The pioneering experimental report on an interaction between psychological stress and the immune system was published by Selye in 1936, showing that chronic stress results in an atrophy of the thymus (148). Since then, numerous epidemiological studies have revealed that severe psychological stress has strong suppressive effects on the immune system (149). Such suppression of the immune system has significant implications on disease susceptibility and progression. Investigations in humans have revealed that chronic stress contributes to many illnesses including cardiovascular diseases and cancer (150-151). Moreover, substantial evidence has linked chronic or recurrent exposure to stress with exacerbation of inflammatory and autoimmunity diseases such asthma, rheumatoid arthritis, multiple sclerosis, inflammatory bowel disease and psoriasis (152-153) and an increased susceptibility to infectious diseases. The relationship between psychological stress and a higher vulnerability to infectious diseases has also been confirmed experimentally by laboratory stressors demonstrating an impaired responsiveness to Hepatitis B (154-157) and influenza virus vaccination (158-159). In addition, it was shown that human volunteers who were inoculated with five different strains of respiratory viruses showed a dose-dependent relationship between stress and clinical symptoms after infection (160).
Interaction of the immune system and the brain - the HPA and SNS axis
The hypothalamus-pituitary-adrenal (HPA) axis and the sympathetic nervous system (SNS) are the two major pathways by which the immune system is modulated during psychological stress (Figure 4).
The sympathetic nervous system (SNS)
The sympathetic nervous system, a major component of the autonomic nervous system, is a fast-acting response to stress that can be detected within seconds after a stress stimulus. The SNS exerts its function by the release of the sympathetic neurotransmitters, the catecholamines. The SNS originates in the central nervous system (CNS) (in nuclei within the brain stem and spinal cord) and gives rise to preganglionic efferent fibres that leave the CNS through the thoracolumbar region of the spinal cord (161). Most of the sympathetic
Introduction
20 preganglionic fibres travel to ganglia located in the paravertebral chains, where they synapse with postganglionic neurons. From there, the long postganglionic neurons extend across most of the body and their nerve terminals innervate nearly every organ in the body including primary and secondary lymphoid organs (162-163). In general, T cell zones, macrophages and plasma cells are richly innervated, while follicular zones of B cells are poorly innervated (162). These postganglionic sympathetic neurons are noradrenergic fibres, meaning that they act by locally releasing norepinephrine (NE).
Additionally, preganglionic sympathetic fibres that end in the adrenal medulla secrete acetylcholine, which activates the secretion of epinephrine (adrenaline) and norepinephrine (NE, noradrenaline) directly into the blood stream, where they exert their action as circulating hormones (161). The adrenal medulla, unlike the postganglionic sympathetic nerve terminals, releases mainly epinephrine, and to a much lesser extend NE. This response is also known as the sympathetic adreno-medullary response of the SNS. Thus, the principal end products of the SNS are catecholamines (NE and epinephrine) with norepinephrine acting both as a neurotransmitter when released via nerve fibres and as a hormone when it is released by the adrenal medulla into the blood along with adrenalin (161).
The hypothalamus-pituitary-adrenal (HPA) axis
The activation of the hypothalamus-pituitary-adrenal (HPA) axis takes place somewhat slower (usually within 3-5 min of stress onset). The main components of the HPA axis are the paraventricular nucleus (PVN) in the hypothalamus, the anterior pituitary gland located at the base of the brain, and the adrenal glands (164). Upon activation of the HPA axis corticotrophin-releasing hormone (CRH) and arginine-vasopressin (AVP) are secreted from the PVN of the hypothalamus into the hypophyseal portal blood, which in turn stimulate the expression of adrenocorticotropin hormone (ACTH) in the anterior pituitary gland. CRH seems to play a permissive role in ACTH secretion, whereas AVP has synergistic or additive effects, but very little ACTH secretory activity on its own. ACTH then circulates in the bloodstream where it acts on the adrenal cortex to induce the expression and release of adrenal steroid hormones, in particular glucocorticoids (GCs) into the blood (for an overview see (165)). Glucocorticoids (cortisol in humans and most mammals, corticosterone in rats and mice) represent the final effector molecules of the HPA axis that mediate their function primarily by acting directly on immune cells, which they access via the blood.
In addition, GCs play an important role in regulating the activity of the HPA axis under basal and stress conditions, by exerting a negative feedback control directly on the pituitary and also on the synthesis and secretion of CRH and AVP. For example GCs potently inhibit pituitary ACTH secretion and down-regulate the action of CRH through binding to their receptors (166). Moreover the secretion and effects of CRH and AVP are influenced by neurotransmitters in the hypothalamus.
21 Figure 4: Signalling between the central nervous system (CNS) and the immune system through the HPA axis and the SNS
The hypothalamus-pituitary-adrenal (HPA) axis exerts most of its influence systemically through a release of glucocorticoid hormones (GC). The first part of this axis, the hypothalamus, is located in the forebrain and neuroendocrine cells in the hypothalamus release corticotropin-releasing hormone (CRH) that stimulates the release of adrenocorticotropic hormone (ACTH) from the anterior pituitary. The adrenal cortex responds to ACTH by releasing glucocorticoid hormones into the circulation. The sympathetic nervous system (SNS) transmits sympathetic information to peripheral targets by releasing norepinephrine (NE) from noradrenergic nerve terminals that end in all primary and secondary lymphoid organs, and/or systemically by releasing epinephrine (along with some NE) from the adrenal medulla into the circulation. Dotted lines represent negative regulatory GC feedback pathways, blue lines represent bi-directional communication of peripheral immune events to the brain that involve the secretion of inflammatory cytokines like TNF-α, IL-1 and IL-6.
(Figure modified from Esther M. Sternberg, Nature Reviews Immunology, 2006.)
Summing up, the CNS regulates the immune system through two major classes of effector molecules: glucocorticoids, which are regulated in the hormonal stress response by the HPA axis, and the catecholamines norepinephrine and epinephrine, which are released either by the sympathetic adreno-medullary system or via postganglionic nerve fibres. Both branches closely interact with each other and have positive reverberating feedback loops at different levels. For instance, reciprocal neural connections exist between the CRH and noradrenergic neurons, with CRH and NE stimulating each other.
Bidirectional communication between the CNS and the immune system
Bidirectional communication between the CNS and the immune system allows the immune system to signal to the brain through neural and humoral routes. In fact, certain cytokines and in particular the pro-inflammatory cytokines TNF-α, IL-1 and IL-6 are known to activate both the SNS and the HPA-axis (167-168). For example TNF-α, IL-1 and IL-6 stimulate hypothalamic CRH and/or AVP secretion, resulting in the secretion of GCs. How these inflammatory cytokines pass the blood-brain barrier to reach the hypothalamic CRH and AVP neurons is unclear. However, this feedback loop constitutes an important mechanism by
Introduction
22 which GCs serve as a regulatory mechanism to prevent excessive activation of the immune response during infection (169-170). Studies in mice using viruses that elicit both strong early pro-inflammatory and later T cell responses (LCMV clones 13 and WE, influenza, HSV-1) have confirmed the release of endogenous GCs, whereas viruses that induce little or no inflammation do not stimulate significant GC induction (171). In this context it has been demonstrated that endogenous secretion of GCs protect the hostagainst cytokine-mediated pathologies during murine CMV infection (172). More specifically it has been shown that if GCs are removed by adrenalectomy, IL-12, IFN-γ, TNF-α, and IL-6 production increases and the mice die due to septic shock. Moreover, experimental evidence in mice have confirmed that an increase of endogenous GC levels plays a protective role for the host during experimentally induced autoimmune encephalomyelitis (173)and arthritis (174).
Glucocorticoid hormones
GC hormones are long known for their immunosuppressive effects and clinically, GCs and their synthetic analogues are used as potent immunosuppressive agents. Many of the immunomodulatory effects of stress have been attributed to the action of GC stress hormones, therefore the cellular and molecular mechanisms of GC-mediated immunosuppression will be introduced in the following paragraphs.
Molecular mechanism of GC-induced immunosuppression
GCs belong to the family of steroid hormones and their action is mainly mediated through binding to the respective cytoplasmic receptors. There are two main receptors in the cytoplasm for GCs, the glucocorticoid receptors (GR) and the mineralocorticoid receptor (MR). GCs have a higher affinity for MR than for GR (175), thus at low levels, GCs bind preferentially to the MR, only at high levels, e.g. during stress exposure, the GRs are occupied (176). The most accepted mechanism by which GCs enter the cell is through passive diffusion facilitated by their relative small size and lipophilic nature. However, it has also been proposed that GCs can mediate their action by binding to membrane-associated glucocorticoid response receptors (mGCR) and that this interaction might participate in a GC-mediated apoptosis (177-179), but the precise mechanism has not yet been identified.
The glucocorticoid receptors (GRs) belong to the nuclear hormone receptor superfamily that are present in the cytoplasm in an inactive state and form multi-protein complexes with hsp90 and other chaperons. Upon GC binding, GRs dissociate from this complex and translocate as a homodimer to the nucleus where they bind via a zinc finger motif in their DNA-binding domain to the glucocorticoid response element (GRE) (180). The bound GR homodimer then modulates gene expression directly by either up-regulation or down-regulation of target genes, depending on the GRE sequence and promoter context. A direct down-regulation of gene expression occurs mainly via binding to so-called negative glucocorticoid response elements (nGRE) (Figure 5).
23 Figure 5: Molecular mechanisms of GC action
Glucocorticoid hormones (GCs) passively diffuse into the cell and exert their effects by binding to membrane-bound glucocorticoid receptors (GR). Alternatively GCs bind to membrane-bound GR receptors.
In the cytosol GCs bind to the GRs resulting in the dissociation of the heat shock protein complex and a translocation of the ligand-bound GR into the nucleus. In the nucleus GRs modulate e.g. cytokine transcription either by directly binding to glucocorticoid response elements (GRE, nGRE) or via interaction with other transcription factors like NFκB.
GRs can also modulate gene expression through protein-protein interaction with other trancription factors such as NFκB (181-182), activator protein 1 (AP-1) (183-185), STAT and nuclear factor of activated T cells (NFAT) (186-188). Different mechanisms have been demonstrated by which GRs act to down-regulate gene expression via repression of NFκB activity. For example, GRs can induce the expression of the inhibitory protein IκB that in turn sequesters NFκB in the cytoplasm, thereby preventing its translocation into the nucleus (189-190). In addition, direct interaction between NFκB and GRs has also been shown to repress NFκB-dependent gene expression (189, 191-193) as well as a competition between GRs and NFκB for limited cofactors such as CREB-binding protein (CBP) (194).
Cellular mechanism of GC-induced immunosuppression
The most general effect of GCs is to inhibit synthesis and/or release of cytokines that promote inflammatory reactions. For example, GCs have been demonstrated to suppress pro-inflammatory cytokines like IL-1β, IL-6, IL-8, IL-12 and TNF-α (170, 195) and up-regulate anti-inflammatory cytokines like IL-4 and IL-10 (195-196). GCs can induce a shift from TCD4+ Th1 cytokine responses (with predominant secretion of IL-2, IFN-γ and TNF-β) to a Th2
Introduction
24 pattern (with predominant secretion of IL-10 and other anti-inflammatory cytokines) (164, 197-199).
Beside their capacity to modulate cytokine secretion, GCs have been shown to affect nearly all other aspects of immune cell functions such as cell trafficking, maturation and differentiation, antigen presentation, proliferation, and effector function (170). For instance, GCs can substantially impairthe function of T cells. They have been shown to reduce TCD8+
cell activation and proliferation (200). In particular, it has been suggested that GCs inhibit early events in T cell activation. For instance, short pre-treatment of Jurkat T cells with the synthetic GC dexamethasone (DEX) has been shown to inhibit the tyrosine-phosphorylation of ZAP-70, stimulated by CD3 cross-linking, whereas high doses of DEX showed opposite effects (201). Inhibition of T cell proliferation by GCs is accompanied by reduced IL-2 production and is thought to be mediated by GR interference with the transcription factor AP-1 (AP-183, 202-203). Modulation of IL-2 receptor α and β chain expression by GCs is also discussed as a potential mechanism how GCs inhibit the proliferation of T cells (204). While most reports indicate that GCs suppress T cell function, other studies demonstrate an enhancement of T cell function after GC exposure. For instance, in vitro treatment of splenic lymphocytes with corticosterone resulted in increased expression of IL-2 receptors after TCR stimulation when corticosterone was added within the first hour of the stimulus (139).
GCs are also potent inducers of apoptosis, and physiological GC concentrations as well as concentrations achieved during pharmacological GC treatment or chronic stress responses can cause the death of immature thymocytes being most prominent within the CD4+ CD8+ TCRlow thymocyte population (205). Massive GC-induced thymocyte apoptosis is a well-known phenomenon in response to chronic stress that contributes to thymic involution under stress (206-207). In contrast, resting peripheral T lymphocytes are comparatively resistant to GC-induced apoptosis (208).
Furthermore, GCs are also discussed to inhibit antigen-presenting cells by suppressing the generation, maturation, and immunostimulatory properties of DCs. For example GCs have been shown to selectively inhibit the expression of the costimulatory molecules CD80 and CD86 and to down-regulate MHC II expression. (209-210). Physiologically relevant GC levels have also been demonstrated to suppress the efficiency of presentation of MHC I/peptide complexes by virus-infected DCs by decreasing the production of antigenic peptides (209, 211).
Many of the above mentioned findings have been elicited experimentally in in vitro systems or by exogenous administration of GCs and their synthetic analogs. It is important to note that pharmacological doses or forms of GCs exert different effects on immune functions than they do under physiological stress conditions. For example, synthetic versions of GCs (e.g dexamethasone) have a significantly longer half-life (212-213) as well as higher affinity for GC receptors (214), therefore exerting much stronger effects than endogenous GCs.
25 Moreover, synthetic GCs do not bind to corticosteroid-binding globulin (CBG), a plasma protein that binds a large proportion of circulating GCs, thereby preventing its translocation to the nucleus (170). Therefore these approaches may not mirror the complex situations during a physiological stress response, making it difficult to extrapolate the outcome of chronic stress-induced alterations on immune reactions from these data.
Regulation of the HPA axis and GC action under chronic stress exposure
Recurrent or prolonged activation of the HPA axis as it occurs under chronic stress exposure can create an environment in which systemically elevated levels of GCs suppress various immune cell functions (169-170). Enhanced circulating GC levels may directly occur through chronic activation of the HPA axis or secondary through a reduced sensitivity of the HPA axis to negative feedback.
However, the action of endogenous GCs is not solely determined by their systemic concentration in the circulation but various factors selectively modulate the sensitivity of immune cells or tissues to GCs under basal conditions and during physiological stress responses (170). Under basal conditions there is a considerable heterogeneity among immune cells and tissues with respect to the expression of GC receptors (GRs), suggesting that different immune compartments and cells exhibit different responsiveness and sensitivity to GCs. Whereas the thymus and lymph nodes display high levels of GRs (with the thymus expressing the highest amounts), GR expression in the spleen is much lower and thymic T cells were found to express higher levels of GRs than T cells isolated from the spleen (215-216).
GR activation can also be modulated selectively during stress exposure. For instance, in a study investigating the effect of restraint stress on GR activation it was found that acute, stress-induced elevation of GCs results in GR activation in the thymus and LNs, whereas receptors in the spleen were not activated (217). Changes in plasma corticosteroid-binding globulin (CBG) in response to high levels of GCs or catecholamines play an important role by modulating GC action under chronic stress exposure. For example chronic stress exposure can be associated with an impaired production of corticosteroid-binding globulin (CBG) and can also reduce the GC binding capacity of CBG, resulting in increased levels of free and biologically active GC hormones (218-219). Rats exposed to chronic social stress were shown to exhibit a significant decrease in plasma CBG levels (218). This decrease in CBG levels led to greater access of free GC hormones to GRs in the spleen than is typically found under basal or acute stress conditions, providing one mechanism by which chronic stress has a greater impact than acute stress on splenic immune functions (218). The accessibility of GCs to its cytoplasmatic receptors can be further modulated after GCs have entered the cell. This occurs through the conversion of GCs from their active 11β-hydroxy- into its inactive 11-keto-form by the 11β-hydroxysteroid dehydrogenase (11β-HSD). The expression of 11β-HSD is known to be tissue-specific and its activity may also be altered distinctively during chronic GC exposure, providing a further mechanism by which tissue-specific GC sensitivity occurs
Introduction
26 (220). Consideration of these factors that regulate GR activity is important in order to understand the differential impact of GCs on immune responses during stress and how GCs can achieve selectivity in influencing specific immune functions in some tissues or cells but not in others.
Animal models of chronic stress and immunity
A key question is to whether and how the immunosuppressive effects of prolonged GC elevation promote and exacerbate the pathophysiology of infectious and other diseases under physiological stress responses. In this regard, numerous studies in animals have investigated the effect of a variety of psychological and physical chronic stress stimuli on the outcome of infectious diseases, with some studies reporting immunosuppressive effects and others showing enhancing effects (170, 221). These studies also provide evidence that a variety of conditions determine the magnitude of stress-related changes on immunity: (a) the duration of the stress-exposure (acute vs. chronic), (b) the timing of the stress exposure relative to immune activation (c) the nature of the stressor, and (d) the level of GCs secreted during a stress challenge (which to some extent in turn dependents on a and c) (221-222).
Social stress
In general, based on the nature of the stress stimulus, stressors can be grouped into different broad categories: (a) physical (e.g. electric, chemical), (b) psychological (e.g. restraint, exposure to novel environment, forced swimming) or (c) social (222). Animal models based on social stress are considered to be a behaviorally valid system to study the effects of relatively natural stressful situations since social stress is a chronic or recurring factor in the life of virtually all humans. Another reason why social stressors are implicated to be highly relevant in comparison to the other categories (223-224) is the observation that a social stressor, because of an often unpredictable nature, will not induce habituation of the stress response and will therefore not induce adaptation (225-228). Chronic social stress is known to stimulate both activation of the pituitary-adrenocortical system (HPA) and the sympathetic adreno-medullary system, resulting in elevated concentrations of free GCs and the catecholamines norepinephrine and epinephrine, respectively (219, 229).
The social disruption stress (SDR) model
‘Social disruption’ stress is a model of chronic social stress and mirrors a daily recurrent experience of stress. This stress model is based on the disruption of an established social hierarchy of group-housed male mice (residents) in their home cage, which is experimentally
‘Social disruption’ stress is a model of chronic social stress and mirrors a daily recurrent experience of stress. This stress model is based on the disruption of an established social hierarchy of group-housed male mice (residents) in their home cage, which is experimentally