[1] 19w Proceedings of the Australian Physiological and Pharmacological Society Symposium: The Hypothalamus FETAL HYPOTHALAMUS-PITUITARY-ADRENAL AXIS ON THE ROAD TO PARTURITION

[1] 122w 1. Activity of the fetal hypothalamus-pituitary-adrenal (HPA) axis waxes and wanes as a function of gestational age.2. In a number of species, including sheep, at the end of gestation there is an increase in HPA activity, as characterized by an increase in fetal plasma glucocorticoids.3. To a certain degree, the hypothalamus, pituitary and adrenal all act autonomously and, therefore, may be thought of as contributing to the initiation of the signal that results in the increase in steroidogenesis before birth.4. Because it integrates sensory information from beyond as well as within the HPA axis and likely triggers developmental changes within the pituitary, the hypothalamus may be a 'first among equals' in being the ultimate source of triggering information for the HPA axis.

[1] 322w The hypothalamus-pituitary-adrenal (HPA) axis plays a key role in preparing mammalian fetuses for extra-uterine life. As currently understood, the primary systemically acting agent of the HPA axis in this regard is cortisol (corticosterone in rats and mice). The glucocorticoid is responsible for stimulating the terminal maturation of various organ systems, including the lungs, liver and gastrointestinal tract, and likely plays a key role earlier in gestation as well. An example of the essential role of glucocorticoids may be seen in their effects on the lungs, where glucocorticoids accelerate morphological changes and stimulate the production of pulmonary surfactant. 1 In the absence of adequate exposure to glucocorticoids, as occurs in homozygous mice that lack corticotropin-releasing hormone (CRH), animals are born with poorly developed lungs and die of respiratory distress shortly after birth. 2 Normal lung develop-ment can take place if the mother of such mice is given corticosterone in her drinking water during pregnancy. In more physiological terms, we know that there is an exponential increase in the concentration of glucocorticoids in the fetal plasma over the last 10% of gestation 3 and one consequence of birth prior to this occurrence is respiratory distress. 4 Often, this can be successfully treated in humans at risk of preterm delivery, as with the transgenic mice, by maternal administration of glucocorticoids. 5 Equally important as the level of glucocorticoids to which the fetus is exposed is the timing of exposure. Precocious exposure of fetuses to glucocorticoids is associated with adverse outcomes that extend into adult life and those outcomes currently known probably represent only the most profound physiological and pathological effects. 6 Given the wide variety of actions of glucocorticoids, it is not unreasonable to speculate that many other developmental changes are dependent on or influenced by glucocorticoids as part of precisely timed sequences of events. Thus, an understanding of the control of glucocorticoids and the HPA axis is integral to the understanding of normal development.

[2] 230w In order to address the question of the origin of the signals that drive the fetal HPA axis, it may be helpful to review and contrast the function of the better-characterized postnatal HPA axis with the fetal counterpart. In our current basic understanding, the HPA axis exists as a multicomponent negative feedback loop. 7 That is, the hypothalamus is the source of several factors, most notably CRH and vasopressin (AVP), that stimulate the corticotrophs of the anterior pituitary to secrete adrenocorticotropin (ACTH). It should be noted that evidence also exists for inhibitory factors of hypothalamic origin. 8 Adrenocorticotropin stimulates growth of the adrenal cortex and the biosynthesis and secretion of a number of steroid hormones, the most important to the present discussion being cortisol (corticosterone in rats and mice). As noted earlier, cortisol has many systemic effects, including acting at the hypothalamus and pituitary to inhibit further activity of the HPA axis (negative feedback). 7 For the purposes of the present report, it is sufficient to consider ACTH as the sole input to the adrenal, with cortisol, CRH and AVP as the only inputs to the pituitary. In contrast, inputs to the hypothalamus are varied. In addition to the inhibitory influence of glucocorticoids, there are diverse inputs from other parts of the brain, which can be grouped together as part of the 'stress' response or as components of circadian rhythm.

[3] 390w The latter two functions reflect much of our understanding of the role of the HPA axis in postnatal animals. Fetal perceptions of 'stress' undoubtedly differ from those of postnatal animals. The fetal HPA axis is probably more appropriately thought of as being involved in the control of overall fetal growth and development. Important changes in the activity of the fetal HPA, as reflected in measurements of plasma concentrations of ACTH and glucocorticoids, are changes that occur over periods of multiple days, 3 in contrast with the daily and hourly movements observed postnatally. 7 Thus, the regulation of the fetal HPA axis is somewhat different from its postnatal counterpart. Two other notable differences between fetal and postnatal HPA axes are changes in the sensitivity of the adrenocortical cells to stimulation by ACTH, which occur during gestation, 9 and the continued secretion of ACTH near the end of gestation, despite the presence of cortisol at concentrations that should be sufficient to inhibit the further secretion of ACTH. 3 Therefore, it may be assumed that the control mechanisms ascribed to the postnatal axis may not necessarily apply to the fetal HPA axis. The purpose of the remainder of the present review is to discuss the signals within the fetal HPA axis that determine its activity. This is not intended to be a comprehensive review of the literature on this subject, but rather a limited examination of the contribution of each of the components of the axis towards development of the axis and the potential to be the source of the signalling that regulates changes in the axis during development. The results discussed here come from experiments in fetal sheep and their tissues. This has been a most useful species for such studies because of the availability of the animals and ability to perform surgery and experiments on fetuses in utero. In the present paper, the focal point will be the increase in fetal adrenal steroidogenesis that occurs near the end of gestation. For purposes of the present discussion, this can be summarized as being represented by the exponential increase in plasma cortisol that occurs in fetal sheep over approximately the last 10-15 days of gestation. Prior to that time, plasma cortisol has remained at or below 5 ng/mL for several weeks and, during the increase, cortisol levels typically increase by at least 10-fold.

[1] 446w Given that factors from the pituitary can drive ontological development of the adrenal cortex, it is reasonable to ask whether changes in pituitary function over gestation occur in a way that would be consistent with a coordinating role over the HPA axis. In this regard, it may be worth focusing on the biosynthetic capacity for ACTH. The peptide is synthesized as a sequence in the much larger precursor POMC, 7 which is also secreted and circulates. 16 As noted earlier, POMC and ACTH exert opposing actions on the fetal adrenal. Thus, any changes in the processing or secretion of various POMC-derived peptides by the pituitary that occur over gestation are physiologically relevant. We do know that the secretion rates of the inhibitory precursors POMC and pro-ACTH relative to that of ACTH(1-39) decrease over the last one-third of gestation, as demonstrated in vivo and in vitro. 16,17 The responses of the fetal pituitary to standard hypothalamic stimulation (CRH or AVP) also changes markedly over gestation. In fetal sheep pituitary in vitro, the maximum response to CRH declines as the maximum response to AVP increases. 18 Additionally, we found that the average secretory response of responding cells undergoes developmental changes. Whereas responses in pituitary cells of fetuses at 110 or 120 days gestation to CRH or AVP can be accounted for solely in terms of recruitment of additional cells to secrete ACTH, by 135 days gestation the response to AVP involves an increased amount of ACTH being secreted by the responding cells as well. 19 In contrast, other series of experiments, designed to directly address whether fetal corticotrophs are inherently different from adult corticotrophs and, thus, likely to be subject to regulation in a different manner, provide results that suggest fetal pituitary cells are probably taking their cues from outside the pituitary. For example, in comparing responses of 108 day fetal and adult sheep cells, we found that unstimulated and CRH-stimulated ACTH secretion in vitro, in the presence and absence of the protein synthesis inhibitor cycloheximide, are virtually identical on a per cell basis (Fig. 2). 20 The change in levels of POMC mRNA in response to 3 h treatment with CRH is also indistinguishable between the two groups of cells (24Ϯ5 vs 27Ϯ7%). 20 Other developmental differences between fetal and adult cells, such as levels of mRNA for the CRH type 1 receptor (CRHR1), can be partially accounted for by reaction to extrapituitary factors, such as cortisol (Fig. 3). 21 Indeed, the corticotrophs of the fetal pituitary, like their adult counterparts, appear to be subject to influence from both poles of the HPA axis, with the presumed developmental changes likely the effect of prolonged exposure to glucocorticoids.

[1] 218w The HPA axis undergoes marked changes over the course of gestation and is, in turn, responsible for promoting the timely maturation of other organ systems. One physiologically critical event in the development and function of the HPA axis is a sustained increase in glucocorticoid synthesis late in gestation. The adrenal cortex undergoes numerous significant changes necessary for the increase in steroidogenic activity. Although the adrenal cortex can apparently act somewhat autonomously in late gestation to produce the increase in plasma glucocorticoids, as demonstrated by experiments with hypophysectomized sheep fetuses, in the intact animal the fetal adrenal is subject to multifactorial input from the pituitary, which likely involves a balance of inhibitory and stimulatory signals. The fetal anterior pituitary also undergoes changes over gestation that result in altered function. These are not autonomous, being primarily responses to hypothalamic and adrenal factors and are probably similar to the modulatory influences that occur in postnatal life. The hypothalamus, in contrast, is essential for the normal development of the HPA axis and is likely the ultimate source of the cues that coordinate the development of the axis, at least in the final one-third of gestation. With all the structures of the HPA axis intact, except for disrupted communication between the hypothalamus and pituitary, normal development of the pituitary and adrenals is halted.

[1] 643w The adrenal is a logical place to search for the key to control HPA axis activity. In many ways, the fetal adrenal cortex appears autonomous. The steroidogenic response to ACTH changes markedly over gestation, with peaks in adrenocortical ACTH responsiveness occurring at 40-49 days of gestation and again at labour (length of gestation is approximately 145Ϯ3 days gestation), with markedly decreased adrenocortical responsiveness in the interim. 9 In studies of twin fetal sheep, we have found that there is a surprising level of dissociation of adrenocortical activity between the fetuses. 10,11 In daily measurements of plasma concentrations of cortisol and ACTH over the last 10 days before labour, we found that the terminal increase in cortisol began earlier and remained significantly higher in one twin than the other. There was no such difference between twin fetuses with regard to the plasma concentrations of ACTH. In in vitro tests performed on cells of the pituitaries and adrenals post-mortem, we found no differences between the twin fetuses in the ACTH secretory responses of the pituitary cells to CRH or AVP. In contrast, the adrenal cells of the twins that experienced the earlier greater increase in cortisol produced significantly greater cortisol responses to ACTH in vitro than those of the sibling twins. 10 Subsequent studies have provided evidence for an increased responsiveness of adrenal cells in vivo, as well. Despite plasma cortisol responses to exogenous ACTH having been indistinguishable between twins prior to separation of the baseline cortisol concentration, once plasma cortisol concentration in one twin passed a threshold level, indicative of the beginning of the terminal increase in cortisol, the cortisol response to exogenous ACTH was consistently greater in the twin with the higher baseline (in the face of what should have been feedback inhibition of the HPA axis). 11 Another line of evidence of independence by the adrenal cortex comes from a series of studies in hypophysectomized fetal lambs. 12,13 Following hypophysectomy at 117 days gestation, fetuses received either saline or synthetic ACTH at a steady low level of infusion. With plasma levels of ACTH below those normally produced during gestation and with no increase in concentration of ACTH, plasma cortisol nevertheless increased at term in hypophysectomized fetuses that received the low level infusion of ACTH in a pattern indistinguishable from that of intact control fetuses. 12 Other studies have demonstrated that the sensitivity to ACTH of the fetal adrenal increases at the end of gestation, even in the absence of a pituitary. 13 This evidence, obtained late in gestation, would suggest that the adrenal cortex is not dependent on signals from the pituitary and, thus, may be the source of signals that determine the activity of the axis. Such may be the case, provided there are no adrenocortical inhibitory signals from the pituitary and no other pituitary cues that may have programmed a sequence of events into the developing adrenal prior to the timing of the hypophysectomy. Following on some earlier studies of the activity of high molecular weight forms of ACTH on adrenocortical function, we measured cortisol responses to purified ACTH precursors, namely pro-opiomelanocortin (POMC) and pro-ACTH, in fetal sheep adrenal cells in vitro. 14 Representative data from that study are presented in Fig. 1. Both POMC and the biosynthetic intermediate pro-ACTH inhibited the cortisol secretory responses to ACTH (with all peptides used at physiological concentrations). Other factors from the pituitary, including fragments of the POMC molecule, may play a role in programming or establishing a development pattern for the adrenal cortex. We have found, for example, that infusion of POMC(1-77), but not POMC(1-49), into fetuses, even at 136-138 days gestation, increases adrenal mass, but Summary of representative data from Schwartz et al. 14 not cortisol output. 15 Evidence such as this is, therefore, consistent with some measure of the separate control of development of adrenal growth and steroidogenesis by the pituitary.

[2] 142w A solid body of evidence implicates the hypothalamus as the key developing component of the HPA axis and the potential ultimate source of ontological cues. In terms of observed changes in the developing hypothalamus relevant to control of the pituitary, it is known that levels of mRNA for CRH and AVP vary as a function of gestational age. The relative inability to directly measure hypothalamic responses per se in vivo or in vitro dictates that, to a large extent, the role of the hypothalamus must be ascertained by measuring the impact of the removal of its input. Two ways in which this is accomplished are studies of the effects of lesioning discrete areas of the hypothalamus or surgical disconnection of the 21 hypothalamus from the pituitary in a way that leaves both structures intact and functional, but severs the communication between them.

[3] 235w Numerous studies have been performed in fetal sheep in which the paraventricular nucleus (PVN) of the hypothalamus is lesioned. With permanent bilateral lesions placed at 106-110 or 118-122 days of gestation, the most apparent physiological change is that the fetus fails to deliver. 22,23 For purposes of study, these fetuses must be delivered surgically approximately 10 days after labour and delivery would normally occur. Lesioning of the PVN prevents normal development of the pituitary and adrenals, suggesting the PVN as a coordinating structure of the HPA axis during the last one-quarter of gestation. 24,25 Quite interestingly, the results of the lesion studies also illustrate how complex the regulation of the pituitary and adrenals must be. As noted above, full removal of pituitary input via fetal hypophysectomy at approximately 115 days, plus low-level replacement of ACTH is sufficient to allow for subsequent normal steroidogenic activity and labour, evidence that the adrenals may be the driving element in the development of the fetal HPA axis. In PVN-lesioned fetal sheep, there are normal circulating levels of ACTH throughout most of gestation, but no increase in ACTH or glucocorticoids at term or evidence of labour. 22 This is likely due to inadequate development of the pituitary (in functions other than the secretion of immunoreactive ACTH) and adrenals. 24,25 Therefore, rather than having an autonomously operating adrenal, the data are consistent with a complex regulatory system, orchestrated by the hypothalamus.

[4] 73w Studies in which the PVN has been left intact, but the physical connection between the hypothalamus and pituitary has been surgically disconnected (HPD), have also provided valuable insights into the function of the HPA axis in fetal sheep. In the HPD procedure, the median eminence is severed, a foam barrier is inserted above the pituitary, the wounds closed and the fetus returned to the uterus. 26 Blood supply to the pituitary is maintained.

[5] 128w The HPD fetuses also demonstrate a relatively flat steroidogenic profile at the time labour and delivery would normally occur; they do not deliver. 27 In contrast with surgical controls, the concentration of ACTH(1-39) does not increase at term in HPD fetuses. 28 This is associated with impaired development of the adrenal cortex. 28 Taken together, the results of experiments with PVN lesions and HPD fetuses indicate the necessity of intact hypothalamic input to the HPA axis, even late in gestation, for normal development to occur. While it is clear that an intact fetal hypothalamus is required for the orchestration of the neuroendocrine events that underlie the normal prepartum increase in plasma cortisol, it is still unclear what inputs to the fetal hypothalamus initiate these events in the sheep.