Clin. exp. immunol. (1977) 27, 1-12.

Network of immune-neuroendocrine interactions

H. BESEDOVSKY* & E. SORKIN Schweizerisches Forschungsinstitut, Medizinische Abteilung,
7270 Davos-Platz, Switzerland

(Received 23 June 1976)

SUMMARY
In order to bring the self-regulated immune system into conformity with other body systems its
functioning within the context of an immune-neuroendocrine network is proposed. This hypo-
thesis is based on the existence of afferent-efferent pathways between immune and neuroendocrine
structures. Major endocrine responses occur as a consequence of antigenic stimulation and changes
in the electrical activity of the hypothalamus also take place; both of these alterations are temporally
related to the immune response itself. This endocrine response has meaningful implications for
immunoregulation and for immunospecificity. During ontogeny, there is also evidence for the
operations of a complex network between the endocrine and immune system, a bidirectional inter-
relationship that may well affect each developmental stage of both functions. As sequels the func-
tioning of the immune system and the outcome of this interrelation could be decisive in lymphoid
cell homeostasis, self-tolerance, and could also have significant implications for pathology.

INTRODUCTION
The network of interactions of systems in higher organisms includes not only self-monitoring and self-
regulatory processes, but also other integrative forms of regulation mediated by the nervous and endo-
crine system. The central function of the immune system is now believed to be the distinguishing of self
from not-self (Burnet & Fenner, 1949). There are reasons for believing that the immune system to a
large extent regulates its own functioning. The last decade has witnessed comprehensive progress in the
various elements comprising immunoregulation (Jerne, 1977). As the immune system is also of critical
importance for the host to cope with the threats and challenges of exceedingly complex internal and
external environments, it is considered by us unlikely that it would function in the completely autonomous
manner thus far proposed. That hormones exert distinctive effects on the immune system is generally
appreciated. However, the operation of afferent and efferent pathways to and from the neuroendocrine
structures has thus far not received any serious consideration in relation to the function of the immune
system. Many kinds of neuroendocrine regulatory functions involve changes manifested in altered peri-
pheral blood levels of certain hormones and in their hypothalamic activity. To conceive of neuroendo-
crine regulatory mechanisms in connection with immune responsiveness, it is first essential to ascertain
that the immune response itself can elicit such hormonal changes.
In this laboratory it has recently been shown that in the course of the primary immune response to
soluble antigens, or to nonreplicating foreign cells, there occur temporal endocrine changes of striking
magnitude which parallel the elaboration of specific antibody-producing cells (Besedovsky et al., 1975).
Furthermore, there have concurrently been observed changes in electrical activity of neurones ('firing
rate') in the ventromedial part of the rat hypothalamus after antigenic stimulation (Besedovsky, Sorkin
Felix & Haes, 1976). Data such as these, attest to the existence of information transmitted to and from
the neuroendocrine system during its response to antigenic stimulation. Also during early ontogeny,
complex bidirectional interactions between the endocrine and the immune systems have been described.
Correspondence: E. Sorkin, Schweizerisches Forschungsinstitut, Medizinische Abteilung, 7270 Davos-Platz, Switzerland.
* Present address: Argentina Facultad de Medicina, Departmento de Fisiologia, Universidad de Rosano,

Rosano 2000.
1
2 H. Besedovsky & E. Sorkin
In this communication we develop the proposition that during development and expression of immune
functions there are permanent interrelations between the components of the immune and neuroendocrine
systems. From such bidirectional interactions it would follow that the actual state of both systems at any
given time is influenced by the integration of their functions. This communication briefly develops the
concept of an immune-neuroendocrine network and identifies some of its implications for immuno-
regulation.
(I) Mutual influence ofendocrine and immune function in ontogeny
The first point that bears emphasizing is the remarkable near-parallel development of the immune and
endocrine systems in various mammalian species during early ontogeny. This can hardly be fortuitous.
Thus, in species such as mice and rats which at birth have only marginal levels of immunoglobulins and
in which immune responses are minimally effective during the first days of extrauterine life (Solomon,
1971), many endocrine mechanisms are also underdeveloped at this time, notably sexual differentiation,
functioning of the hypophysis, thyroid and adrenal glands (Jost, 1969; Jost et al., 1973). There are species
on the other hand such as guinea-pigs (Illingworth et al., 1973; Pals, Reineke & Schaw, 1973; Moog &
Ford, 1957), sheep (Alexander et al., 1971; Bassett, Thorburn & Wallace, 1970; Bassett & Alexander,
1971; Dussault, Hobel & Fisher, 1971), cow (Dubois, 1971), and man (Jost, 1969), which to greater or
lesser degree are more mature immunologically at birth; in these species the aforementioned endocrine
mechanisms are likewise synchronized in a more advanced stage of development. The rabbit provides an
example of a species intermediate between these extremes. In the aforementioned species, a similar time
correlation has also been established between the effect of neonatal thymectomy on the development of
the immune system and the endocrine status at birth.
In this report, it is intended to examine the proposition that there also exists a bidirectional influence
(i.e. each affects the other) between the endocrine and the immune system. The evidence derives from
experimental situations in which a disturbance of this interrelation at one level results in changes at
other levels. Such a disturbance is found in conditions where the development of the immune system is
impaired or alternatively where the endocrine environment is changed. In either case, the other para-
meters change in an essentially parallel fashion.
Four kinds of disturbances of the proposed bidirectionally interacting systems are given as examples:
(1) The germ-free state, i.e. lack of antigenic challenge is manifested in a marked underdevelopment
in the total mass of host lymphoid tissue and depressed immunoglobulin levels; it is also expressed in an
altered endocrine state (Wostmann, 1968). There are reports of thyroid (Pleasants, 1968; Vought et al.,
1972), adrenal (Miyakawa & Ukai, 1970; Miyakawa, 1966) and testicular (Nomura et al., 1973) in-
sufficiency in germ-free mice and rats. In guinea-pigs adrenal hyperfunction has been described (Chak-
hava, 1973).
(2) Neonatally thymectomized mice and congenitally thymusless mice which have a T-lymphocyte
deficiency, display a profoundly disturbed endocrine system. Absence of a thymus in animals kept under
conventional, specific-pathogen-free (SPF) or germ-free condition, results in a degranulation of STH-
producing cells in the adenohypophysis (Bianchi, Pierpaoli & Sorkin, 1971), a delay of puberty in females
(Besedovsky & Sorkin, 1974), persistance of the reticular zone of the adrenal gland (Pierpaoli & Sorkin,
1972), hypothyroidism (Pierpaoli & Sorkin, 1972; Pierpaoli & Besedovsky, 1975), and alterations in blood
levels of gonadal hormones (Pierpaoli & Sorkin, 1972). A number of these parameters can be normalized
only by early thymus implantation (Besedovsky & Sorkin, 1974). The passive transfer of lymphoid cells
makes the recipients immunocompetent, but fails to normalize the aforementioned hormone-dependent
parameters. Endocrine influences of the thymus on other endocrine glands are already expressed during
perinatal life, especially with regard to female sexual function. This finding provides a strong indication
that the thymus is involved in the programming of the neuroendocrine system (Besedovsky & Sorkin,
1974; Pierpaoli & Besedovsky, 1975).
(3) Surgically bursectomized 62-hr-old chicken embryos show the following endocrine alterations
later in embryonic life or at the time of hatching: degranulation of gonadotrophic cells in the hypo-
physis, underdeveloped oviduct, adrenal hypertrophy, low level of corticosterone in chorioallantoid fluid
 Network of immune-neuroendocrine interactions 3
and high level of testosterone in this fluid in females (Besedovsky et al., 1976, submitted for publication).
(4) Changes in hormonal environment produce alterations in acquisition of immune capacity. Apart
from thymic hormones there are numerous examples of hormonal influence on the development of the
immune system (Pierpaoli, Fabris & Sorkin, 1970). An instructive model is the hypopituitary dwarf
mouse with its deficiency in growth hormone and thyrotropin. Cell-mediated immunity as measured by
transplantation criteria is defective; this deficit can be normalized by injection of STH and thyroxine
(Fabris, Pierpaoli & Sorkin, 1971a, b).
The above-mentioned examples of interference at the level of antigenic challenge, development of
central and peripheral lymphoid tissue, or at the level of endocrine functions attest to a complex network
of interaction between the immune and the endocrine systems in ontogeny. Fig. 1 illustrates in a dia-

Brain

Endocrine system

Bursa or Thymus
equivalent

Secondary
lymphoid
tissue

I
Antigen
FIG. 1. Diagram of network of interactions between immune and endocrine systems in ontogeny and in adult
life.

grammatic way possible sequences of these interactions. It is rather predictable that the overall product
of this interaction during different developmental stages could make for a situation which favoured
either tolerance to selfcomponents or immunity.
(II) The integration of the immune and endocrine responses following antigenic stimulation
The fact that hormones influence the immune system is well established. Such findings are entirely
logical in view of the known role of hormonal influence on varied basic cellular functions such as protein
synthesis, control of gene expression, cell replication and the allosteric arrangement of cell membranes,
all of which are basic to physiological processes and have an essential function in the immune perform-
ance. Furthermore, some hormones operate via their capacity to influence the rate of formation of
cyclic 3,5 adenosine monophosphate (cyclic 3,5 AMP), an intermediate which is also known to affect
strongly lymphoid cells and antibody production (Bourne et al., 1974).
In the past, investigations on the involvement of hormones in the immune response have been based on
the administration of hormones or the ablation or blockade of endocrine glands. Depressed or stimulated
immune responses resulted depending on the particular hormone involved, the dose and the timing of its
4 H. Besedovsky & E. Sorkin
administration. It is thus evident, that externally induced changes in the level of certain hormones can
influence immune reactivity considerably. However, the possibility that the immune response itself
could bring about changes in hormone levels has not been previously considered. We have shown that
during the primary immune response of rats to a particulate antigen, sheep red cells (SRBC), or to
soluble antigen Trinitrophenyl-haemocyanin (TNP-Hae), and of mice to TNP-horse red blood cells
(TNP-HRBC), corticosterone levels increased several-fold while there occurred temporal changes in
thyroxine levels.
In Fig. 2b is shown a two- to three-fold increase in serum corticosterone levels above normal at 5,
6, 7 or 8 days after injection of 4 x 10i SRBC into female rats. The fact that no significant changes in
the serum corticosterone level occurred on days 1 and 3 after immunization attests to the fact that the
animals were not stressed by handling or the injection of the cells.

C
E6

aCat
U
LL

U0cL
E~~~~~~~
Cc)
7
S

)I-

0 2 4 6 8 105?7
Days
FIG. 2. Changes in serum corticosterone and thyroxine levels during the immune response to sheep red blood
cells (SRBC) in rats: (a) plaque-forming cells (PFC) x 103 per spleen; (b) corticosterone levels in serum;
animals immunized with SRBC; control injected with rat red blood cells (RRBC); (c) thyroxine levels in
serum; animals immunized with SRBC; a control immunized with RRBC.

A biphasic change in the serum thyroxine concentration was observed in animals treated with the
higher dose of SRBC (Fig. 2c). After an initial increase on day 3, the serum thyroxine decreased by
approximately 30% below normal on days 5-8. Rats injected with the same dose homologous red cells
showed no changes in corticosterone and thyroxine levels at any time. Changes in blood hormone levels
are known to be the main information for the neuroendocrine structures, e.g. the hypothalamus, and this
 Network of immune-neuroendocrine interactions 5
organ regulates the blood hormone level. Accordingly, the observed changes, whatever their cause, can
be considered as also involving a specific neuroendocrine response and attest to the existence of an afferent
pathway from the activated immune system to the central regulatory structures involved in endocrine
regulation.
It is not excluded that the almost infinite diversity of the process of immunization itself, i.e. the great
spectrum of antigenic agents, the range of immunizing dosage and the route of administration, will all
serve to make for diversity in the altered hormonal profiles which result. This would be a consequence
of the qualitative and quantitative spectrum of lymphoid cells brought into play, their interactions and
responses, and their consequences as manifested by elaboration of various classes of Ig, mediators, ag/ab
complexes, etc. Moreover, dissimilar neuroendocrine effects would be expected even in response to the
same antigen differently administered, depending on whether the response it elicits involves primarily
T cells, B cells, macrophages, etc.
(III) Link between the immune response and the hypothalamus
The described temporal endocrine changes which accompany the production of antibodies are in our
view an expression of the response of neuroendocrine structures to signals which derive from activated
lymphoid cells. To obtain direct proof of such a postulated afferent pathway, we looked for significant
changes in the target organ, the hypothalamus. The experimental system utilized was the electrical
activity of individual neurones in the ventromedial nuclei of the hypothalamus of rats after immunization
with sheep red blood cells or TNP-haemocyanin.
The data for animals stimulated with SRBC are summarized in Fig. 3. On day 1 when PFC have not

10 _ -llll

8 T_ ~~~~~~~~~~~~~liii'
2
g 4_
| ~~~~~~~~~~~~~~~~~~lilil~l
:'UlE6 -
j1ii1l
0.-

I<Ol*

- n-29n=5
|s'|| 1~

Dayl Day 5

FIG. 3. Increase in firing rates of neurones of ventromedial nuclei in the rat hypothalamus after i.p. injection
of 5 x 109 sheep red blood cells (a3). Controls injected with saline (o).
yet been produced, the firing frequency of neurones in the antigen-stimulated group was no diffierent
from that of saline controls. On day 5, which represents the peak of IgM PFC formation in spleen, there
was seen a three-fold increase in the firing rate of the neurones. Hypothalamic responses to the other
antigen employed, 250,ug TNP-haemocyanin, were also observed. A greater than two-fold increase in
firing rate was noted.
Data such as these are tangible evidence for the operation of a flow of information from the activated
immune system to the brain, showing that the immune system, just as a number of others, is linked with
6 H. Besedovsky & E. Sorkin
the CNS. Consequently, it seems a reasonable view that the hypothalamus is intimately linked to the
process of immunoregulation in a manner external to the immune system.
Presumably, this neuroendocrine mechanism functions when a critical threshold of lymphoid tissue
activation is reached, sufficient to elaborate products serving as a signal to the hypothalamus which in
turn releases factors that via the hypophysis and its polypeptide hormones closes the circle by its effects
on the immune system.
(IV) Possible nature of link between immune and neuroendocrine response
It is a rational assumption that the primary link between the immune and the neuroendocrine system is
effected by one or another ofthe multiple events known to follow immunization. The afferent and efferent
pathways to and from the hypothalamus could then be visualized as functioning in the following ways:
(1) Antigen. While antigen is obviously the trigger of immune mechanisms, it has so far been unknown
that the neuroendocrine system can be affected as well. It is most unlikely in our view that antigen itself
is the direct cause of the altered blood hormone levels that ensue. For one thing, antigenic recognition
by lymphocytes occurs well before the observed endocrine changes occur (see Fig. 2). For another thing,
while the immune system has an enormous 'library' of lymphocytes recognition sites for 'reading' the
almost infinite array of antigenic determinants, there is no evidence whatsoever that the neuroendocrine
structure has an equally elaborate recognition counterpart for antigens.
(2) Antibody. Since antibodies can cross the blood brain barrier, this specific humoral end product of
antigenic stimulation must also be considered as possible messenger to the neuroendocrine system.
However, no experimental data exist at present to suggest that antibodies to extrinsic antigen can influ-
ence the neuronal activity in the hypothalamus.
(3) Electrical signalfrom peripherical nerves. In contrast to the extensive knowledge on innervation of
different organs or tissues to the best of our knowledge the significance of nerve endings in lymphoid
structures is unknown. Accordingly, it is difficult at present to postulate an afferent or efferent neural
pathway to and from the brain and to understanding the regulatory significance of the known pharmaco-
logical influence on the immune response by mediators of the autonomous nervous system, such as
catecholamin (Hadden, Hadden & Middleton, 1970) and acetylcholin (Hadden et al., 1975) and some
reports of immunological alterations after neurotomy (Kesztyus, 1967) or electrolytic lesions of the brain
(Jankovic & Isakovic, 1973; Isakovi6 & Jankovic, 1973; Stein, Schiavi & Camerino, 1976). However, one
cannot exclude entirely a bidirectional neural pathway between lymphoid tissue and the central nervous
system contributing to immunoregulation.
(4) Hormones. The immune response also elicits changes in peripheral hormone levels, which we believe
can influence immune processes in an efferent way. Whether these hormonal changes are a direct conse-
quence ofthe immune events on the endocrine target glands or secondary to effects of the neuroendocrine
system on these targets is not known. Even if the former were the case, the hypothalamus-hypophysis
axis would still represent the normal manner of controlling the endocrine system.
(5) Mediators. Chemical mediators are released by activated T-lymphocytes and by antibody-producing
B cells during antigen induced proliferation. It is conceivable that one or another of these potent effector
molecules may influence the endocrine target glands either directly or more likely via hypothalamus-
hypophysis. An alternative possibility is that immune complexes via activation of plasma constituents or
cells could initiate a similar sequence of events. It can hardly be a coincidence that histamine and sero-
tonin, low mol. wt cell products also known to mediate the immune response, are present in high concen-
trations in the median eminence of the hypothalamus, the general pathway for all neuroendocrine pro-
cesses (White, 1966; Crawford, 1958). There is also no blood-brain barrier for precursors of histamine
and serotonin in the median eminence (Davison, 1958; Udenfriend, Weissbach & Bogdanski, 1957).
Furthermore, it is now postulated that serotonin (Navmenko, 1973) and probably histamine as well
(White, 1966) are linked (possibly via hypothalamic releasing factors) with ACTH release by the adeno-
hypophysis; this sort of mechanism may be operative in effecting the increased corticosterone blood
levels reported in Fig. 2b. There is no analogous information, at present, regarding the other major
category of mediators of the immune response, the lymphokines.
 Network of immune-neuroendocrine interactions pl

(V) Implications of the bidirectional relationships between immune and endocrine system
The previously discussed data and reasoning strongly support the idea of a complex network between
two host systems, operative from early ontogeny into adult life. Further knowledge of the profile of
hormonal responses and the neuroendocrine lymphoid tissue inter-relationship will be required for a more
definitive interpretation of the physiology and the biologic significance of this interaction. We are
presently ignorant of the extent to which the endocrine changes consequent to the immune response
affect in turn metabolic conditions and cellular activities and consequently the course of the immune
response itself (efferent pathway). The information already at hand would seem sufficient to warrant
some speculation on the ways in which hormonal control might be exerted over key situations such as
immunospecificity, antigenic competition, tolerance and macrophage function.
(1) Immunospecificity. Specificity of the immune response has its basis in the interaction of antigen
with specific receptors on lymphocytes. Following this primary event, a regulatory intervention by
hormones could be exerted at many levels. The specificity of the response is due to high affinity cells, but
antigen will also act on cells of lower affinity as well. Furthermore, non-specific mediators released by
activated cells can activate other cells, unrelated to the antigenic stimulus, and thereby induce a con-
commitant increase in nonspecificity. While the antigen provides the information for a highly specific
response the activation of other cells would lead to a disturbance of the system, introducing background
noise. Such an uneconomical event may be potentially harmful to the host to the extent of enhancing the
probability of autoimmune reactions. Also an excessive expansion of lymphoid cell mass would in effect
raise the concentration of soluble mediators to an undesirable or possibly even dangerous extent. One
way of suppressing such untoward events would be through the action of corticosteroids and other
steroids, which are well known to suppress or even delete lymphoid cells not stimulated by antigen as
well as macrophage function and to inhibit lymphokines synthesis (Wahl, Altmann & Rosenstreich,
1975). Non-stimulated cells (before the onset ofthe immune response) are easily influenced by hormones,
e.g. corticosteroid, but after the antigen administration the course ofcellular events is much more difficult
to modulate by the very same agent (Makinodan, Santos & Quinn, 1970; Claman, 1975). Indeed, the
observed increase of corticosteroid during the immune response may well have this very function of
suppressing a potentially harmful expansion of lymphoid tissue of low or no affinity for the antigen. This
fits well with the observation that adrenalectomy leads to a pronounced increase in spleen weight in mice.
Antigenic challenge of adrenalectomized mice results in a still further weight increase. Controls did not
show any spleen weight increase when antigenically challenged (Kieffer & Ketchel, 1971). In our view,
this is a consequence of prevention of antigen provoked corticosteroid increase, resulting in a deleted
hormonal control.
This hormonally controlled mechanism could also be the key to the situation labelled as antigenic
competition, where two or more antigens are injected simultaneously (or more often sequentially) and the
resultant immune response to either antigen is less than when they are given singly. This phenomenon is
actually an expression of the existence of a regulatory mechanism for suppressing unrelated cells. In our
view, apart from suppressor cells, the hormonal changes following antigenic stimulation are also respon-
sible for this suppression. The following experimental evidence strongly supports our contention.
When sheep red cells (SRBC) are injected into rats previously immunized with TNP-horse red cells,
there occurs a marked depression of the response to SRBC. The first antigen had already evoked a
severalfold increase in corticosterone blood levels at the time the second antigen (SRBC) was injected.
(Table 1). If this increase in corticosterone were responsible for the suppression of the SRBC response,
prior adrenalectomy would be expected to counteract this suppression. The fact is that direct experi-
mental test of this issue shows (Fig. 4) a twenty-fold increase in plaque-forming cells in adrenalectomized
rats injected first with horse RBC and then with sheep RBC. In animals given only SRBC, adrenalectomy
does not change the response appreciably.
In addition to those well-recognized factors that make for immunospecificity such as specific regula-
tion by components of the system (e.g. suppressor T cells, antibodies), this can also be brought about by
hormonal non-specific regulation.
8 H. Besedovsky & E. Sorkin
TABLE 1. Increased serum corticosterone levels during antigenic
competition

Antigen Day of sacrifice Corticosterone

pg/100 ml serum
0 5-81+ 11
TNP/Horse RBC 6 28-7 + 6-1
TNP/Horse RBC 11 11-9 + 1.2
TNP/Horse RBC
and on day 6 SRBC 11 28-4 ± 4-2
SRBC 5 30*4 + 109

Mean+ s.e. 109 TNP/Horse RBC resp. 109 SRBC were injected
i.p. into adult female Holtzmann rats.
The results show that there is an increase in serum corticosterone
level 6 days after immunization with horse red cells, i.e., when the
second antigen, SRBC, was injected. This maintains the high level
of corticosterone and may influence the process of antigenic
competition (see also Fig. 2).

400H b d
Xi
cL

CA
1
L.
0

x
200h

CL

C) I I I I I I
n=8
.-I n=8
n=25 n=19
FIG. 4. Prevention of antigenic competition in adrenalectomized rats: (a) normal rats immunized i.p. with 109
SRBC. Test on day 5; (b) Adrenalectomized rats immunized with 109 SRBC. Test on day 5, (c) Normal rats
injected with 109 HRBC and 6 days later immunized with 109 SRBC. Test for PFC against SRBC on day 5
after SRBC; (d) adrenalectomized rats injected with 109 HRBC and 6 days later immunized with 109 SRBC.
Test for PFC against SRBC on day 5 after SRBC.

The maintenance of specificity by non-specific intervention deserves special mention. Since there is no
known means by which the wide range of antigenic determinants can be specifically recognized by
hormones (non-specific factors), we propose that hormones can enhance specificity by inhibiting or
blocking unrelated or low affinity cells. Similar examples are found in the nervous system, another
highly discriminative network. Lateral or afferent inhibition of surrounding neurones is a well-known
mechanism to eliminate background noise when an impulse is required to be highly discriminative.
Thus, we believe that hormonal regulation can augment or enhance immunospecificity of a response
by suppression of low- or non-specific cellular events without disturbing the requisite clonal expansion
of high affinity cells.
(2) Endocrine status and tolerance. The decision whether an antigenic stimulus will drive the lympho-
 Network of immune-neuroendocrine interactions 9
cyte along the pathway of tolerance or immunity probably also involves the overall endocrine status.
Since variation in hormone levels can augment or diminish the immune response, it is conceivable that
the induction of the tolerant state would be affected by changes in hormonal levels in blood evoked by
antigenic challenge. Such changes may be of different nature and duration at different stages of develop-
ment.
Recognition of self and establishment of tolerance to self components develops during embryonic life
at a time when the maximum number of new antigens are emerging. As already mentioned, there is a
remarkable correspondence in the kinetics and pattern of the development of immune and endocrine
function in various mammalian species during the embryonic period, and a mutual influence on both
systems. Antigen interaction with 'primitive' (not fully competent) lymphoid cells may well induce
changes in hormonal levels which in turn would affect the endocrine status of the animal as observed
during adult life. The outcome of the immune-neuroendocrine network of interactions during ontogeny
can delete or inhibit the selfreacting cells. Regrettably there is little known concerning the nature and
profile of hormones during each stage of embryonic life. It is probable that early endocrine environment
is highly unfavourable for the thymus (and bursa equivalent) to exert their maturational function on
lymphocytes directly or by modifying the endocrine environment. Constant with such an interpretation
is the finding that the thymus of e.g. 14-day-old embryonic mice can establish immunocompetence in
adult mice thymectomized at birth even though these embryos are themselves immunologically immature
(Miller & Osoba, 1963).
Data have been obtained, suggesting that the foetal adrenal gland also plays some role in the generation
of a hormonal environment in ontogeny which is unfavourable for thymus function. There is e.g. an
inverse correlation between the time of acquisition of immunocompetence and involution of the foetal
adrenal gland in diverse mammalian species (Besedovsky, 1971). Furthermore, engraftment of albino
rat adrenal foetal gland, taken during the last week of gestation, effects prolongation of rat skin allograft
survival. Later in ontogeny when ever fewer additional self-antigens are generated a reduction of the
incidence of antigen interaction with primitive lymphoid cells would be expected with diminished
consequences for the endocrine environment. The changes in these conditions would therefore make for
a fuller expression of thymus and bursa function in immunological terms and in their endocrine inter-
relation with other glands. In such a sequence would be created a totally new and different endocrine and
lymphoid cell environment which with emergence from the neonatal state increasingly favours the path-
way to immunity rather than tolerance.
No less than five kinds of tolerance have been described by Medawar and many theories on mechanisms
have been proposed (for review see: Howard & Mitchison, 1975; Katz & Benacerraf, 1974). The influence
of hormones in the induction of adult tolerance has to be considered as a physiological component of that
mechanism (Diener & Lee, 1974). It is here suggested that situations in which administration of antigen
leads to tolerance will be found to be paralleled by concommitant endocrine changes evoked by that very
same schedule of antigen. It will, therefore, be of particular interest to explore the kinetics of tolerance
induction as it relates to hormone status. The present hypothesis would predict a certain parallelism
between antigen induced changes in hormonal responses, in particular of corticosteroids, concommitant
with the acquisition of the state of tolerance.
(3) Relevance of endocrine status for monocyte/macroplage function. Corticosteroids can suppress the
colloid clearance function of the reticulo-endothelial system, prevent accumulation of macrophages in
the delayed hypersensitivity reaction, depress their activation or decrease the number of circulating
monocytes (for a summary, see: Baum, 1975; Claman, 1975). Since reactions of cell-mediated immunity
and tumour-cell inhibition and killing involve an effector function for macrophages (Keller, 1973; Hibbs,
Lambert & Remington, 1972; Evans & Alexander, 1972) and corticosteroid plasma levels are high in
various forms of cancer (Mackay et al., 1971), perhaps in part because of continuous antigenic challenge
of the immune system, it is possible that the hosts macrophage/monocyte response is endocrinologically
disturbed.
In contrast, oestrogens are potent stimulators of reticuloendothelial phagocytic activity. Significantly,
reticuloendothelial activity is higher in female than in male rats and mice and varies with the oestrus
10 H. Besedovsky & E. Sorkin
cycle (reviewed in: Baum, 1971). In view of the aforementioned findings on the susceptibility of phago-
cytic monocytes to glucosteroids, it may be appropriate to now give consideration to the possibility that
a changed endocrine environment induced by antigenic challenge could also affect macrophage function
in its various immunologic manifestations.
(4) Implications for pathology. It is a rational expectation that as the described network of immune-
neuroendocrine interactions has pervasive effects in both directions, this would have special implications
for certain disease states. Thus, endocrine disorders have consequences for the immune system as in
Cushing's disease (Britton, Thoren & Sj0berg, 1975). In the other direction, autoimmune mechanisms
can produce several endocrine diseases (Irvine, 1974). There are also situations where both systems are
simultaneously affected as in thymus- and bursaless animals, in which there is a parallel occurrence of
immune and endocrine derangements (Besedovsky & Sorkin, 1974; Pierpaoli & Sorkin, 1972; Pierpaoli
& Besedovsky, 1975; Besedovsky et al., 1976). A further possibility is that multiple antigenic stimuli over
a protracted period would lead to cumulative alterations in the coupled immune and neuroendocrine
response. Such disruptions could include different kinds of hormonal or immune derangements, in-
cluding for example uncontrolled proliferation of lymphoid tissue and accessory cells. Thus in lympho-
proliferative diseases, viz, in the high leukaemia mouse strain AKR (Metcalf, 1960) and in SJL/J mice
known to develop a high incidence of reticulum cell sarcoma and other neoplasms (Pierpaoli et al., 1974),
endocrine disorders are detected long before clinical disease syndromes are actually expressed. It is also
a notable fact that some animals and humans bearing different kinds of tumours show very similar hor-
monal patterns, raised corticosteroid and depressed thyroxine levels, a pattern expressed briefly in our
animals stimulated with conventional antigens (Shigeru et al., 1966; Galton & Ingbar, 1966; Deshpande
et al., 1969; Marmorston, 1966; Deshpande, Hayward & Bulbrook, 1965; Ghosh, Lockwood & Penning-
ton, 1973; Lancet, 1974; Jensen et al., 1968; for similar changes in pre-illness respiratory infections, see
Mason et al., 1967). Such changes in endocrine pattern in blood and urine of cancer patients are already
utilized for prognosis (Mackay et al., 1971; Hayward & Bulbrook, 1968; Rao & Hewit, 1970). Also in
patients with chronic lymphatic leukaemia (Gallagher et al., 1962; Gallagher et al., 1965), there is an
increase in cortisol in blood.
The common feature of the endocrine changes observed in the above-mentioned diseases are in our
view at least in part related to immune responses. These changed hormonal conditions might have a
significant influence on the course of disease. Thus, it will be of interest to analyse the hormonal profile
in patients with autoimmune disease, lymphoproliferative diseases and cancer in which we expect signifi-
cant endocrine alterations.
This work was supported by the Schweizerischer Nationalfonds zur Forderung der wissenschaftlichen Forschung, grant no
3.600-0.75. We are grateful to Dr H. Laudy for helpful criticism.
The technical assistance of Miss Regula Kellerhals and Mr A. Buhler is gratefully acknowledged.

REFERENCES
ALEXANDER, D.P., BRITTON, H.G., FORSLING, M.L., NIXON, BESEDOVSKY, H.O. & SORKIN, E. (1974) Thymus involvement
D.A. & RATCLIFFE, J.G. (1971) The concentrations of in female sexual maturation. Nature (Lond.), 249, 356.
adrenocorticotrophin, vasopressin and oxytocin in the BESEDOVSKY, H.O., SORKIN, E., KELLER, M. & MULLER, J.
foetal and maternal plasma of the sheep in the latter half (1975) Changes in blood hormone levels during the
of gestation. J. Endocr. 49, 179. immune response. Proc. Soc. exp. Biol. (N.Y.) 150, 466.
BASSETT, J.M. & ALEXANDER, G. (1971) Insulin, growth BEsEmovsKY, H.O., SORKIN, E. FELIX, D. & HAAS, H. (1976)
hormone and corticosterone in neonatal lambs. Normal Hypothalamic changes during the immune response. Sub-
concentration and the effect of cold. Biol. Neonate, 17, 113. mitted for publication.
BASSETT, J.M., THORBURN, G.D. & WALLACE, A.L.C. (1970) BEsEmovsKY, H.O., PEDERNERA, E., PAGANO, E. & Rossi
The plasma growth hormone concentration in the foetal ALORAS, H. (1976) Involvement of bursa of fabricius in
lamb. J. Endocr. 48, 251. chicken endocrine development. Submitted.
BAUM, M. (1975) Effect of administered hormones on the BIANCHI, E., PIERPAOLI, W. & SORKIN, E. (1971) Cytological
immune reaction. Host defence in breast cancer (ed. by changes in the mouse anterior pituitary after neonatal
B. A. Stoll) p. 130. William Heinemann Medical Books thymectomy: a light and electron microscopical study.
Ltd, London. J. Endocr. 51, 1.
BESEDOVSKY, H.O. (1971) Delay in skin allograft rejection in BoURNE, H.R., MELMON, K.L., WEINSTEIN, Y. & SHEARER,
rats grafted with fetal adrenal glands. Experientia, 27, 697. G.M. (1974) Pharmacologic regulation of antibody release
 Network of immune-neuroendocrine interactions 11
in vitro: effects of vasoactive amines and cyclic AMP. Lymphocyte blast transformation. I. Demonstration of
Cyclic AMP, Cell Growth, and the Immune Response (ed. by adrenergic receptors in human peripheral lymphocytes.
Werner Braun, Lawrence M. Lichtenstein and Charles Cell. Immunol. 1, 583.
W. Parker), p. 99. Springer-Verlag, Berlin, Heidelberg, HADDEN, J.W., JOHNSON, E.M., HADDEN, E.M., COFFEY,
New York. R.G. & JOHNSON, L.D. (1975) Cyclic GMP and lympho-
BRITTON, S., THOR#.N, M. & SJ0BERG, H.E. (1975) The cyte activation. Immune Recognition (ed. by A. S. Rosen-
immunological hazard of Cushing's syndrome. Brit. med. thal) p. 359. Academic Press, New York.
J. iv, 678. HAYWARD, J.L. & BULBROOK, R.D. (1968) Urinary steroids
BURNET, F.M. & FENNER, F. (1949) Production of antibodies. and prognosis in breast cancer. Prognostic Factors in
Macmillan, Melbourne. Breast Cancer (ed. by A. P. M. Forrest and P. B. Kunkler)
CHAKHAVA, O.V. (1973) Ascorbic acid and glucocorticoid p. 383. Livingstone, Edinburgh.
levels in relation to underdeveloped lymphatic tissue in HIBBS, J.B., LAMBERT, L.H. & REMINGTON, J.S. (1972)
germ-free guinea pigs. Germfree Research. Biological Possible role of macrophage mediated nonspecific cyto-
Effect of Gnotobiotic Environment (ed. by J. B. Heneghan) toxicity in tumour resistance. Nature: New Biology,
p. 507. Academic Press, New York-London. 235, 48.
CLAMAN, H.N. (1975) How corticosteroids work. 7. Allergy HOWARD, J.G. & MITCHISON, N.A. (1975) Immunological
cdin. Immunol. 55, 145. tolerance. Progr. Allergy 18, 43.
CRAWFORD, T.B. (1958) The distribution of 5-hydroxytrypt- ILLINGWORTH, D.V., PERRY, J.S., ACKLAND, N. & BURTON,
amine in central nervous system of the dog. 5-Hydroxy- A.M. (1973) The maintenance of pregnancy and parturi-
tryptamine (ed. by G. P. Lewis) p. 20. Pergamon Press, tion, in guinea pigs hypophysectomized within 4 days of
Oxford. mating. 31. Endocr. 59, 163.
DAVISON, A.N. (1958) Physiological role of monoaminoxi- IRVINE, W.J. (1974) Autoimmune mechanisms in endocrine
dase. Physiol. Rev. 38, 729. diseases. Proc.'roy. Soc. Med. 67, 449.
DESHPANDE, N., HAYWARD, J.L. & BULBROOK, R.D. (1965) IsAwOviC, K. & JANKovIC, B.D. (1973) Neuro-endocrine
Plasma 17 hydroxycorticosteroids and 17 oxosteroids in correlates of immune response. II. Changes in the lym-
patients with breast cancer and in normal women. 1. phatic organs of brain-lesioned rats. Int. Arch. Allergy, 45,
Endocr. 32, 167. 373.
DESHPANDE, N., JENSEN, V., CARSON, P., BULBROOK, R.D. & JANKOvIC, B.D. & ISAKOVIc, K. (1973) Neuro-endocrine
LEWIS, A.A. (1969) Some aspects of the measurement of correlates of immune response. I. Effects of brain lesions
cortisol production in patients with breast cancer. 3. on antibody production, Arthus reactivity and delayed
Endocr. 45, 571. hypersensitivity in the rat. Int. Arch. Allergy, 45, 360.
DIENER, E. & LEE, K.-C. (1974) B cell tolerance in vitro. The JENSEN, V., DESHPANDE, N., BULBROOK, R.D. & Doous,
fate of the tolerant cell and its control by hormones. T.W. (1968) Adrenal function in breast cancer. Produc-
Immunological Tolerance (ed. by D. H. Katz and B. tion and metabolic clearance of cortisol in patients with
Benacerraf) p. 311. Academic Press, New York. early or advanced breast cancer and in normal women.
DUBOIS, M.P. (1971) Sur l'apparition des secretions hor- J. Endocr. 42, 425.
monales dans hypophyse foetal dans bovins: mise en JERNE, N.K. (1977) The immune system: a web of v-
evidence par immunofluorescence des cellules gonadotrops domains. The Harvey Lectures. (In press.) Academic Press,
thyreotrops. C.R. Acad. Sci. D. (Paris) 272, 1793. New York.
DUSSAULT, J.H., HOBEL, C.J. & FISHER, D.A. (1971) JOST, A. (1969) The extent of foetal endocrine autonomy.
Maternal and fetal thyroxine secretion during pregnancy Foetal autonomy. Ciba Foundation Symposium (ed. by
in the sheep. Endocrinology, 88, 47. G. E. W. Wolstenholme and M. O'Connor) p. 79.
EvANS, R. & ALEXANDER, P. (1972) Mechanism of immuno- Churchill, London.
logically specific killing of tumour cells by macrophages. JOST, A., VIGIER, B., PREPIN, J. & PERCHELLET, J.P. (1973)
Nature (Lond.), 236, 168. Studies on sex differentiation in mammals. Recent Progress
FABRIS, N., PIERPAOLI, W. & SORKIN, E. (1971a) Hormones in Hormone Research, Vol. 29 (ed. by R. 0. Greep). p. 1.
and the immunological capacity. III. The immuno- Academic Press, New York & London.
deficiency disease of the hypopituitary Snell-Bagg dwarf KATZ, D.H. & BENACERRAF, B. (1974) Immunological Toler-
mice. Clin. exp. Immunol. 9, 209. ance. Mechanisms and potential therapeutic applications.
FABRIS, N., PIERPAOLI, W. & SORKIN, E. (1971b) Hormones Academic Press, New York, San Francisco, London.
and the immunological capacity. IV. Restorative effects of KELLER, R. (1973) Cytostatic elimination of syngeneic rat
developmental hormones or of lymphocytes on the tumor cells in vitro by nonspecifically activated macro-
immunodeficiency syndrome of the dwarf mouse. Clin. phages. 3. exp. Med. 138, 625.
exp. Immunol. 9, 227. KESZTYUS, L. (1967) Immunitat und Nervensystem. Akademiai
GALLAGHER, T.F., BRADLOW, H.L., MILLER, D.G., ZUMOFF, Kiad6. Budapest.
B. & HELLMAN, L. (1962) Steroid hormone metabolism in KIEFFER, D.J. & KETCHEL, M.M. (1971) Effects of adrenal-
chronic lymphatic leukemia. 3. cdin. Endocr. Metab. 22, ectomy and antigenic stimulation on spleen weight in
1049. mice. Transplantation, 11, 45.
GALLAGHER, T.F., HELLMAN, L., ZUMOFF, B. & MILLER, LANCET (1974) Editorial. The thyroid, prolactin and breast
D.G. (1965) Steroid hormone metabolism in chronic cancer, p. 908 and Immunological changes in women,
myelogenous leukemia. Blood, 25, 743. p. 909.
GALTON, V.A. & INGBAR, S.H. (1966) Effect of a malignant MACKAY, W.D., EDWARDS, M.H., BULBROOK, R.D. & WANG,
tumor on thyroxine metabolism and thyroid function in D.Y. (1971) Relation between plasma-cortisol, plasma-
the rat. Endocrinology, 79, 964. androgen-sulphates, and immune response in women
GHOSH, P.C., LOCKWOOD, E. & PENNINGTON, G.W. (1973) with breast cancer. Lancet, ii, 1001.
Abnormal excretion of corticosteroid sulphates in patients MAKINODAN, T., SANTOS, G.W. & QUINN, R.P. (1970)
with breast cancer. Brit. med. 7. i, 328. Immunosuppressive drugs. Pharmacol. Rev. 22, 189.
HADDEN, J.W., HADDEN, E.M. & MIDDLETON, E. (1970) MARMORSTON, J. (1966) Urinary hormone metabolite levels
12 H. Besedovsky LIE. Sorkin
in patients with cancer of the breast, prostate and lung. Hormones and the immune response (ed. by G. E. W.
Ann. N.Y. Acad. Sci. 125, 959. Wolstenholme and Julie Knight) p. 126. J. & A. Churchill,
MASON, J.W., BUEsCHER, E.L., BELFER, M.L., MOUGEY, London.
E.H., TAYLOR, E.D., WHERRY, F.E., RICKETTS, P.T., PIERPAOLI, W., HARAN-GHERA, N., BIANCHI, E., MULLER,
YOUNG, P.S., WADE, J., EARLY, D.C. & KENION, C.C. J., MESHORER, A. & BRuE, M. (1974) Endocrine disorders
(1967) Pre-illness hormonal changes in Army recruits as a contributory factor to neoplasia in SJL/J mice.
with acute respiratory infections. Psychosom. Med. 29, 545. 5. nat. Cancer Inst. 53, 731.
METCALF, D. (1960) Adrenal control function in high and PIERPAOLI, W. & SORKIN, E. (1972) Alterations of adrenal
low leukemia strains of mice. Cancer Res. 20, 1347. cortex and thyroid in mice with congenital absence of the
MILLER, J.F.A.P. & OSOBA, D. (1963) The role of the thymus. Nature: New Biology, 238, 282.
thymus in the origin of immunological competence. The PLEASANTS, J.R. (1968) Characteristics of the germ-free
immunologically competent cell, Ciba Foundation Study animal. The Germfree Animal in Research (ed. by Marie E.
Group No. 16 (ed. by G. E. W. Wolstenholme and Julie Coates) p. 113. Academic Press, London, New York.
Knight) p. 62. J. & A. Churchill, London. RAO, L.G.S. & HEWIT, M.L. (1970) Prognostic significance
MIYAKAWA, M. (1966) The morphological characteristics of of a steroid-discriminant function in patients with in-
the germfree mammals, including the results of recent operable lung cancer. Lancet, ii, 1063.
studies on the adrenal in the germfree rats at the depart- SHIGERU, S., TAKEsHI, T., TSUGUO, S., TAKEsHI, T.,
ment of pathology, Nagoya University. I. The morpho- YosHImAsu, O., KEISHI, M. & ToKuICHIRO, S. (1966)
logical characteristics described for the cecum and the Blood corticosterone level of tumor-bearing rats. Gann, 57,
lymphatic tissue in the germfree mammals. International 307.
Congress Microbiol. IX Symposia, p. 291. Ivanorsky Inst. SOLOMON, J.B. (1971) Ontogeny of defined immunity in
Virol. Moscow, Russia. mammals. Foetal and Neonatal Immunology p. 234. North
MIYAKAWA, M. & UKAI, M. (1970) Adrenal gland functions Holland Publishing Company, Amsterdam.
in germfree animals. Jap. 5. clin. Med. 28, 2178. STEIN, M., SCHIAVI, R.C. & CAMERINO, M. (1976) Influence
MOOG, R. & FORD, E. (1957) Influence of exogenous ACTH of brain and behavior on the immune system. Science,
on body weight, adrenal growth, duodenal phosphatase, 191, 435.
and liver glycogen in the chick embryo. Anat. Rec. 128, UDENFRIEND, S., WEISSBACH, H. & BOGDANSKI, D.F. (1957)
592. Increase in tissue serotonine following administration of its
NAVMENKO, E.V. (1973) Central regulation of the pituitary precursors. 5-hydroxytryptamine. 5. biol. Chem. 224, 803.
adrenal complex. Studies in Soviet Science. Consultants VOUGHT, R.L., BROWN, R.A., SIBINOVIE, K.H. & MCDANIEL,
Bureau, New York, London. E.G. (1972) Effect of changing intestinal bacterial flora on
NOMURA, T., OHSAWA, N., KAGEYAMA, K., SAITO, M. & thyroid function in the rat. Horm. Metab. Res. 4, 43.
TAJIMA, Y. (1973) Testicular functions of germfree mice. WAHL, S.M., ALTMANN, L.C. & ROSENSTREICH, D.L. (1975)
Germfree Research (ed. by J. B. Heneghan) p. 515. Inhibition of in vitro lymphokine synthesis by gluco-
Academic Press, New York, London. corticosteroids. 3. Immunol. 115, 476.
PALS, A.J., REINEKE, E.P. & SCHAW, G.H. (1973) Serum WHITE, T. (1966) Histamine in the brain. Handbook of
thyroxine levels in the perinatal guinea pig. Lab. Anim. Experimental Pharmacology, Vol. XVIIIJI, Histamine and
Sci. 23, 511. Anti-Histaminics Part 1. p. 789. Springer-Verlag, Berlin-
PIERPAOLI, W. & BmsEDovwY, H.O. (1975) Role of the Heidelberg, New York.
thymus in programming of neuroendocrine functions. WOSTMANN, B.S. (1968) Defence mechanisms in germfree
Clin. exp. Immunol. 20, 323. animals. Part 1. Humoral defence mechanisms. The
PIERPAOLI, W., FABRIS, N. & SORKIN, E. (1970) Develop- Germfree Animals in Research (ed. by Marie E. Coates)
mental hormones and immunological maturation. p. 197. Academic Press, London, New York.