 U/kg of body weight, the
mean being 1/20 U; b) 7 normal dogs which were terminated
without being duressed previously in any way; c) 9 normal dogs which
were terminated at varying time intervals after a major surgical
procedure (in most instances a pancreatectomy); and d) 8 ventral
hypothalamectomized dogs which were either terminated during some sort
of duress or were found dead. Open bars represent
individual animals; the crossed bars represent the group
averages. Although there was considerable variability surrounding this
datavarying size dogs and the difficulty in quantitating the
magnitude of any particular duressed statethe data graphed clearly
demonstrate a) the absence of adrenal atrophy and b) a normal range in
cholesterol content of these animals. The suspicion of a tendency to
adrenal hypertrophy (on the basis of gross inspection) in the
hypothalamectomized dog is also verified by the composite data.
[Reproduced with permission from A. D. Keller et
al.: Am J Physiol 179:514, 1954 (253 ).]
|
|
In the 1960s, Egdahl (254, 255, 256) performed a large number of experiments
in dogs bearing surgically isolated pituitaries. In animals in which
the brain above the inferior colliculus was removed, the basal adrenal
venous 17-hydroxycorticosteroid output was increased, and this
secretion was augmented further by stimuli such as burning the hind
leg, laparotomy, and constriction of the inferior vena cava (Figs. 14
, 15
, and 16
). From these studies, Egdahl
derived two conclusions: first, "higher areas of the brain tonically
inhibit lower areas which produce and release ACTH-releasing
neurohumors in the dog"; second, "an area which produces an
ACTH-releasing neurohumor is located in the hindbrain. This hindbrain
factor (HBF) acts through the systemic blood stream to effect ACTH
release from the pituitary, or possibly, in the intact animal to
stimulate release of the hypothalamic ACTH-releasing neurohumor"
(254). To test the second hypothesis, Egdahl removed the brain to the
level of the pons, removed the entire spinal cord, and subjected the
animals to bilateral nephrectomy or total abdominal evisceration.
However, none of these procedures prevented the elevation in basal
adrenal venous 17-hydroxysteroid secretion. These findings led Egdahl
to speculate as follows: "On the other hand, if it is suggested that
there is an inhibitory neurohumor as well as an excitatory one
which has been removed by the brain removal procedure, then the
spontaneous activity of the pituitary has been unmasked. The
hypothesis which is proposed in this paper suggests that the excitatory
and inhibitory neurohumoral areas are localised in the
diencephalon-mesencephalon and are controlled by impulses coming from
both the spinal cord and cerebral cortex" (256). This quotation is
likely to be the first, and clearest, enunciation of the possible
existence of both stimulatory and inhibitory hypothalamic factors
involved in the regulation of ACTH release.

View larger version (25K):
[in this window]
[in a new window]
|
Figure 14. The effect of brain removal on adrenal venous
corticosteroid output in the dog. Resting level of corticosteroid
output in this animal ranged from 2.2 to 5 g/min. Burns on five
different occasions resulted in sharp increases in corticosteroid
output. The increase in output following burning was comparable to that
obtained following exogenous ACTH. [Reproduced with permission from
R. H. Egdahl: Endocrinology 66:200216, 1960
(254 ). © The Endocrine Society.]
|
|

View larger version (21K):
[in this window]
[in a new window]
|
Figure 15. Dog C-2 (total brain removal, total spinal cord
removal with isolated pituitary). Fourteen hours after this operation
adrenal cortical hypersecretion was still present. [Reproduced with
permission from R. H. Egdahl: Endocrinology
71:926935, 1962 (256 ). © The Endocrine Society.]
|
|

View larger version (44K):
[in this window]
[in a new window]
|
Figure 16. B-303 (isolated pituitary with brain removal to
pons and bilateral nephrectomy). The kidneys are not essential to the
increased pituitary adrenal activity observed in dogs with isolated
pituitaries. [Reproduced with permission from R. H. Egdahl:
Endocrinology 71:926935, 1962 (256 ). © The Endocrine
Society.]
|
|
In the mid-1960s, Halász and co-workers (257, 258, 259) analyzed the
effects of partial and total deafferentation of the medial basal
hypothalamus (MBH) on pituitary-adrenal function in the rat (Fig. 17
). One to 4 weeks after the surgical
interruption of all the neural afferents to the MBH, plasma CS levels
in the morning and afternoon were significantly increased, and the
normal CS diurnal rhythm was abolished (Fig. 18
). Surgical stress further increased
plasma CS in these animals, and compensatory adrenal hypertrophy
developed after unilateral adrenalectomy. In contrast, interruption of
the lateral, dorsal, and posterior connections of the MBH did not raise
plasma CS levels and did not alter the normal CS diurnal rhythm (258).
A diurnal variation in pituitary ACTH content was also observed in
sham-operated rats, the values being relatively high in the morning and
low in the afternoon (259). Complete deafferentation of the MBH
markedly increased pituitary ACTH levels and abolished the ACTH diurnal
rhythm, but interruption of the dorsal, lateral, and posterior
connections of the MBH did not affect adenohypophysial ACTH content and
did not alter its diurnal rhythm. From these findings, Halász and
co-workers concluded that complete deafferentation of the MBH increased
both the secretion and synthesis of ACTH.

View larger version (76K):
[in this window]
[in a new window]
|
Figure 17. a, d, And g, schematic sagittal drawings of the
hypothalamo-pituitary complex to demonstrate the different
deafferentations of the medial basal hypothalamus. b, e, and h, The
knife cuts as seen from the base of the brain. Heavy
line indicates the cut; broken lines with arrow
indicate the level of the histological sections presented in c.f. - c,
f, i. Microphotographs of rat hypothalamus showing the knife cut
(arrows). c, Complete deafferentation (coronal section);
f, incomplete deafferentation (coronal section); g, frontal cut
(horizontal section) (x25). Abbreviations: AC, anterior commissure;
AL, anterior lobe of pituitary; ARC, arcuate nucleus; ME, median
eminence; MM, medial mamillary nucleus: OC, optic chiasm; ON, optic
nerve; PED, cerebral peduncle; PL, posterior lobe of pituitary; PV,
paraventricular nucleus; SC, suprachiasmatic nucleus; V, third
ventricle; VM, ventromedial nucleus. [Reproduced with permission from
B. Halász et al.:
Neuroendocrinology 2:4355, 1967 (258 ). © Karger,
Basel.]
|
|

View larger version (31K):
[in this window]
[in a new window]
|
Figure 18. Upper, Diurnal variation in plasma
corticosterone levels. Lower left, Pituitary-adrenal
response to stress. Lower right, Adrenal compensatory
hypertrophy in animals with partial or total deafferentation of the
medial basal hypothalamus. S, Sham operated; CD, complete
deafferentation; ID, incomplete deafferentation; FC, frontal cut. The
drawings at the top of this figure show, in schematic
sagittal section, the type of deafferentation (heavy
line) in the groups below. AM, Rats sacrificed in the morning;
PM, sacrificed in the afternoon. Adrenal compensatory hypertrophy: I,
weight of the adrenal removed first; II, weight of the second adrenal
30 days later. [Reproduced with permission from B. Halász
et al.: Neuroendocrinology 2:4355, 1967
(258 ). © Karger, Basel.]
|
|
In the 1960s and 1970s, several groups analyzed the effects of
hypothalamo-pituitary disconnection or pituitary stalk section on
endocrine function in the monkey (260, 261, 262, 263, 264, 265, 266). Kendall and Roth (260)
removed the forebrain or sectioned the pituitary stalk and noted that
adrenal venous 11-hydroxysteroid secretion remained unaltered for
up to 8 h after the surgery, and their observations corroborated
Egdahls findings. Knobil and co-workers (261, 262, 263) studied the
effects of an anterior or complete disconnection of the MBH and noted
that anterior disconnection of the MBH did not alter the cortisol
diurnal rhythm in most animals and had little effect on plasma cortisol
concentrations. In contrast, complete MBH disconnection severely
attenuated, or abolished, the normal cortisol diurnal rhythm and
lowered plasma cortisol concentrations. Ferin and co-workers (264, 265, 266)
studied the effect of sectioning the pituitary stalk and noted that
effective surgery caused an acute and chronic rise in serum PRL levels,
abolished the normal cortisol rhythm, lowered the morning cortisol
levels to the afternoon range seen in intact animals, and markedly
attenuated the cortisol response to hypoglycemia. However, a
morphological analysis of the pituitary revealed that the surgery
caused a variable amount of necrosis (545%) of the pars distalis in
most animals. Taken together, the findings in the primate partly concur
with those obtained in the dog, the rat, and the sheep, and the
discrepancies might be due in part to the anterior pituitary necrosis
produced by pituitary stalk section.
B. Effect of hypothalamo-pituitary disconnection in adult and fetal
sheep
The hypothalamo-pituitary disconnection (HPD) procedure was first
described by Clarke et al. (267), and it essentially removes
the hypothalamic-releasing and release-inhibiting factors from the
hypophysial-portal circulation and severs the innervation to the
intermediate and posterior pituitary lobes. However, in contrast to
stalk section, which often compromises the blood supply to the anterior
pituitary and causes infarction of the gland, HPD preserves the blood
supply to the adenohypophysis. The procedure causes no apparent
histological alteration of the anterior lobe, but results in marked
hypertrophy of the intermediate lobe and atrophy of the posterior
pituitary.
As expected, the lack of GnRH in the hypophysial-portal circulation in
the HPD animal caused a gradual decline in plasma LH and FSH
concentrations, and they became undetectable 1 week after the procedure
(267). By analogy with the hypothalamic-pituitary-gonadal axis, if the
hypothalamus merely stimulated ACTH release, the absence of CRF and AVP
in the hypophysial-portal circulation of HPD-treated animals would also
be expected to reduce plasma ACTH and cortisol concentrations to
subnormal, or undetectable, levels and abolish their response to
stress. However, in 1986, Clarke et al. (268) noted that
basal ACTH,
-MSH, ß-endorphin, and cortisol levels were
significantly elevated in those animals subjected to HPD and concluded
that the ovine pituitary could function in the absence of hypothalamic
influence. In 1988, Engler et al. (269) confirmed the effect
of HPD on the basal levels of these POMC peptides and demonstrated that
HPD virtually abolished their rise in response to an audiovisual
stimulus and insulin-induced hypoglycemia, indicating that CRF and AVP
are essential to mount an adequate pituitary-adrenal response to stress
(Fig. 19
). In addition, these authors
also noted that the direction in which plasma ACTH and cortisol levels
changed after HPD was identical to that of PRL, since
hyperprolactinemia is also a consequence of pituitary isolation in the
sheep, the rat, and the monkey (263, 270, 271, 272, 273). The hypothalamus both
stimulates and inhibits PRL secretion and synthesis, and the dominant
inhibitory regulation is mediated by DA, whereas the stimulatory input
is mediated by neuropeptides such as TRH, vasoactive intestinal
peptide, OT, and the newly discovered PRL-releasing peptide (271, 272, 274, 275, 276, 277). Since the hyperprolactinemia that follows stalk section is
thought to be due to withdrawal of the major inhibitory factor DA (271, 274), it was argued that the increased ACTH secretion by the isolated
anterior pituitary might also be due to withdrawal of the corticotropes
from the influence of a hypothalamic factor that inhibits ACTH
secretion. It was suggested this substance be named corticotropin
release inhibitory factor (CRIF) since neither of the two best
characterized inhibitory substances in portal plasma, namely SST and
DA, seemed to possess the criteria required of such a substance (see
below).
The POMC peptide ultradian rhythm also persists after HPD, indicating
that the ovine pituitary possesses an intrinsic ability to secrete
these peptides in a pulsatile fashion (278). However, as judged by
measurement of pulse amplitude and interpulse interval, the
POMC-peptide rhythm in the HPD-treated animal differs from that in the
intact animal, indicating that the POMC-peptide ultradian rhythm in the
intact animal is the net result of an intrinsic pituitary oscillator
that is being continually modified by the hypothalamus (Fig. 20
). Similar conclusions have been made
by Carnes and co-workers (279, 280), who noted that CRF
immunoneutralization in the rat abolished high amplitude-low frequency
ACTH pulses but did not affect the appearance of low amplitude-high
frequency fluctuations (279). In addition, these authors also noted
that pulsatile ACTH secretion persisted in the PVN-lesioned animal
(280).
In addition to raising the basal plasma levels of ACTH and cortisol,
HPD also increases the POMC mRNA level in the ovine anterior pituitary
(281). These findings strongly suggest that the hypothalamus also
exerts a tonic inhibitory influence over POMC gene expression (Fig. 21
), although it is currently unknown
whether this regulation is exerted at the level of mRNA transcription
or translation.