Septic Shock: A Disruptor of the Hypothalamic–Adrenal Axis as We Know It

The classic model of the hypothalamic–pituitary–adrenal (HPA) axis casts it as a major modulator of physiologic equilibrium: a stress increases hypothalamic AVP and/or CRH secretion, which increases adrenocorticotropin (ACTH) production and secretion, leading to increased adrenal glucocorticoid (GC) production and secretion. Enhanced GC action tampers the immunologic response to stress and decreases harmful cytokines, with the theoretical downside of impairing normal immunologic function. Thus, it is envisioned (somewhat teleologically) that GCs also restore equilibrium/balance by negative feedback inhibition of CRH and ACTH (1).

Septic shock defies this construct and may point the way for a more complete model of the HPA axis which would encompass a richer understanding of intra-adrenal GC action, the role of other GC-stimulating peptides, and trophic effects of ACTH to enforce fasciculata identity.

While the increased cortisol levels in human septic shock are in keeping with classic teaching, ACTH levels in the initial week are suppressed. Investigation of this apparent physiologic dissociation revealed a profound decrease in corticosteroid binding globulin (CBG) and a prolonged cortisol half-life due to reduced cortisol metabolism by 11beta-hydroxysteroid dehydrogenase type 2 (11βHSD2) and A-ring reductases (2), leading to increased biologically active cortisol levels in septic patients compared with healthy control subjects. Additionally, compared with healthy controls, lower amplitude ACTH and cortisol pulse bursts showed a similar ACTH:cortisol dose response but were more irregular and asynchronous (3). While these data suggest that increased biologically active cortisol accounts for a decrease in regulated ACTH secretion, the decreased ACTH stimulus could not account for all cortisol secretion.

Thus, the model of septic shock forces reconsideration of the “rules” of HPA regulation. Two recent papers by Tebrick et al evaluated septic intensive care unit patients and mice in an induced-sepsis model. Compared with murine controls, the earlier study (4) showed the expected increases in circulating corticosterone, hypothalamic corticotropin-releasing hormone (CRH) and arginine vasopressin (AVP), and mRNA adrenal transcripts related to cortisol production (melanocortin receptor 2, melanocortin receptor accessory protein, steroidogenic enzymes, cholesterol transport). The previous decrease in circulating ACTH was confirmed in people and mice; possible explanations included decreased pituitary PC1/3 (the enzyme responsible for proopiomelanocortin [POMC] cleavage to ACTH), and increased Annexin A1 (ANXA1, a known inhibitor of regulated ACTH release). However, the classic stress model did not explain persistent corticosterone synthesis with a progressive distortion of the architecture of the adrenal fasciculata and depletion of cholesterol ester in septic mice.

A subsequent paper reported in Endocrinology (5) used the same models to ask whether hydrocortisone (HC) infused in doses that mimic those used clinically (200–400 mg/24 hours) further aggravates the adrenal phenotype, and whether normal ACTH pulsatility (induced by CRH administration) would maintain normal fasciculata function. Evaluation of the adrenal gland was extended to evaluation of macrophages and their production of tumor necrosis factor α and interleukin 1β. Most of the previous findings regarding circulating POMC, ACTH and cortisol/corticosterone levels, and pituitary PC1/3 were replicated in the placebo group. Also, adrenal explants exposed to equimolar concentrations of ACTH or POMC secreted similar amounts of corticosterone, demonstrating the feasibility of POMC-induced steroidogenesis.

HC effects were similar to those of placebo except for lower levels of circulating ACTH and corticosterone, pituitary ACTH, adrenal mRNA markers of steroidogenesis, and cholesterol ester content, findings consistent with classic understanding of negative feedback and ACTH action. Additionally, the number of intra-adrenal macrophages increased dramatically, as did their known product, tumor necrosis factor α. By contrast, CRH effects were similar to placebo except for higher levels of circulating ACTH and normal levels of PC1/3 (ANXA1 was not evaluated.)

The authors conclude that in septic shock, endogenous CRH maintains POMC production but intrapituitary ACTH levels are low because of reduced PC1/3 levels; circulating ACTH decreases due to GC (feedback) and ANXA1 inhibition of the regulated secretory pathway. Further, intact POMC leaks from the corticotrope via the constitutive secretory pathway to increase circulating levels, which stimulate corticosterone production (and presumably increase markers of cholesterol transport and steroidogenesis).

These findings and conclusions in sepsis upend the classic teaching of pituitary ACTH “control” of adrenal trophic growth and steroidogenesis and suggest questions that require further study.

1. The usual hand-in-hand trophic support of adrenal structure and steroidogenesis was lost. Does that suggest different mechanisms of action for POMC vs ACTH?

2. Compared with healthy animals, circulating corticosterone levels were increased with CRH and placebo administration, yet hypothalamic CRH mRNA only increased in the placebo group. Does this imply feedback by ACTH or CRH on hypothalamic production?

3. What increases pituitary ANXA1 in placebo-treated sepsis? If levels are similar in the other treatment groups, do formyl proteins like lipopolysaccharide (LPS) regulate pituitary ANXA1 levels as they do in the adrenal (6)?

4. Despite similar pituitary POMC and PC1/3 levels (implying similar ACTH production) the HC group had lower circulating ACTH and corticosterone levels than placebo. Does this represent a failure of release due to enhanced negative feedback inhibition of the regulated pathway, with “unmasking” of the importance of ACTH to corticosterone production, or does it reflect a further decrease in CBG?

5. Annexin is involved in macrophage trafficking; in an LPS model of sepsis, adrenocortical ANXA1 and formyl proteins reduce lipid droplets/cholesterol ester and steroidogenesis (6). Is this mechanism involved in the loss of cholesterol ester in sepsis? Do we miss critical players by focusing mostly on endocrine pathways?

6. What is the role of intra-adrenal corticosterone vs circulating GC in regulation of steroidogenic enzymes (7)? To what extent does pulsatile vs constant GC exposure affect this?

7. Does a nonglycosylated POMC have the same actions as the endogenous substance?

8. Would application of an immune modulation model of septic shock identify immunologic drivers of endocrine abnormalities (8)?

Thanks to Dr Téblick and colleagues for these provocative studies. I look forward to additional reports of deviations from our current model of stress, so that we can capture all the different modes of regulation of the HPA axis and align them with findings of our immunology colleagues. Perhaps we will find that the adrenal gland is not a passive victim, but rather an active player in HPA–immune interactions.

Glossary

Abbreviations

ACTH

adrenocorticotropin

ANXA1

Annexin A1

GC

glucocorticoid

HC

hydrocortisone

HPA

hypothalamic–pituitary–adrenal

POMC

proopiomelanocortin

Financial Support

This work was funded by the intramural program of the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, 1ZIA DK075122.

Disclosure Statement

Dr. Nieman receives royalties from UpToDate.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.