Hypothalamic-Pituitary-Adrenal Axis in Healthy Neonatal Foals
ACVIM 2008
Michelle Henry Barton, DVM, PhD, DACVIM; Kelsey Hart, DVM
Athens, GA, USA

Introduction

The hypothalamic pituitary adrenal axis (HPA axis) tightly regulates systemic cortisol levels in both health and disease, and thus plays an integral role in the maintenance of cellular, organ, and whole body homeostasis. The purpose of this review is to specifically discuss the dynamic changes in HPA axis function that occur in healthy neonatal foals during the first week of life.

Components and Actions of the HPA Axis

Both exogenous and endogenous physiologic stressors, such as hypotension, hypoglycemia, pain, tissue injury, and cytokine release, activate the peripheral and central nervous systems. These signals are interpreted and integrated in the hypothalamus, culminating in the release of corticotropin-releasing hormone (CRH). CRH is secreted into the hypothalamic-pituitary portal circulation, where it acts on the anterior pituitary gland to induce the release of adrenocorticotropic hormone (ACTH or corticotropin). ACTH, in turn, enters the systemic circulation and stimulates the adrenal cortices to release cortisol. Cortisol is synthesized on demand from cholesterol in the zona fasciculata and zona reticularis (middle and deep layers) of the adrenal cortex. Cortisol synthesis primarily takes place in the mitochondria and endoplasmic reticulum, wherein cholesterol is converted to pregnenolone, then to progesterone, then to 17-OH progesterone, and finally to cortisol (i.e., hydrocortisone). Once this newly synthesized cortisol is released from the adrenal gland, increased systemic cortisol levels exert a negative feedback effect on the pituitary gland and hypothalamus, resulting in subsequent down-regulation of both CRH and ACTH secretion.1 With an intact HPA axis, systemic cortisol levels are maintained at a level that is appropriate for the existing degree of physiologic stress. Since cortisol is not stored in the adrenal cortices or in peripheral tissues, the circulating level of cortisol is controlled primarily by its rate of biosynthesis. Thus any disruption in cortisol synthesis will immediately result in systemic cortisol insufficiency.

After release from the adrenal glands, cortisol binds to its high affinity plasma protein transporter, corticosteroid binding globulin (CBG or transcortin). Corticosteroid binding globulin is a 50 to 60 kD glycoprotein that is a member of the serine protease inhibitor superfamily. In most species, approximately 90 to 95% of plasma cortisol is bound to CBG, with a small percentage also exhibiting low affinity binding to albumin. However, similar to other lipid-soluble molecules, like thyroid hormones, it is only the unbound or "free" fraction of cortisol that is biologically active. CBG is a negative acute phase protein and thus blood concentrations of CBG decrease during acute stress and inflammation, increasing the pool of physiologically active free cortisol.

Free cortisol enters cells by passive diffusion through the membrane and binds to the cytoplasmic glucocorticoid receptor, which is a member of the ligand-regulated nuclear receptor family. The cortisol-glucocorticoid receptor complex then dimerizes and translocates to the nucleus where it binds to DNA and regulates gene transcription, resulting in a myriad of systemic effects that are vital for adaptation to stress in both health and disease. Essential physiologic responses to the stress of illness include maintenance of blood pressure, provision of energy to tissues, and control of an appropriate inflammatory response.1 Cortisol increases the synthesis and tissue reactivity to catecholamines. In addition, it down regulates the synthesis of the potent vasodilator nitric oxide.2 These combined effects beneficially impact cardiac contractility, vascular tone, and endothelial integrity, all of which are vital for the maintenance of systemic blood pressure. Cortisol also regulates metabolism by increasing hepatic gluconeogenesis and inhibiting the uptake of glucose by adipose tissue. These effects, in concert with cortisol-induced free-fatty acid release from adipose tissue and protein catabolism, mobilize several energy sources needed for host defense, maintenance of organ function, and repair of damaged tissues.2 Finally, the cortisol-receptor dimer acts in the nucleus to inhibit gene expression of proinflammatory mediator transcription factors, such as nuclear factor κβ (NF-κβ ), while transactivating gene expression of anti-inflammatory molecules.3,4 Through these actions, cortisol exerts various anti-inflammatory and immunomodulatory effects by decreasing production and activity of pro-inflammatory cytokines including interleukin-1, tumor necrosis factor (TNF), and γ-interferon, stimulating the release of anti-inflammatory mediators (e.g., soluble TNF receptor, transforming growth factor, and interleukin 10), and inhibiting the activation of pro-inflammatory pathways involving cyclooxygenase-2, inducible nitric oxide synthase, and phospholipase A2.2

Studies in Neonatal Foals

Previous Work

The HPA axis and cortisol response to exogenous ACTH has been evaluated in healthy foals during gestation and the early neonatal period.5-10 Compelling evidence in foals shows that maturation of the HPA axis starts during the last 4 to 5 days prior to parturition and continues into the first one to two weeks of life.6,8,10,11 Serum cortisol levels in foals remain very low until just prior to parturition, which is much later than in other species (e.g., lambs).12 Within 5 to 10 minutes after parturition, plasma ACTH concentrations in newborn foals peak to 250 to 300 pg/ml and decrease to adult levels within 6 hours of life.6 Concurrently, serum total cortisol concentrations peak at approximately 12 to 14 µg/dl within 30 to 60 minutes post-partum and decline within 6 to 12 hours after birth. Serum total cortisol concentrations in foals are significantly less than mature horses during the entire first year of life.13 Furthermore, diurnal variations in cortisol concentrations in foals are not detectable until after the first week of life. The maximal cortisol response to 125 µg of exogenous ACTH (SynacthenTM) given IM in healthy term foals occurs on the day of birth, with peak cortisol concentrations reaching approximately 8-12 µg/dl 60 to 90 minutes after ACTH stimulation. Basal cortisol concentrations and the cortisol response to ACTH decrease in subsequent days after birth, such that by five days post-partum the peak cortisol response to ACTH is between 5 to 7 µg/dl.6,10 There is little information on free cortisol concentrations in foals. However, similar to reports in other neonates, foals appear to have a lower corticosteroid binding globulin capacity and thus may have a greater percentage of free cortisol compared to mature horses, despite lower total cortisol concentrations.14

Further insight into the role and development of the HPA axis in neonatal foals was revealed by studies on prematurity. Premature foals (gestational age <320 days) have low levels of serum cortisol (<3 µg/dl) in the 2 hours after birth, as compared to term foals (12 to 14 µg/dl).7 Concurrently, premature foals showed significant increases in endogenous ACTH levels, peaking at 650 pg/ml at 30 minutes postpartum, as compared to a peak of 300 pg/ml in term foals.7 This low cortisol concentration and delayed, high ACTH peak in premature foals suggests that the failure of cortisol synthesis resides primarily at the level of the adrenal gland, while the central portions of the HPA axis appear intact.10 Synthetic ACTH stimulation tests (125 µg cosyntropin IM) in premature foals support this theory. On the first day of life, premature foals show a poor response to ACTH, with only a 28% increase in plasma cortisol 30 to 60 minutes following stimulation, as compared with a 208% increase in normal term foals.10 In addition, premature foals in this same study showed other abnormalities consistent with cortisol insufficiency, including abnormally low neutrophil to lymphocyte ratios, hypoglycemia, and hypoinsulinemia. Similar findings of HPA axis dysfunction and cortisol insufficiency are reported in premature human infants.15

Even full term neonatal foals appear to lack a fully responsive HPA axis, as was demonstrated by investigations into the cortisol response to insulin and hypotension. Insulin-induced hypoglycemia is a potent stimulus for cortisol release; in fact, the cortisol response to insulin-induced hypoglycemia often is considered the "gold-standard" test for adrenal insufficiency.16 When insulin was given to induce hypoglycemia in 3 to 10 hour old foals, the change from baseline in serum cortisol concentration was less than half of the insulin-induced cortisol response achieved in 7 to 14 day old foals.9 In a comparative study on the responsiveness of the adrenal gland to acute hypotension, the foal once again appears to lag behind developmentally in the neonatal period.5

Thus, the intriguing question that is stimulated by this knowledge is how does the delayed HPA axis development in the neonatal foal impact its ability to respond to stress and illness during the neonatal period? Any stressful event, from exposure to extreme temperatures to severe systemic sepsis, can disrupt the balance of homeostasis. The ability to appropriately regain homeostatic stability is pivotal to survival through that stressful event and the HPA axis plays a vital role in this process. It might seem logical to hypothesize that neonatal HPA axis immaturity could impact and potentially limit the foal's ability to respond to physiologic stress and disease in the neonatal period. Indeed, in recent years critical care specialists in the human arena have given much attention to the role of the HPA axis during critical illness. In people2,17-19 and some veterinary species20,21 evidence is mounting that there is a significant association between severe illness, especially sepsis, and HPA axis dysfunction. This topic will be discussed in detail in the next session. However, before one can assess the HPA axis during illness, a precedent must be set for what is normal during health. Although ACTH, cortisol, and the cortisol response to ACTH was eloquently described in neonatal foals in the 1980s by Rossdale et al, the studies were performed with tetracosactrin (ACTH 1-24; SynacthenTM Novartis) that is not widely available in the United States, and the ACTH and cortisol assays that were used then have been widely replaced by newer non-radioactive immune methodologies. Furthermore, pioneering studies on ACTH stimulation testing in foals were performed using 125 to 250 µg of ACTH which results in blood concentrations of ACTH that are almost 300 times that which can be achieved physiologically.16 Although a supra-physiologic dose of ACTH might be appropriate when testing for absolute adrenal gland insufficiency, lower "physiologically" relevant doses of ACTH would be more appropriate for investigation of relative, transient dysfunction of the adrenal gland such as appears to occur in neonatal foals.16

Current Work in Neonatal Foals

To establish current, clinically relevant HPA axis evaluation protocols in neonatal foals, documentation and validation of the cortisol response to aqueous synthetic 1-24 ACTH (cosyntropin; CortrosynTM Amphastar Pharmaceuticals, Rancho Cucamonga, CA) in neonatal foals was necessary. Fourteen healthy 3-day-old foals were given four different doses (1, 10, 100 and 250 µg) of cosyntropin intravenously and serum total cortisol concentrations were measured by a chemiluminescent immunoassay every 30 minutes for two hours (see Figure 1).22 There were four important conclusions from this study. First, in contrast to findings in other species, neonatal foals do not mount a significant cortisol response to the standard low dose (1 µg) of cosyntropin. Second, neonatal foals do mount a cortisol response to 10 µg of cosyntropin, which is statistically indistinguishable from the cortisol response to 100 or 250 µg of ACTH at 30 minutes. Third, 90 minutes after giving 10 µg of ACTH, cortisol values are statistically indistinguishable from baseline values. And finally, there is no significant difference in the cortisol response to 100 and 250 µg ACTH at either 30 or 90 minutes following administration. These findings suggest a suitable paired low dose/high dose cosyntropin stimulation test using a 10 µg ACTH dose followed 90 minutes later by a 100 µg ACTH dose for complete assessment of adrenal function in neonatal foals (see Table 1).

Next, resting basal cortisol and endogenous ACTH concentrations and the cortisol response to the above suggested low dose/high dose cosyntropin stimulation protocol were tested in 10 healthy foals at four times (birth, 12-24 hours, 36-48 hours, and 5-7 days) during the first week of life.23 Total serum cortisol and endogenous ACTH concentrations were measured by chemiluminescent immunoassays. In addition, the effect of low dose cosyntropin on the subsequent cortisol response to high dose cosyntropin was investigated in 8 foals with a "sham" low dose cosyntropin test (i.e., foals were given saline at time zero, followed 90 minutes later by the 100 µg cosyntropin). Finally, free cortisol concentrations were determined using an ultrafiltration/ligand binding method with radiolabelled cortisol.24 There were several important conclusions from this study. Firstly, in agreement with previous reports, resting basal cortisol and endogenous ACTH concentrations in foals are significantly higher at birth as compared to all other ages assessed and stabilize by 12-24 hours of age. Secondly, comparison of the cortisol data between the sham low dose/high dose cosyntropin and the real low dose/high dose cosyntropin indicates that the low dose of cosyntropin neither primes nor suppresses the cortisol response to the high dose of cosyntropin. Thirdly, healthy neonatal foals show a significant and similar cortisol response to both low and high dose cosyntropin; while the magnitude of the cortisol levels reached significantly decreases with age, the pattern of response to cosyntropin stimulation was similar across all four ages (see Figure 2). Fourthly, compared to human infants and adults, and mature horses, HPA axis activity appears to be substantially lower in neonatal foals. Specifically, the mean plasma ACTH concentration in 4 to 7 day old infants (500 pg/ml)25 is approximately 15 times greater than the mean plasma ACTH concentration of similar age foals (34 ± 23 pg/ml). Basal cortisol values in infants at 4 to 7 days of age (14.8 ± 1.9 µg/dl)25 are roughly seven times greater than age-matched foals (2 ± 0.8 µg/dl). Similarly, the peak cortisol responses after similar weight controlled low dose/high dose ACTH stimulation testing in 4 to 7 day old infants (38.1 ± 5 µg/dl and 84 ± 6.9 µg/dl, respectively)25 are approximately 10 to 15 times the magnitude of response obtained in age-matched foals. Finally, our findings suggest that the percentage of free cortisol is relatively higher in the neonatal foal and does not change significantly in response to exogenous ACTH, despite the resultant increase in total serum cortisol concentration. Specially, at birth the mean percentage of total cortisol that is free in foals is 53%, which decreases to 40% by 5 to 7 days of life.23 This is in contrast to human infants wherein the reported mean percentage of free cortisol is 30% at birth and then gradually decreases to approximately 19% by 3 months of age.26 Although assays for corticosteroid binding globulin not currently available in horses, the free cortisol levels in foals imply that there is less corticosteroid binding capacity and/or nonspecific protein binding of cortisol in the neonatal period, resulting in a higher free cortisol fraction as compared to other species.

The results of these studies in foals indicate that during the first week of life, serum cortisol concentrations are lower and the response to exogenous ACTH is blunted in comparison to other neonatal species and the mature horse. These unique species- and age-related differences in HPA axis function suggest that ACTH stimulation testing protocols need to be age-controlled in the neonatal foal. Furthermore, although the same basic diagnostic criteria for HPA axis dysfunction in people can be applied to foals, the absolute cut-off values for adequate versus inadequate cortisol response to cosyntropin used for people, infants, or mature horses cannot be used in foals and will depend on data established in healthy foals. In addition, considering the low basal and ACTH-stimulated cortisol concentrations in neonatal foals during the first week of life, replacement hydrocortisone therapy during critical illness in neonatal foals warrants further investigation.

Figure 1.
Figure 1.

Serum total cortisol concentrations (µg/dl) in healthy 3- to 4-day old foals before and 30, 60, 90, and 120 minutes after intravenous administration of 1, 10, 100, and 250 µg of cosyntropin. Cosyntropin doses with different letter superscripts have significantly different area under the concentration curves.22
 

Figure 2.
Figure 2.

Serum cortisol responses to a paired low/high dose cosyntropin stimulation test in healthy neonatal foals. 10 µg and 100 µg of ACTH was given at times 0 and 90 minutes, respectively.23
 

Table 1. Proposed low dose/high dose cosyntropin stimulation test in neonatal foals.

Time

Task

0

Obtain blood for determination of baseline serum cortisol concentration.
Give 10 µg ACTH IV.

30 minutes

Obtain blood for determination of cortisol concentration.

90 minutes

Obtain blood for determination of cortisol concentration.
Give 100 µg ACTH IV.

120 minutes

Obtain blood for determination of cortisol concentration.

180 minutes

Obtain blood for determination of cortisol concentration.

References

1.  Guyton A. Textbook of medical physiology, 7 ed. 1986.

2.  Marik P, et al. Chest 2002;122:1784.

3.  Beishuizen A, et al. J Endotxon Res 2003;9:3.

4.  Czock D, et al. Clin Pharmacokinet 2005;44:61.

5.  O'Conner S, et al. Europeon J of Physiology 2005.

6.  Ousey J, et al. Equine Vet J 2004;36:616.

7.  Rossdale P, et al. Equine Vet J 1984;16:300.

8.  Silver M, et al. J Endocrinology 1994;142:417.

9.  Silver M, et al. J Reprod Fert Suppl 1987;35:607.

10. Silver M, et al. Equine Vet J 1984;16:278.

11. Nathanielsz P, et al. J Reprod Fert Suppl 1975;23:625.

12. Brown E, et al. Acta Endocr 1978;88:364.

13. Flisi n ska-Bojanowska A et al. Zentralbl Veterinarmed A 1992;39:641.

14. Irvine C, et al. J Reprod Fert Suppl 1987;35:19.

15. Watterberg K. Seminars in Neonat 2004;9:13.

16. Kozyra E, et al. Ann Pharm 2005;39:691.

17. Annane D, et al. JAMA 2002;21:862.

18. Annane D, et al. JAMA 2000;283:1038.

19. Marik P, et al. Crit Care Med 2003;31:141.

20. Burkitt J, et al. J Vet Intern Med 2007;21:226.

21. Gold J, et al. J Vet Intern Med 2007;21:791.

22. Hart K, et al. J Vet Intern Med 2007;21.

23. Hart K, Barton M. Unpublished data. 2007.

24. Ho J, et al. J Clini Endocrin and Met 2006;91:105.

25. Soliman A, et al. Metabolism 2004;53:209.

26. Rokicki W, et al. Biology of the neonate 1990;57:2.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Michelle Barton, DVM, PhD, DACVIM
University of Georgia
Athens, GA


SAID=27