A History of Psycho-Neuro-Endocrine Immune Interactions in Rheumatic Diseases

Neuroendocrine immunology describes the influence of neuronal pathways with neurotransmitters or endocrine glands with hormones on immune functions. Originally, this portrays the efferent influence of the brain on immune cells making use of connecting factors (neurotransmitters, especially steroid hormones). Nowadays, the bidirectional aspect of the communication between the brain and the immune system is described with the expression psycho-neuro-endocrine immunology, which also reflects the strong influence of the immune system on brain function (with psychological/psychiatric, neuronal and endocrine consequences). Using this description, we briefly summarize the historical development in the field of rheumatology.

Historically, neuronal pathways with the sensory nervous system and the autonomic nervous system (sympathetic; parasympathetic: N. Vagus) were described before hormonal pathways came on stage, which led to the sequence of presentation in this article.

With a look on clinical cases, the concept of neuronal regulation of inflammation is confirmed by reports of patients with hemiplegia and chronic inflammatory diseases, where the paralytic side is protected from inflammation. Cases have been reported in whom hemiplegia manifested long after outbreak of chronic inflammatory rheumatic disease or long before, leading to protection independent of the time point of disease onset. Table 1 demonstrates the respective historical literature. The natural experiment of hemiplegia clearly demonstrates the importance of the nervous system in chronic inflammatory rheumatic diseases.

Since 2000 years (Celsus and Galen), clinicians recognize redness, warmth, swelling, pain, and altered function as immediate cardinal signs of inflammation. Neurogenic vasodilatation reported in 1876 by Stricker and 1901 by Bayliss [24, 25]; the inflammatory axon reflex with reddening skin observed in the 1910s by Bruce and by Breslauer [26‒28]; the flare response reported by Lewis around 1930 with erythema, hyperalgesia, and edema [29]; rediscovery of the antidromic vasodilatory flare response and dorsal root reflex by Chapman [30]; and Kelly’s and Jancsó’s more extended concept of neurogenic inflammation in the 1950 an 1960s [31, 32] were all expressions of the same principle: the proinflammatory influence of sensory nociceptive nerve fibers on acute inflammation and on above-mentioned cardinal clinical signs.

In these early years of scientific inquiry, the focus was mainly on the vasodilatory aspect of inflammation and plasma extravasation that leads to redness, warmth, swelling, pain, and altered function. With the discovery of substance P in 1931 by Ulf von Euler [33] and the later correct allocation to sensory nerve fibers by Lembeck in 1953 and, particularly, by Otskua and Konishi in 1976 [34, 35], the important immunomodulating role of substance P was found, e.g., in experimental arthritis in 1984 by Levine et al. [36]. The discovery of substance P is nicely demonstrated elsewhere [37]. Neurotransmitters of the sensory nervous system and propagation of inflammatory activation of neurons/microglia in the spinal cord are relevant for symmetrical joint symptoms [38, 39].

Neurotransmitters/neuropeptides from sensory nerve fibers like substance P are proinflammatory [40], and upon activation of the sensory nerve fiber substance P is released into the vicinity of the nerve ending (an efferent function of sensory nerve fibers). This is the mode how the sensory nerve fiber can have a proinflammatory effect in the periphery.

Other neuropeptides of afferent sensory nerve fibers and their receptors were discovered like calcitonin gene-regulated peptide (CGRP) in 1982 [41], galanin, glutamate, and others. While substance P is often described as a prototype proinflammatory neuropeptide [40], CGRP might have opposing anti-inflammatory but also vasodilatory activities [42]. Thus, depending on the local amount of released substance P relative to CGRP, the tissue might undergo a proinflammatory or anti-inflammatory reaction (in any case a vasodilatory reaction with more immune cell evasion).

In synovial tissue of patients with rheumatoid arthritis (RA), our group demonstrated a preponderance of substance P-positive nerve fibers over CGRP-positive nerve fibers [43]. This would render sensory nerve fibers proinflammatory by a direct action of substance P on immune cells, which supports the earlier concept of neurogenic inflammation (vasodilation, plasma extravasation, broad immune cell activation).

Another phenomenon in many inflammatory diseases is sensitization of the sensory nervous system (nociceptive pathways) because of peripheral, spinal, and more central sensitization [44]. The brain controls the input arriving through afferent sensory nerve fibers by descending pathways, which has been first described by Sherrington in 1915 [45] and later by Melzack and Wall in 1965 [46‒48]. Missing control or altered control of this input through sensory nerve fibers and upregulation of local inflammatory pathways (microglia) can lead to sensitization of nociceptive pathways leading, e.g., to hyperalgesia [49].

In parallel, nociceptive fibers can be activated/sensitized to release more neuropeptides into the vicinity of the peripheral nerve terminal [50, 51], which is an efferent function of otherwise afferent nerves. Thus, a hyperactive pain system can have a proinflammatory role by releasing substance P and other neurotransmitters locally. This is particularly true when sensory nerve fibers start to sprout under inflammatory conditions as demonstrated in RA [52, 53].

Cytokines play an important role for sensitization in arthritic joints and the spinal cord as recognized in the 1990s [54‒59]. While the role of mitogen-activated protein kinases for sensitization is known for a while [56, 60], recently, therapy with Janus kinase inhibitors also showed the stimulating role of Janus kinase-signal transducer and activator of transcription pathways for sensitization [61]. Importantly, autoantibodies can directly induce hypersensitivity in mice [62, 63]. More central sensitization might be observed through functional magnetic resonance imaging techniques and anti-TNF therapy rapidly ameliorates central nervous system (CNS) sensitization [64].

The term hypersensitivity is used in pain science, neurology (headache), dental medicine (dentine hypersensitivity, temporomandibular disorder), gastroenterology (irritable bowel syndrome), dermatology (itch, lipohyperplasia dolorosa), urology (interstitial cystitis, pelvic pain), psychiatry (autism spectrum disorders), and other medical disciplines. In rheumatology, the word “hypersensitivity” was typically linked to the hypersensitive immune system such as in adverse reactions versus medication, anaphylaxis, contact dermatitis, allergy, or other acute immune reactions. Maybe the presence of the notion of “immune hypersensitivity” slowed down the uptake of the expression “sensory hypersensitivity” in rheumatology. Nevertheless, rheumatologists have treated patients with sensory hypersensitivity for a long time.

Already in 1904, William Gowers coined the expression fibrositis [65], which was correctly renamed into fibromyalgia by Yunus et al. [66] in 1981 because the disease was not inflammatory at all [67]. The primary form of fibromyalgia is characterized by diffuse musculoskeletal aches, pains or stiffness and accompanied by several other symptoms such as tiredness, anxiety, sleep problems, headache, irritable bowel syndrome, and numbness as originally described in the paper of Yunus et al. [66]. Hypersensitivity characterizes primary and secondary fibromyalgia (primary = without any other disease; secondary = with another disease, e.g., systemic lupus erythematosus [SLE]). Again, autoantibodies might play a role in fibromyalgia hypersensitivity that has recently been put forward [68].

The idea of “primary hypersensitivity” of nociceptive pathways led to a third form of pain description next to nociceptive pain (arises from actual or threatened damage to nonneural tissue and is due to the activation of nociceptors) and neuropathic pain (caused by a lesion or disease of the somatosensory nervous system). The new form of pain was called nociplastic pain [69, 70], and it is characterized by “pain that arises from altered nociception despite no clear evidence of actual or threatened tissue damage causing the activation of peripheral nociceptors or evidence for disease or lesion of the somatosensory system causing the pain” [69]. This form of pain has much to do with altered function of pain-related sensory pathways in the periphery and brain.

An important origin for nociplastic pain is early traumatic experiences in childhood or adolescence sometimes reexperienced in adulthood (double hit). It was Freud and Breuer already in 1895, in their much-criticized book on hysteria (German title: Studien über Hysterie), to link earlier experiences in childhood and adolescence with chronic pain in adulthood. The subject of early trauma and later chronic pain has been resumed in the 1950s by George Engel, who described among many affected individuals a patient with chronic painful states and arthritis after childhood abuse/neglect [71].

The first study in rheumatology was carried out in the early 1990s on patients with fibromyalgia [72], a disease with marked sensory hypersensitivity (see above). The authors from McGill University in Montreal found a strong interrelation between childhood trauma and later chronic pain in these patients [72].

A recent book summarized clinical and pathophysiological consequences of early traumatic events including pain, sleep disorders, fatigue, psychiatric diseases, chronic inflammation, and others using more than 1,100 references including historical literature [73]. This book also summarizes the links between early traumatic experience, chronic immune activation, and autoimmunity, an association first described in the year 1978 in the juvenile form of arthritis [74] and in 2004 in the adult form of arthritis [75]. Since these first papers were published, many more similar reports in different rheumatic diseases confirmed this strong interrelationship (see literature in ref. [73]).

Already Galen, 2000 years ago, described the sympathetic trunk in contrast to the vagus nerve and, in addition, he found the superior and inferior cervical ganglia, the semilunar (celiac) ganglia, and the rami communicantes [76]. Jean Riolan in the early 17th century reported that these nerve fibers have their origin in the spinal cord and not in the vagus nerve, which was the general belief since Galen [76]. Willis in 1664 showed the great mesenteric plexus radiating like a sun (solar plexus), and he gave a much better description of the anatomy of the autonomic nervous system [76].

While the term “sympathetic” already existed since Galen described simultaneity of functional change in one organ and “sympathetically” in another organ (particularly, during disease such as in “sympathetic ophthalmia”), the expression “sympathetic” was not allocated to the above-described nervous system or the sympathetic trunk until the appearance of J. B. Winslow, who worked in Paris in the early 18th century. Before Winslow, the nervous structure made up of the sympathetic trunk was called “nervus intercostalis,” and Windslow renamed it into “la grande sympathique” to designate it as the major source of sympathetic outflow [76].

Xavier Bichat in 1800 introduced a physiological and anatomical separation of two nervous systems: organic life (autonomic nervous) versus animal life (somatomotor). He recognized that the sympathetic trunk is a chain of “little brains” (=ganglia) [77]. Friedrich Henle, Benedict Stilling, Claude Bernard, and Charles-Edouard Brown-Séquard discovered regulation of the vessel diameter through sympathetic nerve fibers [77].

Claude Bernard first demonstrated higher centers of the sympathetic nervous system (SNS) in his famous puncturing experiment of the hypothalamus leading to glycosuria, which was blocked by cutting splenic/mesenteric nerves [77]. The first to distinguish SNS and parasympathetic nervous system was Langely in 1921 [78], who also coined the expression “autonomic nervous system.” Excellent reviews and books demonstrate more history on the SNS [76‒80].

The first hormone described was adrenaline as recently depicted [81]. While different groups discovered adrenaline at the transition from the 19th to the 20th century [82‒84], Ulf von Euler found the key neurotransmitter of sympathetic nerve fibers much later in 1940 in the Karolinska Institute [85, 86]. Noradrenaline is the key neurotransmitter in sympathetic nerve fibers because adrenaline – the “hormone” – does not predominate in these nerve terminals. Importantly, von Euler also described the presence of noradrenaline in lymphoid tissue like the spleen [85, 87], which was confirmed for little amounts of administered radioactive adrenaline by later studies by Axelrod et al. [88]. However, at that time, scientists thought that sympathetic regulation in the spleen serves solely vasoregulation and perhaps storage of red blood cells – and nothing else. The term sympathetic neuroimmunomodulation was not born.

The link of the SNS to the immune system was prepared by work of Tohru Ishigami in 1919 by his experiments of adrenaline-regulated opsonization of tubercle bacilli [89]; by Metalnikov and Chlorine by conditioning the immune response in 1926 [90] (now known to be partially related to the SNS [91]); by Andor Szentivanyi studying the role of beta-adrenergic receptors in anaphylactic reaction in the late 1950s [92, 93]; by Draskoci and Jankovic [94] describing thymus involution and suppression of immune responses after systemic reserpine in 1964; by Charles Reed and his team studying the suppressive effect of adrenaline on vaccination response in 1968 (Abstract: [95]); and by Macmanus and Whitfield [96] showing the stimulating effect of cyclic AMP on thymocyte mitogenic responses.

However, the breakthrough work was carried out around the year 1970 because two groups demonstrated the influence of adrenergic agonists and antagonists on phytohemagglutinin (PHA)-stimulated transformation of human lymphocytes [97‒99]. This is the start of mechanistic sympathetic neuroimmunomodulation. Hadden et al. wrote [98]:

The findings were corroborated by many studies in the 1970s that described specific adrenoceptors on the membranes of leukocytes/lymphocytes and sympathetic influence on immune function [100‒110]. A further key to crack open the enigma of sympathetic neuroimmunomodulation was the staining of sympathetic nerve fibers by chemical techniques such as with formalin (fixing and staining) known already in the 1950s/early 1960s [111, 112] and early 1970s using the sucrose-glyoxylic acid method [113]. Using these techniques – not specific for noradrenaline – sympathetic staining of lymphoid organs was fairly possible for the first time [114‒118].

Another step forward was the discovery of tyrosine hydroxylase, the rate-limiting enzyme for adrenaline/noradrenaline biosynthesis [119]. This enzyme also exists in sympathetic nerve fibers and sympathetic denervation reduces the activity of this enzyme in the nerve ending [120]. Tyrosine hydroxylase is transported along the sympathetic axon outside the vesicles/granules [121] using a slow transport system with a velocity of 1–3 mm/day [122].

These findings allowed for specific staining of sympathetic nerve fibers at a time point when respective antibodies/antiserum against tyrosine hydroxylase were first reported in the middle of the 1970s [123, 124]. Both techniques – chemical and immunological – together with the understanding of a crosstalk between immune cells and sympathetic nerve fibers led to clear evidence of anatomical contact sites between nerve fibers and immune cells in lymphoid organs studied by David Felten and his group [125‒129], which was also supported for peptidergic innervation [130, 131].

While (1) noradrenaline exists in sympathetic nerve endings in the spleen, (2) since anatomical contact sites between sympathetic nerve fibers and splenic immune cells were demonstrated, (3) given that release of noradrenaline can be stimulated from splenic nerve terminals, and (4) because adrenergic receptors on immune cells were well known, our group added to this evidence of a real neuroimmune synapse by showing functional crosstalk between nerve ending and immune cell (reviewed in [132]). Using microsuperfusion chambers loaded with intact splenic tissue, we showed that the action of electrically released endogenous noradrenaline modulates macrophage cytokine secretion in the spleen, which can be inhibited by competitive neurotransmitter antagonists at intact anatomical locations (reviewed in [132]). In the model of collagen type II arthritis (CIA), the technique demonstrated that an α-adrenergic influence on immune function comes to the fore, which has a general proinflammatory character [133]. All information showed that in the spleen, a functional neuroimmune synapse exists (Table 2). Now, the time was ripe for the study of the SNS in situations with chronic inflammatory diseases/experimental rheumatology.

In the absence of better therapeutic options, necessity is the mother of invention. In the 1950s, some groups experimented with epinephrine injections in patients with arthritis in order to stimulate the adrenal cortex (i.e., glucocorticoids) but the anti-inflammatory success of this hormone therapy in low doses was marginal compared to glucocorticoids per se [143]. Others tested therapeutic removal of sympathetic lumbar ganglia in patients with RA and osteoarthritis in order to treat painful states in hip and knee joints [23, 144], but this heroic treatment did not stand the test of time.

In the 1970s, Dick et al. [145] found a decreased α-adrenergic vasoconstrictor tone during ¹³³Xenon perfusion of the knee joint in patients with RA as compared to controls and osteoarthritis. This indicated that inflammation impedes the normal α-adrenergic vasoconstrictor tone – a typical vasodilatory response during inflammation in different tissues. This phenomenon was observed by several other groups and, thus, stood the test of time (e.g. [146‒148]).

The first authors that markedly added to the knowledge of the sympathetic influence on experimental arthritis were the group around Jon Levine and Allan Basbaum in San Francisco after they had published their seminal paper on proinflammatory substance P effects in arthritis [36, 149]. Although their clinical paper on guanethidine therapy in RA has never been repeated in larger groups of patients [150], it shows the conviction towards the proinflammatory effect of the SNS. A publication of the same group in the year 1988 stated that the SNS exerts its proinflammatory effects on adjuvant arthritis through the β2-adrenergic receptor while α1/2-adrenergic pathways were not influential [151]. The same group later corrected this statement because an α2-adrenergic suppressive effect on adjuvant arthritis was observed [152]. In the late 1990s, the San Francisco group switched to other possible mediating factors like bradykinin and the vagus nerve.

At approximately the same time, the Rochester group around David Felten, Dianne Lorton, and Denise Bellinger started studies on adjuvant arthritis showing that denervation of the sympathetic influence at local lymph nodes enhanced severity of arthritic changes. They summarized their findings as follows [153]: “These modulatory effects are distinctly different from the effects of sympathetic nerve fibers in the joints themselves (AU: considering the data of Levine et al.).” This was corroborated by the same group in 1996 and 1999 [154, 155].

From the studies of Levine et al. [36, 149] and Lorton et al. [153‒155], one recognizes two important locations of influence: (1) the local influence in the synovial tissue (β2-adrenergic effects support inflammation in the form of vasodilation and plasma extravasation) versus (2) the local influence in draining lymph nodes and spleen (β2-adrenergic effects inhibit inflammation/specific immune responses).

The increasing understanding of the immune system that started with the discovery of cellular surface markers on immune cells (typing of cells) and secreted products such as cytokines (typing of cells) and chemokines in the 1980s changed the understanding of pathophysiology in different experimental models of arthritis. While the earlier applied injections of capsaicin, zymosan, kaolin, Freund adjuvant, carrageenan and others locally into the joint only reflected a very acute inflammatory response with plasma extravasation and local stimulation of innate immune and bystander mechanisms (= a real acute neurogenic inflammation), the auto-antigenic induction of experimental arthritis with bovine collagen type II, methylated bovine serum albumin and other antigens demonstrated the full spectrum of innate and adaptive immune reactions locally and in distant lymphoid organs. The parallel understanding that neurotransmitters of the SNS influence innate and adaptive pathways in very distinct and sometimes opposing ways depending on location and timing, led to a more complex picture of sympathetic neuroimmunomodulation in arthritis. More and more, it became the affair of trained immunologists.

Our group added to this complexity by demonstrating anti- and proinflammatory effects of 6-hydroxydopamine-induced sympathectomy in the CIA model depending on the time point of drug application during the development of chronic arthritis [156, 157]. An early sympathectomy (before onset of CIA and until approximately day 20) always protected the animals from arthritis; however, later sympathectomy on day 55 in the mouse model aggravated CIA.

The differential effect of early and late sympathectomy is probably due to the appearance of catecholamine-producing local cells in the later phases of the disease. While early sympathectomy destroys the proinflammatory neuronal influence on CIA development, late sympathectomy destroys anti-inflammatory catecholamine-producing cells beneficial in CIA [158‒160]. Today, we recognize a multitude of influencing roles of the SNS on chronic inflammation/chronic immune activation through a variety of different pathways (summarized in Table 3). In the following text, some arguments of Table 3 are shown in a historical perspective.

The local influence of sympathetic neurotransmitters will change when sympathetic nerve fibers are reduced in inflamed tissue because it changes local neurotransmitter concentrations. Noradrenaline at low concentrations binds to α-adrenoceptors, at high concentrations it binds to, both, α- and β-adrenoceptors (it is a question of affinity). In the early 1990s, two groups independently showed the loss of sympathetic nerve fibers in synovial tissue of patients with RA using qualitative methods of immunostaining in a small number of RA patients [167, 168].

In extensive quantitative immunofluorescent studies on a great number of RA patients, we corroborated the loss of sympathetic nerve fibers in synovial tissue of RA and other inflammatory diseases [52, 166, 169‒173]. In experimental models of arthritis, others demonstrated a significant loss of sympathetic nerve fibers of splenic regions distant to the entry point (hilus of the spleen) [174]. At the entry point, the density is even higher as if the nerve fibers accumulate there after retraction, which was confirmed by others near the lymph nodes [175, 176].

The loss of sympathetic nerve fibers with lower concentrations of noradrenaline was discussed as an evolutionarily positively selected process to create “zones of permitted inflammation” [162]. Low concentrations of noradrenaline would only activate α-adrenergic but not β2-adrenergic receptors, which is a proinflammatory signal helpful in wound healing and local infection [162].

Notwithstanding the general finding of lower sympathetic nerve fiber density in inflamed tissue, some authors have demonstrated increased innervation toward distinct sites in the tissue. Increased innervation in synovial tissue appeared in experimental arthritis models in the proximity of cartilage [177, 178] and around arterial walls in synovial tissue of RA patients and arthritic animals [179, 180]. The meaning of uneven distribution of sympathetic innervation in inflamed joints is not yet known, but it might be due to different species, timing of sample collection, specific regions in inflamed tissue and divergent methods of quantification. It might also occur as a controlled process in areas where a high sympathetic influence through β2-adrenoceptors serves a particular function (cartilage: support of cartilage degradation to release energy-rich fuels [181]; around vessels: support of β-adrenergically mediated vasodilation instead of α-adrenergically stimulated vasoconstriction to favor immune cell extravasation).

The important function of β-adrenergic receptors for inflammation, mainly the β2-adrenergic receptor, goes back to asthma/anaphylaxis research in the 1960s by Andor Szentivanyi [182], but still today, the β2-adrenergic receptor is given a main role in asthma [183]. In asthma research, the loss of adequate β2-adrenergic signaling remains a challenge because cofactors like inflammatory cytokines, growth factors, respiratory viruses, and certain allergens can disturb the function of the receptor [184]. The first indication of β2-adrenergic receptor dysfunction in asthma goes back into the 1980s [185] and, e.g., TNF plays an important role because it can inhibit β-adrenergically stimulated adenylyl cyclase activity in cultured airway smooth muscle cells [186].

Since the β2-adrenergic receptor has an important anti-inflammatory function in macrophages when acutely challenged with lipopolysaccharide [187], a similar loss in anti-inflammatory capacities of the β2-adrenergic receptor in asthma can lead to deficits in anti-inflammatory sympathetic neuroimmunomodulation, which was confirmed in an animal model of arthritis [188].

Stimulated by asthma research, Christoph Baerwald and his student Matthias Wahle investigated this phenomenon carefully, and they found clear evidence for decreased function of the β-adrenergic receptor in leukocytes from RA patients [189‒193]. This was the first evidence that this phenomenon exists in RA similar than in asthma patients and also known for the β1-adrenoceptor in the heart as demonstrated in aged people [194].

Not many studies worked on β-adrenergic receptor signaling in inflamed tissue of patients with rheumatic diseases such as RA. One study demonstrated an obvious decrease of G protein-coupled receptor kinase 2 (GRK2) but not of GRK 5 in material from RA patients [195]. The disease of adjuvant-induced arthritis led to a decrease of GRK2, GRK3, and GRK6 in splenocytes and mesenteric lymph node cells [196]. This differential regulation of the GRKs leads to activation of the noncanonical β-adrenoceptor signaling pathways with proinflammatory effects through mitogen-activated protein kinases [197].

In 2013, Lorton and Bellinger showed defects in β2-adrenergic receptor signaling with reduced cyclic AMP induction in splenocytes of adjuvant arthritic rats [198]. In 2015, we corroborated the idea/the finding of a malfunctioning β2-adrenergic receptor in mixed synovial cells of patients with RA by observing a GαS-to-Gαi signaling switch [199]. Inhibitors of GRK2 might reverse this switch [200]. All these studies, clearly point to defects of the canonical β2-adrenergic receptor signaling, leading to proinflammatory consequences. Thus, loss of sympathetic nerve fibers plus changes in β2-adrenergic signaling are strong proinflammatory signals.

In the early 1980s, Sato et al. [201] demonstrated the first mechanistic link between inflammation and a heightened sympathetic activity to the heart. These authors studied the influence of pain on the activity of the sympathetic inferior cardiac nerve. They observed a clear increase of the firing rate of these sympathetic nerve fibers upon stimulating pain in joints, skin, and muscles. When the authors acutely provoked joint inflammation by injection of kaolin and carrageenan, the firing rate markedly rose and this surmounted the normal pain response [201, 202]. Others corroborated the inflammation-induced increase in sympathetic activity in the adjuvant arthritis model [203]. Pain and inflammation stimulate the SNS because under both conditions the body requires high amounts of energy-rich fuels [204].

Today, we observe a high sympathetic activity in different chronic inflammatory diseases [205‒213]. This circumstance certainly adds to the increased frequency of cardiovascular diseases in patients with chronic inflammation [214‒216]. A higher sympathetic activity is also related to energy provision to the immune system [165], water retention, and volume overload in patients with rheumatic diseases [217, 218].

For the first time, Galen described the vagus nerve [77]. The role of the vagus nerve as a possible inhibitor of the SNS was demonstrated in 1846 by arresting the heartbeat after electrical vagus nerve stimulation (Ernst Heinrich Weber and Eduard Weber from Leipzig, cited in [77]). Pflüger, Bidder, and the Webers showed the influence of the vagus nerve on the intestine [77].

Chemical neurotransmission of neuronal impulses was studied by Dixon in 1907 (muscarine effects), Lehmann found choline in 1907, and Loewi found acetylcholine as the vagus neurotransmitter in 1921 [77]. Furthermore, John Langely [78] distinguished sympathetic and parasympathetic nervous system and created the expression “autonomic nervous system.” Langley recognized the antagonism between sympathetic and parasympathetic systems.

While the role of the SNS in neuroimmunomodulation has been studied for more than a century [219], investigation of the vagus nerve as a significant factor of neuroimmunomodulation appeared late in the 20th century. Some groups recognized the proinflammatory effects of vagus manipulation or electrical vagus nerve stimulation for neurogenic inflammation in the rat trachea [220‒226]. Mediastinal vagotomy inhibits neurogenic inflammation in the rat bronchial tree [221, 227]. In the field of pulmonary research, the phenomenon of vagally induced neurogenic inflammation is widely known (we can call it textbook knowledge).

At the beginning of the 1990s, two groups independently discovered vagus-mediated sickness behavior including hyperthermia stimulated by injection of proinflammatory substances into the peritoneal cavity [228‒230]. A possible crosstalk of vagus nerve fibers and mast cells was demonstrated in intestinal mucosa [130, 231].

A first indication of a vagally mediated inhibition of inflammation was shown in the model of colitis with trinitrobezenesulfonic acid in rats, and the authors concluded that vagal nerve fibers have a protective role [232], later supported by others [233]. This was challenged by others in the dextran sodium sulfate colitis because the SNS – not the vagus nerve – had the major anti-inflammatory role [170, 234, 235].

Kevin Tracey’s group then produced several publications at the beginning of the 2000s that showed the anti-inflammatory influence of electrical vagus nerve stimulation on systemic inflammation, particularly TNF secretion, which was mediated through the alpha7 nicotinergic acetylcholine receptor (α7nAChR) [236‒239]. Interestingly, the spleen is an important relay station relevant for these vagally mediated effects [240, 241].

According to Tracey’s group, sympathetic nerve fibers to the spleen should be responsible for anti-inflammatory vagal effects mediated in the spleen through acetylcholine-producing cells [242], but others criticized this unilateral interpretation because of the fundamental role of sympathetic nerve fibers to the spleen and the absence of cholinergic nerve fibers [243‒246]. In addition, data in humans with vagus nerve stimulation or vagotomy did not fit to the concept of vagal anti-inflammation in the early years of inquiry [247‒252], which might depend on the fact that investigated subjects did not suffer from a chronic inflammatory disease. Subsequent studies in patients with chronic inflammatory diseases provided a new picture (next paragraphs).

After several reviews of Kevin Tracey and his group, many researchers of chronic inflammation took up the idea of the anti-inflammatory vagus nerve influence. The theory also entered the field of rheumatic diseases. Soon, it was shown that an agonist to the α7nAChR had anti-inflammatory influence in synovial cells of patients with RA and in the CIA model [253‒256]. In addition, vagus nerve stimulation in the rat CIA model showed some anti-inflammatory activities [257], supported in zymosan arthritis in rats through vagal and central nervous pathways that activate (!) the SNS [258, 259]. This opened the question whether, or not, the vagus stimulation exerts anti-inflammatory effects through retrograde central processing and augmentation of downstream sympathetic influences. This question is still unanswered.

With vagus nerve stimulation in humans, researchers obtained mixed results with positive anti-inflammatory effects in Crohn’s disease, RA, and psoriatic arthritis [260‒265] but negative results in other studies in rheumatic diseases, which might depend on stimulation techniques [266, 267]. All studies in patients were carried out in an uncontrolled fashion, which still inhibit the authors of this review from drawing firm conclusions due to well-known placebo effects.

Although many different hormones have been studied in rheumatic diseases, this brief history can only cover the most important ones. We focus on glucocorticoids, androgens, progesterone, estrogens, and the Vitamin D hormone.

Philip Showalter Hench (1896–1965) was an American rheumatologist. Together with his colleague Edward C Kendall and the Swiss chemist Tadeus Reichstein, he was awarded the Nobel Prize for Medicine in the year 1950 for the discoveries on glucocorticoid hormones and their favorable effects in RA.

Already in the late 1920s, Hench observed the favorable effects of jaundice in a patient with RA. At the time, this was a great surprise because RA remission was a medical curiosity. Between 1929 and 1934, he collected data of 16 patients with RA who developed jaundice that ameliorated the crippling disease [268]. Today, we know that in the bile, some biliary acids with the typical steroid hormone structure can have anti-inflammatory activities [269]. However, many therapeutic approaches with bile or compounds thereof – Hench expected to discover the healing “substance X” in the bile – were not successful.

At the same time, he observed the ameliorating effect of pregnancy in women with RA, and he speculated on a “common denominator substance X” favorable in jaundice and pregnancy [270]. He wrote: “It does not seem illogical to suppose that the agents responsible for both these phenomena are closely related, perhaps identical, and if the agent is a chemical substance, it would appear that it is neither bilirubin nor a strictly female sex hormone.”

During the hunt for “substance X,” Philip Hench also recognized that other inflammatory diseases like psoriasis arthritis, asthma, hay fever, Addison’s disease, and even migraine were sometimes relieved during pregnancy and/or jaundice. “Substance X was unspecific and bisexual.”

In the early 1940s, he started a collaboration with Edward Kendall “in a laboratory (at Mayo Clinic) a few yards away.” Edward Kendall and colleagues – biochemists – already isolated several adrenal steroid hormones in the late 1930s [271], but administration of them to humans lasted years because the substances were difficult to isolate from extracts. In the year 1948, in a collaboration of Edward Kendall and the American company of Merck and Co., Inc., enough compound E (cortisone) was available to treat a woman “badly crippled with RA” [272]. More patients followed with highly favorable results because the biologically inactive cortisone is reactivated to active cortisol.

The similar effects of adrenocorticotropic hormone were linked to the adrenocorticotropic hormone-induced secretion of adrenal glucocorticoids. This was the breakthrough! In Switzerland, concurrently, Tadeus Reichstein also discovered the different adrenal hormones.

Today, we well know that therapeutic glucocorticoids ameliorate inflammation in autoimmune diseases and other inflammatory diseases [273]. The anti-inflammatory effects of glucocorticoids stood the test of time. However, we also know that glucocorticoids must be administered at low doses up to a maximum of 5 mg prednisolone per day to induce a sufficient effect [274] and to prevent severe side effects [275‒277]. Administration according to circadian rhythms with a nightly increase of low-dose glucocorticoids at 2:00 a.m. has shown several beneficial effects [278‒280].

Although low doses of glucocorticoids below 5 mg prednisolone/day seem to be relatively safe, the subject is still strongly debated due to adverse events [281]. In a world of biological and targeted synthetic disease-modifying antirheumatic drugs, rheumatologists try to avoid the use of glucocorticoids over longer periods. The consensus accepted by almost all rheumatologists is to use the lowest dose for the shortest time possible to deliver the required benefit [281].

The known inadequate secretion of the hypothalamic-pituitary-adrenal axis in patients with chronic inflammatory diseases has been extensively described earlier [282, 283]. This discrepancy of lowered glucocorticoid availability and action on one side and increased inflammation on the other side has been called “disproportion principle” [284].

Testosterone and other androgens were discovered in the early 1930s by two groups from the Netherlands [285] and Germany [286]. For the work on the discovery of sex hormones, Adolf Butenandt and Leopold Ruzicka received the Nobel Prize in the year 1939.

The first study on 17-ketosteroids (e.g., androstenedione, dehydroepiandrosterone [DHEA]) in ankylosing spondylitis and RA dates back to the year 1947 [287]. However, the authors of this early report did not see marked differences in urine concentrations of 17-ketosteroids compared to controls. Already 3 years later, however, in 1950, the lower excretion of 17-ketosteroids was demonstrated in RA patients [288], which was confirmed in many later investigations in the 1980s [289‒293].

After the impressive success of therapy with glucocorticoids by Hench and colleagues [272] (see above), some researchers used other steroid hormones like testosterone in the treatment of RA [143, 294, 295]. In the study by Ishmael et al. [294] in patients with RA, it was reported that intramuscular testosterone propionate led to decided improvement in 65% of patients within 24 h, but this was challenged by Guest et al. [143]. Several little uncontrolled therapy studies with testosterone were reported between 1950 and 1970, but results were mixed. Later, animal studies using androgens showed favorable effects in different animal models of SLE and arthritis [296‒300].

In 1991, in an open-labeled study on testosterone undecanoate in male patients with RA, one of us (M.C.) demonstrated the anti-inflammatory effects as a significant reduction in IgM rheumatoid factor concentration, lower number of affected joints and decreased daily intake of NSAIDs [301]. In a double-blind placebo-controlled study in RA patients, the group of Bijlsma et al. [302] showed that testosterone had several beneficial effects [303]. Similarly, in SLE, androgens such as DHEA were tested with some positive effects on disease activity and bone quality [304‒308], which was recently systematically summarized [309]. These studies stimulated some rheumatologists until this day to use androgens such as DHEA in an off-label therapy.

Already in the 1950s, conversion studies with precursor androgens were reported in vivo in an RA patient and guinea pigs and in vitro in perfused liver [310]. Since androgens can give rise to estrogen production by the aromatase complex, the question of conversion of androgens to other hormones was important. Using original synovial material from RA patients, it was demonstrated that androgens are increasingly converted to estrogens in patients with RA [311]. Nevertheless, some androgens like androstenedione and testosterone even inhibited androgen-to-estrogen conversion in mixed RA synovial cells and stimulated the production of more potent 5α-reduced androgens [312]. Thus, testosterone may still be an interesting treatment option in patients with RA with generally low levels of this hormone, particularly in men. Several anti-autoimmune effects of androgens on diverse immune cells were recently summarized [313]. Androgen and estrogen receptors were discovered in synovial tissue and cells of RA patients in the early 1990s [314, 315], later confirmed by others [316].

The important role of proinflammatory cytokines in inhibiting androgen production was confirmed during biological disease-modifying therapy because, e.g., tocilizumab (inhibits IL-6 receptor) increased serum levels of the potent precursor androstenedione, also observed relative to cortisol and 17-hydroxyprogesterone [317], and anti-TNF therapy increased the precursor DHEA sulfate [318]. The loss of systemic androgens in chronic inflammatory diseases was interpreted as a mechanism of energy distribution to an activated immune system by the breakdown of glycogenic amino acids in skeletal muscles with the consequence of sarcopenia [165].

Similarly stimulated by Hench’s publication in 1949, progesterone was used in several very small studies in RA around 1950, but this therapy was not successful [319‒321]. Guest et al. [143] used the progesterone precursor pregnenolone and in combinations with testosterone showed some beneficial effects in open studies. A group from Chile, in the 1980s, reported on some beneficial effects of intra-articular progesterone injections in RA patients [322].

In studies in experimental arthritis, progesterone alone had no effect on the clinical course of the disease [323‒325]. When progesterone was given together with 17β-estradiol, the favorable effect was stronger than with the estrogen alone [324], which was challenged by others [325]. In an inflammation-induced cartilage degradation model, progesterone treatment had no beneficial effects in vivo, but it blocked IL-1-induced cartilage degradation in vitro [326]. Others confirmed the beneficial effect of progesterone in a model of destructive cartilage invasion by activated synovial fibroblasts in vitro [327], which serve as an RA model.

Progesterone treatment showed positive effects in the lupus model of New Zealand rats [328]. The favorable anti-inflammatory role of progesterone in vitro has been reviewed [329].

Due to the strong women-to-men preponderance in many chronic inflammatory diseases, estrogens were in the focus from the start of their discovery in the 1930s with the assumption that they could have rheumatic disease-stimulating effects. From the late 1940s – after the publication on glucocorticoids by Hench et al. [272] – some groups treated patients with chronic inflammatory diseases using estrogens. Most studies were carried out in small numbers of patients without randomization or placebo control. Some even tried to treat SLE with strong estrogens like diethylstilbestrol [330], a substance that was banned in the USA in 1971 due to severe side effects. Soon, it was clear that estrogens are not an adequate therapy for chronic rheumatic diseases.

The discussions on unfavorable effects of oral contraceptives in chronic inflammatory rheumatic diseases started in the 1960s, particularly in the direction of SLE. It was summarized in a publication by Chapel and Burns in 1971 [331]. They recommended to prevent treating SLE women with oral contraceptives. This was the situation with the early generation of oral contraceptives where estrogen and progesterone doses were relatively high. Particularly, thromboembolic events were in the focus of considerations. Case reports in this direction appeared all the time during the late 1960s until 1980.

The first studies on the risk of increased susceptibility to chronic rheumatic diseases in normal women taking oral contraceptives appeared at around the same time [332]. The authors concluded that “results obtained fail to demonstrate that normal women using oral contraceptives are at a greater risk of developing rheumatic symptoms or serological changes than a similar group of non-users.” After many studies into this subject [333‒335], today, we know that in SLE, no influence of oral contraceptives exists in mild disease, whereas a higher risk of flares and complications are present in patients with antiphospholipid antibodies or preceding thrombosis. In RA, modern oral contraceptives might even be protective, and they certainly do not exacerbate the disease (summarized in [313]).

In 1977, Kunkel’s group in New York established an interesting link between estrogens and autoimmune diseases because they related the estrogen excess and androgen deficit in Klinefelter’s syndrome to the susceptibility of SLE [336]. Later reports confirmed this association in SLE and other diseases [337‒339]. Somewhat later Bob Lahita of the same group found high urinary 16α-hydroxylated estrogens [340], which have strong immunostimulating and proliferative effects later confirmed by others [341, 342]. The role of sex hormones in SLE was reviewed for the first time in 1981 [298].

In the same period, early studies in New Zealand lupus mice found an influence of sex hormones on disease outcome, and estrogen antagonists and androgens were demonstrated to be favorable in this model [297, 343]. However, estrogen antagonist therapy with tamoxifen failed in SLE patients in 1984 [344]. In models of RA like the CIA model, however, estrogen therapy had beneficial effects because incidence and severity of the disease were suppressed presented by Rikard Holmdahl and Hans Carlsten from Sweden [324, 345, 346].

The success of estrogens in RA models prompted the first therapy with modern estrogens in a small group of RA patients carried out by Hans Bijlsma’s group [302]. The authors reported some beneficial effects, which were later challenged by other investigators [347]. However, the beneficial influence of estrogens on bone density was confirmed in RA patients [348].

More and more discussions appeared around the use of hormone replacement therapy in postmenopausal women. After a multitude of epidemiological studies carried out between 1990 and 2020, we can summarize that the risk of developing SLE is somewhat increased and that there is a higher incidence of mild to moderate flares in SLE. In RA patients, hormone replacement therapy is protective (summarized in [313]).

At around 1990, the clear difference of sex hormone influence in diseases such as SLE (estrogens aggravate) in contrast to RA (estrogens can suppress) was state of the art. However, within the same diseased animal, estrogens sometimes exacerbated and sometimes ameliorated typical manifestations of autoimmunity [349]. The dichotomous effect was a mainstay of scientific discovery because it explained the diverse effects of sex hormones within the same animal/disease/patient.

While some author groups spoke for the beneficial effects of estrogens in RA/experimental arthritis others argued strongly against it. We recall a NeuroEndocrine Immune Study Group meeting at the beginning of the 2000s of the American College of Rheumatology during which the two groups were present. Understanding the dichotomous effects of estrogens on different basic immune functions can help explain the contrasting situation [350]. In the years between 1980 and 2005, many stimulating effects of estrogens on B-cell function were discovered (e.g., antibody production by a Japanese group in the early 1980s [351]). During the same years, many inhibitory effects of estrogens on T cell and macrophage function were similarly reported (summarized in [350]).

Today, we know that one and the same disease like RA can be triggered by a prevailing influence of B cells (rituximab leads to remission) or – in contrast in another patient – by a prevailing influence of T cells (abatacept leads to remission). Even within an animal model two contrasting estrogen-sensitive pathophysiologies can exist [349]. This is similar to many chronic inflammatory diseases where either rituximab or abatacept leads to remission. From a clinical standpoint, whether the B cell or the T cell is dominant does not play a role because the disease has very similar clinical manifestations. Thus, the old B-cell and T-cell schools of disease pathophysiology were both correct, and who was correct simply depended on the prevailing immune mechanism in the affected patient.

It was hypothesized that B-cell diseases and T-cell diseases exist under the same clinically observable entity (summarized in [350]). If this is the case, estrogens with their stimulating effect on B cells would exacerbate B-cell diseases, whereas it is opposite for the T-cell disease/macrophage disease. The theory includes the following standpoint: when diseases during reproductive years are more frequent in women than men, they are estrogen-stimulated B-cell diseases. While this theory can unify the opposite standpoints, it has not yet led to a clinical characterization of subgroups of hormone-dominant immune pathophysiologies that can be used to treat patients.

In general, women are more affected by these diseases than men. We called it the women-to-men preponderance [352]. Why does it happen in the first place? Estrogens serve a fantastic immune supporting role, which is relevant during the reproductive years, during which women are at a higher risk to develop infectious diseases than men are (during cohabitation, pregnancy, postpartum, breastfeeding). At the expense of a higher risk of autoimmune diseases, estrogens protect women during episodes with a higher danger of infections.

Rickets, the English disease, or scientifically correctly rachitis is a condition of weak or even soft bones appearing in young children leading, e.g., to bowed legs, stunted growth, bone pain, and other skeletal abnormalities. In an article of 2018, it is described that rickets was a widespread phenomenon 2000 years ago in Roman times [353]. Francis Glisson (1597–1677) described the condition in much detail using strong expression like “… the whole bony system … flexible like wax that is rather liquid, so that the flabby and toneless legs scarcely sustain the weight of the superimposed body, so that the tibiae yield to the weight of the fabric pressing down on them from above and become bent…” [354].

Usually, we link the problem of rickets to Victorian England because it was widespread at that time. Thus, it is not surprising that English scientists were strong in this special field of inquiry. In 1906, F. Gowland Hopkins postulated a dietary factor highly important to overcome rickets [355]. In 1919, Edward Mellanby produced a model of rachitis by feeding young dogs with a low-fat diet, and he diagnosed the disease radiologically. In addition, he successfully treated the condition with cod-liver oil [356], which was later confirmed by Harriette Chick and colleagues in rachitic children at the University Kinderklinik in Vienna [357].

McCollum et al. [358] at Johns Hopkins University, Baltimore, took a great step toward discovering the mediator of these effects. They found that there exists a fourth substance next to the already known vitamin A, B, and C, in fat preparations or cod-liver oil, which they called vitamin D (the fourth letter in the alphabet). In the next years, seminal papers of Hess/Weinstock, Steenbock, Windaus/Hess, and Rosenheim/Webster more and more drew nearer to the molecular structure of vitamin D [359‒363]. All teams in London, Göttingen, and New York worked together “according to a friendly arrangement” [364]. Finally, the structure of the important provitamin was discovered simultaneously by Windaus/Hess in Göttingen/New York and Rosenheim/Webster in London [362, 365].

To summarize the discovery of vitamin D, it remains surprising that only Adolf Windaus received the Nobel Prize in Chemistry in 1928 “for the services rendered through his research into the constitution of the sterols and their connection with the vitamins” declared on the official site of the Nobel Prize homepage (https://www.nobelprize.org/prizes/chemistry/1928/summary/). A share of the prize between Windaus, Hess, and Rosenheim would have been more appropriate when the subject of the price would have been vitamin D alone.

In the late 1960s and early 1970s, the first evidence appeared indicative of a vitamin D receptor (VDR) [366‒368]. The receptor had a generalized tissue distribution beyond the classical target organs of the intestine (calcium uptake), bone (calcium incorporation), and kidneys (calcium reabsorption) [369]. In 1983, a 1,25-dihydroxyvitamin D3 receptor macromolecule was detected in peripheral mononuclear leukocytes and T and B lymphocytes from normal humans [370]. The gene for the VDR has been discovered in the later 1980s [371, 372]. Interestingly, over the past 3 decades Haussler, Pike and colleagues have demonstrated that 1,25-(OH)2 vitamin D3 works through a VDR-mediated mechanism that involves many coactivators and repressors or directly interact with and regulate hundreds of genes in the entire body (partly by epigenetic mechanisms) [373].

Clinical and translational studies finally showed the major effects of vitamin D3 that include bone metabolism, protection against bacterial and viral infections and immune system modulation [374, 375]. Late at the beginning of the XXI century, several findings indicated a complex interplay between viral infections and vitamin D, including the induction of an antiviral state, functional immunoregulatory features, interaction with cellular and viral factors, induction of autophagy and apoptosis, and genetic and epigenetic alterations [376]. The most recent evidences for the antiviral effects of vitamin D were shown from 2020 onwards in the context of the SARS-Cov-2 infection (COVID-19) and the effects of low vitamin D serum levels on COVID-19 severity [377].

Another set of extra-endocrine effects exerted by vitamin D3 concerned its immunoregulatory activities, first shown in 1981 with vitamin D-induced in vitro differentiation of mouse myeloid leukemia cells and vitamin D3 metabolism-dependent phagocytic functions of macrophages [378, 379]. In 1984, the combination of the presence of the VDR in human T lymphocytes and the effect of vitamin D3 (1,25-dihydroxyvitamin D3) in inhibiting the growth-promoting interleukin-2 produced by human T lymphocytes activated in vitro represented the first evidence of vitamin D3 as an immunoregulatory hormone [380].

As for the role of vitamin D in rheumatology, until November 2023 almost 6,000 publications (PubMed) have investigated it, reporting progressive achievements in basic and clinical research regarding the role of vitamin D3 in immune-mediated and autoimmune pathological conditions [381, 382]. From a practical point of view, data have confirmed that keeping serum 25(OH)D concentrations above 50 ng/mL (125 nmol/L) all year long reduces the risk for community outbreaks of infections, sepsis, and autoimmune disorders [383].

Several autoimmune disease can directly affect the CNS such as multiple sclerosis, Guillain-Barré syndrome, neuromyelitis optica, autoimmune hypophysitis, and some others. A peripheral immune cell is usually the stimulator of the local autoimmune problem in the brain. Other autoimmune disease such as neuropsychiatric SLE can severely affect the CNS through brain-reactive autoantibodies and CNS vasculitis leading even to overt psychosis [384]. These diseases/phenomena are not included in our presentation.

Although separation of organic CNS diseases based on central immune reactions of aggressive peripheral immune cells in contrast to “functional disorders” based on peripheral inflammation is difficult, we focused here on the latter bystander disease sequelae like fatigue, anxiety, insomnia, and depression. The observation of fatigue, anxiety, insomnia, and depression in patients with chronic inflammatory diseases goes back into the 1930s and even before (summarized in [385, 386]). In the 1960s and 1970s, smaller studies correctly identified anxiety and depression in patients with RA [387‒391] or SLE [386]. However, clinicians did not know the epidemiologic dimension of these sequelae for many years.

This markedly changed with the introduction of health scoring systems that also included questions towards anxiety and depression like the “Arthritis Impact Measurement Scales (AIMS)” in 1980 [392, 393], the “Stanford Health Assessment Questionnaire Functional Disability Index (HAQ)” [394], or the “short-form health survey (SF-36)” [395]. From 1990 onwards, with the use of scoring systems, many epidemiological studies clearly confirmed the frequent presence of fatigue, anxiety, insomnia, and depression [396‒402]. Anxiety, insomnia, and depression were linked to disease activity and pain [403‒405].

Similarly, fatigue was not a clinical manifestation that would have been recognized until the middle of the 1980s [406‒408]. Neurasthenia as it was called in much earlier times – from the 19th century onwards – had a strong negative connotation (e.g., [409]). Many rheumatologists recognized fatigue in the context of fibromyalgia, a disease that was not palpable for the typical physician in the 1980s/1990s or before. Modern systematic reviews and meta-analyses demonstrated a prevalence for fatigue in chronic rheumatic diseases in the range of 40–60% (healthy population 9–20%) and a prevalence for depression in the range of 10–20% (healthy population: 2–4%). Nowadays, the clinical problem is obvious.

Benjamin Hart was the first to have described sickness behavior in animals [410, 411]. Key characteristics are lethargy, sleepiness and depression, anorexia and less thirst, weight loss, reduced grooming and bodily activity, and he recognized that “it is a highly organized behavioral strategy … at times critical to the survival of an individual” [411]. He also mentions that the sickness response saves energy for making fever and installing an adequate immune reaction. He hypothesized that IL-1 is the essential central and peripheral cytokine in these very acute responses [411].

IL-1 was the cytokine of interest in finding an important sensory role of afferent vagus nerve fibers in 1994. Two groups independently discovered vagus-mediated sickness behavior including fever responses elicited by injection of proinflammatory substances into the peritoneal cavity [228‒230, 412]. The group of Robert Dantzer and Rosemarie Bluthé focused on IL-1 in their mechanistic studies of sickness behavior [413, 414]. The endogenous IL-1 receptor antagonist blocked IL-1 mediated effects in the CNS [415] (10,437). IL-1 also disturbed sexual behavior in rats injected with IL-1, and this disturbance was aggravated by parallel administration of TNF [416]. This latter phenomenon clearly points to the redundancy of cytokines in stimulating sickness behavior.

At the beginning of the 2000s, the link between cytokines and depression was more and more established. One study in human subjects was groundbreaking because the group of Andy Miller from Atlanta, GA, USA, showed that interferon alpha therapy in 40 melanoma patients led to severe depression, particularly, in those patients on parallel placebo therapy in contrast to parallel paroxetine [417]. The interferons entered the stage of sickness behavior.

In rheumatology, anti-TNF therapy was just approved for the treatment of patients with RA [418]. In 2004, in RA patients, the reduction of fatigue by anti-TNF strategies was first described in a progress report [419], and this was supported by a report on IL-1 receptor antagonist (anakinra) [420, 421]. The first clear indication of a highly beneficial effect on fatigue of biologic disease-modifying antirheumatic therapy was given in a study in psoriasis patients in 2006 [422] and in RA in the same year [423]. Soon other similar studies followed to confirm these favorable effects in other chronic inflammatory diseases.

We recently summarized the pathophysiologic underpinnings of cytokine-induced fatigue making use of many basic research studies (historical literature in this reference [424]). Robert Dantzer, in a review article of 2008, described the pathways from high inflammatory load to fatigue, anxiety, and finally major depression as a continuum with the end result of “decompensation of the mechanisms that regulate sickness behavior” [425]. Here, decompensation means breakdown of physiological mechanisms, unfortunately, leading to pathology. Recall that sickness behavior is an evolutionary positively selected physiological strategy of the body to preserve energy stores for the activated immune system [165]. An important aspect in decompensation might be adverse life events experienced in childhood and adolescence as described above [73].

Evolutionary medicine is a relatively new approach in psycho-neuro-endocrine immunology in rheumatic diseases that started in the later 2000s. This is not totally correct because aspects like sickness behavior (see above) have been seen through the lens of evolutionary medicine already in the late 1980s [411].

Combining evolutionary medicine with bodily energy regulation, however, results in a strong umbrella theory that can teach us much about the clinical manifestations of chronic inflammatory diseases [165, 204, 426‒432]. The model will not be repeated here in detail due to space constraints. In short, the theory says that physiological responses of the immune system and the brain – the two dominant selfish organ systems – were positively selected during evolution for transient, non-life-eliminating inflammatory episodes like infection or wound healing. In chronic inflammatory diseases, our body uses the same basically physiological responses for a too long time.

Suffice it to say that disease sequelae like fatigue, anxiety, depression, insomnia, anorexia, malnutrition, volume overload due to increased water retention, cachexia, cachectic obesity, insulin resistance and hyperinsulinemia, dyslipidemia, inadequate secretion of cortisol relative to inflammation, loss of androgens and higher conversion of androgens to estrogens, hypogonadism, local adipose tissue in inflamed regions, high sympathetic activity, hypertension, loss of local sympathetic nerve fibers, decreased parasympathetic activity, sprouting of sensory nerve fibers and sensory sensitization, inflammation-related anemia, and inflammation-related osteopenia can be explained using the theory. Considering evolved energy trade-offs helps us understand how an energy imbalance can lead to these disease sequelae [165, 204, 426‒432].

A brief history on psycho-neuro-endocrine immunology in the field of rheumatology unfortunately cannot address all relevant aspects of one specific field, and the authors are very aware of this shortcoming. Therefore, the reader must see this review as a viewpoint through the biased eyes of the authors. We apologize that we missed several publications that have paved the way to psycho-neuro-endocrine immunology of rheumatic diseases. In addition, including publications for an article like this one must necessarily stop at a certain time point, which in our situation is somewhere around the late 2000 years (but some papers were also included from the 2010s). We know that important articles have appeared thereafter.

For example, former critics now recognize dysfunction of the hypothalamic-pituitary adrenal axis as a relatively normal thing in patients with RA (the disproportion principle was explained in subchapter 3.5 Hormonal Pathways – Glucocorticoids). They base the new understanding on the hyporesponsiveness of the adrenal gland and on tissue glucocorticoid resistance [433]. Adrenal androgen deficiency became a well-known fact in clinical medicine, and many clinicians treat their patients with androgens off-label when serum levels of testosterone are low (but big studies do not exist). We have not reported on prolactin whose proinflammatory role for chronic inflammatory diseases was described early [434]. Similarly, no report was given on the role of vasoactive intestinal peptide, an important neuropeptide with anti-inflammatory activities [435, 436].

The hypogonadal situation in men and women with decreased ovarian reserve was not part of our presentation, although many studies since 2010 demonstrated this phenomenon. Investigations on thyroid abnormalities are still in a state of infancy although the low T3 syndrome is often present in our patients. Natriuretic hormones were not discussed, although we know that serum levels are often elevated indicating sympathetic activity, cardiac involvement, or volume overload due to water retention. We omitted mentioning the role of corticotropin-releasing hormone, melatonin, and α-MSH in chronic inflammation.

Finally, the literature on fatigue, anxiety, depression, and insomnia became innumerable since the year 2010. Most of it has not been reported here because the general pathophysiologic ideas were known before. Importantly, with the focus on fatigue, clinicians started to understand the broad ideas of psycho-neuro-endocrine immunology that entered practical rheumatology today.

When we look into the future, we still wish to see more therapies, which emanate from the field of psycho-neuro-endocrine immunology. The pharmaceutical companies addressed some of the ideas, but developments into this direction are still scarce.

The authors declare that they do not have any conflicts of interest related to the contents of the work.

There was no specific funding for this work.

R.H.S. created tables and wrote the draft version of the manuscript. M.C. discussed and corrected the text and drafted the part on vitamin D hormone.