V

Glucocorticoids and
immune function:
physiological relevance
and pathogenic potential
of hormonal dysfunction
Thomas Wilckens
The more knowledge accumulates on the molecular
action of glucocorticoids, the less appears to be known
about the physiological relevance of these data.
Mechanisms that determine bioactMty of
glucocorticoids are largely neglected in studies on their
molecular actions in immunoregulation. However,
alterations of these mechanisms may contribute to the
pathogenesis of acute or chronic inflammation, ranging
from septic shock to autoimmune disease, or even
acquired immuno deficiency syndrome. In this article,
Thomas Wilckens attempts to challenge the
long-standing dogma of glucocorticoids being only
immunosuppressive in their action, and suggests a
physiological role in which they are a prerequisite for a
coordinated immune response.

Glucocorticoids are at the centre of attention as a result of

their anti-inflammatory properties, in addition to their
potential role in the development of autoimmune disease 1,2. The molecular basis for their action has been
reviewed by Barnes and Adcock3; this principally
includes the binding of the ligand to an intracellular
glucocorticoid type II (type II) receptor, and the formation
of a complex, which then binds to a so-called glucocorticoid response element. Glucocorticoid response elements
are found abundantly in diverse genes encoding
cytokines as well as other mediators or cellular regulatory
elements of inflammation and immune responses in general. A recent additional function of glucocorticoids is
their ability to directly interact with various transcription
factors that have important roles in immune responses.
Glucocorticoids are released upon stimulation of the
immune system by various stimuli via a neuroendocrine-immune network, and because of this it has
been proposed that their role in this context is to prevent
immune reactions from 'overshooting'. Thus, glucocorticolds are generally thought to be immunosuppressive;
this is further substantiated by in vivo and in vitro data, for
example, the regulation of cytokine gene expressionL
However, the majority of these studies used synthetic
ligands for glucocorticoid receptors or, if an endogenous

1995, Elsevier Science Ltd

ligand was used for a comparison of efficacy, it was often

found that the endogenous glucocorticoid was less
potent. Overall, the importance of these findings may
have been underestimated with respect to the characterization of the physiological role of glucocorticoids. Here,
it may be noted that valuable information might be lost if
only synthetic ligands are used to study the physiological
role of glucocorticoids, particularly when concerned with
the immune-neuroendocrine network.
For example, suppression of interleukin 6 (IL-6) activity by dexamethasone as well as by hydrocortisone has
been reported, yet the endogenous glucocorticoid cortisol
was 25 times less effective than dexamethasone4. Thus,
IL-6 suppression by cortisol only occurred at concentrations that are not reached in vivo. Similar findings were
reported for the IL-1 receptor antagonist protein 5. In vitro,
endotoxin induced IL-1 receptor antagonist protein gene
expression was blocked by both dexamethasone and cortisol, again cortisol being less potent. Both reports contrast
observations in vivo in models of experimental endotoxaemia or septic shock, where levels of IL-6 and IL-1Ra are
found to be high despite elevated levels of endogenous
cortisol5,6. However, the authors conclude from these
studies that the physiological role of glucocorticoids may
be one of general suppression of cytokine induction during inflammatory processes. It should be noted that interpolation of the cortisol concentrations used in vitro with the
physiological concentration range of free cortisol, only
very minor changes in cytokine levels may be physiologically relevant, even though specific delivery systems may
regulate local hormone levels that are currently not
measurable. Defining a physiological concentration range
for dexamethasone or other synthetic corticosteroids is
not relevant for physiological functioning.
Studies concerned with the in vivo regulation of
cytokines during conditions of stress demonstrated di~screpancies between the proposed general suppression of
cytokine induction under conditions of elevated corticosteroid levels. Thus, in a state of elevated endogenous
glucocorticoids induced by chronic stress, splenic IL-lc~
levels, as well as plasma IL-6 levels, were unaffected in
their response to intraperitoneal administration of endotoxin, whereas the level of tumour necrosis factor-c~ in
plasma was significantly reduced in the hypercortisolaemic state:. However, other groups have reported su ppressed cytokine induction after pretreatment of humans
or animals with glucocorticoids1. Taken together these
results suggest that some cytokines may well be suppressed by endogenous glucocorticoids whereas others
are not; this may depend on the type of glucocorticoid, the
concentration used, and bioavailability of the glucocorticoid, as well as the specific tissue being studied.

Mechanisms regulating the pharmacokinetics,

bioavailability, and biological effects of
glucocorticoids
Regulatory influences, which may have direct clinical
implications are discussed (see Box 1).

TiPS-June 1995(Vol. 16)

T. Wilckens,
ResearchFellow,
Institute for Hormoneand
Fertility Research,
University of Hamburgr
Grandweg 64. 0-22529
Hamburg, Germany

1 9 3

Box 1. M

at the

higl'dy

"

as the
is in ~

the

te glucocorticoid re

~ m e n t is
on the c
of AP-1, as well
id
1. The
on of AP-I
regulated by the activation of different

(i).
KB
The p65 subunit of the NFKB is a target for both the

relation is

(6).

Contras~g the separate effects of glucocorticoids on AP-1 or NFsB, a synergistic inter-

not

Refm
3

Bioavailability
A large proportion (90-97%) of the endogenous ligands
(for example, cortisol, corticosterone) in circulation are
bound to transcortin (or corticosteroid binding globulin,
CBG) and to a lesser extent to albumin. In contrast, with
synthetic ligands such as dexamethasone (with the exception of prednisolone) only a small proportion binds to
CBG (Ref. 8). Thus, the so called 'physiological' effects of
glucocorticoids observed in vitro (which often become
significant at concentrations greater than 10~ M) do not
necessarily accurately reflect free physiological concentrations of bioavailable and bioactive cortisol9. As shown
in a more recent study, free cortisol was reported to be
between 11-43 nM, a concentration that is between 5-9.3%
of the total cortisol concentration 10.Physiological concentrations of glucocorticoids, similar to those found in relevant target tissues may be more accurately reflected by
the concentration of free cortisol detectable in saliva and
in aqueous humour u,12. However, salivary cortisol is
usually 50% lower than free plasma cortisol as saliva contains an enzyme that can metabolize cortisol. There are
reports that indicate free cortisol to remain in a normal
range in some disease states 8, but at present it is not
known precisely how free glucocorticoid levels are

1 9 4

TiPS -

June 1995 (Vol. 16)

R.I., Gualberto, A., JeweU,C. M , , ~

A. S, Jr (1995)Mol. Call.Biol. 15, 943-953

J. A. and

regulated in inflammation or during other immunological 'stressor' responses. It has also been suggested that
CBG acts as a transport protein for glucocorticoids to the
inflamed regions, and this raises the possibility that tissue
cortisol levels could exceed those of free cortiso113, a theory, which at present appears very difficult to prove. Considering that CBG is not only influenced by oestrogens 8,
but is also downregulated by IL-6 (Ref. 14), it would
appear that such a simple feedback loop may have a profound impact on the outcome of a local or even systemic
inflammatory reaction. Mutations of CBG have been
described in humans, but it is unclear whether other variations of CBG exist which could be of clinical importance s.

Two and a 'half' receptors for glucocorticoids

Type II receptors are thought to be a major determinant
of apoptotic cell death, for example, in T cells. In human
leukaemic T-cells, glucocorticoids upregulate type II receptors, whereas in a B-cell line they downregulate mRNA
expression, indicating tissue specificity in glucocorticoid
receptor regulation 15. One possible explanation for these
observations is that endogenous glucocorticoids display an
affinity for the mineralocorticoid (type I) receptor, which is
approximately 10-fold higher than the type II receptor. This

extracellular
compartment

glucocorticoid
corticosteroid binding globulin

r ' e 'tor

Io J
nucleus
<~
~ ( 4 ~

'~

oestrogen

<~[~>

~ / i r ~ )
<~

AP-,

trdden

(7)NF.B

simple

DNA

glucocorticoid
response element

Fig. Thevariousstepsinvolvedinthedeliveryof glucocorticoidstothedifferentglucocorticoidresponseelements,andthedifferentinteractionsthat

canoccur.Numbersin parenthesesreferto the differentstagesthatarediscussedin the boxtext. 11#HSD,11-[3-hydroxysteroiddehydrogenase;
hsp90,heatshockprotein90;NFKB,nuclearfactorKB;p50andp65,p50andp65subunitsof NFKB.

At the molecular level, type I and type II receptors may

means that only under conditions where glucocorticoids
are elevated during the circadian peak or to concentrations compete at the glucocorticoid response element, dependassociated with stress, will type II receptors be occupied, ing on the type of response element (either simple or comi.e. after saturation of type I receptors 16. Dexamethasone posite); for example, cortisol may have opposing effects if
binds preferentially to type II receptors. Type I and type II bound to the different receptors, type I receptors acting as
receptors show not only diverse patterns of co-localization a functional antagonist to type II receptors 19.
in various tissues, but also influence their own expression 17.
Recently, evidence for a role for both receptors in
In tissues that express both receptors such as the spleen, immunomodulation has been provided; in adrenaltreatment with the type I receptor agonist aldosterone was ectomized rats supplemented with corticosterone to
able to prevent the upregulation of type II receptors that maintain low concentrations (predominantly type I reoccurs following adrenalectomy, whereas in the thymus, ceptor occupation), corticosterone enhanced in vitro
expressing only type II receptors, no effect was observed. concanavalin A-induced T-cell proliferation, whereas
The physiological relevance of these studies becomes evi- stress-related levels of corticosterone, which also occupy
dent at low endogenous corticosteroid levels, as at these type II receptors, were completely suppressive 20. In furconcentrations only type I receptors will be occupied. It was ther experiments Wiegers and co-workers demonstrated
further demonstrated that target tissue responsiveness distinguishable effects of exposure duration, dose, and
depends both on receptor expression and endogenous cor- receptor type occupation on T-cell mitogenesis in vitro
tisol levels, and therefore varies between tissues through- using specific agonists and antagonists 21.
Heterodimerization of type I and type II receptors has
out the circadian rhythm as well as under conditions of
stress 18. It is worthwhile noting that the relative ratio of been reported22; however, whether this mechanism has an
type I : type II receptors seemed to be of minor influence, a impact on immunoregulation requires further investifact that may be of particular importance in immune- gation. Recently, a second type II receptor, termed type
related tissues or cells that express only very low amounts lip (the other now being named type II~), has been
characterized 23. The type IIJ3 receptor is the product of
of type I receptors.

TiPS

June1995(Vol.16) 1 9 5

alternative splicing and, at present, does not bind any

known ligand, but it is able to act as a functional antagonist of type IIc~receptors at the glucocorticoid response
element; type II]3 receptors are expressed in all target
tissues for glucocorticoids. The potential new endogenous, tissue-specific regulator of glucocorticoid action
awaits further investigation with respect to its clinical
potential.

Access of glucocorticoids to type I and type H

receptors by 11-~-hydroxysteroid dehydrogenase
11-6-hydroxysteroid dehydrogenase (11[3HSD) is the
key enzyme in extrahepatic conversion of 11q3-hydroxysteroids such as cortisol and prednisone to their inactive
metabolites 24,25.In the kidney, 1113HSD stops the type I
receptor (a target for aldosterone), from being occupied
by glucocorticoids. In the kidney, 1113HSD also accounts
for a low prednisolone:prednisone ratio 26. 11[3HSD is
abundantly found in various tissues including the gastrointestinal tract, kidneys, liver, spleen, skin and lungs.
Thus, it can regulate the efficacy of endogenous glucocorticoids at the tissue level. It should be noted that two
different types of 116HSD have been found in human
colon27.Only 116HSD expressed in aldosterone target tissues (ones with type I receptors) can metabolize dexamethasone 24. If 116HSD was a target for regulation by
cytokines, as observed with dehydroepiandrosterone sulphatase 28, this would provide a basis for a localized regulatory mechanism of glucocorticoid bioavailability during
inflammation. One subtype of 1113HSD appears to be
localized to neuroendocrine cells of the lamina propria
which have been implicated in the pathogenesis of inflammatory bowel diseases29; thus, extensive glucocorticoid
metabolism by 116HSD could result in chronic inflammation. It has been shown that lymphocytes of rheumatic
patients display enhanced glucocorticoid metabolism30;
however, at present it is uncertain as to whether these
lymphocytes express 11f~HSD. In the clinic, inhibitors of
11[3HSD enhanced the local effect of glucocorticoids in
patients with psoriasis and eczema31.

Tissue specific expression of heat shock

protein 90
Heat shock protein 90 (hsp90) is complexed with the
free glucocorticoid receptors and hence can interfere with
the binding of glucocorticoids to these receptors. Hsp90
also shows tissue specific expression 32. In an animal
model of systemic lupus erythematosus (SLE), a tissue
specific (spleen) increase of hspg0 preceeded the onset of
disease 33. Hsp90 also regulates the trafficking of type II
receptors between the cytoplasm and the nucleus, that is,
receptor recycling, it can influence glucocorticoid efficacy34.So far it is unknown whether the observed increase
in hsp90 in the animal model is causative, but it does
reflect similarities with a subgroup of patients suffering
from SLE (Ref. 35). If autoimmunity against hsp90,
namely interfering with receptor recycling, could be
demonstrated as for other nuclear factors, this would

1 9 6

TiPS - J u n e 1995 (Vol. 16)

T
possibly explain some tissue-specific chronic inflammatory processes.

Competition for the same transcription factor

With respect to the gene regulation of IL-6, it was
shown recently that nuclear factor KB (NFKB) is a target
for the glucocorticoid type II receptor as well as the oestrogen receptor36,37. Earlier studies suggested competition
between steroid receptors for nuclear factors38.At present
it is not known whether mechanisms such as competitive
inhibition or synergism between these two steroid hormones may play a role at the cellular level in gene regulation, for example, with cytokines.

Antagonismof glucocorticoid effects by cytokines

Several studies have shown that the effects of glucocorticoids can be antagonized by individual cytokines and, to
more closely resemble in vivo situation, combinations of
cytokines. Thus, a single injection of transforming growth
factor-[3 (TGE-{3) is able to counteract glucocorticoidinduced impairment of wound healing39. Inhibition of
T-cell proliferation by glucocorticoids is also inhibited by
IL-1, IL-6 and interferon-~/(IFN-y)4;when administered in
combination, these agents act synergistically. IFN-y also
antagonizes the glucocorticoid-mediated suppression of
IL-1[3synthesis in human monocytes41.In addition, a combination of IL-2 and IL-4 may influence the binding affinity of glucocorticoids for type II receptors42.

Augmentation of immune responses by

glucocorticoids
Glucocorticoids not only exert suppressive, but also permissive actions during immune responses43.These include
the upregulation of IL-6 receptors and their signal transduction component gp130 (Ref. 44), upregulation of IL-1
receptors, which may have anti-inflammatory potential45.
Furthermore, glucocorticoids have been considered a
necessary prerequisite in the induction of the acute phase
response 4~. The production of a variety of proteins and
proteases during an inflammatory response, or following
tissue damage in general, is considered part of a highly
complex defence mechanism to re-establish homeostasis.
Full initiation of the acute phase response requires synergism between IL-6-1ike cytokines and glucocorticoids for
the induction of several but not all, of the so called 'acute
phase proteins'.
These findings suggest that the physiological role of
glucocorticoids is not merely immunosuppressive, but
rather immunomodulatory. Their anti-inflammatory
properties arise from their role in the coordination of the
immune response.

Concluding remarks
The effects of glucocorticoids are determined by highly
spedfic regulatory mechanisms, which may also display
tissue specifldty. Further, the effects will differ depending
on whether an endogenous or synthetic ligand is being
studied. Thus, the physiological role of glucocorticoids in

V
relation to immune reactions on the basis of in vitro as well
as in vivo data cannot be solely based on the use of synthetic
ligands without comparison with the endogenous corticosteroid. Therefore, studies on glucocorticoid effects in tissues obtained from patients may need repetition to compare endogenous versus synthetic ligands in terms of
efficacy, use of physiological concentrations, and exposure
duration. The latter will be difficult, since for various synthetic glucocorticoid preparations, the pharmocokinetics
and pharmacodynamics are not yet fully characterized.
Estimation of tissue levels by the measurement of bioactive
compounds and metabolites in saliva may contribute to
definition of more accurate criteria than the one used at
present, namely 'dose-equivalence', a criterion base mainly
on the development of side-effects rather than their therapeutic efficacy 47. A necessary step would involve reevaluation of in vitro and in vivo results by the use of spedfic
type I or type II receptor agonists and antagonists. In vitro,
it will be of importance to choose an appropriate control,
as withdrawal following synthetic ligands may lead to
hyperresponsiveness, as demonstrated with IL-113(Ref. 48).
Future directions
Disturbances of the mechanisms described on a systemic
or localized basis bear a great potential for immunological
dysfunction. Thus, it is likely that many immunological diseases that are accompanied by inflammation may involve
a component related to the actions of glucocorticoids. For
example, overexpression of a tissue-specific subtype of the
enzyme 1113HSD or type II13receptors may be the basis for
localized steroid resistance. Similar results would be
expected by alterations in the expression levels of type II or
type I receptors (or both) in individual tissues.
Characterization of the events and their underlying
mechanisms that occur at different physiological glucocorticoid levels, and the regulation of these levels during
immune reactions, may help to understand the development of disease and generate new ideas for pharmacological targeting of neuroendocrine-immune functions.
Further, the role of glucocorticoids needs to be characterized in disease states that involve developmental changes
in neuroendocrine regulation, such as juvenile chronic
arthritis. In the clinic the combination of physiological concentrations of glucocorticoids with other agents, such as
cytokines, could potentially enhance endogenous defence
mechanisms. Synergism of IL-4 and glucocorticoids, for
example, in downregulation of IL-1 levels and upregulation of IL-1 type II receptors combines anti-inflammatory actions while counteracting the cellular autoimmune
response; this could be a feasible intervention in the treatment of rheumatoid arthritis45,49.
Finally, now that the interactive communication
between the endocrine and the immune system is generally accepted, the central role of gtucocorticoids in this network requires that the mechanisms which determine
glucocorticoid efficacy in immunomodulation need to be
further characterized. This will require close interdisciplinary collaborations.

Selected references
1 Chrousos, G. P. (1995) New Engl. J. Med. 332, 1351-1362
2 Wick, G., Hu, Y., Schwarz, S. and Kroemer, G. (1993) Endocr. Rev. 14,
539-563
3 Bames, P. J. and Adcock, I. (1993) Trends Pharmacol. Sci, 14, 436-441
4 Waage, A., Slupphaug, G. and Shalaby, R. (1990) Eur. ]. Immunol. 20,
2439-2443
5 Arzt, E. et al. (1994) Endocrinology 134, 672-677
6 Santos, A. A. et al. (1993) Arch. Surg. 128, 138-144
7 Wilckens, T. et al. (1994) Intensive Care Med. 20 (Suppl. 58), 225
8 0 r t h , D. N., Kovacs, W. J. and Debold, C. R. (1992) in Williams Textbook of Endocrinology (Wilson, J. D. and Forster, D. W., eds),

pp. 489-620, Saunders

9 Mendel, C. M. (1989) Endocr. Rev. 10, 232-274
10 Lent~es, E. G. et al. (1993) Clin. Chem. 39, 2518-2521
11 Kirschbaum, C. and Hellhammer, D. H. (1994) Psychoneuroendocrinology 19, 313-333
12 Knisely, T. L., Hosoi, J., Nazareno, R. and Granstein, R. D. (1994)
Invest. Ophthalmol. Visual Sci. 35, 3711-3723
13 Hammond, G. L., Smith, C. L., Paterson, N. A. and Sibbald, W. J.
(1990) J. Clin. Endocrinol. Metab. 71, 34-39
14 Bartalena, L., Hammond, G. L., Farsetti, A., Flink, I. L. and Robbins, J.
(1993) En&crinology 133, 291-296
15 Denton, R. R., Eisen, L. P., Elsasser, M. S. and Harmon, J. M. (1993)
Endocrinology133, 248--256
16 Dallman, M. F. (1993) Trends Endocrinol. Metab. 4, 62-69
17 Miller, A. H. et al. (1993) Endocrinology 133, 2133-2140
18 Spencer, R. L., Miller, A. H., Moday, H., Stein, M. and McEwen,
B. S. (1993) Endocrinology 133, 1941-1950
19 Funder, J. W. (1993) Science 259, 1132-1133
20 Wiegers, G. J., Croiset, G., Reul, J. M. H. M., Holsboer, F. and
de Kloet, E. R. (1993) Am. I. Physiol. 265, E825-E830
21 Wiegers, G. J., Reul, J. M. H. M., Holsboer, F. and de Kloet, E. R. (1994)
Endocrinology135, 2351-2357
22 Trapp, T., Rupprecht, R., Castren, M., Reul, J. M. H. M and
Holsboer, F. (1994)Neuron 13, 1457-1462
23 Bamberger, C. M., Bamberger, A. M., De-Castro, M. and Chrousos,
G. P. I. Clin. Invest. (in press)
24 Funder, J. W. (1994) Steroids 59, 164-169
25 Frey, F. J., Escher, G. and Frey, B. M. (1994) Steroids 59, 74-79
26 Escher, G., Frey, F. J. and Frey, B. M. (1994) Endocrinohrgy 135,
101-106
27 Whorwood, C. B., Ricketts, M. L. and Stewart, P. M. (19~)4)
Endocrinology135, 2533-2541
28 Hennebold, J. D. and Daynes, R. A. (1994) Endocrinology 135, 67-75
29 Mezey, E. and Palkovits, M. (1992) Science 258, 1662-1665
30 Klein, A., Buskila, D., Gladman, D., Bruser, B. and Malkin, A. (19q0)
]. Rheumatol. 17, 30-33
31 Teelucksingh, S. et al. (1990) Lancet 335, 1060-1063
32 Vamvakopoulos, N. O. (1993) Mol. Cell. Endocrinol. 98, 49-54
33 Faulds, G. B., [senberg, D. A. and Latchman, D. S. (lq94) I. Rheuma
tol. 21, 234-238
34 Kang, K. I. et al. (1994) Proc. Natl Acad. Sci. USA 91,340~344
35 Conroy, S. E., Faulds, G. B., Williams, W., Latchman, D. S. and
Isenberg, D. A. (1994) Br. J. Rheumatol. 33, 9Z'k-926
36 Ray, A. and Prefontaine, K. E. (1994) Proc. Natl Acad Sci. LISA 91,
752-756
3?' Ray, A., Prefontaine, K. E. and Ray, P. (1994) ]. Biol. Chem. 269,
12940-12946
38 Meyer, M. E. et al. (1989) Cell 57, 433-442
39 Beck, L. S. et al. (1993)J. Clin. Invest. 92, 2841-2849
40 Almawi, W. Y. et al. (1991) J. Immunol. 146, 3523-3527
41 Geiger, T., Arnold, J., Rordorf, C., Henn, R. and Vosbeck, K. (1993)
Lymphokine Cytokine Res. 12, 271-278
42 Kam, J. C., Szefler, S. J., Surs, W., Sher, E. R. and Leung, D. Y. M
(1993) ]. hnmunol. 151, 3460-3466
43 Munck, A. and Naray-Fejes-Toth, A. (1992) Mol. Cell. Endocrinol. ~)0,
C1-C4
44 Schooltink, H., Schmitz-Van-de-Leur, H., Heinrich, P. C. and RoseJohn, S. (1992) FEBS lx'tt. 297, 263-265
45 Colotta, F. and Mantovani, A. (1994) Trends Pharmacol. Sci. [5,
138-139
46 Baumann, H. and Gauldie, J. (1994) Immunol. Today 15, ?'4-80
47 Barth, J., Moellman, H. W., Wagner, T., Hochhaus, G. and
Derendorf, H. (1994) Deutsch. Med. Wschr. 119, 1671-1676
48 Baybutt, H. N. and Holsboer, F. (1990) Endocrinoloyy 127, 476-480
49 Miossec, P. (1993) Eur. Cytokine Net. 4, 245-251

TIPS

June 1995 (Vol. 16)

Acknowledgements
The author gratefully
acknowledges the
training and support of hqs
teachers, namely
S Reichlin, M f Dallman
and C A Dinarello The
author thanks his
co{leagues at the IHf- for
theh critical comments
and discussions of this
article The author thanks
S Rupprecht for her
continued support

1 9 7