Origin of ultradian pulsatility in the hypothalamic-pituitary-adrenal axis

Abstract

The hypothalamic-pituitary-adrenal (HPA) axis is a neuroendocrine system that regulates the circulating levels of vital glucocorticoid hormones. The activity of the HPA axis is characterized not only by a classic circadian rhythm, but also by an ultradian pattern of discrete pulsatile release of glucocorticoids. A number of psychiatric and metabolic diseases are associated with changes in glucocorticoid pulsatility, and it is now clear that glucocorticoid responsive genes respond to these rapid fluctuations in a biologically meaningful way. Theoretical modelling has enabled us to identify and explore potential mechanisms underlying the ultradian activity in this axis, which to date have not been identified successfully. We demonstrate that the combination of delay with feed-forward and feedback loops in the pituitary-adrenal system is sufficient to give rise to ultradian pulsatility in the absence of an ultradian source from a supra-pituitary site. Moreover, our model enables us to predict the different patterns of glucocorticoid release mediated by changes in hypophysial-portal corticotrophin-releasing hormone levels, with results that parallel our experimental in vivo data.

Figure 1.

Regulation of HPA axis activity. The hypothalamic PVN receives circadian inputs from the SCN and homeostatic/stress inputs from the brain stem and limbic areas. The PVN projects to the median eminence where it releases CRH into the portal circulation. This passes to corticotrophs in the anterior pituitary which release ACTH from pre-formed granules into the venous circulation. This ACTH reaches the adrenal cortex where it activates the synthesis and secretion of CORTisol (in man) or CORTicosterone (in the rodent). CORT in turn feeds back to inhibit the release of ACTH from the anterior pituitary, and to a lesser extent, CRH from the hypothalamus.

Figure 2.

Experimental data demonstrating the ultradian glucocorticoid rhythm underlying the classic circadian profile. Levels of blood corticosterone were recorded over a 24 h period in two individual male Sprague–Dawley rats. Blood samples were collected every 10 min using an automated blood sampling system. Grey bars indicate the dark phase (19.15–05.15 h). Adapted from Spiga et al. (2007).

Figure 3.

Response of the pituitary–adrenal system to constant CRH drive. Units of all hormone levels are arbitrary. (a) Different combinations of constant CRH drive and delay can lead to two qualitatively different responses. On one side of the transition curve, the pituitary–adrenal system responds with constant levels in ACTH and CORT. On the other side of the transition curve, the pituitary–adrenal system responds with pulsatile fluctuations in the levels of ACTH and CORT, despite the fact that the CRH drive is constant. In the region of pulsatile response, the frequency of the pulses is indicated by the colour bar. (b–d) Model predictions for ACTH (blue) and CORT (black). Each time series was computed with the same delay (10 min), but different levels of constant CRH drive, as indicated by the three points in (a).

Figure 4.

Period of CORT pulses inside the pulsatile region. Period of ultradian CORT rhythm computed for different values of the adrenal delay T_lag (min) and different levels of CRH drive (arb. units). For all four values of the delay, we observe ultradian pulses with a physiological period. See also the colour bar in figure 3a.

Figure 5.

Response of the pituitary–adrenal system to circadian and ultradian patterns of CRH drive. Units of all hormone levels are arbitrary. (a) Different combinations of constant CRH drive and delay can lead to two qualitatively different responses. On one side of the transition curve, the pituitary–adrenal system responds with constant levels in ACTH and CORT. On the other side of the transition curve, the pituitary–adrenal system responds with pulsatile fluctuations in the levels of ACTH and CORT, despite the fact that the CRH drive is constant. In the region of pulsatile response, the frequency of the pulses is indicated by the colour bar. (b) Experimental data demonstrating an increase in pulse amplitude during the circadian peak. Adapted from Spiga et al. (2007). (c) Model prediction for a noisy circadian CRH drive close to (but below) the pulsatile region, as indicated by the corresponding arrow in (a). Response demonstrates NICOs during the peak of the circadian CRH drive. Computed with a delay of 9.4 min. (d) Model prediction for a circadian CRH drive in the pulsatile region, as indicated by the corresponding arrow in (a). Response demonstrates increased pulse amplitude during the peak of the circadian CRH drive. Computed with a delay of 15 min. (e) Model prediction for ultradian pulses of CRH drive in the pulsatile region, as indicated by the corresponding arrow in (a). Response demonstrates a frequency in CORT governed by the pituitary–adrenal system and not by the frequency of the CRH forcing. Computed with a delay of 12 min. (f) Model prediction for ultradian pulses of CRH drive in the region of constant response, as indicated by the corresponding arrow in panel (a). Response demonstrates a frequency in CORT that is governed by the frequency of the CRH forcing. Computed with a delay of 12 min.

Figure 6.

Effect of subchronic treatment with a GR antagonist on the 24 h corticosterone profile. (a) Data points represent mean levels of blood corticosterone measured from individual male Sprague–Dawley rats injected twice a day for 5 days with either the GR antagonist Org 34850 (10 mg kg⁻¹, subcut., n = 7, grey dots) or VEH (5% mulgofen in 0.9% saline, 1 ml kg⁻¹, subcut., n = 7, black dots). Blood samples were recorded over a 24 h period and collected every 10 min using an automated blood sampling system. Also shown are curves numerically fitted to the two datasets, demonstrating an increase in amplitude during the circadian peak under the influence of Org 34850. Grey bar represents the dark phase (19.15–05.15 h). Adapted from Spiga et al. (2007). (b) Model simulations show the response of the system to circadian CRH both with (grey) and without (black) a GR antagonist. Infusion of a GR antagonist increases the amplitude of the ultradian glucocorticoid rhythm during the peak of the circadian CRH drive together with a minor increase in ultradian frequency (grey). Computed with a delay of 15 min.