Authors: Markus J. Schwarz
Categories: Basic Research
Source: Dialogues in Clinical Neuroscience
Recent research has overcome the old paradigms of the brain as an immunologically privileged organ, and of the exclusive role of neurotransmitters and neuropeptides as signal transducers in the central nervous system. Growing evidence suggests that the signal proteins of the immune system - the cytokines - are also involved in modulation of behavior and induction of psychiatric symptoms. This article gives an overview on the nature of cytokines and the proposed mechanisms of immune-to-brain interaction. The role of cytokines in psychiatric symptoms, syndromes, and disorders like sickness behavior, major depression, and schizophrenia are discussed together with recent immunogenetic findings.
Keywords: psychoneuroimmunology, cytokine, schizophrenia, major depression, sickness behavior, immunogenetics
The first evidence suggesting an interaction between the brain and the immune system came from studies performed 80 years ago by Russian investigators. Derived directly from a Pavlovian perspective on the conditioning of behavioral and physiological responses, a conditioned stimulus (CS) was repeatedly paired with injections of foreign proteins. Subsequent exposure to the CS, alone, was purported to induce antibody production in addition to a conditioned increase in a variety of non-specific defense responses.^1^ As this research attracted very little attention outside the Soviet Union, the commonly accepted beginning of research in the field of psycho-neuroimmunology (PNI) is associated with the experiments of Ader, who was studying taste aversion conditioning in rats in the 1970s. Conditioned animals that were reexposed to a CS, previously paired with the immunosuppressive effects of cyclophosphamide showed an attenuated antibody response to sheep red blood cells.^2^ The results of these initial experiments demonstrated that the immune system was subject to classical conditioning.
A more complex research direction in the field of PNI was the study of behavioral influences on immunity, starting in the 1950s with the research on stress and infectious disease.^3^ During the 1970s, Besedovsky was beginning to systematically investigate the neuroendocrine-immune system network with his studies on the effects of immune responses on neural and endocrine function.^4^ Felten described the direct contact of noradrenergic sympathetic nerve fibers with lymphocytes and macrophages.^5^ He showed that these nerve fibers were localized in specific compartments of lymphoid organs, forming close, synaptic-like neuroeffector junctions with T lymphocytes and macrophages.^6^ These “hard-wired” connections between the brain and the immune system have since been shown to be a major route for behavioral and central cytokine influences on immune function. They are, thus, a cornerstone for a mechanistic understanding of the signaling between the nervous system and immune system.
All these investigations demonstrated the influence of the central nervous system (CNS) on the immune function. However, this is only half the truth, as the brain-immune interaction is bidirectional. The old paradigm of the brain as an immunologically privileged organ may have inhibited the research of the immune system's action on brain and behavior. Meanwhile, it is commonly accepted that immune cells enter the brain even under normal, nonpathological conditions, and that all kinds of brain cells - neurons, glial and endothelial cells - are sensitive to the transmitters of the immune the cytokines.
Direct evidence for the neural activities of cytokines was first obtained after injections of various cytokines systemically or into the cerebral ventricles (intracerebroventricular [ICV]). These studies established that cytokines can activate the hypothalamus-pituitary-adrenal (HPA) axis,^7,9^ induce fever,^10^ prolong slow-wave sleep,^11^ reduce food^12^ and water intake,^13^ and decrease motility^14^ These effects were evident not only in experimental animal, but also in humans who received cytokine injections for cancer treatment.^15,16^ The most tested cytokine with regard to brain-immune interactions is interleukin-1 (IL-1), although other cytokines such as the tumor necrosis factor alpha (TNF-α),^17,18^ interferon (IFN),^19,20^ IL-2,^21^, IL-6,^22^ and IL-12^23^ can all induce one or several of the above responses.
A large number of studies have investigated the ways in which the cytokines influence brain function. Although there are still a lot of open questions, the following sections will try to give a short overview of the current knowledge of this part of PNI.
The term cytokine defines a large group of nonenzymic proteins that act as hormones. Their actions, as well as their target cell populations, are both diverse and overlapping. Once released into the environment, cytokines traverse small distances to ligate their high-affinity receptors via either autocrine or paracrine fashion. Upon ligation of the cytokine receptor, a signaling cascade is triggered resulting in an alteration in gene transcription by the target cell.
The pleiotropic features of cytokines have led to a confusing nomenclature because they were frequently named for their biological activity. It therefore repeatedly happened that a single cytokine was described by several names. Another problem in the nomenclature of cytokines emerged from their redundancy of action. In the 1970s, two terms were the term cytokine defined the large group of protein transmitters of the immune system. The term interleukin tried to list all known cytokines in numerical order. At the time this article was written, the list of ILs ended at IL-27.^24^ Unfortunately, some historical groups like the IFNs, TNF, lymphotoxins (LT), transforming growth factor beta (TGFβ), leukocyte inhibitory factor (LIF), and most of the colony stimulating factors (CSFs) remained to be known by their old names.^25^ This is particularly confusing, for example, the IFNs are divided into two totally different type I interferons (IFN-α, IFN-β and others) and type II interferon IFN-γ. On the other hand, the new group of chemokines was defined after IL-8 and several related cytokines were determined to form a distinct cytokine family.
According to the complexity and diversity of the cytokines, there are several possibilities for grouping them. They can be grouped by structural similarities, clustered chromosomal localization of their encoding genes, or principally similar functional features.
The typical cytokine is a glycosylated monomeric peptide of about 150 amino acids. Others are homodimers (eg, IL-5, M-CSF) or homotrimers (TNF-α and LT-α), or heterodimers (IL-12) or heterotrimers (LT-α/β).^26^ Three-dimensional structure studies have shown that many otherwise nonhomologous cytokines adopt similar conformations. These structural features of some cytokines permit their grouping into families. Members of the large IL-2/IL-4 family (including IL-2 to IL-7, IL-9, IL-11, IL-12 p35, IL-13, IL-15, type I and type II IFNs, and CSFs) share a common tertiary architecture characterized by bundles of four antiparallel α-helices in a spatially similar arrangement.^26-28^
Two important cytokine families show distinct
The cytokines are not members of a single gene superfamily. Remarkably few similarities have been noted in their primary nucleotide or amino acid sequences, and their genes are, for the most part, scattered throughout the genome. However, some chromosomal regions where cytokine coding genes are clustered are known. Most interestingly, some of these chromosomal regions seem to be associated with psychiatric disorders, especially schizophrenia.
One cluster of genes coding for members of the IL-10 family is located on chromosomal regions 1q32. These are the cytokines IL-10, IL-19, IL-20, and IL-24.^34^ This region is of major interest in genetic schizophrenia research, as several linkage studies identified a susceptibility locus for schizophrenia there.^35-37^ Indeed, a recent study points to the IL-10 gene itself as a susceptibility gene for schizophrenia.^38^
With exception of IL-18, the members of the IL-1 family are encoded by closely linked genes on the long arm of chromosome 2.^31^ An association of polymorphisms in the genes coding for IL-1α, IL-1β, and IL-1ra with schizophrenia was reported by Katila and colleagues.^39^ Although this finding was not replicated, another group described bifrontal temporal gray matter volume deficits and generalized white matter tissue deficits in schizophrenia patients who were carriers of a distinct polymorphism in the IL-1β gene.^40^
The cytokine gene cluster on chromosomal region 5q23-35 contains genes coding for IL-3, IL-4, IL-5, IL-9, IL-12 p40, IL-13, GM-CSF, and others.^26^ This is of particular interest, as linkage studies pointed to a possible susceptibility locus for schizophrenia in this chromosomal region.^41^ The same region was recently identified as containing the susceptibility gene for Crohn's disease.^42^
The genes coding for three members of the TNF family, TNF-α, TNF-β, and LT-β, are located in an immunologically important the HLA-III region, which is embedded between the HLA-II and the HLA-I region on the short arm of chromosome 6. Again, genome scans have indicated a linkage of this chromosomal region with schizophrenia.^43^ Boin and colleagues reported the association of a functional single nucleotide polymorphism in the TNF-α gene with schizophrenia,^44^ but a replication study performed by our group could not confirm this finding.^45^
The type I IFNs are encoded on chromosomal region 9p22. This cluster contains about 15 closely linked functional IFN-α and IFN-ω genes in addition to a single IFN-β gene.^46^ All members of the IFN-α/β family (IFN-α, IFN-β, IFN-ω, and IFN-τ) show at least 30% homology in their amino acid sequence.^25^ No conclusive data are available indicating a susceptibility gene for a psychiatric disorder in this chromosomal region.
Cytokines are most commonly grouped by their functional similarities, although this kind of categorization is highly arbitrary with regard to their pleiotropy. One of the most prominent concepts used to discriminate two distinct ways that the specific immune system can react on environmental stimuli is the classification of T helper 1 (Th1) and T helper 2 (Th2) cell diversity. This classification is based on the cytokine production patterns of T helper cells and reflects the polarization of the immune answer to either a cell-mediated (Thl) or a humoral (Th2) immune response.^33^
Th1 cells mainly produce IFN-γ, IL-2, IL-12, IL-18, and TNF-β, while Th2 cells principally secrete IL-4, IL-5, IL-6, IL-10, IL-13, and TGFβ. TNF-α and IL-10 are commonly characterized as Th2-like cytokines, although they are synthesized by both Th1 and Th2 cells.^47,48^ IL-12 and IL-4 are essential for the development of Th1 and Th2, respectively.^49^ The Th1 system promotes cell-mediated immune responses against intracellular pathogens, whereas the Th2 system helps B cell maturation and promotes humoral immune responses against extracellular pathogens. Th1 and Th2 cytokines antagonize each other in promoting their own type of response, while suppressing the other type of helper cell. Which system will dominate over the other depends on the relative timing and ratio of IL-4 to IFN-γ and IL-12.^50-52^ Figure 1 may help to understand the balance between Th1 and Th2.
Figure 1. The balance between Th1 (cell-mediated) and Th2 (humoral) response of the adaptive (specific) immune system. IL, interleukin; IFN-γ, interferon gamma; Th1 , T helper 1; Th2, T helper 2.
Such a polarized development of T cells happens not only on the peripheral level, but also in the CNS. Although initiation of T-cell responses is unlikely to occur within the CNS, T cells and monocytes will be massively recruited if pathogens are placed into the cerebral ventricles.^53^ Perivascular macrophages, owing to their location close to the blood-brain barrier (BBB), can stimulate T cells to proliferate and secrete Th1 cytokines.^54^ Following extravasation into the CNS parenchyma, T cells also interact with intrinsic CNS cells, particularly microglia and astrocytes.^53^ Microglia progressively acquire a clear-cut macrophage phenotype in response to CNS injuries,^55^ and can induce the production of Th1 cytokine IL-12^56,57^ and of Th2 cytokines such as IL-10 and TGFβ.^53^ Astrocytes are also potential sources of TGFβ, which inhibits MHC II (major histocompatibility complex II) and ICAM-1 (intercellular adhesion molecule 1) expression in macrophage/microglia.^58^ Microglia and astrocytes also secrete chemokines that may affect the recruitment of Th1 and Th2 cells. In sum, a complex network between microglia, astrocytes, and T cells is involved in the balance between Th1 and Th2 systems, which in turn might have impact on immune responses within the CNS.
One of the major reasons why the brain has long been defined as an immunologically privileged organ is the presence of a tight barrier between the brain and the the BBB. To our current knowledge, there are five pathways via which cytokine signals may cross the BBB.
There is evidence for an active, saturable, and specific transport system for certain cytokines across the BBB. By the use of radiolabeled cytokines in animal experiments, cytokines like IL-1, TNF-α, and IL-2 were demonstrated to be transported across the BBB.^59-62^ These experiments suggest that active transport plays a significant role in getting cytokines across the BBB. One limitation, however, is that the absorption of labeled cytokines into the brain tissue may not reflect the transport of cytokines across the BBB, but the binding of cytokines to the BBB. Some data show that the majority of intravenously (IV) infused radiolabeled IL-1α can be found on brain endothelial cells,^63^ or on the surface and pinocytotic vesicles of the brain endothelia shortly after injection.^64^
The second pathway is that cytokines may affect the C the circumventricular organs (CVOs), which possess a leaky BBB. The CVOs are midline structures bordering the 3rd and 4th ventricles and are the only areas of the brain at are outside the BBB. CVOs are characterized by the small size, high permeability, and fenestrated capillaries. These barrier-deficient areas are recognized as important sites for communicating with the cerebrospinal fluid and between the brain and peripheral organs via blood-borne products. CVOs include the following structures^65,66^:
Lesions of the OVLT suppressed intraperitoneal (IP) lipopolysaccharide (LPS)-induced fever^67,68^ and removal of AP-blocked IL-1-induced c-fos expression in the paraventricular nucleus,^69^ indicating the important role of these CVOs in transmitting the peripheral cytokines into the brain. However, there are also controversial results, showing the opposite effect.^70,71^ The discrepant results may be attributable to the extent of the lesion and the different doses of LPS and IL-1 used in these studies. Altogether, it seems that low doses LPS and IL-1 may specifically affect the CVOs and high doses of LPS and IL-1 may gain access to CNS at other sites.^72^
The third pathway for cytokines to engage the CNS is the vagus nerve. Numerous studies have been published demonstrating the involvement of vagus nerve in peripheral cytokine-induced CNS responses. One of the first observations was that peripheral LPS-induced hyperalgesia can be blocked by vagotomy, indicating that afferent vagal pathways innervate specific regions of the brain as a key connection between peripheral cytokines and the CNS.^73^ Others reported the role of the vagus nerve in inducing fever,^74^ activating the HPA axis, depleting norepinephrine in the hypothalamus,^75^ prolonging slow- wave sleep,^76^ and suppressing food-motivated behavior.^77^ Thus, there is major evidence that vagal afferents are important for conveying signals generated from IP injection of low doses of LPS. The role of vagal afferents is more important for the behavioral depression that develops in response to peripheral immune stimuli than for the induction of fever and activation of the HPA axis.^78^
The fourth pathway is that peripheral immune stimuli may induce the production of cytokines by cells of the BBB, which then secret cytokines into the brain parenchyma. In situ hybridization studies showed that the cells of the BBB respond to peripheral immune stimulation by producing IL-1,^79^ IL-6,^80^ and TNF-α.^81^ Thus, during systemic immune challenge, production of cytokines by cells of the BBB may result in widespread cytokine activity in the entire CNS. This is consistent with a report that the IL-1 bioactivity can be found in all brain regions after high-dose peripheral LPS injection.^82^
Local action of cytokines at many brain sites may actually be mediated via the receptors on endothelial cells. This binding induces another important effect of peripheral cytokines on cells of the BBB: the induction of cyclooxygenase-2 (COX-2), a rate-limiting enzyme of prostaglandin synthesis. Predominantly IL-1 and TNF-α induce the expression of COX-2 in endothelial cells of the BBB.^83,84^
As many cytokine-induced CNS effects can be blocked by COX inhibitors,^85,86^ the cytokine-induced COX-2 activity in BBB cells may represent a central mechanism of cytokine-CNS interaction.
Finally, it has to be considered that cytokines may enter the brain via infiltrating leukocytes. It has long been known that leukocytes may enter the brain under both normal and pathological conditions.^87^ In normal brain, scattered and random crossing of the BBB by leukocytes provide immune surveillance for the CNS.^88^ Under pathological conditions such as bacterial meningitis, activated leukocytes expressing inflammatory cytokines may infiltrate the brain.^89,90^ Additionally, CNS action of cytokines may weaken the BBB, promoting an increased infiltration of cytokine producing leukocytes.
The above mechanisms for the entrance of cytokines into the brain highlight the limitations of measuring peripheral levels of cytokines in neuropsychiatrie disorders. Cytokine levels in the blood (ie, serum or plasma) may reflect the systemic immune status and have been established as useful clinical markers in septic shock, inflammatory disorders, or cancers,^91^ but cannot conclusively clarify the cytokine expression within the CNS.
The specificity of the response to cytokines is provided by their receptors. Thus the expression of cytokine receptors is necessary for signal transmission of the cytokines entering the brain. Table 1 lists some selected cytokines and their receptors that are expressed within the CNS.
As the known effects of cytokines on the brain physiology are extremely numerous and complex, we will just give a few examples.
The effect of IFNs on neurons starts very early during brain development, where they regulate neuronal migration and differentiation.^92^ In vitro and in vivo studies showed the modulating effect of IFNs on the production of prolactin^93^ and - of particular interest regarding psychopathology - on the catecholaminergic, dopaminergic, serotonergic, and glutamatergic neurotransmitter systems, eg, the induction of transcriptional activity of the serotonin (5-hydroxytryptamine [5-HT]) transporter.^94-97^
TNF-α regulates the secretion of norepinephrine in the brain.^98^ Peripheral administration of TNF-α induces the cerebral tryptophan content^99^ and the synthesis of 5-HT and dopamine.^100^
There is experimental evidence that IL-1 can activate the 5-HT transporter thereby increasing the reuptake of 5-HT from the synaptic cleft.^101^ Furthermore, IL-1 enhances nonrapid eye movement (NREM) sleep and activates the serotonergic system.^102^ It has also been reported that the NREM sleep enhancement induced by IL-1 is partially inhibited by brain 5-HT depletion,^103^ suggesting that this IL-1 effect is partly mediated by the serotonergic system. Observations that the biological activities of IL-1 and 5-HT overlap to a large extent suggest that interactions between these two systems may be relevant to the manifestation of behavior under a variety of conditions. In contrast to IL-1, the Th2 cytokine IL-10 reduces NREM sleep.^104^
IL-2 can affect gene expression, neuronal activity, and neurotransmitter release in brain regions subserving sleep, memory and cognition, locomotion, and neuroendocrine function. IL-2 modulates the neurotransmission of acetylcholine, dopamine, and norepinephrine in a biphasic manner.^105^ It appears to be a potent and specific regulator of neurotransmission in frontal cortex, hippocampus, striatum, and hypothalamus.^106^
IL-6 is produced by neurons, astrocytes, and microglia.^107^ This cytokine promotes neuronal differentiation and survival,^108^ and modulates the neurotransmitter systems summarized above.^109-111^ Several studies have investigated the influence of IL-6 on the production, release, and metabolism of 5-HT. Peripherally administered IL-6 increases the concentrations of tryptophan and the 5-HT metabolite 5-hydroxyindole acetic acid (5-HIAA) in the brain,^112-114^ and it has been proposed that the interaction between IL-6 and brain 5-HT is a complex process.^115^
Recent studies have demonstrated a new type of neural activity of cytokines. IL-1 appears to act on neurons in hippocampus and amygdala to inhibit long-term potentiation and weaken synaptic strength.^116,117^ A contrary effect was demonstrated for TNF-α. Astrocytes continuously release TNF-α to control synaptic strength. The group led by Beattie has demonstrated that TNF-a induces the expression of glutamatergic receptors of the AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazole pro prionate) type on neuronal axons.^118^ Inhibition of the TNF-α signal by specific antibodies or soluble receptors lead to a reduced AMPA receptor expression. These cytokine activities may be involved in the ability of cytokines to alter the neural processes of learning.^119,120^
Besides the direct action of cytokines on brain cells, a biochemical link between cytokines and 5-HT is provided by the IFN-γ-controlled tryptophan metabolism. The essential amino acid tryptophan is the precursor of two distinct metabolism pathways, leading to the products 5-HT or kynurenine (Figure 2) The enzyme indoleamine-2,3-dioxygenase (IDO) metabolizes tryptophan to kynurenine, which is then converted to quinolinic acid by the enzyme kynurenine hydroxylase. Both IDO and kynurenine hydroxylase are induced by IFN-γ. The activity of IDO is an important regulatory component in the control of lymphocyte proliferation.^121^ It induces a halt in the lymphocyte cell cycle due to the catabolism of tryptophan.^122^ The Th2 cytokines IL-4 and IL-10 inhibit the IFN-β-induced tryptophan catabolism by IDO.^123^ The enzyme IDO is located in several cell types including monocytes and microglial cells.^124^ Thus, an IFN-yinduced, IDO-mediated decrease in CNS tryptophan availability may lead to a serotonergic deficiency.
Figure 2. The essential amino acid tryptophan is converted either into the neurotransmitter serotonin, or into the neuroactive metabolite kynurenine, which is further degraded to quinolinic acid. The rate-limiting enzyme in the kynurenine pathway, indoleamine2,3-dioxygenase (IDO), and kynurenine hydroxylase (KYNHydrox) are activated by the cytokine interferon gamma (IFN-γ).
The IL-1 receptor was the first cytokine-related structure detected in the brain.^125,126^ In 1988, the group led by Dantzer demonstrated the functional relevance of IL-1 and its receptor in the IL-1 injected into the brain of rats induced conditioned taste aversion.^127^ In the same year, the effect of another cytokine, TNF-α, was introduced, when the group of Plata-Salaman showed the suppressing effect of IL-1 and TNF-α on feeding behavior.^128^ Later, it was shown that peripherally administered LPS induces the expression of IL-1α, IL-1β, TNF-α, and IL-6 in brain macrophages and microglia.^129-131^ This effect seems to be a key mechanism in the induction of non-specific sickness symptoms including fever, anorexia, hyperalgesia, and the so-called sickness behavior, which is characterized by weakness, malaise, Hstlessness, cognitive impairment, depressed mood, lethargy, and reduced feeding behavior.^132^
The first indication of a role for cytokines in sickness behavior in humans came from clinical trials in which recombinant or purified cytokines were used to treat specific cancers or chronic viral infections such as hepatitis B or C. The syndrome produced by these inflammatory mediators resembled those seen in patients with major depression (MD).^133^
The first report on the sickness behavior-inducing effect of cytokines was published by Smedley and colleagues, who treated patients with advanced locally recurrent breast cancer with a high dose (160 MU/week) of IFN-α.^15^ Within 1 h of administration, they observed influenzalike symptoms, which 1 week later were superseded by lethargy, anorexia, and nausea, with a consequent loss of weight in most patients. Other side effects included profound somnolence, confusion, and paresthesia. Low-dose IFN-α therapy (3-5 MU three times a week) induces less severe psychiatric symptoms such as irritability and depression accompanied by impaired concentration, lack of motivation, sleep disturbances, and decreased libido.^134^ Depressive symptoms induced by IFN-α or IL-2 therapy were described to be related to a decreased tryptophan availability.^135^
Not only sickness behavior, but also schizophrenia-like symptoms including agitation, cognitive impairment, disorientation, delusions, and hallucinations are induced by IL-2 and IFN-α.^136,137^ Denicoff and colleagues were the first to report dose- and time-related psychiatric side effects in cancer patients treated with recombinant IL-2 that ranged from brief to severe agitation and combat iveness, requiring antipsychotic therapy.^138^
Besides the observation in patients suffering from malignancies or chronic inflammatory diseases, experimental data in healthy humans confirmed that cytokines, particularly TNF-α and IL-6, induce depressed mood, anxiety, and memory impairment.^139^
The observations described above led to the hypothesis that sickness behavior may serve as a model for the immune-related pathophysiology of major depression (MD).^132^ In fact, there is a large body of evidence for an altered immune response in depressed patients.
As described above, IFN-γ is a characteristic marker of Th1 cells. IFN-γ is produced in higher amounts by lymphocytes of patients with MD than by those of healthy controls,^140^ and higher plasma levels of IFN-γ in depressed patients, accompanied by lower plasma tryptophan availability, were described.^141,142^ This gives additional evidence for a possible link between the Th1 -like cytokine IFN-γ and the IDO-related reduction in 5-HT availability in the CNS of depressed patients. Given a functional relationship among the Th1 -dominated immune system, the serotonergic system, and MD, antidepressant therapy should be adequate to induce a Th1 to Th2 shift. There are indeed some reports demonstrating the potency of antidepressants to significantly reduce the IFN-γ/IL-10 ratio in vitro^143^ and to suppress the Th1 response in patients.^144^
The most frequently investigated immune parameter in patients suffering from MD is IL-6. Most of the publications report a marked increase of in vitro IL-6 production^145^ or serum IL-6 levels in depressed patients.^146-151^ Since IL-6 is a prominent marker of monocyte activity, a predominant activation of the monocyte/macrophage system in MD was hypothesized.^152^ IL-6 may be involved in the modulation of the HPA axis.^153^ Activation of the HPA axis is one of the best-documented changes in MD.^154^ Furthermore, the relationship between psychological or physical stress and an enhanced IL-6 secretion in the peripheral immune system seems to be well established.^155-158^ Impaired stress coping is often observed in depressed patients. Thus, the high number of reports of elevated peripheral IL-6 levels in MD patients may be related to psychological stress.
On the other hand, there is evidence for a relationship between high peripheral IL-6 levels and elevated CNS 5-HT availability. IV or IP administration of IL-6 induced not only an activation of the HPA axis, but also an increase in brain tryptophan and 5-HT metabolism, whereas the norepinephrine metabolism was unaffected.^113^ Accordingly, IL-6 seems to mediate the activation of the HPA axis and the 5-HT CNS after administration of the endotoxin LPS.^112^ Thus, elevated plasma levels of IL-6 do not fit with the hypothesis of a 5-HT deficiency in MD. Rather, it should be recognized that an inherent heterogeneity exists in the etiology of depression and different neurotransmitter systems may be disturbed.
On the basis of the commonly accepted idea that MD may be a heterogeneous group of disease entities, the group of Arolt and Rothermundt investigated the difference between melancholic and nonmelancholic MD regarding their cytokine expression patterns.^159^ They detected profound differences between these diagnostic nonmelancholic patients showed increased counts of leukocytes, lymphocytes, and natural killer (NK) cells in the acute stage of disease and after 2 and 4 weeks of treatment. However, their in vitro production of the cytokines IL-2, IL-10, and IFN-γ was unchanged compared with that of healthy controls. Melancholic patients on the other hand demonstrated normal cell counts, but a decreased in vitro production of IL-2, IFN-γ, and IL-10 during the acute stage of disease. Following clinical improvement, cytokine production patterns normalized in these patients.
A pathophysiological role of cytokines is also discussed in the other major psychiatric disorder, schizophrenia. The reports of the psychotic symptoms inducing effects of IL-2 in cancer patients attracted attention of this Th1-like cytokine to the immunopsychiatric schizophrenia research. Early studies reported elevated IL-2 levels in cerebrospinal fluid of schizophrenia patients,^160,161^ but others failed to replicate these intriguing findings.^162-165^ IL-2 levels in serum were reported to be either increased^166^ or decreased.^167^
A significant decrease in the production of IL-2 by peripheral lymphocytes is one of the best-replicated immunological findings in schizophrenia.^168,169^ Some data suggested that decreased IL-2 production is associated with acute illness in patients who produce elevated amounts of autoantibodies, or in patients with later age at onset,^170-172^ though there are again some contradictory findings.^173,174^
The group led by Arolt repeatedly found a markedly decreased in vitro production of IFN-γ, but the association with psychopathological variables was not consistently replicable.^168,175,176^
Mittleman et al examined the cerebrospinal fluid of juvenile first-onset schizophrenia patients in comparison to juvenile patients with obsessive-compulsive disorder or attention deficit hyperactivity, and reported a reduction to undetectable levels of IFN-γ in most of the schizophrenia patients, in contrast to the measured levels in the other neuropsychiatrie patients. On the other hand, levels of the Th2 cytokine IL-4 were only detectable in cerebrospinal fluid of juvenile patients with schizophrenia, but were too low to be detected in that of the control groups.^165^
As stated above, the IL-10 gene was identified as a possible susceptibility gene for schizophrenia. it is remarkable that administration of clozapine at a 10^-4^ M concentration markedly reduces the in vitro production of IL-10 by stimulated lymphocytes.^177^ This may indicate the possible relevance of IL-10 in the pathophysiology of schizophrenia. IL-10 was in fact reported to be increased in the serum of schizophrenia patients.^178^ Two out of three in vitro stimulation studies showed no difference between schizophrenia patients and healthy controls,^174,179^ while one described a highly significant increase in IL-10 production of chronically ill schizophrenia patients, compared with healthy controls, with medium levels in the subgroup of paranoid schizophrenia patients.^180^
Peripheral administration of IL-6 induces increased dopamine and 5-HT turnovers in the hippocampus and frontal cortex of rodents, without influencing the metabolism of norepinephrine.^114^ Within the brain, IL-6 is produced not only by glial cells, but also by neurons.^181^ Thus, IL-6 is of interest in schizophrenia research. A remarkable number of publications report significantly increased serum IL-6 levels in schizophrenia patients. Some of them additionally found an association with duration of illness, negative symptoms, and treatmentresistant schizophrenia.^150,166,182,187^ However, these data are limited by several confounding factors influencing serum IL-6 levels, such as smoking, gender, age, body mass index, and ongoing infections, as well as clozapine treatment.^188,189^
In summary, the whole body of data on cytokines in schizophrenia indicates a relatively reduced production of Th1-like cytokines and a more pronounced production of Th2-like cytokines. This cytokine profile, together with the numerous findings of increased levels of circulating antibodies and other immunological data, prompted us to the hypothesis of a Th1 to Th2 shift in schizophrenia.^190,191^ A cytokine dysbalance like a Th2 predominance may be related to a disturbed neurodevelopment and brain maturation, as it is proposed as pathomorphologic correlate of schizophrenia.^192^
The interdisciplinary approach of PNI has led to an integrative view of the immune system and the nervous system. Meanwhile, it is commonly accepted that not only does the CNS influence the immune reaction, but also that the immune system, particularly via its hormones- the cytokines - acts on brain function and behavior. There is ample evidence for the contribution of cytokines in psychiatric symptoms, syndromes, and disorders, and the involvement of the immune system fits to other commonly accepted etiopathological concepts like the neuro-developmental hypothesis of schizophrenia.
Genetic research gives further evidence for the possible involvement of the cytokine system especially in schizophrenia. However, the exact mechanisms of (inter) action must be elucidated in further investigations. Immunopsychiatrists may learn from somatic disorders like the systemic lupus erythematosus (SLE), an inflammatory disease affecting many organ systems including the CNS. The CNS affection in SLE encompasses a wide spectrum of neurological and psychiatric features including dementia, anxiety, depression, and psychosis,^193^ and the causative role of cytokines, predominantly TNF-α, for the neuropsychiatrie symptoms of SLE was proposed.^134^ Another aspect for future research derives from first therapy approaches in psychiatric disorders based on immunological considerations. The report of the therapeutic efficacy of a COX-2 inhibitor in schizophrenia^194^ has particularly demonstrated the importance of immunological research in psychiatric disorders.
Thus, the new paradigm of brain-immune interaction appears to evoke new research and treatment strategies.