Background: The neuro-endocrine regulation of immune functions is based on a complex network of interactions. As part of this series of articles, we refer here to immune-sympathetic interactions that are triggered by different types of immune challenge. Summary: We mention the initial hypothesis that led to the proposal that the sympathetic nervous system (SNS) is involved in immunoregulation. We next refer mainly to our initial work performed at a time when most immunologists were concentrated in clarifying aspects of the immune system that are essential for its regulation from within. The first approach was to explore whether immune responses to innocuous antigens and superantigens can elicit changes in the activity of the SNS, and their potential relevance for the regulation of the activity of the immune system. The following step was to explore whether comparable immune-SNS interactions are detected in different models of diseases with immune components, such as parasitic and viral infections and autoimmune pathologies. Key Messages: We pose some general considerations that may at least partially explain seemly discrepant findings, and remark the importance of interpreting immunoregulatory effects of the SNS together with other neuro-endocrine inputs that simultaneously occur when the activity of the immune system changes. Finally, we provide some arguments to re-consider the use of the expression “reflex” in immunology.

This special issue deals with the present state of research and recent progress in our knowledge on the sympathetic control of the immune system. We believe that it might also be of interest, especially for young readers, to get a personal, rather historical, view on how we started thinking on the possibility of the existence of interactions between the immune and sympathetic nervous systems (SNSs) that would lead to immunoregulation. Thus, we have chosen to focus on our early work in this particular field and finalize by commenting some aspects that might explain apparently controversial findings.

Our early work oriented to explore whether neuro-endocrine mechanisms are involved in immunoregulation was performed when most immunologists were concentrated in understanding how the immune system is regulated from “within.” As often in science, it was not easy to deal with an “emerging” field that intended to link disciplines considered as completely independent. In light of our present knowledge, the sophisticated functioning of the immune system was rather primitive at that time. It might be enough to mention as example that the T-cell receptor was not characterized yet neither was it known how immunologic diversity is acquired. The main focus in immunology was on B cells; the idiotypic-anti-idiotypic network was being proposed, and macrophages were considered as some kind of accessory cells of relative low relevance for the development of an immune response. It is therefore not surprising that most of the best-known immunologists were rather reluctant to accept the proposal that the immune system could be also regulated from “without.”

Of course, it is always difficult and risky to define when a given scientific field started and this work would have not been possible without previous knowledge acquired by others. A detailed review by Elenkov and colleagues, which includes a large list of early references until 2000, is strongly recommended to the reader [1].

Our initial hypothesis was that, if the immune response is subject to sympathetic regulation, there should be an active exchange of signals between the immune system and the SNS. This implies that immune activation should elicit autonomic responses capable of affecting the intensity and course of the immune response, and that this interaction should occur under conditions in which one system does not harm the other. Thus, it was also necessary to deal with the confounding factor that, under natural conditions, immune responses are often associated with tissue damage and altered organ functions, and that illness, as stressor, can per se elicit neuro-endocrine responses. To circumvent these issues, we chose to immunize animals with innocuous antigens that can elicit a strong adaptive immune response without causing disease. A classical model used by immunologists was to immunize animals with sheep red blood cells (SRBC), which permits to include controls that received the same number of syngeneic red blood cells. Based on this hypothesis, we studied whether the content of noradrenaline (NA), the main sympathetic neurotransmitter, was affected in the spleen of rats immunized with SRBC. Shortly preceding the peak of the specific antibody response to this antigen (3–4 days after injection), splenic NA levels were reduced by 40–50%, without changes in spleen weight [2]. At later time points, although the spleen was enlarged in immunologically high responder rats, NA levels were still significantly decreased regardless of whether the results are expressed as concentration or per total spleen [3]. This is important because it excludes an eventual dilution of NA concentration due to an increased spleen weight. In successive studies, we found that the duration of the decrease in splenic NA content depends on the intensity of the immune response, which is more prolonged in immunologically high responder animals [3]. This is most likely due to the markedly reduced (about 70%) turnover rate in the synthesis of the neurotransmitter in the spleen of SRBC-immunized rats [4]. In addition to direct effects on splenic immune cells caused by this decrease, another consequence of reduced sympathetic activity would be a decreased splenic vascular tonus, which would lead to increased blood flow and influence the contact of immune cells with circulating antigens [5].

Another model that we used was to explore if prolonged exposure to non-pathogenic environmental antigens could elicit changes in SNS activity. For this purpose, we compared NA concentration and content in lymphoid organs of adult specific-pathogen-free (SPF) and germ-free (GF) rats [6]. GF animals are completely free of microbiota, including that typically found in the gut, lungs, mouth, and other mucosa, and lack a systemic response to this flora [7]. SPF animals are free of a specific list of pathogens and are raised under strict procedures to prevent the introduction of pathogens.

As known since a long time [8], we confirmed that IgG blood levels were several-fold higher in SPF than in GF in the same animals used for catecholamine determinations. The spleen, thymus, and lymph nodes of SPF rats contained about 50% less NA than GF animals, but no differences were observed in the stomach and segments of the intestine without Peyer patches. These results indicate that a permanent reduction of NA levels is not compensated by local changes in tyrosine hydroxylase activity, the rate-limiting enzyme in NA synthesis, or an eventual increased production of catecholamines by lymphoid cells. Thus, a most likely possibility is that the reduced NA content in lymphoid organs of SPF rats is a consequence of their permanent contact with non-pathogenic environmental antigens. A central involvement is indicated by the fact that the NA and adrenaline (A) content in the adrenal glands, which are not in direct contact with lymphoid cells, is also reduced in SPF rats as compared to GF.

Another approach to explore whether immune cells can affect sympathetic nerve activity in lymphoid organs was to study the consequence of a permanent absence of T cells. For this purpose, we investigated the development of splenic sympathetic innervation in athymic, nude Foxn1n (nu/nu) mice at different times during ontogeny. While no differences were detected at birth, higher splenic NA levels were found in 7-, 14-, and 21-day-old athymic mice as compared to their normal thymus-bearing littermates Foxn1n/Foxn1+ (nu/+). Thymus transplantation or thymocyte inoculation into newborn nude mice resulted in splenic NA levels comparable to those of normal nu/+ mice. Histochemical studies fully confirmed these differences [9]. In a later work, we found that, in parallel to splenic sympathetic alterations, the concentration of brain-derived neurotrophic factor was significantly higher in the spleen and hypothalamus of athymic mice as compared to heterozygous nu/+ controls [10]. This finding may contribute to explain the increased splenic innervation of T-cell-deficient mice. All alterations in nu/nu mice were abrogated by thymus transplantation at birth. These results suggest that T cells or their products can induce (1) a decrease in the number and activity in splenic sympathetic nerve fibers; (2) a decrease in NA content in the hypothalamus, which, in turn, may influence the descending neural pathways associated with the SNS; and (3) changes in neurotrophin concentration in the spleen and hypothalamus. Taken as a whole, the results commented above led us to conclude that:

  • The specific immune response to innocuous antigens elicits a decrease in splenic sympathetic activity, a finding that allows to dissociate changes in the SNS induced by the immune response itself from eventual effects of a disease.

  • Permanent exposure to non-pathogenic environmental microorganisms causes a decrease in catecholamine content in the spleen, thymus, lymph nodes, and adrenal glands.

  • The absence of T lymphocytes results in higher NA content and increased sympathetic innervation in the spleen, effects that are reversible by T-cell inoculation or thymus transplantation at birth.

Based on these conclusions, the next step was to evaluate the effect of endogenous NA on the immune response. We approached this aspect by exploring the effect of local surgical denervation of the spleen and general permanent depletion of sympathetic fibers using the neurotoxin 6-hydroxy-dopamine (6-OHDA) on the immune response to SRBC [2]. The number of cells producing specific antibodies to SRBC in the spleen of rats in which the organ was locally denervated was increased by about 70% as compared to sham-operated animals. In rats that were chemically denervated at birth with 6-OHDA and immunized with SRBC at adulthood, the number of B cells producing specific anti-SRBC antibodies was slightly increased. However, when chemical denervation at birth was combined with adrenalectomy during adulthood, the number of these cells was more than doubled.

Although a decrease in splenic NA concentration during the immune response was observed in all models that we had used, not all results reported later agree on the immunological consequences of this sympathetic response. In our view, these discrepancies are in many cases only apparent when the models and procedures used by different authors are carefully analyzed. Among the variables that may serve to explain the different results are (1) when animals were denervated in relation to the injection of the antigen (before, simultaneously, or after); (2) the type of denervation (local or systemic); (3) the time when and which sympathomimetics or antagonists were injected (also it is not the same to inject NA or specific agonists); (4) whether it is possible that the effect observed is direct or indirect; (5) if the read-out of the outcome of an effect evaluates a specific or an innate immune response; (6) the type and subtype of adrenergic receptor targeted; (7) if immune stimulation resulted in disease and, if so, which type; (8) in which species, at which age, and in which sex the experiments were performed. Even the impact that circadian catecholamine rhythms have on lymphoid cell circulation, recruitment, and redistribution has to be considered [11, 12].

We mention only a few selected examples to illustrate this point. As described, we found that local surgical denervation of the spleen in adult rats that were immunized with SRBC 5 days after the operation resulted in an increased specific immune response [2]. Some years later, Hall and colleagues reported the opposite results [13]. However, in these experiments, general sympathectomy was induced chemically by 6-OHDA administered 48 and 24 h before SRBC injection into adult mice, a procedure that resulted in a decreased specific immune response. At least part of the discrepancy could be later explained by the demonstration that 6-OHDA has a direct inhibitory effect on immune cells [14], and that this neurotoxin induces an initial liberation of NA and an increase in corticosterone levels [15].

Another example derives from a more recent report showing that sympathectomy can result in opposite effects when immune cells are stimulated by mitogens or by specific antigens. While the concentration of IL-2 and IFN-γ is increased in the supernatant of cultures of spleen cells obtained from denervated animals in response to antigen-specific (keyhole limpet hemocyanin; KLH) stimulation, it is decreased in response to the T-cell mitogen concanavalin A, as compared with the amounts produced by spleen cells from control mice [16]. This finding also may explain seemly contradictory results using KLH as antigen but combined with lipopolysaccharide (LPS) [17].

Other variables such as the level of expression of adrenergic receptors in different lymphoid cell populations at different states of activation also serve to illustrate this point. For example, β2-adrenergic receptor expression is increased in anti-CD3-activated murine Th1 cells, but not in activated Th2 cells [18], and human peripheral memory CD8+ T cells express significantly higher level of β2-receptors than naive cells [19]. These patterns of receptor expression result in a distinctive sensitivity to NA and therefore in a preferential deviation of the immune response toward a particular type.

An example at molecular levels is the report that β-adrenergic receptor signaling in immune cells follows the cAMP-PKA pathway and predominantly leads to immune suppression. However, this effect can be desensitized and inflammation can be promoted if the signal follows another non-canonical intracellular pathway via β-arrestin 1 and 2 [20, 21].

After our initial findings, we decided to continue the study of immune-SNS interactions using another type of immune challenge. In the eighties, immunologists started to be very interested in the particular type of immune response induced by superantigens because they can elicit a T-cell response without antigen processing and presentation. Superantigens are classified as exogenous, which are derived from Gram+ bacteria, such as staphylococcal enterotoxins, toxic shock syndrome toxin, mycoplasma arthritis, and streptococcal pyrogenic exotoxins, and endogenous, which are cell membrane proteins encoded by certain viruses that infect mammalian cells, such as the mammary tumor virus. The in vivo response to superantigens does not necessarily culminate in immune effector functions, but may induce clonal unresponsiveness. By binding to specific Vβ chains of the T-cell receptor, superantigens can activate up to 30% of the mature T-cell pool and induce an extensive proliferation that ends in apoptosis of the stimulated cells and to a state of anergy of the remaining cells (for review [22]). It has been shown later that superantigens also stimulate B cells by binding to complementary determining regions [23].

We have used as model the superantigen staphylococcal enterotoxin B (SEB), which induces a significant in vivo stimulation and clonal expansion of splenic Vβ8+T cells 2 days after injection. This is later followed by a substantial decrease in the number of these T cells, and the remaining ones are unresponsive when re-challenged with the superantigen in vitro (anergy). IL-2 concentration in plasma, as well as its production upon re-exposure of spleen cells to SEB in vitro, is markedly increased 2 h after in vivosuperantigen administration and strongly decreases on the following days [22]. We confirmed that all these immunological changes with the same kinetics were observed in the same animals used to evaluate several non-immune parameters in SEB-injected mice [24]. Among them, we found that this type of immune response is paralleled by biphasic changes in splenic sympathetic activity [24]. NA concentration in the spleen was significantly increased 2 h after SEB injection, but reduced by about 60% on day 2, as compared to simultaneously vehicle-injected controls. On day 4, at an early stage of the anergic phase, the decrease in the percentage of splenic CD4Vβ8 cells was paralleled by a marked decrease in splenic NA concentration. No comparable changes were detected in the kidney, which is not affected by SEB and was used as a control organ. Although the concentration of corticosterone, NA, and A in plasma was also markedly increased 2 h after SEB injection, their levels remained within the normal range thereafter. These studies provided the first evidence of an immunoregulatory cross-talk between sympathetic nerves and superantigen-activated immune cells, and indicated that the sympathetic response induced by SEB may have immunoregulatory implications.

To explore whether sympathetic innervation has any biological relevance for this particular type of immune response, the experiments were repeated in mice that had been treated with 6-OHDA 5 days before SEB injection. As expected, splenic NA concentration was decreased by more than 93% in denervated animals at the time of SEB injection. SEB-induced spleen cell proliferation and IL-2 production in vitro and IL-2 concentration in blood were significantly decreased in sympathetically denervated mice primed with the superantigen in vivo as compared to intact SEB-injected mice. The clonal deletion of CD4Vβ8 that occurs in the spleen of SEB-injected intact animals is not observed in animals that had been chemically sympathectomized prior to the administration of the superantigen, but rather the relative percentage of these cells is significantly increased.

Since apoptosis is considered the main process that leads to SEB-induced clonal deletion of proliferating specific Vβ cells, our results led us to study if NA can induce apoptosis in lymphoid cells, a possibility that, in fact, we confirmed [25]. In further studies, we found that regulatory T (Treg) cells are among the targets of NA-induced apoptosis, and that NA reduces the frequency of Foxp3+ cells and Foxp3 mRNA expression, the transcription factor necessary to induce a regulatory phenotype in murine T cells. These effects are mediated by β2-adrenergic receptors in a concentration- and time-dependent manner [26]. In agreement with this finding, chemical sympathectomy significantly increased the percentage of Treg cells, and a partial recovery in the splenic concentration of the neurotransmitter led to Treg percentages comparable to those of intact controls. Furthermore, the concentration of splenic NA negatively correlated with the frequency of CD4+Foxp3+ Tregs. Interestingly, single Foxp3+ Tregs were localized in the proximity of NA-producing nerve fibers. Taken together, these results indicate that the SNS can also play an immunomodulatory role by decreasing the Treg population, an effect that might be important for the control of immunological tolerance by the nervous system.

Breneman and colleagues predicted that, based on our findings that splenic NA concentration is inversely correlated with the degree of immunological activity [6] and that the magnitude of the reduction is proportional to the intensity of the immune response [3], a decrease in NA levels might be anticipated in animals with high immunological activity [27]. They confirmed this prediction by showing that noradrenergic innervation and splenic NA content are decreased in MRL lpr/lpr mice, which develop an autoimmune lymphoproliferative disease that resembles systemic lupus erythematosus (SLE) in humans.

As mentioned above in our studies with the superantigen, we found that NA can induce lymphoid cell apoptosis, and that the mechanism involved is Fas-independent [28]. These results also prompted us to study SNS-immune interactions in lpr/lpr mice, which lack functional Fas (CD95) expression and are resistant to Fas ligand (CD178)-mediated apoptosis, a critical mechanism for the maintenance of peripheral tolerance. Since, although at a much later age, the congenic MRL +/+ strain used by Breneman and colleagues as control also develops an autoimmune disease, we used lpr/lpr mice on a C57 background to evaluate NA concentration in the spleen and kidney of male and female from birth until animals were 1 year old, in parallel to several parameters that indicate the progression of the disease in autoimmune mice. Early in ontogeny (less than 24 h old), the concentration of NA was significantly increased in the spleen of C57Bl lpr/lpr male mice compared with normal C57Bl littermates. However, splenic sympathetic innervation gradually and markedly declined in male and female lpr/lpr mice as the disease progressed [25]. No differences between lpr/lpr and congenic controls were detected in the kidney. Immunohistochemical studies detecting tyrosine hydroxylase confirmed the profoundly decreased noradrenergic splenic innervation in adult lpr/lpr mice. As known, adult lpr/lpr mice have markedly increased IgM blood levels. While splenic NA concentration positively correlated with IgM blood levels in normal mice of all ages studied, there was an inverse correlation in lpr/lpr mice when the disease was overtly manifested.

To study the relevance of the sympathetic alteration for the progression of the disease, we experimentally advanced the loss of noradrenergic fibers that occurred naturally during adult life in lpr/lpr mice by performing neonatal sympathectomy. The concentration of IgM and IgG2a in blood was markedly higher in denervated lpr/lpr mice than in intact lpr/lpr mice, and the appearance of lymphadenopathy was accelerated. Furthermore, although neonatal denervation did not affect the life span of normal animals, it shortened significantly the survival time of lpr/lpr mice [25]. 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 supports the hypothesis that the SNS can modulate the expression of lymphoproliferative diseases.

In collaboration with other colleagues, we also used another model of autoimmune disease to explore the status of sympathetic innervation, namely, induction of experimental arthritis in susceptible Dark Agouti (DA) rats by injection of collagen in Freund’s incomplete adjuvant [29]. A significant reduction in the density of sympathetic nerve fibers in the proximity of the joints was observed shortly after the appearance of symptomatic arthritis, confirming previous results in mice and in patients with rheumatoid arthritis [30, 31]. No comparable changes were observed in collagen-injected Piebald Virol Glaxo (PVG) rats, which are resistant to the induction of arthritis, experimentally induced autoimmune encephalomyelitis, and thyroiditis [32]. Our studies further showed that the gradual loss of noradrenergic fibers in the joints of arthritic rats is paralleled by increased hypothalamic NA content, that there is a dissociation between hypothalamic cytokine gene expression and noradrenergic activity in the brain, and that there is a lack of sustained stimulation of the stress axes [29]. As a whole, the results indicate a disruption in the communication between afferent immune messages to the central nervous system and main efferent anti-inflammatory pathways under control of the brain during collagen-induced arthritis. It is interesting to add that Baerwald and colleagues [33] had previously reported that CD8+ T cells obtained from the blood and synovial fluid of patients with rheumatoid arthritis express less β2-adrenergic receptors than healthy controls.

To extend our studies on SNS-immune interactions to other types of pathologies, we used mice infected with Trypanosoma cruzi, the parasite that causes American trypanosomiasis (Chagas disease) in humans and is transmitted by an infected insect. About 6–7 million people worldwide are estimated to be infected with this parasite [34], and around 70 million people are at risk of contracting this disease [35].

An acute disease can be triggered in mice by subcutaneous injection of few parasites. We have used the Tulahuén strain of T. cruzi injected subcutaneously into C57BL/6 J mice [36]. A sexual dimorphism was observed, males being more susceptible than females. All males were dead 30 days after inoculation of the parasite. Around 50% of the infected females died around the same time, but for reasons that we could still not clarify, the rest survive the infection. Profound thymic atrophy, with preferential loss of CD4+CD8+ cells, parasitemia, which is considerably larger in males than in females, amastigote nests and inflammatory foci in the heart, a large splenomegaly, and increased blood levels of IFN-γ, TNF-α, IL-1β, and IL-10 were observed in all infected mice. A marked decrease in splenic NA concentration and content, despite a 6–7-fold increase in the weight of the spleens of infected mice, was detected 17 days after inoculation of the parasite. This decrease was observed in males and females, but it was proportionally more marked in males. An extreme reduction in sympathetic nerve fibers in the spleen of infected mice was confirmed by immunohistochemical fluorescent detection of tyrosine hydroxylase. No difference between control and infected mice was observed in NA and A concentration in blood.

As in the model of autoimmune lymphoproliferative disease described above, we studied the effect of advancing the spontaneous denervation of infected mice by inoculating the parasite into adult mice that had been denervated at birth. Chemical denervation did not significantly affect the concentration of specific IgM and IgG2a antibodies to T. cruzi, as compared with non-denervated mice, neither worsened myocarditis, but resulted in increased parasitemia and IL-6 and IFN-γ blood levels in infected male and female mice. Perhaps more important was the fact that, although denervation per se does not affect the life span of normal mice, all denervated, infected males died before non-denervated infected animals, and all denervated infected females were dead 30 days after infection.

The results illustrate the relevance of the SNS for the outcome of a parasitic infection. Furthermore, they indicate that the residual sympathetic activity in lymphoid organs of infected mice can still contribute to modulate the immune response to the parasite, as shown by increased mortality and production of proinflammatory cytokines in sympathetically denervated, infected mice.

We have also performed another study using this model of parasite infection to analyze the contribution of adaptive immunity to these responses by comparing several immune and non-immune parameters in recombinase activator gene 1 (RAG-1)-deficient mice, which lack mature B and T lymphocytes, and normal C57Bl/6 mice [37]. Higher parasitemia, increased IL-1β and IL-6 blood levels, less marked changes in the weight of lymphoid organs, no cardiomegaly, and earlier mortality were observed in RAG-1-deficient following infection as compared with normal, infected mice. As expected from our previous results in mice lacking mature T cells, basal NA concentration was higher in the spleen of non-infected RAG-1-deficient than in C57Bl/6 mice. However, splenic NA concentration and content were decreased in both strains after infection. These results indicate that cells other than mature T or B cells or their products can contribute to affect SNS activity during infection with this parasite.

Also in collaboration with other colleagues, we investigated the role of the SNS during a viral infection [38]. The model used was inoculation of Friend virus, a single-stranded RNA retrovirus, in C57Bl mice, which are less susceptible to this virus than other mouse strains and recover from the infection. The spleen is a main reservoir of the Friend virus, and infected splenic cells were detected 2 days after inoculation, peaked after 5 days, and decreased thereafter. NA concentration and content were transiently but massively decreased in the spleen of infected mice 10 days after infection. Although the levels of the neurotransmitter tended to recover, they did not reach those of non-infected controls until the end of the studies on day 25, when the viral load was also markedly decreased but the virus was still present in this lymphoid organ. The decrease in NA levels was paralleled by a significant upregulation of the expression of monoamine oxidase A and catechol-O-methyltransferase in the spleen. Inhibition of these enzymes, which blocked NA degradation and prevented the NA decrease induced by the virus, significantly reduced the viral load. In agreement with these findings, chemical sympathectomy prior to inoculation of the virus aggravated the acute infection and extended the duration of the disease. These findings demonstrate that the SNS also plays an important role during viral infections.

In summary, the results described indicate that the SNS plays a protective role during the diseases studied. However, SNS effects are dampened during several pathological conditions because NA fibers in immune organs are reduced, as shown in the animal models of parasitic and viral infections and autoimmune disease used. Comparable observations have been reported in humans during terminal stages of sepsis [39] and certain inflammatory diseases (e.g., arthritis [31]). There is also evidence that sympathetic effects cannot be fully expressed due to alterations in the capacity to increase the expression of β2-adrenergic binding sites in lymphoid cells in patients with SLE [40]. The models and results summarized in previous sections are schematically represented in Figure 1.

Fig. 1.

Different types of immune stimulation elicit a decrease in splenic NA concentration. The figure schematically represents changes in splenic NA concentration (c) observed in parallel to the immune response (b) elicited by a given antigenic stimulation or immune status (a). No absolute values are given for these parameters. Dashed lines indicate basal NA levels. The original data from which these diagrams derive can be found in the references indicated in square brackets. The immune response, triggered either by an innocuous, conventional antigen [2] or a superantigen [24], or by environmental non-pathogenic microorganisms [6], elicits a decrease in NA levels in the spleen. This decrease is more marked in immunologically high responder animals [3] and is also observed in athymic mice following T-cell inoculation or thymus implantation [9, 10]. Decreased splenic NA concentration is also observed in models of pathologies that involve immune activation, such as an autoimmune lymphoproliferative disease [25], and a parasitic [36] or viral [38] infection.

Fig. 1.

Different types of immune stimulation elicit a decrease in splenic NA concentration. The figure schematically represents changes in splenic NA concentration (c) observed in parallel to the immune response (b) elicited by a given antigenic stimulation or immune status (a). No absolute values are given for these parameters. Dashed lines indicate basal NA levels. The original data from which these diagrams derive can be found in the references indicated in square brackets. The immune response, triggered either by an innocuous, conventional antigen [2] or a superantigen [24], or by environmental non-pathogenic microorganisms [6], elicits a decrease in NA levels in the spleen. This decrease is more marked in immunologically high responder animals [3] and is also observed in athymic mice following T-cell inoculation or thymus implantation [9, 10]. Decreased splenic NA concentration is also observed in models of pathologies that involve immune activation, such as an autoimmune lymphoproliferative disease [25], and a parasitic [36] or viral [38] infection.

Close modal

The previous sections summarize mainly our studies showing that stimulation of the immune system elicits changes in sympathetic nerve activity that can, in turn, affect immune functions. Using different models, we also describe that disturbances in regulatory sympathetic signals can contribute to pathology. At this point, we would like to comment on a couple of other aspects that we consider also important. We want to remark that, as in previous sections, and because the literature is at present so extensive, only few examples are quoted to illustrate these aspects, and many relevant studies are not included here.

We have discussed above some possible reasons that might explain apparent discrepancies in the results reported on the outcome of SNS effects on an immune response. A first consideration is that the situation is even more complex when immune-SNS interactions are studied in models of disease. For example, Dimopoulos and colleagues showed that pretreatment of rabbits with a β1-AR antagonist before infection with the Gram-negative bacterium Pseudomonas aeruginosa improves survival and decreases bacterial load [41]. These results agree with a previous report from Straub and colleagues [42], who also showed that chemical sympathectomy prior to infection of mice with the same bacterium decreases its dissemination. However, when denervated mice were infected with Staphylococcus aureus, a Gram-positive bacterium, the opposite results were obtained; namely, sympathectomy resulted in increased bacterial burden. The results might be explained by the different effects of the SNS on the particular pattern of cytokine expression induced by Gram-positive and Gram-negative bacteria [42]. These examples indicate that the SNS may play a differential role depending on the immune mechanisms involved in the control of the invading pathogen. An interesting review by Bucsek and colleagues discusses this aspect, including not only bacterial infections but also viral and parasitic infections, and autoimmune diseases, such as multiple sclerosis, SLE, and inflammatory bowel disease [43]. The authors conclude that, in general, adrenergic stimulation, particularly when it is chronic, favors the progression of infections by suppressing immune functions, while a deficient adrenergic input contributes to autoimmune diseases. It is beyond the aim of this article to review the extensive literature available in this subject. Rather, the scope is to point out the importance of a careful analysis of the models used to draw conclusions about the role of the SNS in immunoregulation.

A further aspect that should be considered is that the SNS does not operate isolated from other neuro-endocrine systems that are also affected when the activity of the immune system changes. For example, our early studies showed that the immune response to the same innocuous antigen (SRBC) also induces a decrease in NA turnover rate in the brain [44], affects the electrical activity of hypothalamic neurons [45], and elicits an increase in endogenous corticosterone and biphasic changes in thyroxine blood levels [46]. Thus, the decrease in NA concentration in lymphoid organs that were observed following SRBC immunization coexists with all these changes. A comparable situation concerning glucocorticoids and the SNS is provided by the superantigen stimulation model. It has been reported that injection of the superantigen SEB results in an acute and short-lasting increase in corticosterone blood levels [47]. Besides confirming this result, we have shown, as already commented, that the immune response to SEB also elicits biphasic changes in splenic NA concentration [24].

Well-known actions of the combination of neural and endocrine agents emphasize the importance of considering effects of the SNS on immunity within an integrative context that ponders the impact of other physiological systems. The classical example that glucocorticoids can affect the activity of tyrosine hydroxylase can illustrate this statement. It is also known that glucocorticoids can affect the expression of adrenergic receptors in many different cell types, including lymphocytes [48]. Cooperative effects of glucocorticoids and catecholamines favoring type 2 cytokine expression have also been described [49]. In general, glucocorticoids have also permissive effects on NA actions by inhibiting catecholamine-degrading enzymes and catecholamine reuptake, or by increasing the binding capacity of β-adrenergic receptors. However, glucocorticoids can also inhibit some sympathetic functions under certain conditions. For example, endogenous glucocorticoids decrease catecholamine synthesis and release during immobilization-induced stress [50]. References to these and many other related findings can be found in an extensive review by Sapolsky and colleagues [51]. The effect of other hormones and other components, e.g., co-transmitters present in sympathetic fibers that are also released upon sympathetic nerve stimulation, such as ATP and neuropeptide Y [52], and the involvement of afferent signals [53], should be added to this complex scenario.

A third consideration is that the immune response during infection is conditioned by strategies of certain pathogens to survive, some of which are related to their preference to home in particular niches and/or organs with particular innervation characteristics. Interwoven with all the variables mentioned, other conditions, such as superimposed effects of stress, which can also differ depending on whether it is acute or chronic, sexual dimorphism, and inter-species differences can influence the outcome of bi-directional sympathetic-immune interactions and, in fact, the whole network of immune-neuro-endocrine interactions. It is clear that this complexity does not only apply to infectious diseases, but also apply to other pathologies that involve the participation of the immune system, such as autoimmune and inflammatory “sterile” diseases, and even cancer. Thus, it is practically impossible to predict the final effect of a single component of the network, e.g., the SNS, isolated from other inputs.

Finally, a more general and conceptual consideration is related to the expression “sympathetic-immune reflex” that we have used to term these interactions when we started investigating them [2]. At that time, we proposed that the change in NA concentration during the immune response may represent the efferent limb of a reflex mechanism triggered by antigen-stimulated cells. We are considering to abandon this expression because it might be rather reductionist. The analogy between the host response to immune activation or inflammation and a neural reflex is in our view not strictly correct according to the accepted concept in classical physiology.

Simple neural reflexes are in general acute responses that happen very quickly (seconds to minutes, such as axon and postural reflexes, to a few hours, such as digestive reflexes), and they operate as the occasion demands. Many of the neuro-endocrine responses that are observed during specific immune responses, including changes in SNS activity, persist for days. Besides corticosterone and catecholamines, these responses include changes in other neuro-endocrine mediators, and, among other effects, the immune response can also alter cytokine expression in the brain and cause behavioral alterations and changes in intermediate metabolism, all of which elicit a complex counterregulatory response integrated by the central nervous system. These observations also apply to inflammation. Several papers that use the concept of “reflex” as expression of an immunoregulatory effect of autonomic nerves have used as model administration of the endotoxin LPS. However, LPS inoculation does not mimic an infection with Gram-negative bacteria because, under natural conditions, the concentration of this endotoxin increases as the pathogen proliferates in the host during the incubation and the overt expression of the disease. Furthermore, as in the case of an antigenic challenge, LPS administration induces many changes at peripheral, central, and metabolic levels that are coordinated by the brain. In addition, it has also to be considered that inflammation, even if acute, is a process that lasts days and it frequently includes also wound repair. Thus, in view of the present knowledge, it might be better to return to the expression “autonomic immunoregulation” as part of an immune-neuro-endocrine network of interactions.

In conclusion, neuro-endocrine immunoregulatory responses, including those of the SNS, need to be evaluated during each particular pathology. This postulation implies that the immunoregulatory outcome of a host response might be different depending on the present conditions and even on the past history of each individual. We believe that future research in this field should be done using this integrative criterium. The simultaneous evaluation of neuro-endocrine changes during the development of severe diseases would hopefully provide better bases for therapeutic interventions. Although this might seem too ambitious, this knowledge would reflect the actual status of the patient and could lead to a more personalized medicine. Fortunately, such an integrative criterium is today feasible due to the availability of technical resources that allow to detect deviations from heath at multiple levels.

The authors have no conflicts of interest to declare.

The work of the authors quoted in this contribution has been supported by different grants from the Swiss National Science Foundation and the Deutsche Forschungsgemeinschaft.

H. Besedovsky conceptualized and wrote the first draft of this manuscript and collected the work published by other groups. A. del Rey contributed to write the draft, edited the final version, and designed the figure.

1.
Elenkov
IJ
,
Wilder
RL
,
Chrousos
GP
,
Vizi
ES
.
The sympathetic nerve: an integrative interface between two supersystems: the brain and the immune system
.
Pharmacol Rev
.
2000
;
52
(
4
):
595
638
.
2.
Besedovsky
HO
,
del Rey
A
,
Sorkin
E
,
Da Prada
M
,
Keller
HH
.
Immunoregulation mediated by the sympathetic nervous system
.
Cell Immunol
.
1979
;
48
(
2
):
346
55
.
3.
del Rey
A
,
Besedovsky
HO
,
Sorkin
E
,
Da Prada
M
,
Bondiolotti
GP
.
Sympathetic immunoregulation: difference between high- and low-responder animals
.
Am J Physiol
.
1982
242
1
R30
3
.
4.
Besedovsky
HO
,
del Rey
AE
,
Sorkin
E
.
Immune-neuroendocrine interactions
.
J Immunol
.
1985
135
2
750s
4s
.
5.
Rogausch
H
,
del Rey
A
,
Kabiersch
A
,
Besedovsky
HO
.
Interleukin-1 increases splenic blood flow by affecting the sympathetic vasoconstrictor tonus
.
Am J Physiol
.
1995
268
4 Pt 2
R902
8
.
6.
del Rey
A
,
Besedovsky
HO
,
Sorkin
E
,
da Prada
M
,
Arrenbrecht
S
.
Immunoregulation mediated by the sympathetic nervous system, II
.
Cell Immunol
.
1981
;
63
(
2
):
329
34
.
7.
Koopman
JP
.
Germ-free animals
.
Encyclopedia of immunology
2nd ed
Elsevier
.
1998
. p.
990
2
.
8.
Gustafsson
BE
,
Laurell
CBG
.
G Gamma globulins in germ-free rats
.
J Exp Med
.
1958
;
108
(
2
):
251
8
.
9.
Besedovsky
HO
,
del Rey
A
,
Sorkin
E
,
Burri
R
,
Honegger
CG
,
Schlumpf
M
et al
.
T lymphocytes affect the development of sympathetic innervation of mouse spleen
.
Brain Behav Immun
.
1987
;
1
(
2
):
185
93
.
10.
Jouda
J
,
Wildmann
J
,
Schäfer
M
,
Roggero
E
,
Besedovsky
HO
,
del Rey
A
.
T cells affect central and peripheral noradrenergic mechanisms and neurotrophin concentration in the spleen and hypothalamus
.
Ann N Y Acad Sci
.
2012
;
1261
:
18
25
.
11.
Dimitrov
S
,
Benedict
C
,
Heutling
D
,
Westermann
J
,
Born
J
,
Lange
T
.
Cortisol and epinephrine control opposing circadian rhythms in T cell subsets
.
Blood
.
2009
;
113
(
21
):
5134
43
.
12.
Scheiermann
C
,
Kunisaki
Y
,
Lucas
D
,
Chow
A
,
Jang
JE
,
Zhang
D
et al
.
Adrenergic nerves govern circadian leukocyte recruitment to tissues
.
Immunity
.
2012
;
37
(
2
):
290
301
.
13.
Hall
NR
,
McClure
JE
,
Hu
SK
,
Tare
NS
,
Seals
CM
,
Goldstein
AL
.
Effects of 6-hydroxydopamine upon primary and secondary thymus dependent immune responses
.
Immunopharmacology
.
1982
;
5
(
1
):
39
48
.
14.
Josefsson
E
,
Månsson
JE
,
Blennow
K
,
Tarkowski
A
.
Immunomodulating and anti-inflammatory properties of the sympatholytic compound 6-hydroxydopamine
.
J Neuroimmunol
.
1994
;
55
(
2
):
161
9
.
15.
Kruszewska
B
,
Felten
SY
,
Moynihan
JA
.
Alterations in cytokine and antibody production following chemical sympathectomy in two strains of mice
.
J Immunol
.
1995
;
155
(
10
):
4613
20
.
16.
Callahan
TA
,
Moynihan
JA
.
Contrasting pattern of cytokines in antigen- versus mitogen-stimulated splenocyte cultures from chemically denervated mice
.
Brain Behav Immun
.
2002
;
16
(
6
):
764
73
.
17.
Zhang
X
,
Lei
B
,
Yuan
Y
,
Zhang
L
,
Hu
L
,
Jin
S
et al
.
Brain control of humoral immune responses amenable to behavioural modulation
.
Nature
.
2020
;
581
(
7807
):
204
8
.
18.
Ramer-Quinn
DS
,
Baker
RA
,
Sanders
VM
.
Activated T helper 1 and T helper 2 cells differentially express the beta-2-adrenergic receptor: a mechanism for selective modulation of T helper 1 cell cytokine production
.
J Immunol
.
1997
;
159
(
10
):
4857
67
.
19.
Slota
C
,
Shi
A
,
Chen
G
,
Bevans
M
,
Weng
NP
.
Norepinephrine preferentially modulates memory CD8 T cell function inducing inflammatory cytokine production and reducing proliferation in response to activation
.
Brain Behav Immun
.
2015
;
46
:
168
79
.
20.
Chhatar
S
,
Lal
G
.
Role of adrenergic receptor signalling in neuroimmune communication
.
Curr Res Immunol
.
2021
;
2
:
202
17
.
21.
Freedman
NJ
,
Shenoy
SK
.
Regulation of inflammation by beta-arrestins: not just receptor tales
.
Cell Signal
.
2018
;
41
:
41
5
.
22.
Herman
A
,
Kappler
JW
,
Marrack
P
,
Pullen
AM
.
Superantigens: mechanism of T-cell stimulation and role in immune responses
.
Annu Rev Immunol
.
1991
;
9
:
745
72
.
23.
Deacy
AM
,
Gan
SK
,
Derrick
JP
.
Superantigen recognition and interactions: functions, mechanisms and applications
.
Front Immunol
.
2021
;
12
:
731845
.
24.
del Rey
A
,
Kabiersch
A
,
Petzoldt
S
,
Besedovsky
HO
.
Involvement of noradrenergic nerves in the activation and clonal deletion of T cells stimulated by superantigen in vivo
.
J Neuroimmunol
.
2002
127
1–2
44
53
.
25.
del Rey
A
,
Roggero
E
,
Kabiersch
A
,
Schäfer
M
,
Besedovsky
HO
.
The role of noradrenergic nerves in the development of the lymphoproliferative disease in Fas-deficient, lpr/lpr mice
.
J Immunol
.
2006
;
176
(
11
):
7079
86
.
26.
Wirth
T
,
Westendorf
AM
,
Bloemker
D
,
Wildmann
J
,
Engler
H
,
Mollerus
S
et al
.
The sympathetic nervous system modulates CD4(+)Foxp3(+) regulatory T cells via noradrenaline-dependent apoptosis in a murine model of lymphoproliferative disease
.
Brain Behav Immun
.
2014
;
38
:
100
10
.
27.
Breneman
SM
,
Moynihan
JA
,
Grota
LJ
,
Felten
DL
,
Felten
SY
.
Splenic norepinephrine is decreased in MRL-lpr/lpr mice
.
Brain Behav Immun
.
1993
;
7
(
2
):
135
43
.
28.
del Rey
A
,
Kabiersch
A
,
Petzoldt
S
,
Besedovsky
HO
.
Sympathetic abnormalities during autoimmune processes: potential relevance of noradrenaline-induced apoptosis
.
Ann N Y Acad Sci
.
2003
;
992
:
158
67
.
29.
del Rey
A
,
Wolff
C
,
Wildmann
J
,
Randolf
A
,
Hahnel
A
,
Besedovsky
HO
et al
.
Disrupted brain-immune system-joint communication during experimental arthritis
.
Arthritis Rheum
.
2008
;
58
(
10
):
3090
9
.
30.
Härle
P
,
Möbius
D
,
Carr
DJ
,
Schölmerich
J
,
Straub
RH
.
An opposing time-dependent immune-modulating effect of the sympathetic nervous system conferred by altering the cytokine profile in the local lymph nodes and spleen of mice with type II collagen-induced arthritis
.
Arthritis Rheum
.
2005
;
52
(
4
):
1305
13
.
31.
Straub
RH
.
Autoimmune disease and innervation
.
Brain Behav Immun
.
2007
;
21
(
5
):
528
34
.
32.
Wolff
C
,
Straub
RH
,
Hahnel
A
,
Randolf
A
,
Wildmann
J
,
Besedovsky
HO
et al
.
Mimicking disruption of brain-immune system-joint communication results in collagen type II-induced arthritis in non-susceptible PVG rats
.
Mol Cell Endocrinol
.
2015
;
415
:
56
63
.
33.
Baerwald
CG
,
Laufenberg
M
,
Specht
T
,
von Wichert
P
,
Burmester
GR
,
Krause
A
.
Impaired sympathetic influence on the immune response in patients with rheumatoid arthritis due to lymphocyte subset-specific modulation of beta 2-adrenergic receptors
.
Br J Rheumatol
.
1997
;
36
(
12
):
1262
9
.
34.
World Health Organization (Internet)
Chagas disease (American trypanosomiasis)
. (cited 2023 May 31). Available from: https://www.who.int/health-topics/chagas-disease#tab=tab_1.
35.
Pan American Health Organization (Internet)
. Chagas Disease (cited 2023 May 31) Available from: https://www.paho.org/en/topics/chagas-disease.
36.
Roggero
E
,
Pérez
AR
,
Pollachini
N
,
Villar
SR
,
Wildmann
J
,
Besedovsky
H
et al
.
The sympathetic nervous system affects the susceptibility and course of Trypanosoma cruzi infection
.
Brain Behav Immun
.
2016
;
58
:
228
36
.
37.
Roggero
E
,
Wildmann
J
,
Passerini
MO
,
del Rey
A
,
Besedovsky
HO
.
Different peripheral neuroendocrine responses to Trypanosoma cruzi infection in mice lacking adaptive immunity
.
Ann N Y Acad Sci
.
2012
;
1262
:
37
44
.
38.
Bloemker
D
,
Mollerus
S
,
Gibbert
K
,
Dittmer
U
,
Del Rey
A
,
Schedlowski
M
et al
.
Inhibition of catecholamine degradation ameliorates while chemical sympathectomy aggravates the severity of acute Friend retrovirus infection in mice
.
Brain Behav Immun
.
2016
;
54
:
252
9
.
39.
Hoover
DB
,
Brown
TC
,
Miller
MK
,
Schweitzer
JB
,
Williams
DL
.
Loss of sympathetic nerves in spleens from patients with end stage sepsis
.
Front Immunol
.
2017
;
8
:
1712
.
40.
Pawlak
CR
,
Jacobs
R
,
Mikeska
E
,
Ochsmann
S
,
Lombardi
MS
,
Kavelaars
A
et al
.
Patients with systemic lupus erythematosus differ from healthy controls in their immunological response to acute psychological stress
.
Brain Behav Immun
.
1999
;
13
(
4
):
287
302
.
41.
Dimopoulos
G
,
Theodorakopoulou
M
,
Armaganidis
A
,
Tzepi
IM
,
Lignos
M
,
Giamarellos-Bourboulis
EJ
et al
.
Esmolol: immunomodulator in pyelonephritis by Pseudomonas aeruginosa
.
J Surg Res
.
2015
;
198
(
1
):
175
84
.
42.
Straub
RH
,
Pongratz
G
,
Weidler
C
,
Linde
HJ
,
Kirschning
CJ
,
Glück
T
et al
.
Ablation of the sympathetic nervous system decreases gram-negative and increases gram-positive bacterial dissemination: key roles for tumor necrosis factor/phagocytes and interleukin-4/lymphocytes
.
J Infect Dis
.
2005
;
192
(
4
):
560
72
.
43.
Bucsek
MJ
,
Giridharan
T
,
MacDonald
CR
,
Hylander
BL
,
Repasky
EA
.
An overview of the role of sympathetic regulation of immune responses in infectious disease and autoimmunity
.
Int J Hyperthermia
.
2018
;
34
(
2
):
135
43
.
44.
Besedovsky
H
,
del Rey
A
,
Sorkin
E
,
Da Prada
M
,
Burri
R
,
Honegger
C
.
The immune response evokes changes in brain noradrenergic neurons
.
Science
.
1983
;
221
(
4610
):
564
6
.
45.
Besedovsky
H
,
Sorkin
E
,
Felix
D
,
Haas
H
.
Hypothalamic changes during the immune response
.
Eur J Immunol
.
1977
;
7
(
5
):
323
5
.
46.
Besedovsky
H
,
Sorkin
E
,
Keller
M
,
Müller
J
.
Changes in blood hormone levels during the immune response
.
Proc Soc Exp Biol Med
.
1975
;
150
(
2
):
466
70
.
47.
Gonzalo
JA
,
González-García
A
,
Martínez
C
,
Kroemer
G
.
Glucocorticoid-mediated control of the activation and clonal deletion of peripheral T cells in vivo
.
J Exp Med
.
1993
;
177
(
5
):
1239
46
.
48.
Abraham
G
,
Schusser
GF
,
Ungemach
FR
.
Dexamethasone-induced increase in lymphocyte beta-adrenergic receptor density and cAMP formation in vivo
.
Pharmacology
.
2003
;
67
(
1
):
1
5
.
49.
Salicrú
AN
,
Sams
CF
,
Marshall
GD
.
Cooperative effects of corticosteroids and catecholamines upon immune deviation of the type-1/type-2 cytokine balance in favor of type-2 expression in human peripheral blood mononuclear cells
.
Brain Behav Immun
.
2007
;
21
(
7
):
913
20
.
50.
Kvetnanský
R
,
Fukuhara
K
,
Pacák
K
,
Cizza
G
,
Goldstein
DS
,
Kopin
IJ
.
Endogenous glucocorticoids restrain catecholamine synthesis and release at rest and during immobilization stress in rats
.
Endocrinology
.
1993
;
133
(
3
):
1411
9
.
51.
Sapolsky
RM
,
Romero
LM
,
Munck
AU
.
How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions
.
Endocr Rev
.
2000
;
21
(
1
):
55
89
.
52.
Svensson
E
,
Apergis-Schoute
J
,
Burnstock
G
,
Nusbaum
MP
,
Parker
D
,
Schiöth
HB
.
General principles of neuronal Co-transmission: insights from multiple model systems
.
Front Neural Circuits
.
2018
;
12
:
117
.
53.
Udit
S
,
Blake
K
,
Chiu
IM
.
Somatosensory and autonomic neuronal regulation of the immune response
.
Nat Rev Neurosci
.
2022
;
23
(
3
):
157
71
.