This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Engler, D. Articles by Kola, I. Search for Related Content PubMed PubMed Citation Articles by Engler, D. Articles by Kola, I. Pubmed/NCBI databases Compound via MeSH Substance via MeSH Hazardous Substances DB DOPAMINE

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

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

Top Abstract I. Introduction II. Hypothalamic Stimulation of... III. Hypothalamic Inhibition of... IV. CRIF: A Consideration... V. Future Directions References

 

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-(178–199)

E. Other substances

V. Future Directions

	I. Introduction

Top Abstract I. Introduction II. Hypothalamic Stimulation of... III. Hypothalamic Inhibition of... IV. CRIF: A Consideration... V. Future Directions References

 
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

Top Abstract I. Introduction II. Hypothalamic Stimulation of... III. Hypothalamic Inhibition of... IV. CRIF: A Consideration... V. Future Directions References

View larger version (18K): [in this window] [in a new window]   Figure 2. Effect of synthetic CRF ( and •), purified native ovine CRF ( and ), and 8-Br-cAMP ( and ) on secretion of ACTH (solid line) and ß-endorphin-like immunoactivity (dashed line) by rat anterior pituitary cells in primary culture. Concentrations of synthetic and purified native CRF were determined by amino acid analysis. [Reproduced with permission from W. Vale et al.: Science 213:1394–1397, 1981 (9 ). © American Association for the Advancement of Science.]

View larger version (19K): [in this window] [in a new window]   Figure 3. Concentration-dependent effects of CRF and AVP on ACTH. Ovine anterior pituitary cells (3.3 x 105) were incubated for 4 h with the indicated concentrations of CRF (A) or AVP (B) and ACTH release (), total ACTH accumulation (), and intracellular ACTH content (•) were determined. Results are the means ± SE of triplicate determinations from four experiments. Error bars have been omitted when they were smaller than the symbol size. [Reproduced with permission from J.-P. Liu et al.: J Biol Chem 265:14136–14142, 1990 (15 ).]

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 83:1921–1925, 1986 (51 ).]

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-R_2ß, which differ at their N-terminal domain. CRF-R₂ is a 411-amino acid protein cloned from the rat brain (63), whereas CRF-R_2ß 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 15:6340–6350, 1995 (67 ).]

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 341:197–205, 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 116:1–29, 1996 (88 ). © Elsevier Science.]

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 196:271–285, 1981 (129 ). © Wiley-Liss, Inc., a division of John Wiley & Sons, Inc.]

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 36:165–186, 1983 (111 ). © Karger, Basel.]

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 24:138–153, 1985 (149 ). © Wiley-Liss, Inc., a division of John Wiley & Sons, Inc.]

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 (30–35 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 241:138–153,1985 (149 ). © Wiley-Liss, Inc., a division of John Wiley & Sons, Inc.]

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).

View larger version (23K): [in this window] [in a new window]   Figure 11. A representative example of the secretion of CRF, AVP, three POMC-peptides [ACTH, ir-ß-endorphin (ir-ß-EP), and ir--MSH (ir-- MSH)], and cortisol in a conscious ewe. Shown also are the secretory responses to a 3-min audiovisual stress (hatched area) and insulin-induced hypoglycemia (arrow). The triangles () depict significant hormone pulses. [Reproduced with permission from D. Engler et al.: Neuroendocrinology 49:367–381, 1989 (170 ). © Karger, Basel.]

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 93:1439–1450, 1994 (185 ). © The American Society for Clinical Investigation.]

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

Top Abstract I. Introduction II. Hypothalamic Stimulation of... III. Hypothalamic Inhibition of... IV. CRIF: A Consideration... V. Future Directions References

 
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 rabbit’s lymphopenic response to restraint and cold exposure while exerting little effect on this animal’s 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 data—varying size dogs and the difficulty in quantitating the magnitude of any particular duressed state—the 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 179:5–14, 1954 (253 ).]

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 66:200–216, 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 71:926–935, 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 71:926–935, 1962 (256 ). © The Endocrine Society.]

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 2:43–55, 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 2:43–55, 1967 (258 ). © Karger, Basel.]

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).

View larger version (25K): [in this window] [in a new window]   Figure 19. The effect of a 3-min audiovisual emotional stress on plasma POMC peptide and cortisol levels in the hypothalamopituitary-disconnected (HPD) ewe. The time of exposure to the stress is designated by the hatched area. The sham-operated animals are represented by closed symbols and continuous lines, and the HPD ewes are represented by closed symbols and broken lines. a, ACTH (); b, Ir-ß-endorphin (); c, ir--MSH (•); and d, cortisol (). [Reproduced with permission from D. Engler et al.: Neuroendocrinology 48:551–560, 1988 (269 ). © Karger, Basel.]

View larger version (28K): [in this window] [in a new window]   Figure 20. The POMC peptide and cortisol ultradian rhythm in four hypothalamopituitary disconnected (HPD) ewes. In this figure, the POMC peptides are displayed in the upper panels, and cortisol is shown in the lower panels. Significant hormone pulses are designated by . •, ACTH; , ir-ß-endorphin; , ir--MSH; , cortisol. [Reproduced with permission from D. Engler et al.: Endocrinology 127:1956–1966, 1990 (278 ). © The Endocrine Society.]

View larger version (77K): [in this window] [in a new window]