Neuroimmune Signaling: Cytokines and the CNS

Abstract

Cytokines are important signaling molecules synthesized by immune cells in peripheral tissues and the blood, and by glial cells and other brain-resident cells in the central nervous system (CNS). One characteristic feature of cytokines is their functional redundancy and pleiotropism; a wide variety of cell types respond to cytokines, thereby regulating systemic homeostatic functions including host responses to infection, immune system signaling, inflammation, and trauma. This chapter presents cytokine–CNS interactions on three levels: chemicals, cells, and their coordinated involvement at the systems level to alter behavior.

It is well established that cytokines are mediators of both innate and adaptive immunity and that they signal through ligand binding to cell surface receptors, activating JAK–STAT or MAPK pathways. Cytokines target cells with cognate receptors on their cell surface, which typically results in the recruitment of other immune cells and secretion of more cytokines. In health, cytokines in the periphery and the CNS are not expressed or are expressed in low concentrations, but in response to immune challenge, peripheral cytokines and immune cells like T cells can pass through a protective blood–brain barrier (BBB) using several active transport systems. In addition to peripheral cytokines crossing the BBB to affect the CNS, cells like microglia and brain mast cells secrete cytokines on the brain side of the BBB, a major consequence of which is neuroinflammation.

At the systems level, cytokines of peripheral origin can modulate the CNS by acting on the hypothalamic–pituitary–adrenal (HPA) axis, or via afferent inputs of the vagus nerve, which mediate communication between the peripheral immune system and the CNS through cholinergic signaling. Active cell-cytokine signaling can produce a sickness phenotype that has been demonstrated to alter emotion, cognition, and other behaviors by increasing depression, decreasing spatial learning, altering sexual behavior, and suppressing aggression.

Due to their varied and important involvement in many physiological processes, cytokine activity is very tightly regulated. As such, aberrant cytokine signaling correlates to several pathophysiological states, particularly those associated with inflammatory neurological disorders, some of which are briefly discussed in this chapter.

Appendix: Cytokine Families

Interleukins

Interleukins (literally meaning, “between leukocytes”) were the first cytokines to be studied and are often called the original or “prototypic” inflammatory cytokine. They have long been known to mediate communication between cells to regulate cell growth, differentiation, and motility – a role that extends far beyond mounting an immune response, although this is by far its most important contribution. In the body, the physiological function of interleukins, which are designated numerically, can be described with some clarity. For example, interleukin-1 (IL-1) and IL-2 are primarily responsible for activating B cells and T cells; IL-2 is a stimulant of B- and T-cell growth and maturation; IL-12 initiates a larger proportion of cytotoxic T cells and NK cells to be produced, and IL-1 and IL-6 are important mediators of inflammation. Direct administration of IL-1 into the brain induces sickness behaviors and typically requires a lower dose than peripheral administration.

The IL-1 family of cytokines comprises 11 secreted, soluble factors, composed of three related proteins that are the products of separate genes. Of these, IL-1α, IL-1β, IL-18, and IL-33 are known to play an important role in host defense and immune system signaling. IL-1α and β have similar signaling functions by binding with the cell surface receptor IL-1RI. Some IL-1 ligands (e.g., pro-IL-1-α) are formed as biologically active precursors, whereas others (e.g., pro-IL-1-β) must first be cleaved by a caspase protease.

In the periphery of the body, monocytes and macrophages are the primary source of IL-1, but in the healthy CNS, several endogenous brain immune cells express low levels of all members of the IL-1 family. Circulating immune cells that invade the BBB during inflammation are a source of brain IL-1, but it is mainly produced in the hypothalamus by microglia. Although the cellular source is still debated and not much is known about its regulation, microglia are considered to be an early source of IL-1 once there is infection or injury. Astrocytes, oligodendroglia, neurons, cerebrovascular cells, and circulating immune cells also produce IL-1; numerous targets of IL-1 signaling include T and B cells, monocytes, and microglia.

IL-1 acts on virtually all cell types in the body, including the brain. Much of what is known about the regulation of IL-1 expression in the brain has come from studies conducted on peripheral monocytes; nevertheless, similar mechanisms are thought to be operating in the CNS. The quickest production of de novo IL-1 synthesis comes following trauma, and there are, of course, several points of regulation available at the levels of transcription, translation, and posttranslation. Since its discovery over 30 years ago, it is now known to be produced by many cells – from the immune and nonimmune system alike – and to signal through the IL-1 type I receptor, IL-1R1. It has been demonstrated that IL-1 has the ability to activate expression of corticotrophin-releasing hormone (CRH) in rat hypothalamic explants, an effect that is blocked by nonselective COX inhibition.

IL-1 signaling transmission can occur through cells even with a very low receptor number (10²) because of a cascade that serves to amplify signaling. Once the ligand (i.e., IL-1) binds its receptor, it recruits an AcP and the immune adaptor protein MyD88. Ligand occupancy results in phosphorylation of the intracellular domain of the receptor. Through IL-1R-associated kinase (IKRAK) complexes with the AcP, other proteins such as TRAF6 are activated. More recent evidence showing IL-1 acting independently of IL-1R1 in the brain suggests that other receptors have yet to be identified. All members of the IL-1 family can be expressed in the brain.

Nevertheless, it is yet unclear whether IL-1 has a major role in the healthy CNS, since normal levels are almost non-detectable. Moreover, IL-1 KO mice undergo seemingly normal development. Despite this, evidence does seem to suggest a normal functioning role for endogenous IL-1 in the form of physiologic sleep regulation. For example, central administration of IL-1 was found in rabbits to promote non-rapid eye movement sleep, and IL-1 type I receptor KO mice reportedly sleep fewer hours than controls. Along with IL-1, the pro-inflammatory mediator TNF-α has also been shown to modulate mammalian circadian rhythms, appearing to stimulate the suprachiasmatic nucleus (SCN) in response to LPS injection.

IL-1 signaling has been implicated in diseases of chronic neurodegeneration, such as MS, PD, and AD. Despite conflicting in vivo and in vitro data, IL-1β is considered a promising clinical target in several neuropathologies.

IL-6 , another important multifunctional cytokine, displays both pro-inflammatory and anti-inflammatory effects, notably mediating an acute phase response to infection. Signaling occurs through its membrane-bound receptor IL-6R in so-called classical signaling and through β-subunit gp130 in “trans-signaling.” Signaling through gp130 is shared with several other cytokines. Most anti-inflammatory signaling occurs through the classical signaling pathway; conversely, pro-inflammatory actions of IL-6 have been attributed to trans-signaling. Additionally, brain endothelial cells secrete IL-6 from the luminal surface in response to application of LPS.

In a mouse model of depression , IL-6 has been shown to increase susceptibility to depression, suggesting that inflammatory T-helper cells promote depressive behavior. In practice, anti-cytokine therapy has been used with caution, because septic shock is an overwhelming concern. One well-described mechanism for severe depression involves IL-6 activation of IDO, which can produce tryptophan catabolites, thereby decreasing the availability of 5-HT.

Yet another important interleukin in the CNS is IL-2, which is known to be a T-cell growth factor , since it is produced mainly by T cells and stimulates T- and B-cell growth and differentiation. Unlike IL-1 and IL-6, IL-2 does not appear to be transported across the BBB by a saturable mechanism in normal rodents. One observation in support of this claim is that IL-2 is rapidly degraded in the brain, and there is further consideration of a possible IL-2 receptor further slowing transport ability across the BBB.

In some cases, another interleukin, IL-15, can substitute for IL-2 function, and although they often share the same receptors for signaling, at this time, it is not yet clear whether IL-2 or IL-15 is preferentially signaling in the CNS, but microglia express IL-2 and IL-15 at both the transcript and protein level.

IL-18 is a putative member of the IL-1 family and is produced by microglia and astrocytes, primarily causing further activation of microglia, but also affects T, B, and NK cells. Despite sharing significant sequence homology and structure, IL-18 does not appear to have shared biological function with other IL-1 members. At least in the rat brain, there appears to be constitutive rather than facultative expression of IL-18. Additionally, in vitro, IL-18 has been shown to display effects opposite to those of IL-1, namely, reducing CRH expression after short-term incubation in primary cultures of rat cortical microglia and astrocytes. IL-18 signaling has been demonstrated in the CNS, with localization to the hypothalamus, pituitary, and adrenal cortex (suggesting neuroendocrine function); further localization in the cortex, striatum, hippocampus, and cerebellum suggests additional physiological roles.

Chemokines

Chemokines – chemotactic cytokines – are the largest family of cytokines in humans; their name comes from “kinos,” which is Greek for movement. Inducible chemokine expression is generally modulated by pro-inflammatory stimuli. Their main function is to stimulate chemotaxis of macrophages, neutrophils, and other lymphocytes to sites of injury, damage, or infection. In this way, chemokines regulate leukocyte infiltration of the BBB during inflammatory and infectious disease.

Chemokines are seemingly involved in all pathologies that have an inflammatory component and are employed in host-defense mechanisms. CX3CL1/CX3CR1 signaling is responsible for microglia phagocytosis of neurons under both healthy and pathological conditions and during development contributes to pruning of synaptic circuitry in the CNS. As a result of decreased CX3CL1/CX3CR1 signaling during development in a murine model, researchers report autism-like behavior, with deficits in social interaction and increases in repetitive behaviors. Moreover, CXCL1 can modify inflammatory cytokine production, changing the neuroprotective activity of microglia in several neurological diseases, including ALS, focal cerebral ischemia, PD, MS, and AD.

Chemokines are small proteins of approximately 80 amino acids in size that are classified into four main subfamilies based on the positioning of two of their four conserved cysteine resides; these constitute a motif near the N-terminus: α (CXC), β (CC), δ (CX3C), and γ (XC). The α chemokine family includes IL-8, GRO, and NAP2, which have primary effects on neutrophils. In the β subfamily, there is not an intervening amino acid between conserved cysteine resides; members include MCP-1 and RANTES, which exhibit primary effects on monocytes, although MCP-1 and RANTES also are known to induce histamine release from basophils and activate CD4⁺ and CD8⁺ T cells. There are two receptors each for α and β chemokines, which do not apparently cross-compete.

Chemokines exert their effects through interactions with specific seven-membrane/transmembrane spanning G-protein-coupled receptors (GPCRs) on target cells. Specific roles of receptor phosphorylation sites on chemokine receptors have not yet been elucidated. Recent studies in the CNS demonstrate that chemokines and their receptors are constitutively active in glial and neuronal cells of normal adult brains, suggesting their role as a neurotransmitter.

Expression of fractalkine/CX3CL1 is the best known neuronally derived chemokine and is highly expressed throughout the CNS. In a mouse model of diet-induced hypothalamic inflammation, CX3CL1 was rapidly induced in local neurons, and its inhibition was shown to impair the expected phenotypes of obesity and glucose intolerance. In other studies, it has been suggested that CX3CL1 is involved in protective plasticity. For example, Sheridan and colleagues (2014) reported upregulation of CX3CL1 in rat hippocampus 2 h after spatial learning in a water maze test and found that in primary dissociated cultures, CX3CL1 reduced glutamate-mediated intracellular calcium rises in neurons and glia . The discovery of a unique CX3CL1 receptor on microglial cells (CX3CR1) has also led to the hypothesis that there are specific microglia–neuron interactions that might be mediated through CX3CL1/CX3CR1. Many of the studies designed to test this hypothesis have revealed that the phagocytic activity of microglia can be modulated by CX3CL1/CX3CR1. In the healthy brain, microscopy studies using cx3crf ^GFP/+ mice reveal a dynamic microglial–neuronal relationship, in which microglia branches reach out toward neurons to find harmful secreted products or damage, poised to influence neuron survival.

Cytokine signaling through receptors initiates responses leading to cell proliferation that can involve either Ras signaling or the activation of protein tyrosine kinases. After conversion from GDP-Ras to GTP-Ras, Raf is activated, triggering MAP kinase cascade for transcriptional control via cis-acting sequences such as AP1. On the other hand, Jak kinases respond to chemokine signaling through a physical association with the intracellular domains of the receptor; this leads to tyrosine phosphorylation of one or more receptor complex chains.

Growth Factors

TGF-βs are a superfamily of cytokines comprised of over 50 members, widely recognized for their multifunctional and contextual role in key events of development, disease, and repair. The name TGF comes from their ability to transform cell lines by inducing anchorage-independent growth. Since their initial discovery, additional functions have been discovered: cell cycle control, extracellular matrix formation, hematopoiesis, angiogenesis, chemotaxis, as well as immune functions.

There are four subfamilies of TGF-β, grouped according to sequence similarity: the GDNF, TGF-β, activins, and DVR group, more commonly referred to as bone morphogenic proteins (BMPs). One common structural feature of all these subfamilies is a cysteine knot motif, which is also shared by neurotrophins, PDGF, and glycoprotein hormones.

GDNF, a molecule widely acknowledged as a neurotrophic factor, is a member of a distant TGF-β subfamily that comprises GDNF, neurturin, persephin, and artemin/neublastin. In the presence of TGF-β, GDNF is an extremely powerful trophic factor for neurons in the PNS and CNS. Although they have several structural features in common with TGF-β, they do not signal through TGF-β serine/threonine kinase receptors but use a complex that combines the tyrosine kinase c-Ret and a GPI-anchored α-receptor.

TGF-β signals through several receptor types, based on their size, and are classified as type I, II, or III (53 kDa, 70–100 kDa, 200–400 kDa, respectively). Type I and II receptors are serine/threonine kinase receptors that serve to transduce signal from all members of the TGF-β family, with the exception of the GDNF family. Signaling of TGF-β from the cell membrane to the nucleus can be illustrated with the Smad proteins, whose discovery provided a breakthrough in the understanding of TGF-β-mediated signal transduction.

In adult CNS and PNS, TGF-β2 and TGF-β3 relativities are widely distributed. In the rat, TGF-β1 is restricted to meninges, while TGF-β2 and TGF-β3 appear in neurons and glia (astrocytes and Schwann cells but not oligodendrocytes). With the exception of motoneuron and aminergic nuclei, most thalamic and hindbrain regions are without TGF-β-immunoreactive cell bodies. By far, the highest levels of TGF-β-immunoreactive cells are in the cortical layers, retina, piriform cortex, hippocampus, and ventral spinal cord. Several studies suggest that TGF-β staining in the human brain is less robust than in rat brain but nevertheless has overlapping expression profiles. For example, glial cells exhibit weak staining of TGF-β in the human brain. Consistent with their spatial distribution in the nervous system, TGF-β has been shown to exert multiple effects both in vitro and in vivo. Although TGF-β is not a neurotrophic factor per se, it is critically involved in the regulation of neuronal survival.

Tumor Necrosis Factor

TNF can cross the BBB without disruption when produced by macrophages in response to insult or injury and is a sufficient mediator of both local and systemic inflammation. In the CNS, it acts as pro-inflammatory cytokine and also regulates temperature and the HPA axis of the neuroendocrine system . Moreover, TNF amplifies and prolongs inflammatory responses (and sometimes damage) through the activation of other cells that release cytokines like IL-1. For example, fever is produced by TNF-α in the presence of IL-1β.

Receptors for TNF are by and large type I integral membrane glycoproteins and include p75 and p55 TNF receptors (for TNF-α and TNF-β), Fas, and CD30 (a clinical marker for Hodgkin’s lymphoma), among a number of others.

Since TNF-α can be protective or destructive, there are many affecting factors that can modulate its effect on neurons. Circulating levels of TNF-α, for example, can indicate which paradoxical role is present. Although it plays a critical role in brain development, higher levels are associated with several neurodegenerative disorders. For example, in MS, increased TNF-α can destroy protective myelin in the brain. As a macrophage-derived cytokine, TNF-α engages with cell death receptors to kill compromised cells and if excessively produced results in septic shock.

Increases in TNF transport occur in an animal model of MS, EAE, have been shown not to be dependent on disruption of the BBB but rather on the saturable component of TNF transport.

Interferon

Although initially discovered for their role in antiviral host immune responses, interferon-α (IFNα) is an important mediator to both the innate and adaptive immune response. CNS-related effects include increased excitatory neuronal potential, temperature regulation, circadian rhythm, sleep, and emotion. Direct associations between IFNα and CNS diseases came out of research in humans that found that IFNα therapy often resulted in neurotoxicity and psychiatric complications, such as anxiety and depression. Rodent studies also support systemic IFNα treatment that causes neuropsychiatric via altered monoamine neurotransmission.

IFN signals through the common type I IFNα/β receptor (IFNAR), which results in receptor dimerization and phosphorylation of receptor proteins and transcription factors STAT1 and 2 by Jak. Once phosphorylated, STAT1/2 heterodimerize, migrate to the nucleus, and associate with IFN regulatory factor-9 (IRF-9) to form IFN-stimulated gene factor 3 (ISGF-3). Finally, through interactions with an ISRE in the promoter region of an IFNα responsive gene, IFNα effects are transduced. The most highly expressed IFNα-induced genes include ubiquitin-specific proteinase 18 (USP18) and STAT1. In mice, Wang and colleagues (2008) found a direct effect on the CNS and neurons of systemic IFNα, which has potential direct implications for how we understand the adverse actions of IFNα in human neurological disorders.

A variety of cells including astrocytes, microglia, and neurons from the CNS express IFNα upon direct viral stimulation, but the major source of IFNα comes from peripheral dendritic cells. LPS, a gram-negative cell wall product, is also a robust stimulator of IFNβ. Genes encoding IFNs and their receptors are expressed in the brain under both normal and pathological conditions.

IFNα is associated with a number of human CNS diseases, including CNS lupus and schizophrenia. As an FDA-approved agent used to treat chronic hepatitis B and C, as well as some leukemias, and lymphomas, IFNα has been demonstrated to produce considerable CNS side effects. It has been reported that patients undergoing chronic IFNα therapy experienced neurobehavioral disturbances ranging from memory loss and depression in as high as 50 % of patients.

Colony-Stimulating Factors

CSFs are secreted glycoproteins and bind to the surface of hemopoietic stem cells that causes them to differentiate. Signaling through CSF receptors is necessary for microglia viability and is neuroprotective in models of traumatic brain injury. In brain damage caused by ischemia, granulocyte-macrophage CSF (GM-CSF) causes a robust microglial cell response followed by neuronal survival. Conversely, absence of CSFs results in loss of microglia, grossly disrupted development of the brain, and specific olfactory deficits.

Neurotrophins

The neurotrophin protein family is comprised of a group of four small, structurally, and functionally related soluble growth factors that includes NGF, BDNF, neurotrophin-3 (NT3), and neurotrophin-4/5 (NT4/5). Biosynthesis and secretion of neurotrophins (also referred to as neurotrophic factors) in the mature nervous system are regulated by neurons. Compared to other growth factor families, which typically have a large range of target cells (including nonneuronal cell types), neurotrophins act on distinct neuronal populations, which provides them a potential advantage for therapeutic use.

During embryonic and postnatal development, neurotrophins regulate and shape neuronal survival and differentiation. Beyond a developmental context, there is strong support from studies performed in animals and cell culture that indicates that these proteins have broader biological contexts than simply regulating cell death – they also regulate processes underlying neuronal plasticity and neuronal structural integrity. In the cerebral cortex, neurotrophic factors stimulate both the growth and pruning of dendritic branches. In the adult brain and PNS, neurotrophins trigger or refine fast synaptic responses and neurite morphology, seemingly being made available to local tissue on demand, which influences higher-order functioning like behavior, learning, memory, and cognition.

Neurotrophins strongly inhibit or delay degenerative processes in several in vitro and in vivo models of neurodegenerative diseases such as AD, PD, and ALS. In clinical trials it was hoped that these proteins would inhibit degeneration of nerve cells, but to date these trials have been disappointing. All major neuronal populations affected by AD (e.g., cholinergic cells in the basal forebrain) respond to at least one neurotrophin.

An important yet nevertheless not completely understood observation about neurotrophic factors is that the response of neurons to these proteins can be quite diverse; as a result, it is commonly noted that neurotrophic signaling depends on cellular context. For this reasons, results obtained from investigations using one type of neuron cannot be reliably extended to other types. There are distinct yet overlapping areas of the brain that are responsive to specific neurotrophins. Cholinergic forebrain neurons respond weakly to BDNF, but strongly to NGF, whereas spinal motor neurons will respond strongly to BDNF but find adverse effects in the presence of NGF.

These proteins function as homodimers, and while the overall structure is shared, differences exist in terms of their respective patterns of charged basic or acidic residues, forming characteristic loops that then participate in receptor binding on a target nerve cell. Of note, neurotrophins tend to bind nonspecifically to cell membranes through hydrophilic and hydrophobic interactions. This limits their ability to diffuse through tissues like the brain parenchyma. Furthermore, individual neurotrophic factors have different diffusion rates; for example, it has been demonstrated that NGF has a higher diffusion rate than BDNF. Overall, these are important considerations when thinking of using these proteins as potential therapies.

Neurotrophic factors are known to signal by binding to two different classes of transmembrane receptor proteins found on responsive neurons. The first class of receptor is a type of receptor tyrosine kinase, the Trk protein, of which there are three subclasses (A, B, C); in many aspects Trk receptors are typical tyrosine kinases: ligand binding causes receptor dimerization, leading to autophosphorylation of the complex at defined tyrosine residues. Despite sharing extensive primary sequences in their intracellular domains, cellular responses to all three Trk receptors are diverse. The second class of receptor, p75, binds to all neurotrophins and is a member of tumor necrosis factor receptor (TNF-R) family of receptors. It is often repeated that Trk receptors form binding sites of higher affinity than p75, but these observations come from in vitro studies that have not yet been conclusively backed up by primary cell culture or by in situ work. Furthermore, the situation is even more complicated, as p75 can also bind to all three Trk receptors, which can modify tyrosine kinase activity as well as ligand specificity.

Neurotrophins and their receptors are very often expressed in malignant gliomas; recent evidence targeting Trk receptors in brain tumor-initiating cells suggests that inhibiting Trks could be a promising combinatorial treatment of malignant glioma.

BDNF is an extremely potent neurotrophic factor. Mature BDNF is generated by proteolytic cleavage of proBDNF. Immature BDNF (proBDNF) signals through the NGF receptor p75NTR and plays a role in hippocampal plasticity; mature BDNF activates TrkB. TrkB is widely expressed in the mature CNS and is developmentally regulated. Numerous studies in cell culture demonstrate that signaling through TrkB increases neuronal survival and hypertrophy in dependence on PI3-kinase and MEK signaling pathways. Overall, BDNF is an important factor in the generation, survival, and remodeling of cells in the brain.

Expression of BDNF and signaling is influenced by a variety of stimuli: calcium influx from synaptic activity, brain injury such as ischemia, and inflammation. Seizure activity increases BDNF expression throughout the hippocampus, and BDNF-TrkB signaling increases mossy fiber sprouting, thereby contributing to epileptogenesis.

BDNF has many roles in CNS function of relevance to health of the organism, such as a successful HPA stress response, and is closely involved in the development of neural circuits across many levels, from stem cell survival, to the maturation of refinement of developing circuits. When BDNF is directly infused into the brain in the dentate gyrus of the hippocampal formation or the subventricular zone (two areas of the brain where adult neurogenesis exists in mammals), the number of adult-born neurons increases. Work in mice indicates that BDNF is not necessarily a survival factor for neurons but rather is a differentiation factor.

Recent investigations of autism spectrum disorder (ASD) have found that BDNF levels are often increased in animal models of ASD and may underlie macrocephaly, or brain overgrowth, that is often reported to be overrepresented in ASD children but is not present at birth. At this point, it is not clear whether more neurons and/or glia are generated during the brain growth process in ASD, but trophic factors have been proposed to be involved.

Neuropoetins

Neuropoetins have a very well established role in the regulation of neuronal, glial, and immune responses in response to disease or injury; a subset of these cytokines are also involved in nervous system development through neurogenesis and modulation of stem cell fate. More recent evidence suggests that neuropoietic cytokines might contribute to the etiology of psychiatric disorders.

One neuropoetin, leukemia inhibitor factor (LIF), promotes self-renewal of ESCs via the JAK–STAT pathway in mice, but not in humans, at least in vitro. However, LIF is necessary for long-term growth of human embryonic neural stem cells and might also contribute to NSC differentiation into neurons, astrocytes, and oligodendrocytes. Although the self-renewal process of primitive neural stem cells is under the control of unknown factors, it appears that LIF, possibly in conjunction with ciliary neurotrophic factor (CNTF), induces changes later in development. For example, it has been shown that LIF can induce expression of Gfap. GFAP is an intermediate filament protein expressed by a number of different CNS cells, including astrocytes, and ependymal cells. Both LIF and CNTF promote the switch from nonadrenergic to cholinergic function in cultured sympathetic neurons. LIF modulates neuropeptide levels in sensory neurons; CNTF supports ciliary neuron survival in the CNS. Additionally, LIF has a well-documented role in injury-induced neurogenesis, such as in the case of spinal cord regeneration.