Physiology of psychoneuroimmunology: A personal view

Abstract

This article offers a personal view on how the concept of the existence of a network of immune-neuro-endocrine interactions has evolved in the last 30 years. The main topic addressed is the relevance of the exchange of signals between the immune, endocrine and nervous systems for immunoregulation and brain functions. Particular emphasis is given to circuits involving immune cell products, the hypothalamus-pituitary-adrenal axis and the sympathetic nervous system. The operation of these circuits can affect immune functions and the course of inflammatory, autoimmune and infectious diseases. We also discuss increasing evidence that brain-born cytokines play an important role in brain physiology and in the integration of the immune-neuro-endocrine network.

Full text

Named Series: Twenty Years of Brain, Behavior, and Immunity
Physiology of psychoneuroimmunology: A personal view
Hugo O. Besedovsky
*
, Adriana del Rey
Department of Immunophysiology, Institute of Physiology and Pathophysiology, Deutschhausstrasse 2, 35037 Marburg, Germany
Received 4 August 2006; received in revised form 20 September 2006; accepted 22 September 2006
Abstract
This article offers a personal view on how the concept of the existence of a network of immune–neuro–endocrine interactions has
evolved in the last 30 years. The main topic addressed is the relevance of the exchange of signals between the immune, endocrine and
nervous systems for immunoregulation and brain functions. Particular emphasis is given to circuits involving immune cell products,
the hypothalamus–pituitary–adrenal axis and the sympathetic nervous system. The operation of these circuits can affect immune func-
tions and the course of inflammatory, autoimmune and infectious diseases. We also discuss increasing evidence that brain-born cytokines
play an important role in brain physiology and in the integration of the immune–neuro–endocrine network.
Ó2006 Published by Elsevier Inc.
Keywords: Immunophysiology; Immune–neuro–endocrine interactions
1. Introduction: general aspects and definitions
Not long ago (at least for the time-scale in science),
when we wanted to introduce an article on interactions
between the immune, the endocrine and the nervous sys-
tems, we referred to an ‘‘emerging’’ field. This type of intro-
duction is no longer justified since this area of research is
now firmly established. The appearance of BBI 20 years
ago was a great pillar for its establishment. Before, it was
difficult to convince reviewers of journals that focused on
a given specialty that boundaries between disciplines are
relative and some times self-imposed.
We believe that this special series of BBI is important
because it provides an historical perspective that may serve
to influence future developments and trends. Regarding
our contribution to this issue, we want to clarify that, in
our view, we have been asked to cover the probably broad-
est aspect of the field; thus, our contribution is far from
being an attempt to review the progresses in the ‘‘Physiol-
ogy of Psychoneuroimmunology’’. We shall emphasize
some aspects of Immunophysiology such as active interac-
tions between the immune system (IS) and the hypothala-
mus–pituitary–adrenal (HPA) axis or the sympathetic
nervous system (SNS), and the role of peripheral and cen-
tral cytokines as mediators of these interactions. Interac-
tions with other endocrine and autonomic systems are
not or only marginally mentioned. Aspects such as stress,
sleep and behavior are considered in other articles of this
series. We have written part of this article as a sort of a
‘‘scientific autobiography’’ since, as ‘‘old timers’’, we
believe that to share our own experience could be of inter-
est for the younger generation that will certainly provide
the future breakthroughs that the field needs. We apologize
because, due to space limitations, the work of many col-
leagues is not quoted and not even mentioned. With some
exceptions, references from before 1987, and also many
between 1987 and 1996, are omitted. These references can
be found in earlier reviews (Besedovsky and del Rey,
1996; Besedovsky and Sorkin, 1977).
We also want to clarify an aspect that makes the ‘‘es-
sence’’ of Physiology. There is no doubt that the IS is a
physiologic homeostatic system that, within certain limits,
contributes to the constancy and integrity of the organism
(preservation of self and neutralization of danger). Howev-
er, a source of confusion could be that while immune
responses are physiologic responses expected to be
0889-1591/$ - see front matter Ó2006 Published by Elsevier Inc.
doi:10.1016/j.bbi.2006.09.008
*
Corresponding author. Fax: +4964212868925.
E-mail address: besedovs@mailer.uni-marburg.de (H.O. Besedovsky).
www.elsevier.com/locate/ybrbi
Brain, Behavior, and Immunity 21 (2007) 34–44
BRAIN,
BEHAVIOR,
and IMMUNITY

maximally efficient during infectious/inflammatory diseas-
es, they sometimes contribute to pathology. Thus, although
the control and regulation of the different systems of an
organism is at the core of Physiology, the unique condition
in the case of immunoregulation is that such a physiologic
process operates simultaneously and interwoven with path-
ological events. Another point to remark is that ‘‘Physiol-
ogy’’ (as indicated by its etymology, derived from the
Greek Physis: nature), is a discipline that deals with natural
processes. Thus, for example, any finding of the effect of a
hormone on a given immune parameter using a pharmaco-
logical approach should be followed by studies of whether
such effect is also observed under natural, physiological or
pathophysiological, conditions.
2. Before 1987: the ‘‘old times’’
Most epistemologists agree that the acquisition of scien-
tific knowledge is preceded and nourished by deductive
intuitive views. The work on immune–neuro–endocrine
interactions 35 years ago was largely based on intuition
since it was based on the belief that the IS, as other phys-
iologic systems, should also be subject to the integrative
control of neuro–endocrine mechanisms. In fact, the data
available was relatively scarce and based, for example, on
the effect of manipulating certain brain areas and endocrine
mechanisms on some immune processes that, in many
cases, are not adaptive, such as acute hypersensitivity.
Receptors for only few hormones, like insulin, or neuro-
transmitters, like adrenergic agents, were identified or sus-
pected on immune cells. There was also some evidence
showing interactions between neuro–endocrine and
immune mechanisms during ontogeny, as it is the case of
the effect of the thymus on the maturation of sexual func-
tions. Nevertheless, the information available was just
enough for the initial formulation of the hypothesis that
immune responses are subject to a level of neuro–endocrine
regulation. As with other physiologic regulations, neuro–
endocrine immunoregulatory mechanisms must be based
on the operation of information channels between immune
cells and the nervous and endocrine systems. Because both
innate and adaptive immune responses involve different
cells and mediators at different stages, their extrinsic regu-
lation should be based on well-synchronized neural and
endocrine changes. Such changes should, in turn, be capa-
ble of modifying the activity of immune cells at defined step
of the immune response.
To approach the above mentioned hypothesis experi-
mentally, it was necessary to show that: (1) the neuro–en-
docrine changes that occur during the immune response
are not a consequence of the disease or of the stress of
being sick; and (2) the immunologically induced neuro–en-
docrine responses can affect the functioning of the IS. Fol-
lowing these criteria, it was possible to demonstrate that
glucocorticoid blood levels are increased in a threshold
dependent manner during the course of specific immune
responses to innocuous antigens. We also showed that
the increase in corticosterone levels during the response
to an antigen can interfere with the response to a second,
unrelated antigen. This finding, which provided an expla-
nation to the phenomenon of antigenic competition, indi-
cated that the endocrine change observed was relevant
for immunoregulation. Interestingly, while trying to induce
antigenic competition in vitro, we also observed that low
doses of glucocorticoids could increase the number of cells
producing specific antibodies. The question that arose was
how two distinct bodily systems, the immune and endo-
crine systems, can exchange information. To approach this
question, we stimulated immune cells with mitogens or
antigens in vitro and found that cell-free supernatants from
these cultures contained factors capable to stimulate the
HPA axis (Besedovsky et al., 1981), an effect that was med-
iated by the pituitary gland. We denominated this material
‘‘glucocorticoid-increasing factor’’ (GIF). Fortunately, we
invested limited efforts to try to purify this factor, since,
as it will be mentioned later, several cytokines share the
capacity to stimulate the HPA axis. This finding led to
the proposal of the operation of an immune-HPA axis
immunoregulatory circuit. Close to the time when these
studies were published, Smith and colleagues reported that
glucocorticoids interfere with the production of the T cell
growth factor (as IL-2 was termed at that time) and of
other immune products. It was also shown that changes
in endogenous levels of glucocorticoids can inhibit immu-
noglobulin production. Furthermore, it was found that
resting immune cells are more sensitive to the inhibitory
effect of glucocorticoids than activated cells. On these
bases, we postulated that the immune–HPA axis circuit
may have the function of preventing the excessive expan-
sion of cells with low affinity for the antigen and of those
cells that are recruited under the polyclonal influence of
lymphokines. In this way, the specificity of the immune
response would be preserved and even improved. It is con-
ceivable that this circuit, by impeding a cumulative exces-
sive expansion of lymphoid and accessory cells, plays a
role in preventing autoimmune and lymphoproliferative
diseases. Experimental and clinical examples showing the
relevance of the HPA axis–immune circuit are mentioned
later in this article. The proposed immune–HPA axis cir-
cuit that controls an over production of immune products
is in line with the concept formulated by A. Munck (who
was the first to characterize the glucocorticoid receptor)
that the essential and general function of glucocorticoids
is to control the overshoot of locally produced mediators.
Around 1986, some cytokines became available in pure
and recombinant form, making it possible to test their
capacity to induce neuro–endocrine responses. IL-1 was
the first cytokine that was shown to activate the HPA axis
(Besedovsky et al., 1986). The described interconnection
involving peripheral immune mechanisms and endocrine
responses under brain control led to the view of ‘‘long-
loop’’ immune–neuro–endocrine circuits (Fig. 1, blue lines)
and was subject of intense investigation during the follow-
ing decades.
H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44 35

In many conditions, the SNS operates in a coordinated
fashion with the HPA axis. It seemed, therefore, logic to
study whether the SNS also participates in immunoregula-
tion. On this basis, we explored whether noradrenaline
(NA) concentration in lymphoid organs is affected during
the immune response to innocuous antigens. The results
showed that a decrease in the splenic concentration, con-
tent, and turnover of NA, the main sympathetic neuro-
transmitter, precedes the peak of the immune response to
innocuous antigens. We interpreted such neural response
as a way to relieve immune cells from a tonic inhibitory
effect of NA. However, this interpretation has to be taken
with caution under the light of later findings (see below).
Although the activity of the SNS is integrated centrally,
the pioneer work of Felten and Bulloch showing that
immune cells and noradrenergic nerve fibers are in close
contact in lymphoid organs led us to propose the existence
of short-loop immunoregulatory neural circuits (Fig. 1,
green lines).
During the eighties, we also started studying the possi-
bility that the presence of neoplastic cells, either when they
are transplanted, de novo induced or developing spontane-
ously, could induce endocrine responses. Clear endocrine
alterations, which occurred before the tumor became pal-
pable and without any overt sign of disease, were noticed
during tumor growth. The alterations involved not only
steroid but also pituitary hormones such as prolactin and
growth hormone, which, as shown by the groups of Berczi
and Kelly, are known to support immune responses.
The hypothalamus controls numerous pituitary and
autonomic functions. On this basis, we searched for evi-
dence that hypothalamic neurons could change their activ-
ity during a peripheral immune response. Our studies
coincided in time with the evidence provided by Ader
and Cohen that immunosuppression can be behaviorally
conditioned (Ader and Cohen, 1975). This finding provided
an irrefutable proof that an intact brain can affect immune
functions and further rationale to the search for bi-direc-
tional channels of information between the IS and the
brain.
We evaluated in the same animal both the immune
response and the rate of firing of individual neurons in sev-
eral hypothalamic nuclei and at various intervals after
injection of innocuous antigens. A clear increased in the
rate of firing of neurons of the ventromedial hypothalamic
nucleus was detected in parallel to the immune response to
Psycho-sensorial
Stimuli Behaviour
Antigenic Stimuli Immune response
CNS
Immune cell products
Endocrine
System
Peripheral
Nervous
System
immune
neuro
endocrine
products
Immune System
immune
neuro
endocrine
products
long-loop interactions
peripheral interactions
central interactions
homeostatic adjustments
Fig. 1. The immune–neuro–endocrine network. The immune, endocrine, and nervous systems can interact at multiple levels. Schematically, these
interactions can be classified as long (blue lines) and short (green lines) loops. There are also interactions within the brain (red circle). The branching
arrows in brown represent other homeostatic consequences of these interactions. The operation of these interactions can be modified by stimuli acting
primarily on the immune system or on the brain, for example psychosocial stimuli. The outcome of immune–neuro–endocrine interactions can result in
immune and/or behavioral modulation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
paper.)
36 H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44

these antigens. These results provided direct evidence that
the brain receives information from the IS under non-
stressful conditions in rats under narcosis, using two differ-
ent non-infectious antigens and a conventional methodolo-
gy to evaluate neuronal activity. It was also found that NA
turnover rate in hypothalamic neurons was inhibited dur-
ing the immune response to sheep erythrocytes, an effect
that is likely mediated by products released by immune
cells (Besedovsky et al., 1983).
The evidence showing that the immune response can, by
itself, elicit neuro–endocrine responses and that products
derived from immune cells can mediate these responses
led us to propose that the IS acts as a peripheral receptor
organ able to transmit information to the brain about
responses to external or internal antigenic stimuli. At this
time, Blalock and Smith discovered that immune cells
can produce ‘‘pituitary hormones’’ (Blalock et al., 1985).
Also since brain cells can produce cytokines originally
described as immune products, they stressed that there
was a common usage of ligands and receptors of immune
and neuro–endocrine mediators, and called the IS ‘‘the
sixth sense’’.
This period ended with a very good input to the field: a
plenary symposium on immune–neuro–endocrine interac-
tions was included for the first time at an international con-
gress of Immunology (6th International Congress of
Immunology, July 1986, Canada). Although this Sympo-
sium was held in parallel to another one dealing with the
‘‘hot’’ immunological finding of the MHC restriction, a
large audience enthusiastically attended ‘‘ours’’.
3. 1987–1996
An almost explosive increase in the number of scientists
committed to research in immune–neuro–endocrine inter-
actions occurred during the decade 1987–1996. BBI was
launched not only as consequence of the broad interest in
this field but also contributed to its growth.
The understanding of the immune–HPA axis circuit was
deepened by studies showing that many cytokines share the
capacity to stimulate this axis. IL-1, IL-2, IL-3, IL-6, IL-8,
IL-11, IL-12, TNF, INFc, and GM-CSF are among the
cytokines that can integrate such glucocorticoid-mediated
feedback. The site of action of the cytokines was a matter
of discussion. It is now clear that the acute effect of cyto-
kines is basically exerted at the level of the hypothalamus
via the release of CRH and to some extend also AVP.
However, during prolonged situations effects at pituitary
and adrenal levels were also observed (for review Besedov-
sky and del Rey, 1996; Turnbull and Rivier, 1999). The
understanding of the relevance of the HPA axis–immune
feed back circuit during disease begun during the period
1987–1996 with three remarkable findings. It was reported
by different groups that: (1) the stimulation of the HPA
axis induced by LPS is to large extent mediated by cyto-
kines. Although it was known that LPS administration
results in increased glucocorticoid levels, this effect was
considered a consequence of the septic shock induced by
the endotoxin; (2) the immune–HPA axis loop is altered
in chickens that develop autoimmune thyroiditis spontane-
ously (Schauenstein et al., 1987); and (3) Lewis rats, which
are prone to inflammatory diseases such as rheumatoid
arthritis, have a deficient corticosterone response to IL-1
(Sternberg et al., 1989). We shall come back to this issue
in the next section. It should also be mentioned that it
was found that, besides the HPA axis, other endocrine sys-
tems are affected during the immune response (for review
Besedovsky and del Rey, 1996).
The cellular levels at which the SNS can exert regulatory
actions were further characterized during this period (for
review Heijnen and Kavelaars, 1999). The main cells target
of noradrenergic neurotransmitters appear to be immature
and mature thymocytes, thymic epithelial cells, T lympho-
cytes, macrophages, mast cells, plasma cells, and entero-
chromaffin cells. Evidence was found that NA and
adrenaline, by stimulating the b
2
-adrenoreceptor-cAMP-
protein kinase A pathway, inhibit the production of type
1/proinflammatory cytokines, e.g. IL-12, TNFaand IFNc.
Stimulation of b-adrenergic receptors inhibits antigen-pre-
senting cells and Th1 cells, but stimulate the production of
type 2/anti-inflammatory cytokines such as IL-10 and
TGFb. Thus, endogenous catecholamines may cause a
selective suppression of Th1-mediated inflammation and
cellular immunity, favoring in this way humoral immunity
and also protect the host from the detrimental effects of
proinflammatory cytokines and other products of activated
macrophages (for review and references Elenkov et al.,
2000; Sanders, 2006). Synergistic effects of glucocorticoids
and catecholamines were found (Elenkov et al., 2000).
There were some controversies about the effect of surgi-
cal sympathetic denervation and chemical depletion of NA
stores on immune responses. We are personally biased in
this matter because we found that both procedures result
in an increased immune response to sheep erythrocytes,
an effect that agrees with other reports using different anti-
gens. However, studies using 6-hydroxy-dopamine to
deplete catecholamines showed, in some cases, immuno-
suppressive effects when tested shortly after the administra-
tion of the neurotoxin. Our interpretation is that the acute
NA release caused by nerve terminal damage and the
marked increase in glucocorticoid levels induced by this
procedure exert an inhibitory effect on immune cells. It is,
however, clear to us that denervation is a very extreme pro-
cedure that, although reveling tonic effects of sympathetic
neurotransmitters on immune cell activation, differentia-
tion and survival, does not allow definitive conclusions
regarding immune responses that are phasic, dynamic,
and stage-depending processes. For more information on
sympathetic immunoregulation obtained during 1987–
1997 see reviews (Elenkov et al., 2000; Sanders, 2006).
Although cholinergic innervation was found in the thymus
but not in peripheral lymphoid organs such as the spleen,
effects of parasympathetic mediators on immune processes
were reported. For example, Rinner and Schauenstein
H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44 37

observed an inhibitory effect of the cholinergic agonist car-
bachol on an in vitro induced immune response to SRBC.
Peptidergic innervation was also found to integrate
potential immunoregulatory short-loop circuits. Most lym-
phoid organs also receive sensory peptidergic innervation
that is mainly confined to the parenchyma, and the most
abundant peptides found are tachykinins (substance P,
neurokinin A), calcitonin gene-related peptide (CGRP),
and vasoactive intestinal polypeptide/peptide histidine iso-
leucine (VIP/PHI). In addition, evidence was found that
some of these peptides co-exist with catecholamines, and
that there is a close spatial relationship between peptidergic
nerve fibers and mast cells, T cells and macrophages. Pep-
tidergic nerves also appear to be sparse in pure B cell
regions. Neuron-mast cell contacts are relatively often seen
in all lymphoid organs, with the exception of the spleen
(Weihe et al., 1991).
Although some evidence was provided earlier, it became
well established during 1987–1997 that immune cytokines
influence complex mechanisms that involve a variety of
neuronal circuits such as thermoregulation, food intake,
sleeping patterns and behavior. These effects will not be
discussed here and references can be found in specific
reviews. We shall only mention a few examples. It became
clear that there is not a single endogenous pyrogen since
several endogenous substances, such as IL-1, IL-6, IL-8,
IFNc, IFNband GM-CSF can induce fever. Also several
cytokines can inhibit food intake, among them are IL-1,
IL-6, IL-8 and TNFa. The capacity to increase slow wave
sleep is also shared by different cytokines such as IL-1, IL-
2, IFNc, and TNFa, an effect that is moderated by IL-4,
IL-10 and IL-13. Several cytokines are known to exert pro-
found effects on behavior, e.g. learning and explorative and
avoidance behavior. Some of these actions are likely to
occur at CNS levels and mediated by IL-1 since i.c.v.
administration of IL-1ra blocks such effects (Kent et al.,
1992). Many of the mechanisms integrated at brain levels
are under control of catecholaminergic and serotonergic
brain neurons. Simultaneously with A. Dunn, we reported
that IL-1 stimulates NA turnover rate in the brain. In the
case of noradrenergic neurons the effect was not only
restricted to the brain but was also detected at low spinal
cord levels, suggesting a neural pathway for IL-1 effects
in the CNS. There was also evidence of the relevance of
the stimulation of central noradrenergic neurons for the
effect of IL-1 on the HPA axis and fever (for review Bese-
dovsky and del Rey, 1996; Turnbull and Rivier, 1999).
An aspect that raised great interest during this period
was to understand how immune signals can reach the brain
and affect brain functions. The evidence showed that
humoral and neural pathways are involved. Humoral path-
ways can convey information to the brain either via cir-
cumventricular organs or via the endothelial/ glial
interphase in brain vessels. The groups of Dantzer, Maier
and Watkins, and Blatteis showed the existence of a neural
pathway by demonstrating the important role of vagal
afferences in the transmission of cytokine signals to the
brain (for review Dantzer, 2004; Maier et al., 1998). In gen-
eral, the data available indicated that the humoral route
may be followed during immune processes that result in
high levels of cytokines in the circulation while the neural
route seems to predominate when these mediators are
released locally in tissues with vagal innervation. In addi-
tion, both routes could operate simultaneously or sequen-
tially during immune responses. Towards the end of the
period 1987–1996, an important event for the field of
brain-immune interactions was the demonstration original-
ly derived from the laboratories of Bartfai and Dantzer
that activation of peripheral immune cells by LPS induces
cytokine production in the brain. Initially, we had some
concerns about these results because the dose of LPS used
can disrupt the blood–brain barrier or cause brain altera-
tions as consequence of the endotoxic shock. However,
when the studies were done using a dose of LPS that does
not affect the blood–brain barrier and does not cause overt
symptoms of sickness behavior, we also detected cytokine
gene expression in the brain (Pitossi et al., 1997), thus con-
firming and extending previous work. The expression of the
genes for IL-1b, IL-6, TNFa, and IFNcis increased follow-
ing peripheral administration of LPS. The onset of tran-
scription and the peak of mRNA accumulation depend
on the cytokine and the brain region studied. IL-1band
IL-6 expression is preferentially increased in the hypothal-
amus and hippocampus, while TNFaexpression is more
marked in the thalamus–striatum. In the conditions men-
tioned, these cytokines are less inducible in the brain cor-
tex. No correlation between cytokine gene expression and
the density of vascular structures in a given brain area
was detected, neither was a preferential cytokine expression
observed in brain areas that include the circumventricular
organs.
In our view, the neuro–endocrine responses triggered by
activated immune cells involve powerful hormones and
neurotransmitters that can affect not only immune mecha-
nisms but that also contribute to homeostatic adjustments
necessary during diseases in which the IS is activated. In
the following, we shall refer to the effect of cytokines in
controlling blood flow in lymphoid organs and in setting
the conditions for adequate provision of energy to immune
cells.
A main physiologic function of the SNS is the control of
blood pressure and tissue blood flow. This process is rele-
vant for the effectiveness of the immune response. Indeed,
immune cells need to circulate to reach the sites where
inflammatory and infectious processes occur. The coupled
blood and lymphatic circulatory systems provide the routes
for immune cell circulation while adhesion molecules, che-
mokines, integrins and other locally produced mediators
control their homing and mobilization. In addition, the
spleen is inserted in the circulatory system in a way that
favors the clearance, uptake and retention of microorgan-
isms and their contact with immune cells. In rodents, this
organ is deprived from lymphatic circulation and the
control of blood flow is strictly based on the sympathetic
38 H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44

vascular tonus. As mentioned, we found that a decrease in
sympathetic activity precedes the peak of a specific immune
response. Paradoxically, it was shown that administration
of IL-1 and LPS can increase SNS activity. This effect
was not compatible with the increase in blood flow and
accumulation that we have detected in lymphoid organs
during immune process, which is likely based on a decrease
in the sympathetic control of the vascular tonus. There is
however an explanation for these seemingly contradictory
findings. IL-1 produced in lymphoid organs selectively
increases blood flow by interfering locally with the sympa-
thetic tonus, causing a redistribution of blood flow. While
stimulation of the splenic nerve in vivo decreases splenic
blood flow, this effect is completely abrogated in animals
treated with LPS or IL-1 as consequence of a postjunction-
al inhibition of NA release. Studies in vitro showed that IL-
6 and TNF exert a comparable effect on NA release by
sympathetic nerve fibers. These data reinforce the previous-
ly discussed evidence that activated immune cells are less
exposed to the effects of noradrenergic nerves and suggest
a mechanism by which pro-inflammatory cytokines could
deviate blood supply to the sites where immune processes
take place (Rogausch et al., 1997). This redistribution of
blood supply would favor the contact between immune
cells and antigens, a process that is essential for an efficient
immune response.
Glucose is the main source of energy for the brain and
for most peripheral tissues, including immune cells. In
addition, essential immune processes such as endocytosis,
phagocytosis, increased cell turnover, clonal expansion,
production of numerous mediators and generation of effec-
tor cells and molecules are very expensive in terms of ener-
gy. We have shown that low, sub-pyrogenic doses of IL-1
induce a profound, long lasting, insulin-independent hypo-
glycemia in mice (del Rey and Besedovsky, 1989; del Rey
and Besedovsky, 1992). This effect, which is also observed
in insulin-resistant animals, develops in mice against
increased levels of counter regulatory hormones such as
catecholamines, glucocorticoids, and glucagon. There is
also evidence that the hypoglycemic effect of IL-1 can be
triggered at central levels since intracerebroventricular
administration of the cytokine induces a reduction in
glucose blood levels (del Rey et al., 1998). However, the
most surprising effect is observed when mice and rats are
challenged with a glucose load several hours after a single
intraperitoneal injection of IL-1 (for review Besedovsky
and del Rey, 1996). In this situation, it is clearly seen that,
following a transient elevation of glucose levels in blood, its
concentration returns to the previously reduced levels and
IL-1-injected animals remain hypoglycemic for several
hours more. These findings strongly indicate that IL-1
changes the rigid set point that characterizes glucose
homeostasis. These changes are linked to the capacity of
IL-1 to induce glucose transport and oxidation, e.g. in
adipose cells and fibroblasts, and also to an effect on
glucoregulatory mechanism under brain control (del Rey
and Besedovsky, 1992). The combination of local and
central effects of IL-1 would serve to deviate glucose to
lymphoid organs and inflamed/infected tissues to satisfy
the high cost of energy of immune responses (Fig. 1,
branching arrows in brown).
4. 1997–2006
From 1997 on, the research in this field was mainly ori-
ented to search for the physiological and pathological rele-
vance of immune–neuro–endocrine interactions and to
better understand the molecular basis underlying these
interactions.
The importance of the immune–HPA axis regulatory
circuit in controlling inflammatory and autoimmune pro-
cesses was firmly established (Besedovsky and del Rey,
2006; Sternberg, 2006). Considerable experimental and
clinical evidence underscores the relevance of the cyto-
kine–HPA axis feedback circuit during autoimmune and
infectious diseases. For example, adrenalectomy or block-
ade of glucocorticoid receptors aggravates disease and
increases mortality in animal models of rheumatoid arthri-
tis, multiple sclerosis, Hashimoto’s thyroiditis and septic
shock. Decreased response of the HPA axis or reduced sen-
sitivity to glucocorticoids is also implicated in human
pathologies such as rheumatoid diseases, multiple sclerosis,
Sjogren’s syndrome, allergic asthma, athopic skin disease,
inflammatory bowel disease and fibromyalgia (for review
Besedovsky and del Rey, 2006; Sternberg, 2006). The
defects, detected in so many pathologies, are now explained
by a better knowledge of glucocorticoid actions that can
affect a plethora of inflammatory mediators. It has to be
added that many effects of glucocorticoids are modulated
by the adrenal androgen DHEA. In some cases, particular-
ly during chronic diseases, the cortisol/DHEA ratio reflects
the efficacy of the cytokine–HPA axis circuit. This circuit
and its relevance in pathophysiology are schematically rep-
resented in Fig. 2 as an example because, at present, this is
probably the best established neuro–endocrine immunoreg-
ulatory circuit.
A novel aspect linked to immunoregulation by norad-
renergic nerves derives from the finding that NA induces
apoptosis in lymphoid cells via stimulation of b-adrenergic
receptors. This finding led us to study whether pro-apopto-
tic effects of NA can affect immune responses that result in
apoptotic-mediated specific T cell deletion (e.g. the immune
response to superantigens) or during immune pathologies
in which Fas-Fas ligand-mediated apoptosis is defective.
Superantigens, like staphylococcal enterotoxin B (SEB),
induce a strong proliferative response followed by clonal
deletion of a substantial portion of defined VbT cells.
The remaining cells display in vitro anergy. We found that
the immune response to SEB was paralleled by biphasic
changes in the activity of the SNS. Furthermore, sympa-
thetic denervation resulted in decreased SEB-induced cell
proliferation and IL-2 production, and impeded the specif-
ic deletion of splenic CD4Vb4 cells observed in intact ani-
mals without affecting anergy. These studies indicated that
H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44 39

the proapoptotic effect of NA can be expressed in vivo and
is relevant during the immune response to superantigens.
When studying the mechanisms by which NA induces
apoptosis of lymphoid cells we found that this process is
independent of a functional Fas. Thus, we used lpr/lpr
mice, which lack functional Fas (CD95) expression and
are, therefore, deficient in a critical mechanism for the
maintenance of peripheral tolerance, to study whether nor-
adrenergic nerves can affect the expression of the lympho-
proliferative, autoimmune disease that they develop
spontaneously. Early in ontogeny, the concentration of
NA is significantly increased in the spleen of lpr/lpr mice
but splenic sympathetic innervation gradually declines as
the disease progresses. Furthermore, IgM blood levels
and splenic NA concentration inversely correlate when
the disease is overtly manifested. Neonatal sympathecto-
my, which experimentally advances the loss of sympathetic
denervation that occurs spontaneously during adulthood,
Brain
Adrenals
Glucocorticoids
inflammatory/
immune
responses
Hypophysis
ACTH
CRH
AVP
cytokines
-Vascular responses
-Production/action of
pro-inflammatory
cytokines and
chemokines
-APC
-Prostaglandins
-NO
-Complement fractions
-Adhesion molecules
-Integrins, selectins
-Vasoactive factors
-Anti-inflammatory
mediators:
Annexin 1
(lipocortin 1)
MAPK
Phosphatase 1
IL-10
SOCS-3
-Switch Th1/Th2
-Phagocytosis of
apoptotic cells
Anti-inflammatory/immunoregulatory responses
1
2
3
3
3
4
The cytokine-glucocorticoid feedback circuit
Brain
Adrenals
inflammatory/
immune
responses
ACTH
CRH
AVP
cytokines
1
2
3
3
4
Alterations in the cytokine-glucocorticoid feedback circuit
Animal models
-rheumatoid arthritis
-multiple sclerosis
-Hashimoto’s thyroiditis
-septic shock
-acute viral and bacterial infections
-chronic parasitic infections
Human pathologies
-rheumatoid diseases
-multiple sclerosis
-Sjogren’s syndrome
-allergic asthma
-athopic skin disease
-inflammatory bowel disease
-fibromyalgia
a) sensitivity
Glucocorticoids
b) sensitivity
A
B
Hypophysis 3
Fig. 2. The cytokine–HPA axis feed back circuit. This circuit is given as an example of immunoregulatory neuro–endocrine mechanisms. (A) Following
certain inflammatory and immune responses, cytokines, particularly pro-inflammatory cytokines, can stimulate the HPA axis at different levels. As
consequence, increased levels of glucocorticoids affect inflammatory and immune processes by down or up regulating several mechanisms or the
production of mediators. (B) Disruption of the cytokine–HPA axis feed back circuit can aggravate the course of certain diseases. Experimental and clinical
evidence indicates that the cytokine–HPA axis feed back circuit is altered in human pathologies and in animal models of certain diseases either because
there is a reduced response of the HPA axis to cytokines (a) or because of the development of glucocorticoid resistance (b). APC: antigen-presenting cells;
NO: nitric oxide; MAPK: mitogen-activated protein kinases; SOCS: suppressors of cytokine signaling.
40 H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44

results in a markedly increased concentration of IgM and
IgG2a in blood, accelerates the appearance of lymphade-
nopathy and shortens significantly the survival time of
lpr/lpr mice. These data show that, in addition to defects
in the Fas pathway, an altered sympathetic innervation in
lpr/lpr mice also contributes to the pathogenesis of the
autoimmune disease, and strongly support the hypothesis
that the SNS can modulate the expression of lymphoprolif-
erative diseases (del Rey et al., 2002; del Rey et al., 2006).
An overall appraisal of today’s view of sympathetic
immunoregulation can be summarized as follows. The evi-
dence available shows that almost all mechanisms involved
in an immune response can be affected by noradrenergic
neurotransmitters (for review Elenkov et al., 2000; Sanders,
2006). Indeed, NA can inhibit or stimulate an immune
response depending on the dose of agonist applied and
on the type of adrenergic receptor stimulated. The effect
of NA also depends on the type of stimulus that triggers
the immune response, the subset of cells affected, and, most
importantly, at which step of the response, lymphoid and/
or accessory cells are exposed to neurotransmitters. Among
the processes directly or indirectly affected by sympathetic
neurotransmitters are antigen presentation and the expres-
sion of costimulatory and adhesion molecules, lymphoid
cell activation, cytokine production, clonal expansion and
deletion, immunoglobulin production, and the generation
of cytotoxic cells. The fact that the simultaneous stimula-
tion of adrenergic receptors and the T cell receptor affects
common intracellular signaling pathways may explain
why adrenergic agonists can affect so many mechanisms
involved in an immune response.
The interest in understanding the relevance of the other
branch of the autonomic nervous system, the parasympa-
thetic system, increased in the last years. There are indica-
tions that vagal nerve efferents exert a protective role
during endotoxic shock. The evidence is based on the fact
that bilateral cervical vagotomy aggravates the decrease
in blood pressure caused by a lethal dose of LPS and that
stimulation of the efferent vagal fibers moderates this effect.
The conclusion that the vagus protects the host from the
endotoxic shock was also indicated by the fact that the
increase in cytokines such as TNFaoccurs in vagotomized
animals while stimulation of the efferent vagus reverses this
effect. Furthermore, using a model of carrageenan-medi-
ated inflammation, it was shown that activation of nicotin-
ic receptors either by vagus nerve stimulation or by
cholinergic agonists significantly inhibits the release of
pro-inflammatory cytokines and blocks leukocyte migra-
tion. In vitro experiments using human macrophages con-
firmed that acetylcholine and cholinergic agonists inhibit
the release of TNF, IL-1 and IL-18 in response to endotox-
in and that this effect is exerted at post-transcriptional lev-
els. The pharmacologic studies clearly indicate that
stimulation of nicotinic cholinergic receptors can be
involved in the control of excessive inflammatory responses
(for review Tracey, 2002). This and previous studies consti-
tute an important aspect of the research in immune–neuro–
endocrine interactions that will most likely show the rele-
vance of the parasympathetic nervous system for immuno-
regulation also during conditions less extreme than lethal
endotoxic shock.
The relevance of neuropeptide release during inflamma-
tory autoimmune processes has become more evident and
this topic has been subject of several reviews. Some neuro-
peptides, particularly SP, are clearly pro-inflammatory
while others, such as VIP and PACAP, are anti-inflamma-
tory. These effects have been found in different models of
disease, such as septic shock, rheumatoid arthritis and
MS (Delgado et al., 2004; Jessop, 2002). It still needs to
be clarified to what extent the stimulation of the HPA axis
that results from the administration of these neuropeptides
contributes to such effects (Nussdorfer and Malendowicz,
1998).
There is now considerable evidence that brain-born
cytokines can affect CNS mechanisms. Cytokines have
been shown to be ‘‘sleep factors’’ and to affect both non-
rapid-eye-movement and REM sleep. Brain levels of IL-1
and TNF correlate with sleep propensity; for example,
their levels increase after sleep deprivation. Furthermore,
immune neutralization of IL-1 or blockade of its receptors
in the brain affects slow wave sleep (Obal and Krueger,
2003), indicating that endogenous IL-1 and TNF are part
of a complex biochemical cascade regulating sleep. It was
also shown that IL-1 expression can be induced in the brain
during stress (for review Besedovsky and del Rey, 1996)
and that some of the symptoms of ‘‘sickness behavior’’
are integrated by cytokines produced in the brain. Sickness
behavior refers to a coordinated set of subjective, behavior-
al, and physiological changes that develop in sick individu-
als during the course of an acute infection. These changes
are caused by effects of IL-1 and other pro-inflammatory
cytokines on brain cellular targets. Indeed, interference
with the effects of these cytokines in the brain abolishes
the expression of certain symptoms of sickness behavior.
This evidence indicates the role of pro-inflammatory cyto-
kines in orchestrating sickness behavior during acute dis-
eases (Dantzer, 2004).
The possible role of cytokines produced in the CNS on
brain physiology and their contribution to the integration
of immune–brain interactions at central levels was
approached in the last years. An essential question was
whether an increase in neuronal activity in a ‘‘healthy’’
brain could affect the local production of cytokines. Direct
evidence derives from the demonstration that pre-synaptic
stimulation of defined neurons, as it happens during phys-
iologic conditions, can control the local production of
cytokines by glial cells and neurons. Long-term potentia-
tion (LTP) of synaptic activity in the hippocampus has
served as model to approach this issue. A clear increase
in IL-1bgene expression, triggered by glutaminergic neu-
rons via NMDA receptors, was observed in hippocampal
slices and in freely moving rats during the course of LTP
(Schneider et al., 1998). More recently, we have observed
that the IL-6 gene is also over expressed during in vivo
H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44 41

and in vitro LTP (Balschun et al., 2004). These data consti-
tute the first evidence that cytokine gene expression in the
brain can be triggered by a pre-synapticaly induced
increase in the activity of a discrete population of neurons.
We and others have studied to what extent cytokines
produced in the brain during LTP can affect synaptic plas-
ticity and performance. At this stage, it is necessary to dis-
tinguish between studies based on exogenous
administration of cytokines and those that focus on the
effects of cytokines endogenously produced by brain cells.
There is a vast literature showing that exogenous in vivo
and in vitro administration of cytokines can affect LTP
induction and synaptic plasticity (for references Besedov-
sky and del Rey, 1996). These studies, although important
from the pharmacological point of view, cannot reliably
reveal the effect of cytokines produced in the brain under
natural conditions. In fact, LTP is a complex phenomenon
that involves a number of receptors and mediators that
influence its inducibility, establishment, and maintenance
in different ways. In particular, the maintenance of LTP
is protein synthesis-dependent and involves the activation
of genes in a given sequence and the release of their prod-
ucts in certain quantity. Thus, it is almost impossible to
mimic the time-dependent effect of an endogenously pro-
duced cytokine by its exogenous administration. For exam-
ple, as it is discussed below, lL-1, a cytokine that inhibits
LTP when administered exogenously, contributes to the
consolidation and maintenance of this process when it is
produced endogenously.
Using the specific IL-1 receptor antagonist (IL-1ra), we
found that blockade of IL-1 receptors, both in vivo and in
hippocampal slices, results in the inhibition of LTP mainte-
nance. This effect is reversible and occurs only when the
antagonist is administered after LTP is triggered, that is
at a time when, according to the studies mentioned above,
increased IL-1 levels are expected. Studies in Type 1 IL-1
receptor knock out mice are in line with this finding (Avital
et al., 2003). We have recently found that, in contrast to the
supportive effect of IL-1, IL-6 contributes to the extinction
of a well-consolidated LTP (Balschun et al., 2004). Collec-
tively, these results strongly suggest that IL-1band IL-6
can control the maintenance of LTP in the brain, a process
that is assigned a role in memory formation and in certain
types of learning. Furthermore, these studies provide evi-
dence for a physiologic, neuromodulatory role of cytokines
originally described as immune mediators.
As in the case of LTP, the effects of cytokine adminis-
tration on learning, memory, and behavior in general,
have been extensively investigated (for review Anisman
et al., 2005). Again, these studies are undoubtfully of
pharmacological relevance but may not reflect the physio-
logical effect of cytokines, which is the main scope of this
article. Thus, only possible physiologic effects of endoge-
nous cytokines on memory and learning are discussed
below. As mentioned above, a transient blockade of
endogenous IL-1 in hippocampal slices and in the brain
of freely moving rats results in inhibition of LTP mainte-
nance. Considering that it is currently accepted that LTP
underlies certain forms of memory, it was predicted that
this process would be inhibited in animals in which IL-1
effects cannot be manifested. This is the case of IL-1
receptor type I knockout mice (Yirmiya et al., 2002).
These mice display significantly longer latency to reach
a hidden platform in the spatial version of the water maze
test and exhibit diminished contextual fear conditioning,
but behave similarly to control animals in hippocampal-
independent memory tasks. Blockade of IL-1 receptors
in the brain of normal animals following a learning task
(Morris water maze) causes hippocampal-dependent mem-
ory impairment. These results suggest that IL-1 signaling
within the hippocampus plays a critical role in learning
and memory processes (Avital et al., 2003). It is worth
noting that in the previously mentioned studies, the
impediment of IL-1 signaling was induced after the train-
ing procedure. In contrast, there is evidence that blockade
of IL-1 effects prior training by using an adenovirus vec-
tor expressing IL-1ra causes an improvement of both
short-term and long-term memory retention scores (Depi-
no et al., 2004). However, as mentioned, endogenously
produced IL-1 during learning significantly contributes
to memorize an established task.
The role of IL-6 endogenously produced in the brain has
also been studied. As discussed above, IL-6 is produced
during LTP. Blockade of endogenous IL-6 after hippocam-
pus-dependent spatial alternation learning resulted in sig-
nificant improvement of long-term memory (Balschun
et al., 2004). Furthermore, IL-6 KO mice exhibited a facil-
itation of radial maze learning over 30 days, in terms of
lower number of working memory errors (Braida et al.,
2004).
Taken these results together, it appears clear that,
although having an opposite role, endogenous IL-1 and
IL-6 produced in the ‘‘healthy’’ brain, are important in
the control of synaptic plasticity and in the hippocampal
processing of memory. The mechanism underlying the
role of IL-1 and IL-6 on these processes is still unknown
but recent data indicate the involvement of NFk-B, a
transcription factor that mediates the production and
effects of multiple cytokines (for review Meffert and Bal-
timore, 2005).
In conclusion, there are examples that cytokines pro-
duced in the healthy brain can contribute to brain physiol-
ogy by controlling neuronal activity affecting in this way
neuro–endocrine control systems, the set point for the reg-
ulation of essential homeostatic mechanisms and intrinsic
functions of the CNS like memory and learning. There is
also evidence that brain-born cytokines coordinate physio-
logic adjustments during disease and neuro–endocrine
immunoregulatory responses (Fig. 1, red circle). Cytokines
appear to play a dual role during brain pathology, on one
hand they can control the local immune response by cyto-
kine–neuronal interactions and on the other hand they can
behave as effector mediators capable to contribute to the
disease.
42 H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44

5. 2007: Waiting for the beginning
This subtitle is clearly optimistic since it pretends to
stress the almost infinite work left to the new generations.
For example, we often referred to neuro–endocrine immu-
noregulatory responses that are elicited following stimula-
tion of the IS. This is an oversimplification that derives
from the experimental models used initially to uncover
immune–neuro–endocrine interactions. There is no ‘‘one’’
immune response: there are probably as many types of
immune responses as pathological conditions. While
immunologists make distinctions between different aspects
of natural immunity and type of adaptive immune
response, e.g. Th1/Th2-mediated responses, in real condi-
tions immunity is based on interwoven mechanisms that
in many cases are the product of a trade-off between
immune cells and the causal agent of a disease, e.g. a micro-
organism, and, why not, the therapeutic intervention used.
During each stage of a disease, pathogen-associated molec-
ular patterns (PAMPs) mediated mechanisms are activated,
particular combinations of immune cells and their specific
products are detected and a cocktail of other mediators
are found in the circulation and in inflamed/infected tis-
sues. Thus, neuro–endocrine responses during disease and
their immunoregulatory outcome are expected to be differ-
ent and would require investigation at all levels. Further-
more, today we could state that local, short-loop
immune–neuro–endocrine interactions are likely to occur
probably in all organs or tissues during health and during
the course of a disease.
An important breakthrough in the last decade derived
from the exploration of the complex intracellular machinery
involved in the response of a cell following stimulation of a
given receptor. For example, the different protein-kinases,
second messengers, transcription factors, post-transcrip-
tional events and intracellular modulators can be points of
encounter during the simultaneous or sequential activation
of receptors by immune and neuro–endocrine derived
ligands on immune and non-immune cells. This would result
in an almost infinite potential synergism or antagonism
between the effects of these ligands under normal and
pathological conditions. However, from the physiological
point of view, the investigation of immune–neuro–endocrine
interactions at cellular level should not be isolated from
the upstream homeostatic events that result in changes
in the production and liberation of different ligands and in
the expression of their receptors on the target cell.
Today it would not sound exaggerated to state that almost
all pathologies have an inflammatory/immune component
that interacts with neuro and endocrine mechanisms. In fact,
immune cells, particularly antigen-presenting cells, are
present in all tissues where they are exposed to hormones
and neurotransmitters.
As stated in the introduction, it is difficult to find the
limit between physiological or pathological situations when
one refers to the IS. We wonder whether the different
denominations given to our field are somehow self-imposed
restrictions. For example, should a field of research devot-
ed to study the psychoneuroimmunology of skin diseases
be denominated psychoneuroimmunodermatology? This
would apply to all medical specialties and also to Biology
(psychoneuroimmunobiology?) and we could continue
adding all possible acronyms to the denomination of our
field. What are we really doing when we study immune–
neuro–endocrine interactions in each organ? In our view,
what we are doing is a serious attempt to integrate the pres-
ent knowledge in biology and medicine based on a multi-
disciplinary approach. This attempt serves to unify data
obtained using systemic approaches with data derived from
the implementation of refined technologies that allow the
analysis of a phenomenon at molecular levels. In our view,
our work serves to reestablish the balance between deduc-
tion and induction, analysis and synthesis, which is, in fact,
the essence of science. The need to reestablish such balance
derives from the actual predominance of analytic
approaches that sometimes degenerate in reductionist
views. Such predominance probably derives from the
immense technologic progress of the last decades that
results in the temptation (sometimes favored by grant agen-
cies and high impact factor journals) to invest resources in
understanding molecular mechanisms isolated from their
physiological and pathophysiological significance. Indeed,
even if all products that serve as messengers for intra, extra,
and intercellular communication would be identified, we
would not be able to decipher all the messages they could
convey. A messenger can be purified but not the message
that it carries, which depends on a large variety of temporal
and spatial conditions and on the present state of the target
that receives the information. Thus, the question that
comes to our mind is: has the time arrived to re-name
our field by simply calling it ‘‘Integrative Biology and
Medicine’’?
References
Ader, R., Cohen, N., 1975. Behaviorally conditioned immunosuppression.
Psychosom. Med. 37, 333–340.
Anisman, H., Merali, Z., Poulter, M.O., Hayley, S., 2005. Cytokines as a
precipitant of depressive illness: animal and human studies. Curr.
Pharm. Des. 11, 963–972.
Avital, A., Goshen, I., Kamsler, A., Segal, M., Iverfeldt, K., Richter-
Levin, G., Yirmiya, R., 2003. Impaired interleukin-1 signaling is
associated with deficits in hippocampal memory processes and neural
plasticity. Hippocampus 13, 826–834.
Balschun, D., Wetzel, W., Del Rey, A., Pitossi, F., Schneider, H.,
Zuschratter, W., Besedovsky, H.O., 2004. Interleukin-6: a cytokine to
forget. FASEB J. 18, 1788–1790.
Besedovsky, H.O., del Rey, A., 1996. Immune–neuro–endocrine interac-
tions: facts and hypotheses. Endocr. Rev. 17, 64–102.
Besedovsky, H.O., del Rey, A., 2006. Regulating inflammation by
glucocorticoids. Nat. Immunol. 7, 537.
Besedovsky, H.O., del Rey, A., Sorkin, E., 1981. Lymphokine-containing
supernatants from con A-stimulated cells increase corticosterone blood
levels. J. Immunol. 126, 385–387.
Besedovsky, H.O., del Rey, A., Sorkin, E., Da Prada, M., Burri, R.,
Honegger, C., 1983. The immune response evokes changes in brain
noradrenergic neurons. Science 221, 564–566.
H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44 43

Besedovsky, H.O., del Rey, A., Sorkin, E., Dinarello, C.A., 1986.
Immunoregulatory feedback between interleukin-1 and glucocorticoid
hormones. Science 233, 652–654.
Besedovsky, H.O., Sorkin, E., 1977. Network of immune-neuroendocrine
interactions. Clin. Exp. Immunol. 27, 1–12.
Blalock, J.E., Harbour-McMenamin, D., Smith, E.M., 1985. Peptide
hormones shared by the neuroendocrine and immunologic systems. J.
Immunol. 135, 858s–861s.
Braida, D., Sacerdote, P., Panerai, A.E., Bianchi, M., Aloisi, A.M., Iosue,
S., Sala, M., 2004. Cognitive function in young and adult IL
(interleukin)-6 deficient mice. Behav. Brain. Res. 153, 423–429.
Dantzer, R., 2004. Cytokine-induced sickness behaviour: a neuroimmune
response to activation of innate immunity. Eur. J. Pharmacol. 500,
399–411.
del Rey, A., Besedovsky, H., 1989. Antidiabetic effects of interleukin 1.
Proc. Natl. Acad. Sci. USA 86, 5943–5947.
del Rey, A., Besedovsky, H.O., 1992. Metabolic and neuroendocrine
effects of pro-inflammatory cytokines. Eur. J. Clin. Invest. 22 (Suppl.
1), 10–15.
del Rey, A., Kabiersch, A., Petzoldt, S., Besedovsky, H.O., 2002.
Involvement of noradrenergic nerves in the activation and clonal
deletion of T cells stimulated by superantigen in vivo. J. Neuroimmu-
nol. 127, 44–53.
del Rey, A., Monge-Arditi, G., Besedovsky, H.O., 1998. Central and
peripheral mechanisms contribute to the hypoglycemia induced by
interleukin-1. Ann. N. Y. Acad. Sci. 840, 153–161.
del Rey, A., Roggero, E., Kabiersch, A., Schafer, M., Besedovsky, H.O.,
2006. The role of noradrenergic nerves in the development of the
lymphoproliferative disease in Fas-deficient, lpr/lpr mice. J. Immunol.
176, 7079–7086.
Delgado, M., Pozo, D., Ganea, D., 2004. The significance of vasoactive
intestinal peptide in immunomodulation. Pharmacol. Rev. 56, 249–290.
Depino, A.M., Alonso, M., Ferrari, C., del Rey, A., Anthony, D.,
Besedovsky, H., Medina, J.H., Pitossi, F., 2004. Learning modulation
by endogenous hippocampal IL-1: blockade of endogenous IL-1
facilitates memory formation. Hippocampus 14, 526–535.
Elenkov, I.J., Wilder, R.L., Chrousos, G.P., Vizi, E.S., 2000. The
sympathetic nerve—an integrative interface between two supersystems:
the brain and the immune system. Pharmacol. Rev. 52, 595–638.
Heijnen, C.J., Kavelaars, A., 1999. The importance of being receptive. J.
Neuroimmunol. 100, 197–202.
Jessop, D.S., 2002. Neuropeptides: modulators of immune responses in
health and disease. Int. Rev. Neurobiol. 52, 67–91.
Kent, S., Bluthe, R.M., Dantzer, R., Hardwick, A.J., Kelley, K.W.,
Rothwell, N.J., Vannice, J.L., 1992. Different receptor mechanisms
mediate the pyrogenic and behavioral effects of interleukin 1. Proc.
Natl. Acad. Sci. USA 89, 9117–9120.
Maier, S.F., Goehler, L.E., Fleshner, M., Watkins, L.R., 1998. The role of
the vagus nerve in cytokine-to-brain communication. Ann. N. Y.
Acad. Sci. 840, 289–300.
Meffert, M.K., Baltimore, D., 2005. Physiological functions for brain NF-
kappaB. Trends Neurosci. 28, 37–43.
Nussdorfer, G.G., Malendowicz, L.K., 1998. Role of VIP, PACAP, and
related peptides in the regulation of the hypothalamo–pituitary–
adrenal axis. Peptides 19, 1443–1467.
Obal Jr., F., Krueger, J.M., 2003. Biochemical regulation of non-rapid-
eye-movement sleep. Front. Biosci. 8, d520–d550.
Pitossi, F., del Rey, A., Kabiersch, A., Besedovsky, H., 1997. Induction of
cytokine transcripts in the central nervous system and pituitary
following peripheral administration of endotoxin to mice. J. Neurosci.
Res. 48, 287–298.
Rogausch, H., del Rey, A., Kabiersch, A., Reschke, W., Ortel, J.,
Besedovsky, H., 1997. Endotoxin impedes vasoconstriction in the
spleen: role of endogenous interleukin-1 and sympathetic innervation.
Am. J. Physiol. 272, R2048–R2054.
Sanders, V.M., 2006. Interdisciplinary research: noradrenergic regulation
of adaptive immunity. Brain Behav. Immun. 20, 1–8.
Schauenstein, K., Fassler, R., Dietrich, H., Schwarz, S., Kromer, G.,
Wick, G., 1987. Disturbed immune–endocrine communication in
autoimmune disease. Lack of corticosterone response to immune
signals in obese strain chickens with spontaneous autoimmune
thyroiditis. J. Immunol. 139, 1830–1833.
Schneider, H., Pitossi, F., Balschun, D., Wagner, A., del Rey, A.,
Besedovsky, H.O., 1998. A neuromodulatory role of interleukin-
1beta in the hippocampus. Proc. Natl. Acad. Sci. USA 95, 7778–
7783.
Sternberg, E.M., 2006. Neural regulation of innate immunity: a coordi-
nated nonspecific host response to pathogens. Nat. Rev. Immunol. 6,
318–328.
Sternberg, E.M., Hill, J.M., Chrousos, G.P., Kamilaris, T., Listwak, S.J.,
Gold, P.W., Wilder, R.L., 1989. Inflammatory mediator-induced
hypothalamic–pituitary–adrenal axis activation is defective in strepto-
coccal cell wall arthritis-susceptible Lewis rats. Proc. Natl. Acad. Sci.
USA 86, 2374–2378.
Tracey, K.J., 2002. The inflammatory reflex. Nature 420, 853–859.
Turnbull, A.V., Rivier, C.L., 1999. Regulation of the hypothalamic–
pituitary–adrenal axis by cytokines: actions and mechanisms of action.
Physiol. Rev. 79, 1–71.
Weihe, E., Nohr, D., Michel, S., Muller, S., Zentel, H.J., Fink, T., Krekel,
J., 1991. Molecular anatomy of the neuro–immune connection. Int. J.
Neurosci. 59, 1–23.
Yirmiya, R., Winocur, G., Goshen, I., 2002. Brain interleukin-1 is
involved in spatial memory and passive avoidance conditioning.
Neurobiol. Learn Mem. 78, 379–389.
44 H.O. Besedovsky, A.del Rey / Brain, Behavior, and Immunity 21 (2007) 34–44