Neurons as Immunomodulators: from Rapid Neural Activity to Prolonged Regulation of Cytokines and Microglia

Abstract

Regulation of the brain’s neuroimmune system is central to development, normal function, and disease. Neuronal communication to microglia, the primary immune cells of the brain, is well known to involve purinergic signaling mediated via adenosine triphosphate (ATP) secretion and the cytokine fractalkine. Recent evidence shows that neurons release multiple cytokines beyond fractalkine, yet these are less studied and poorly understood. In contrast to ATP, cytokines are a class of signaling molecule that are much larger, with longer signaling and farther diffusion. We posit that neuron-expressed cytokines are an essential mechanism of neuron-microglia communication that arises as part of both normal learning and memory and in response to tissue pathology. Thus, neurons are underappreciated immunomodulatory cells that express diverse immunomodulatory signals. While neuronally sourced cytokines have been understudied, new technical advances make this a timely topic. The goal of this review is to define what is known about the cytokines expressed from neurons, how they are regulated, and the effects of these cytokines on microglia. We delineate key knowledge gaps and needs for new tools to define and analyze neuronal roles in immunomodulation. Given that cytokines are central regulators of microglial function, a broad new body of work is required to illuminate functional links between these neuronally expressed cytokines and sustained and transient microglial function.

Introduction

Regulation of the brain’s neuroimmune system is central to development, normal function, and disease. Microglia are the resident immune cell population of the brain and play essential roles in health, neurological diseases, and injury, with functions spanning development, learning and memory, neurodegenerative disease, traumatic brain injury, stroke, and others (1–3). Under homeostatic conditions, exquisite communication between neurons and microglia is essential for development and memory consolidation (4). Neurons signal to microglia to prune specific underused synapses to regulate neuronal connections and communication (5–7). Disease conditions can lead to loss of coordinated function, such that microglia aberrantly phagocytose neurons and synapses and release molecules that can damage cells (8, 9). Thus, it is essential to understand mechanisms of neuronal-microglial communication because crosstalk between these cells plays crucial roles in homeostasis and pathogenesis.

Well established mechanisms of neuronal signaling to microglia include secretion of ATP that signals to microglia through purinergic receptors, and secretion of the cytokine/chemokine, fractalkine, that signals to microglia via its receptor CX3CR1 (10–12). Recent evidence shows that neurons release multiple cytokines beyond fractalkine, yet these are less studied and poorly understood. Neurons release cytokines with both pro- and anti-inflammatory properties and apparently with both autocrine and paracrine signaling functions (13–16). Pro- and anti-inflammatory cytokines have the capacity to stimulate or suppress microglial reactivity and immune activity, respectively, with broad functional implications. In contrast to ATP and its byproduct adenosine diphosphate (ADP), cytokines comprise a class of signaling molecule that are much larger (~6–100kDA) and are predicted to possess more sustained signaling properties (tens of minutes-to-hours) (17) with farther diffusion spanning tens of cellular diameters, depending on the tissue (18, 19). Although most abundantly studied in the context of immunity, cytokines are pleiotropic signaling molecules that activate essential pathways involved in diverse physiological functions, such as tissue remodeling in angiogenesis (20) and embryonic patterning (21, 22), survival signaling and trophic support (23), cell death signaling (24, 25), and regulation of plasticity (26). Thus, cytokines provide an essential mode of communication required for homeostatic tissue function and response to pathogens or injury that goes well beyond traditional concepts of immunity. Understanding the roles and regulation of neuronally sources cytokines is crucial to elucidating key underappreciated effects of neural signaling.

Separate classes of cytokines spur functional changes in target cells or trigger cell migration. Cytokines induce immune cells into specific phenotypic programs to maintain homeostasis or take actions associated with acute immunity, allergic response, chronic activation, and many other functions, which may be tissue- and disease- specific (27–31) (Fig. 1). In the brain, cytokines trigger microglia to engulf synapses or pathogens, for example (32–34). Other cytokines are classified as chemokines, which trigger immune cell migration in a directed fashion toward a particular location, e.g., a site of injury or infection (35, 36). Some cytokines, such as colony stimulating factor 1 (CSF1), promote immune cell survival (maintenance), proliferation, and pro-inflammatory responses (37–39). While cytokine release in the brain is most commonly attributed to microglia, multiple lines of evidence reveal that neurons also release cytokines, as we review below (13–16). Because of the diverse roles that cytokines play in modulating microglial function, it is essential to understand how they are expressed in homeostatic and disease states. However, the cytokines expressed by neurons, and their functions, have received little attention to date.

Figure 1. Neurons secrete cytokines that signal to diverse cells.

(A) Neurons (purple) secrete multiple species of cytokine signaling molecules that diffuse within the tissue and interact with diverse other cell types: astrocytes (blue), microglia (yellow), etc. (B) Extracellular cytokines (purple circles) bind with extracellular receptors (blue) on both neurons (autocrine) and other cell types (paracrine) to activate multiple intracellular signaling cascades that regulate membrane trafficking, survival, calcium signaling, apoptosis, and more. Created in BioRender. Wood, L. (2024) BioRender.com/b27w440.

While neuronally sourced cytokines have been understudied, new technical advances make this a timely topic. Traditional approaches, such as single nuclear RNA sequencing (snRNAseq), have proven challenging for studying cytokines from neurons. Cytokines are often lowly expressed in neurons at the transcript level, especially under homeostatic conditions, making it difficult to measure up- or down- regulation of neuronal cytokine expression. Histological analysis of cytokines is either ambiguous as to cellular source, for extracellular cytokines, or may mainly label the intracellular or trans-membrane form, which may or may not be active for signaling making it difficult to define the source of extracellular cytokines (40–42). New technologies, including cell-type specific proteomic labeling, hold the potential to define the interactome of neuronally sourced cytokines with receptors spanning neurons and other cell types. These technologies are in their infancy and have had limited application to study neuronal cytokines, perhaps due to conceptual gaps between fields that study neuronal activity, typically dynamic on the scales of millisecond, and cytokines and microglia, typically changing on the scale of hours or longer.

The goal of this review is to define what is known about the cytokines expressed from neurons and the effects of these cytokines on microglia. We contrast the physical and signaling properties of neuronally derived cytokines with other neuronal signaling. We analyze the substantial body of evidence pointing to critical roles for neuronal cytokine expression in the regulation of microglial functions in homeostasis and disease. We discuss the shared signaling pathways that regulate cytokine expression and synaptic plasticity and consider the potential physiological relevance of this overlap. We conclude by identifying key knowledge gaps that are primed to be addressed using emerging cell-type specific technologies. We hope that delineating these knowledge gaps will spur new tools to define and analyze neuron roles in immunomodulation.

Differences in the Spatial and Temporal Scales of Cytokine and Neurotransmitter Signaling

Neurons secrete multiple molecules that regulate other cells, including neurotransmitters and cytokines. Neurotransmitters and cytokines are often studied separately because they operate on different spatial and temporal scales (Fig. 2). Here, when discussing spatial and temporal scales of signaling molecules, we are referring to the distance that the molecule travels from release to receptor and the time from the stimulus that induced the signal to the duration of its effects. Classic neuron-to-neuron ionotropic neurotransmission signals within and around the synapse on a scale of about 20 nanometers (43). Because the neurotransmitters travel over small distances, they act on a time scale of hundreds of microseconds to tens of milliseconds (44). In contrast, cytokines can act at the synaptic scale, but also diffuse over longer distances across many cell lengths (18, 19) and even signal to the periphery via entry into the blood stream (45–50). These processes by which cytokines are transported over longer distance takes more time and typically cytokines are thought to act over hours-to-days. For example, the neurotransmitter glutamate is released and acts within and around the synaptic cleft (nanometers), with effects lasting from about 100 microseconds to 4 milliseconds on ionotropic receptors (51, 52). In contrast, the cytokine fractalkine exerts sustained signaling via both soluble (90kDa) and membrane-tethered forms that collectively enable signaling for minutes, hours, or longer (14, 53–55). Thus, the effects of cytokines in response to a stimulus can span from nanometers-to-meters and minutes-to-days, while the effects of ionotropic neurotransmission span nanometers and milliseconds. Both neurotransmitters and cytokines have their signaling terminated after being taken up via endocytosis followed by recycling for reuse (56–59). However, the timescales of cytokines recycling are not well understood in the brain. Similarly, the timescales and spatial extent of cytokine secretion have not been well-defined in the brain.

Figure 2: Differences in spatial and temporal scales of neural activity, cytokine signaling, and microglia.

(A) Left: Neurons (purple) communicate via millisecond-precise action potentials and neurotransmitter signaling. Center: Cytokine signaling can arise from neurons and affect microglial (yellow) morphology and function over much slower timescales on the order of minutes to hours. Right: These neuron-microglial interactions ultimately result in tissue remodeling including synaptic pruning and pathogen clearance or accumulation over hours, years, and lifetimes. Because neural activity and cytokines act on different timescales and with different modalities, they are often studied separately. Created in BioRender. Wood, L. (2024) BioRender.com/e78x989. (B) Breaking down the neuroimmune systems into large categories of signaling molecules (left) and cells (right), we note key differences and overlap in their spatial and temporal scales. While neurotransmitters are typically fast and local, cytokines are typically slower and span local and distal locations. Here we are defining signaling spatial scales as how far the signal travels from its release (for extracellular signals) or activation (for intracellular signals) to its receptor or other site of action. Signaling temporal scales span from the initial stimulus to the end of a signal’s effects after the stimulus ceases. We define the cellular timescales as spanning from the initial response to the stimulus to how long the response lasts after stimulus offset and the cellular spatial scale as how far a cell projects processes or travels. For example, neuronal processes can project locally within a region (microns) as well as from the brain to spine (meters), while microglia processes project locally but microglia also move around the brain (centimeters). For the purposes of this illustration, we excluded some neurotransmitters that also act like hormones, and we included synaptic changes separate from neurons themselves.

Cytokines and neurotransmitters that affect metabotropic receptors, like G-protein coupled receptors, are more similar, both in their temporal scales and downstream effects. Both cytokines and neurotransmitters can signal through G-protein coupled receptors (e.g., certain chemokine receptors and metabotropic glutamate receptors) (60). G-protein coupled receptors in turn activate intracellular phospho-protein signaling cascades, which have longer lasting effects on the cell, on the order of minutes to hours, although the exact receptors affected differ for each signaling molecule (61–63). This intracellular phospho-protein signaling can lead to changes in gene transcription and translation, ultimately causing changes in receptors and other mediators of the same signaling pathway (64, 65). This phospho-protein signaling is central to synaptic transmission and plasticity in neurons and in glial responses to the extracellular environment, including transitions of microglia from homeostatic into pro-inflammatory phenotypes (61, 66–70).

While neurotransmitters are only released by neurons, cytokines are released by multiple cell types. Release of cytokines arises from activation of multiple phospho-protein pathways, including the mitogen activated protein kinase (MAPK) pathway and the nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB) pathway (71–73). Both MAPK and NFκB, are present in neurons and microglia (74–77). Indeed, in neurons, these pathways respond to synaptic activity, potentially linking synaptic inputs in neurons to downstream cytokine release by neurons (74–77). We previously showed that different frequencies of electrical neural activity affect these phospho-protein signaling pathways and cytokine expression, revealing a previously unappreciated connection between specific frequencies of neural activity and neuroimmune signaling (78). These different patterns of cytokine expression may signal to microglia. Indeed, we found different frequencies of electrical neural activity have distinct effects on microglia morphology and transcription (78).Our findings build on discoveries showing that driving neural activity within the gamma frequency, specifically 40 Hz, transforms microglial morphology (79–82). However, little else is known about how specific patterns of neural electrical activity affect neuronal cytokine production and release. This divide between studies on neural activity and cytokines is partly due to the differences in timescales between patterns of electrical activity, cytokine production, and cytokine’s effects on microglia.

Many gaps remain in our understanding of the spatial and temporal scales of cytokine signaling in the brain. First, cytokines can be either membrane-bound or soluble, and they may be active in either one of these states, e.g., in the case of fractalkine, which is probably the best studied neuron-released cytokine (54). While the trans-membrane and soluble forms of fractalkine have increasingly well-established roles in homeostasis and disease, the roles of soluble and membrane-bound forms of other cytokines, such as TNF-α, are less well understood (53–55). Because soluble cytokines diffuse freely, while membrane-bound forms are anchored to the cell, these different forms of the same cytokine have very different spatial extents, with broad implications for which cells and receptors are accessible for signaling. Second, the subcellular location of cytokine release from neurons has received little attention. While neurotransmitters are released into the synaptic cleft, tens of nanometers from another neuron, cytokines may be released from other parts of the cell and have much farther to travel. However, the exact locations of release and distances traveled are unclear and may depend on the cytokine and neuron subtype. These gaps largely arise from limited methods to label cytokines in their tissue context and in their active form. Typical assays for cytokines are in tissue homogenate using assays such as ELISA (18). Histological labeling in tissue is possible, but this also labels cytokines within the cytoplasm or membrane-bound inactive pro-forms (83), which may not be involved in active signaling. Additionally, the signal-to-noise ratio is often poor. Thus, careful characterization of the scales of neuronal cytokine signaling are sorely needed and likely require new tools to label and track cytokines and their binding in brain tissue. These key unknowns point to strong needs for new bioengineering tools and analysis of cytokines in the nervous system.

Neurons Express Diverse Cytokines

Cytokines are well established intercellular signaling molecules expressed throughout the brain, with diverse functions ranging from control of microglia and infiltrating peripheral immune cells to regulation of angiogenesis. All major brain cell types are known to express cytokines, including microglia, astrocytes, endothelial cells, oligodendrocytes, and neurons (84–91). Myenteric and submucosal neurons in the gut express the cytokines IL-6 (92) and IL-18 (93) as part of the enteric neuroimmune system as reviewed in Wang et al. (93). However, until recently, only fractalkine, encoded by CX3CL1, has been widely attributed to secretion from neurons within the central nervous system (11, 12). We know little about the roles of neuronally sourced cytokines in the brain and more about the role of cytokines from other cells acting on neurons. Importantly, single cell and single nucleus RNA sequencing approaches have shed limited insight into the cellular origins of cytokines because cytokines are lowly expressed at the transcript level.

Within the context of healthy brain function, fractalkine (also known as CX3CL1) is well established to be involved in development and learning and memory (11, 12). Fractalkine primarily acts through its receptor, CX3CR1, which is exclusively expressed by microglia in the parenchyma, although, it is also expressed by T cells and monocytes in the periphery (94). Importantly, fractalkine is expressed in both soluble and membrane tethered forms. Neuronal secretion of the soluble form acts as a chemokine to recruit distant microglia, while the trans-membrane form promotes microglial adhesion to synapses for the engulfment of that synapse (12). This cytokine- and microglia-mediated synaptic engulfment is crucial to normal development to remove inactive and unnecessary synapses. This same process can go awry in neurodegenerative diseases to remove needed synapses. Thus, fractalkine is an essential regulator of neuronal-microglial crosstalk that is essential for development and homeostatic tissue regulation during learning and memory. Several excellent reviews on the known and emerging roles of fractalkine in development, function, immunity, and tissue repair are available on this topic (95–97).

Beyond fractalkine, multiple studies have found that neurons express other cytokines with diverse functions spanning pro-inflammatory (IL-1α, IL-6), anti-inflammatory (IL-13), chemotactic (CXCL1, CCL2), and trophic (IL-6) properties (Fig. 3). Indeed, neuronal cytokines also contribute to pathogenesis of multiple diseases. For example, neurons have been shown to secrete the chemokine CXCL1 in herpes simplex virus 1 (HSV-1) infection-induced encephalitis (98). CXCL1 and CCL2 appear to control recruitment of monocyte and neutrophil peripheral immune cells that may exacerbate disease severity. Additionally, CXCL1 stimulates ERK and PI3K signaling pathways as well as neuronal tau phosphorylation in Alzheimer’s disease (99). Thus, neurons secrete chemokines in response to infection that can also exacerbate Alzheimer’s disease pathology.

Figure 3: Neuronally sourced cytokines signal to other neurons, astrocytes, and microglia.

Multiple studies have revealed that neurons (purple) secrete cytokines (purple circles) capable of signaling to neurons and to other cell types, including astrocytes (blue) and microglia (yellow). Created in BioRender. Wood, L. (2024) BioRender.com/f33o743.

Neurons secrete pro-inflammatory cytokines with the capacity to stimulate glial reactivity, which can be protective or damaging, depending on the context. For example, recent work established that primary rat neurons secreted Interleukin 1α (IL-1α) in response to treatment with the anticancer drug oxaliplatin. IL-1α knockdown led to reduced neuron viability during co-culture with astrocytes (100). Intrathecal administration of IL-1α to rats treated with the chemotherapeutic oxaliplatin reduced mechanical and thermal hypersensitivity, i.e., pain. Ex vivo analysis revealed that IL-1α reduced glutamate release caused by oxaliplatin in the spinal cord and increased GFAP labeling. Thus, neuronal release of IL-1α plays an inhibitory role in chemotherapy-induced pain. Another study found that whisker stimulation, which induces neuronal electrical activity, increased expression of neuronal IL-1β in the rat somatosensory cortex (101). Epilepsy-associated malformations of the cortex were associated with increased neuronal IL-1β, suggesting that IL-1β may contribute to epileptogenicity (102). Interleukin 2 (IL-2), a pro-inflammatory cytokine known for recruiting T-cell lymphocytes, has been found to be elevated in Alzheimer’s disease. (103). IL-2 has potential protective roles via its ability to stimulate astrocyte reactivity and to recruit astrocytes to Aβ plaques with associated reduction of soluble and insoluble Aβ levels (103). IL-2 has also been found to co-label with NeuN in brain regions associated with sensorimotor gating (104), suggesting it is neuronally sourced. In another study of focal cortical dysplasia (FCD), elevated IL-2 and its receptor, IL-2R, were observed in FCD samples and particularly expressed in hypertrophic neurons (105). Thus, these reports together illustrate that neuronal pro-inflammatory cytokines are a key element of neuronal function and may be involved in both neuroprotection and neuronal dysregulation.

Neurons also secrete anti-inflammatory cytokines, which act in opposition to immune activation, including inhibiting pro-inflammatory cytokines. For example, IL-9 and its receptor, IL-9R have both been shown to be expressed by neurons in vitro (88). In neuron culture studies, IL-9 was found to protect neurons against apoptosis by downregulating the pro-apoptotic factor, Bax. Moreover, both IL-13 and its receptor IL-13Ra1 have been discovered to be neuronally and synoptically localized in rodent and human brains (106). Activation of IL-13 signaling increased NMDAR and AMPAR levels and synaptic activity. Additionally, IL-13 was found to protect against excitotoxic cell death (106). Collectively, these findings indicate that neurons secrete cytokines with classically defined anti-inflammatory properties and they exert pleotropic support of neuronal function and health.

While neurons secrete cytokines, few studies have examined how specific patterns of neural activity affects neuronal cytokine production. Signaling pathways involved in cytokine production are activated in response to synaptic inputs (74–77), suggesting that the patterns of synaptic inputs could alter neuronal cytokine production. This is especially important to understand because extensive research has shown that different frequencies of neural activity are associated with different brain functions (107). Our own work has shown that inducing 40 Hz or 20Hz frequencies of neural activity via sensory stimulation increases or decreases cytokine expression respectively. Furthermore, 20Hz and 40Hz sensory stimulation differentially affect microglia morphology and microglia transcription, with 20Hz inducing a more homeostatic or surveillant phenotype and 40Hz inducing a more engulfing phenotype (78). Furthermore driving 40 Hz neural activity recruits microglia into an engulfing state in mouse models of Alzheimer’s disease (108–110). These studies emphasize the profound connection between neuronal activity, cytokines and brain immune activity, and motivate the need for further work to directly interrogate neuronally sourced cytokines.

While prior methods have focused on individual cytokines, new methods enable a broader and more holistic definition of cytokine expression in healthy and disease states. In particular, we and collaborators used a novel ribosomal biotin ligase method driven by Camk2a-Cre, to profile a panel of 32 neuronal pro-inflammatory, anti-inflammatory, and chemotactic cytokines in the mouse cortex, thalamus, pons medulla, and cerebellum (111). Importantly, 31 out of 32 of these cytokines were detectable in the biotinylated neuron-derived signal. Moreover, the signals of five of these were greater than 50% neuron-derived (IL-10, IL-2, TNF-α, IL-1α, RANTES) in homeostatic animals. This and related approaches can be more broadly applied to define the cytokine proteome and interactome. These studies emphasize that we do not yet have a clear picture of the breadth of cytokines sourced from neurons nor of the cell types on which these cytokines act. Moreover, we do not yet understand the neuronal sub-populations that are capable of expressing cytokines and how expression of each cytokine changes from homeostasis to disease. Thus, there is an essential need for state-of-the-art strategies to define the breadth of neuronally-derived cytokines in homeostasis and brain disease.

Dual Roles for Kinases in Neural Plasticity and Immune function

Although the mechanisms that govern neuronally sourced cytokines are not well studied, cytokine production and expression in multiple cell types is controlled by canonical phospho-protein signaling pathways including NFκB and MAPK, which are expressed in neurons (64, 65, 71–73). These pathways are activated by upstream receptor tyrosine kinases (RTKs), pattern recognition receptors (PRRs), and G-coupled protein receptors as well as by influx of calcium (Fig. 4) (71–77). Kinase pathway activity leads to phosphorylation of transcription factors, such as AP-1 and NFκB, that enable them to translocate to the nucleus, in turn driving transcription of diverse cytokines, including those discussed in the prior section (71–73, 112).

Figure 4: MAPK and NFκB intracellular signaling have dual roles in cytokine release and synaptic plasticity.

The MAPK pathway (left, green) is composed of a series of serine/threonine MAP kinases that function to phosphorylate, or activate, the next kinase. In the canonical NFκB pathway (right, brown) phosphorylation of inhibitor of κB kinase (IKK) complex leads to phosphorylation of NFκB, which, in turn, translocates to the nucleus to induce transcription. In neurons, these pathways respond to signaling induced by calcium influx from synaptic activity as well as other extracellular signaling (grey box above). This signaling cascade leads to gene transcription and translation (grey oval below). MAPK and NFκB signaling leads to transcription of genes that ultimately lead to cytokine release. These signaling pathways also lead to transcription of genes for synaptic proteins that are inserted into the membrane to alter synaptic strength. Other effects include cell growth and proliferation, apoptosis, and inflammation. Thus, these pathways in neurons have dual functions in both synaptic plasticity and immune signaling. Created in BioRender. Singer, A. (2024) BioRender.com/x30m922.

These same signaling pathways involved in cytokine production also play essential roles in synaptic plasticity and learning and memory (76, 77, 113–129). Using genetic manipulations and pharmacological inhibition in multiple animal models, prior work has shown that MAPK and NFκB are essential for intact learning and memory, with these behavioral effects thought to be mediated via synaptic plasticity, the strengthening or weakening of synaptic signaling between neurons (76, 113–119, 126). Thus, MAPK and NFκB play dual roles in immune regulation and learning, with the potential for crosstalk between these functions. Indeed, recruitment of both immune function and plasticity have been shown during development, when synaptic pruning is high, and during injury, when brain tissue undergoes drastic remodeling (4, 123, 130–132). However, due to differences in how cytokine signaling and synaptic plasticity are assessed, these roles are often studied separately. Thus, the potential crosstalk between these signaling pathways in plasticity and immune function has not been well-elaborated in normal brain function and in disease.

Importantly, activation of both MAPK and NFκB in neurons results from synaptic inputs that arise from neural electrical activity (74–77). Thus, these pathways link neural activity patterns, synaptic plasticity, and cytokine release. Synaptic inputs arriving at a high rate in neurons depolarize the cell and open voltage-sensitive calcium channels leading to calcium influx (133, 134). This calcium influx in turn phosphorylates MAPK and NFκB pathways (74–77). The activation of these pathways ultimately leads to gene transcription and translation (Fig. 4) (135, 136). The effects of synaptic inputs on MAPK and NFκB have been discovered in studies examining mechanisms of synaptic plasticity. Signaling molecules within the MAPK pathway, including mitogen activated protein kinase (MEK) and extracellular signal-regulated kinase (ERK), have been shown to respond to high frequency synaptic inputs and cause increases in AMPAR signaling and long-term potentiation— the persistent strengthening of synapses between neurons (74, 75, 120–122). Using causal manipulations of MAPK signaling, prior studies have shown this pathway is required for intact learning and memory, providing foundational evidence that synaptic strengthening plays in role in learning (113–116, 118, 119). While less studied than MAPK, NFκB has also been shown to respond to synaptic inputs and is required for learning and memory, with evidence that it acts via calcium/calmodulin-dependent protein kinase II (CAMKII) (76, 77, 126, 127).

While neuronal studies of MAPK and NFκB have largely focused on their roles in synaptic plasticity, the known role of these pathways in cytokine release and immune regulation underscores a similar potential in neurons. Like microglia, which have dual roles in synaptic pruning and pathology clearance, MAPK and NFκB also play dual roles in these processes (112, 125, 137). Such overlapping mechanisms to recruit these functions may have important physiological consequences. For example, after injury damaged neurons must be cleared (138, 139). Remaining neurons then form new connections to maintain circuit function after some neurons are lost. Neurons are similarly cleared in chronic diseases, such as Alzheimer’s (140). Thus, both clearance of damaged cells or debris and enhanced plasticity play key roles in recovering from damage and coordinating these processes via overlapping molecular signaling may be advantageous. Elucidating the dual roles of these pathways in synaptic plasticity and cytokine release will reveal new ways in which neurons regulate brain immune function. The simultaneous recruitment of plasticity and immune pathways likely has important implications for learning and disease that are yet to be discovered. Furthermore, new tools to simultaneously control plasticity and immune pathways could be the foundation for new therapeutic strategies.

Conclusion

Neuronal cytokine expression is an essential, but understudied mechanism of neuron-microglial communication. Understanding the diversity of neuronal cytokines and their actions is crucial because this crosstalk is essential to normal development, learning and memory, and multiple neurological diseases. Importantly, neuronal release of cytokines could be an important mechanism by which neural activity is transduced into immune signals. Neuronal release of cytokines represents a fundamentally different form of neuronal signaling, with different spatial and temporal characteristics than traditional neurotransmitters. Recent studies have revealed a plethora of neuronally sourced cytokines spanning pro- and anti-inflammatory functions and chemotaxis, yet much is still unknown about their effects. Interestingly, intracellular signaling pathways in neurons that regulate synaptic plasticity also regulate cytokine expression, however the intersection of these two functions is rarely studied. Part of the challenge to understanding neuronal cytokine function is bridging two bodies of literature with different approaches and terminologies. Another challenge is technological limitations that have restricted our ability to measure and manipulate cytokines from specific neuronal sources (Table 1).

Table 1:

Essential questions in neuronal cytokine biology.

Essential Question	Technical Challenges
What are all of the players in the neural cytokine interactome? Need to identify:  • All neuronally secreted cytokines  • All receptors for neuronally secreted cytokines on neurons, microglia, and other brain cells	Limited sensitivity of current proteomic approaches and labeling of cytokines arising from a specific cell-type.
What are the trigger mechanisms of neuronal cytokine secretion?	Lack of real-time reporters for cytokine expression and secretion.
What are the relevant time scales of neuronal cytokine actions? Need to know:  • Duration of action on target cells  • Dynamics of cytokine transcription, translation, and release as a function of stimulus	Lack of real-time reporters for cytokine expression, secretion, and binding.
What are the relevant spatial scales of neuronal cytokine actions? Need to locate:  • Release, e.g., soma versus synapse  • Receptors on neurons or glia and proximity to release location  • Cytokine range: bound versus free floating cytokines	Limited sensitivity and specificity of spatial transcriptomics and proteomics. Lack of conformation-specific antibodies. Lack of real-time reporters.

While there is growing evidence for neuronally sourced cytokines and their role in the regulation of brain function, there is still much we do not understand about this form of neuronal signaling (Table 1). At a basic level, we do not know the breadth of cytokines that arise from neurons nor the exact sub-cellular locations where they are released. A full accounting of these cytokines, their receptors, and their release and receptor locations are needed to obtain a holistic picture of where these signals arise and act. Here we focused on the effects of these cytokines on microglia, but cytokines also signal to astrocytes, neurons, endothelial cells, and more (84–91). Because cytokines cross the blood brain barrier (45–50), they can act outside the brain, however it is unclear if neuronally sourced cytokines are produced at levels high enough to elicit effects outside the nervous system. Given the abundance of neurons in the periphery, including sensory and motor neurons and the enteric nervous system, there is also much to learn about how cytokines sourced from these neuronal compartments affect their local environments and, potentially, the brain. Furthermore, we need to elucidate the potentially diverse mechanisms and pathways that trigger neuronal release of different cytokines, which will go a long way to understanding the functional role of the cytokines. For example, cytokines that are released in response to rapid synaptic inputs could play a role in plasticity, while cytokines that are released in response to neuronal damage likely play a role in tissue remodeling.

Ultimately, we need new tools to specifically measure and manipulate neuronal-sourced cytokines to determine their roles in brain function, behavior, and disease. These new tools, yet to be established, combined with cutting-edge application of existing methodologies will open this nascent field, with rich potential to illuminate novel mechanisms of brain homeostasis and disease.