Although connections between the immune and nervous systems have long been recognized, the precise mechanisms that underlie this relationship are just starting to be elucidated. Advances in sensory biology have unveiled novel mechanisms by which inflammatory cytokines promote itch and pain sensations to coordinate host-protective behavioral responses. Conversely, new evidence has emphasized the importance of immune cell regulation by sensory neurons. By focusing on itch biology and how it has been informed by the more established field of pain research, we highlight recent interdisciplinary studies that demonstrate how novel neuroimmune interactions underlie a diversity of sensory, inflammatory, and infectious diseases.
The skin is one of the first lines of defense against chemical, mechanical, microbial, and thermal insults. Although the immune system is an essential component of cutaneous immunity, it is increasingly evident that the sensory nervous system also plays a critical role in host defense. By evoking sensations such as pain and itch, the organism can immediately sense danger and rapidly initiate a protective behavioral response. Additionally, emerging evidence suggests sensory neurons further aid the body’s response to potentially harmful agents by directly modulating immune cell function through the release of mediators such as neuropeptides. In this review, we highlight recent advances in our understanding of how the sensory nervous system responds to and, in turn, regulates the immune system in the setting of cutaneous inflammation and immunity. Specifically, we will focus on itch biology and the studies in pain that helped to inform this burgeoning field of neuroimmunology (for more comprehensive reviews on pain and itch, see Refs. 1–5).
Neurophysiology of itch
The primary somatosensory neurons that innervate the skin are pseudounipolar, which means that their axons consist of two branches: one that synapses at the dorsal horn of the spinal cord (or brainstem) and another that innervates the peripheral tissue. Its cell body is housed in a peripheral ganglion, either the trigeminal ganglion, if the nerve innervates the face/oral cavity, or the dorsal root ganglion (DRG), if it innervates elsewhere in the body. Itch is sensed following the activation of receptors at peripheral sensory nerve endings in the skin, which ultimately triggers the opening of nonspecific cation channels such as transient receptor potential A1 (TRPA1) or TRPV1 (Fig. 1) (6). The resulting membrane depolarization, if sufficient, results in the opening of voltage-gated sodium channels, such as NaV1.7 and NaV1.8, which initiate and propagate the action potential (Fig. 1) (7). These signals are then rapidly transmitted by the peripheral sensory neuron to the spinal cord, where neurons of the CNS relay the signal to the brain to evoke sensory perception. Yet, although sensory neurons primarily transmit afferent information from the skin to the CNS, they can also release mediators (e.g., neuropeptides) in an efferent manner (such as through an axonal reflex) to communicate with other cell types in the peripheral tissue.
Delineating the precise identity of skin-innervating neurons involved in itch transmission is a highly active area of investigation, as somatosensory neurons are remarkably diverse. Traditionally, somatosensory neurons are classified based on their size, conduction velocity, and degree of myelination. Itch is believed to be largely transmitted by small, unmyelinated, slow-conducting c-fibers; although thinly myelinated, Aδ fibers may also play a role (8). Both of these fibers also transmit pain. Itch was originally believed to be mediated by the same neurons that signaled pain, with the intensity of neuronal firing coding which signal is transmitted (9). Indeed, both itch- and pain-sensory neurons employ many of the same ion channels to transmit their signals, including NaV1.7, NaV1.8, TRPA1, and TRPV1 (10). However, through the discovery of gastrin-releasing peptide receptor (GRPR), Mas-related G protein-coupled receptor (Mrgpr), and the neuropeptide Nppb, it is now recognized that there are specialized pathways that can distinctly mediate itch in the periphery (4, 5, 11–14). Currently, the expression of these hallmark receptors and neuropeptides is used to classify itch-sensory neurons. Recent studies employing single-cell RNA-sequencing of DRG neurons have begun to unveil a more comprehensive classification of itch-sensory neurons, proposing the existence of a number of different subsets (15, 16). However, further work is needed to assess the in vivo functionality of these subsets and what modalities are distinct, shared, or synergistic.
Cytokines and itch
The immune response is organized into specialized effector modules that are tailored to combat different types of pathogens. Type 1 immunity is broadly used to combat infections involving intracellular bacteria and viruses and is characterized by the production of the effector cytokines IFN-γ and/or TNF-α. The production of IL-17A and/or IL-22 is a hallmark of type 3 immunity, which is specialized for extracellular bacterial and antifungal defense. Finally, parasitic infections, along with noxious environmental substances, result in the generation of a type 2 immune response driven by the production of IL-4, IL-5, and IL-13 (17). In addition to coordinating these specialized immune responses, several cytokines also modulate sensory perception and behavior, another key aspect of host defense (18, 19). The early discovery that the canonical proinflammatory cytokine IL-1β can induce pain led to a significant paradigm shift in our understanding of how the immune system participates in sensation and behavior (20). Subsequently, over the past decade, a number of cytokines spanning these specialized immune responses have been discovered to elicit either pain or itch by directly binding to their receptors expressed on sensory neurons. This provokes the hypothesis that the specific sensory response that is evoked depends on the type of pathologic insult and the resulting immune response that is generated.
Building on the initial discovery that IL-1β can induce pain in vivo, a number of studies have identified additional cytokines that can also elicit pain by modulating sensory neuronal signaling (1). For example, in a rodent model of bone cancer–induced pain, IL-6 was found to critically regulate neuronal hyperexcitability as well as increased sensitivity to pain (hyperalgesia) (21). TNF-α was also found to be important in mediating hyperalgesia, specifically in the setting of nerve injury (22). Finally, in another study, injection of IL-17A was found to induce pain in the joints of rats in a TNF-α– and IL-6–independent manner, and treatment with an anti–IL-17 Ab reduced pain symptoms in a model of arthritis independently of effects on joint swelling (23). Taken together, a number of cytokines associated with type 1 and type 3 immune responses, including IL-1β, IL-6, TNF-α, and IL-17A, have been found to modulate neuronal signaling in various models of pain behavior (Fig. 1) (1). However, it is important to note that much of the early literature on neuroimmune regulation of pain is derived from the study of other tissues than the skin, such as the bone, joint, and nerve.
In contrast to pain, which typically involves deeper structures, itch is predominantly confined to the skin. Similar to pain, a number of different cytokines have been found to mediate itch; however, in contrast, the itch-associated cytokines thus far identified are all associated with a type 2 immune response. Poised to respond to environmental insults, keratinocytes are key initiators of a host-protective immune response through the production of alarmins or epithelial cell–derived cytokines, such as IL-33 and thymic stromal lymphopoietin (TSLP). These cytokines potently activate a variety of both innate and adaptive immune cell populations, which results in a robust type 2 immune response characterized by the production of IL-4 and IL-13. However, in addition to driving type 2 immunity, it was recently shown that IL-33 and TSLP can directly activate sensory neurons to evoke itch (24, 25). However, whether these cytokines are the key mediators of itch in type 2 inflammatory skin disorders such as atopic dermatitis (AD) remains to be clearly defined. Thus, epithelial cell–derived cytokines, in response to epidermal stress or disruption, have the capacity to simultaneously and rapidly activate both innate immune responses and scratching behavior.
Downstream of IL-33 and TSLP, a number of cell populations are elicited to produce IL-4 and/or IL-13, including basophils, eosinophils, group 2 innate lymphoid cells (ILC2s), mast cells, and Th2 cells (26–29). These effector cytokines, in addition to their well-known role in promoting barrier inflammation, were recently shown to modulate itch responses in mice (30). IL-31, which is predominantly produced by Th2 cells, also is an important mediator of itch in vivo (31, 32). However, in contrast to IL-4 and IL-13, IL-31 may not play a prominent role in driving cutaneous inflammation. IL-31–deficient mice, in addition to having a reduction in scratching behavior compared to controls, appear to have similar levels of skin inflammation in a model of contact hypersensitivity (33). In support of this concept, anti–IL-31RA mAb treatment (nemolizumab) appeared to preferentially target symptoms of itch rather than inflammation in AD patients in a recent phase 2 clinical trial (34). In contrast, inhibition of the shared receptor subunit for IL-4 and IL-13 (anti–IL-4Rα mAb [dupilumab]) resulted in a dramatic reduction in both overall disease severity (i.e., cutaneous inflammation) as well as itch in phase 3 clinical trials for AD (35). In light of the complex network of cytokines involved in promoting type 2 skin inflammation and itch, futures studies will be required to determine how these cytokines come together to specifically modulate itch in the setting of different inflammatory skin disorders.
Broadly, cytokines that underlie type 1 and/or type 3 immune responses, such as IL-1β, IL-6, TNF-α, and IL-17A (1, 21–23), have been associated with pain, whereas those associated with a type 2 immune response, such as IL-4, IL-13, IL-31, IL-33, and TSLP, involve itch. Additionally, many diseases associated with type 2 inflammatory features are highly pruritic, such as AD, acute and chronic urticaria, and prurigo nodularis (36). Although some skin conditions associated with type 1 and/or type 3 immune responses are pruritic, such as allergic contact dermatitis, psoriasis, and superficial fungal infections, whether effector cytokines specifically associated with these types of immune responses can act as pruritogens remains poorly defined and is an exciting area of inquiry. However, based on the current body of work in sensory neuroimmunology, we speculate that specialized immune responses specifically evoke the protective behavioral response of either pain or itch, depending on the environmental stimulus. Pain responses appear to be more commonly associated with bacteria when aversion to movement may be needed to minimize the spread of infection (e.g., sepsis) and promote healing, whereas the scratching response to itch sensation may aid in the expulsion of larger ectoparasites and noxious environmental substances.
Acute versus chronic itch
Although acute itch is likely a protective behavioral response, chronic itch is a highly debilitating medical disorder (37). A current focus of the itch field is identifying pruritogens, molecules that directly activate sensory neurons to induce itch. A standard technique used to identify such molecules is the injection of a putative pruritogen intradermally into the skin. Potential pruritogens are often injected into the cheek skin to allow researchers to distinguish itch from pain behavior, which are defined as hind limb scratching and front paw wiping, respectively (38, 39). Although this technique has been extremely valuable in identifying key pruritogens, it is important to note that it is an acute itch model. Scratching bouts are evoked within several minutes of introducing the stimulant into the skin and typically last for under an hour. Thus, although it is a very powerful and efficient technique, it may have potential limitations in defining important mediators of chronic itch: when spontaneous scratching commences independently of acute stimuli. These include both genetic and chemically induced models of chronic itch conditions, such as allergic contact dermatitis, dry skin, and AD, in which itch can last from days to weeks (30, 31, 40–42).
A classic example in which a potent mediator of acute itch may not play a key role in driving chronic itch is histamine. Although histamine is a canonical pruritogen that was used to validate the intradermal cheek model (38), antihistamines are generally poorly efficacious in many chronic itch disorders, such as AD (43). Conversely, IL-4 and IL-13 are poor acute pruritogens, and yet, they are critical drivers of chronic itch in the setting of AD-like disease in mice through their direct activity on sensory neurons. This appears to be due to the ability of these cytokines to sensitize neurons to other pruritogens like histamine, IL-31, and TSLP (30). Cytokines thus may have additional direct roles in modulating itch beyond the immediate induction of itch signaling. Collectively, these studies demonstrate that investigating models of both acute and chronic itch are important for providing novel insight into the biology of itch mediators and their clinical implications.
Cytokine signaling in neurons: JAKs
The intracellular signaling pathways downstream of cytokine receptor binding on sensory neurons and how they mediate specific sensations is an exciting area of research that is just starting to be elucidated. One signaling pathway that was found to alter neuronal excitability downstream of IL-1β and TNF-α is phosphorylation of NaV1.8 via p38 MAPK (44–46). However, in contrast to the IL-1 and TNF families, it is well known that, in immune cells, many cytokine receptors use JAKs to activate STAT transcription factors. Similarly, it appears that neurons use JAKs. For example, JAK phosphorylation has been found to be elevated in neurons following spinal cord injury in mice (47). However, the functional significance of JAK signaling in sensory neurons was not entirely clear until more recent studies in itch biology.
Oetjen et al. (30) demonstrated that neuronal JAK1 is a critical mediator of chronic AD–associated itch in vivo. However, the signaling pathways that are activated in neurons by JAKs remain poorly defined. In contrast to classical JAK/STAT signaling in lymphocytes, it is likely that cytokine-mediated stimulation of sensory neurons results in STAT-independent effects. For example, cytokine stimulation of isolated DRG neurons by IL-4 and IL-13 results in rapid neuronal activation, as indicated by a calcium influx within seconds to minutes following cytokine application (30). Given the dependence of this calcium response on TRP channels, we speculate that JAKs may either directly or indirectly influence calcium influx into the cell by modulating these channels. How this occurs is an exciting area of inquiry. JAKs additionally appear to have long-term effects on the excitability of neurons such as by altering the expression of TRPA1 and membrane trafficking of TRPV1 (21, 48). These JAK-dependent alterations in trafficking and transcription may be additional ways by which pain and itch sensitization occurs in the periphery and how symptoms such as allodynia and alloknesis can develop. Thus, we hypothesize that, in addition to rapidly altering ion channel function through STAT-independent phosphorylation events, there may also be STAT-dependent transcriptional events that underlie the chronicity of pain and itch. We anticipate that investigating JAK signaling in neurons may unveil novel biochemistry with regard to JAK signaling and have broad implications for the treatment of both chronic pain and itch disorders.
The significance of neuronal JAK1 signaling in itch is corroborated by clinical trials in which topical JAK inhibitors have demonstrated rapid anti-itch effects within 24–48 h of initial treatment in AD patients prior to any detectable effect on clinical skin inflammation (49, 50). Additionally, in a small proof-of-concept study, we found that even in a chronic itch condition in which there is an absence of noticeable skin inflammation [chronic idiopathic pruritus or generalized pruritus of unknown origin (51)], patients experienced a rapid reduction in itch symptoms in response to a JAK inhibitor (tofacitinib) (30). The importance of neuronal JAK1 in mediating itch is further reinforced by the finding that patients with germline JAK1 gain-of-function mutations develop chronic pruritus that is selectively responsive to treatment with a JAK inhibitor (ruxolitinib) over broader anti-inflammatory agents (i.e., systemic steroids) (52). Collectively, these studies show that investigating cytokine-neuronal interactions can lead to novel therapeutic insights that can be exploited for treatment of itch, pain, and possibly other sensory disorders.
Sensory neurons drive inflammatory skin disorders
In contrast to other barrier surfaces, somatosensory neurons are the primary source of neuropeptides in the skin. Historically associated with signaling in neural tissues, it is now recognized that these small peptides can also act on other cell types, such as keratinocytes and immune cells, provoking the hypothesis that sensory neurons directly and critically regulate skin inflammation (53). This hypothesis is supported by a number of observational studies that document that patients with inflammatory skin disorders such as AD and psoriasis experienced disease resolution in body parts that had a loss of innervation (54–57). Similarly, in murine models of psoriasis, surgical denervation, and chemical ablation of NaV1.8+TRPV1+ sensory nerve fibers also resulted in the improvement of disease severity (58–60). This improvement may be due to, at least in part, the loss of IL-23 production from dermal dendritic cells (60). Identifying which neuropeptides critically regulate these inflammatory processes is currently a highly active field of neuroimmunology.
Thus far, the neuropeptides substance P (SP) and calcitonin gene–related peptide (CGRP) have been strongly implicated in skin inflammation. Psoriasis-like inflammation in denervated skin can be restored by intradermal delivery of SP and CGRP (58). Specifically, SP appears to critically regulate immune cell recruitment, whereas CGRP promotes epidermal hyperplasia (acanthosis) in this setting (58). Similarly, in the context of AD, sensory neurons have been shown to regulate human keratinocyte proliferation in a CGRP-dependent manner (61). Although CGRP and SP have been the primary neuropeptides linked to skin inflammation, additional neuropeptides have been found to critically modulate inflammation at other barrier surfaces, although their role in the skin has yet to be clearly demonstrated. For example, recent studies have shown that ILC2s are regulated by neuromedin U (NMU) released from cholinergic neurons in the gut (62, 63) and somatosensory afferents in the lung (64). Vasoactive intestinal peptide (VIP) is another neuropeptide found to activate lung ILC2s upon its release from cholinergic neurons of the vagal nodose ganglia, which, along with somatosensory neurons, also innervate the lung (65). Given that both NMU and VIP are expressed by somatosensory neurons, whether these neuropeptides also play an important role in regulating cutaneous inflammation is an exciting area of future inquiry (64, 66, 67). One study suggests that VIP can alter the capacity of Langerhans cells to present Ags and thus inhibit the generation of Th1 cell responses (68). Finally, although the primary focus of research in the skin thus far has concentrated on the role of neuropeptides, there is emerging interest in the ability of small molecule neurotransmitters in mediating cutaneous inflammation.
Recent studies have found that sensory neurons, like immune cells, are able to directly sense and respond to microbes. One example is LPS, a major component of Gram-negative bacteria and a key endotoxin that binds TLR4. Along with being immunostimulatory, LPS also directly activates sensory neurons to modulate both pain and itch signaling. Although LPS injection is known to induce pain but not itch (69), one study found that TLR4 signaling promotes histamine-mediated itch by potentiating TRPV1 activity (70). Although studies have shown that sensory neurons detect LPS directly through TLR4 (71–73), others have found that LPS can directly stimulate sensory neurons in an TLR4-independent manner through mechanical perturbation of the neuronal membrane, resulting in the activation of TRPA1 (74, 75). Collectively, these studies demonstrate how bacterial endotoxins can directly manipulate the peripheral nervous system to modulate sensation.
Pore-forming toxins (PFTs) are another class of virulence factors that, in contrast to classical endotoxins, are produced by Gram-positive bacteria, which commonly cause bacterial skin infections such as cellulitis and necrotizing fasciitis. Notably, these infections are strongly associated with disproportionate levels of pain. Recent studies have shown that these PFTs can penetrate and activate sensory neurons directly to evoke pain (76, 77) (Fig. 2). Specifically, α-hemolysin, phenol-soluble modulins, and the leukocidin γ-hemolysin AB from Staphylococcus aureus (76) and streptolysin S from Streptococcus pyogenes (77) were shown to critically mediate bacterial infection–associated pain in the skin through direct neuronal activation. Thus, it appears that, in addition to stimulating the host inflammatory response, the bacterium can directly influence pain sensation and behavior through a variety of toxins, likely independently of the generation of pain-associated cytokines such as IL-1β, IL-6, IL-17A, and TNF-α. Future studies may reveal many more mechanisms by which bacteria and other microorganisms are sensed by sensory neurons and if these signals can modulate itch in addition to pain.
Sensory neurons mediate immunity
Similar to their contribution to inflammatory skin diseases, neurons also play an important role in modulating protective immunity. Recent studies have demonstrated that intact peripheral sensory innervation is critical for optimal production of IL-23 to drive antifungal immunity to Candida albicans. Administration of CGRP was sufficient to overcome the effects of denervation in this context (78). A study by Maruyama et al. (79) suggests neurons detect C. albicans through the binding of neuronal dectin-1 to β-glucan, and this may stimulate the release of CGRP (Fig. 2). Whether neurons can sense other fungi and modulate immunity in a similar fashion remains an intriguing question. Similar to C. albicans, CGRP is also released by neurons upon S. pyogenes infection (Fig. 2). However, whereas CGRP appears to be protective in the setting of C. albicans, it plays a detrimental role in S. pyogenes infection by suppressing neutrophil recruitment and antibacterial immunity (77). Ultimately, these studies bring forth a paradigm in which the sensory nervous system is both capable of sensing microbes directly and, in turn, shaping host immunity.
The skin harbors a vast neuroimmune network that is poised to provide a rapid, coordinated response to a variety of environmental insults. Recent advances covered in this review suggest that specialized protective immune modules may also encode highly specific sensory and behavioral responses. In the setting of pain, the withdrawal reflex (acute pain) and/or aversion to movement (prolonged pain) may help to promote wound healing, prevent the spread of infection, and conserve host metabolic resources. In contrast to pain, the scratching reflex may help to promote the expulsion of macroparasites, toxins, and environmental irritants (80). Thus, scratching the skin in response to itch may parallel the “weep and sweep” responses promoted by type 2 inflammation in the intestine and airway. Finally, not only does the cytokine milieu dictate the sensation and behavioral response generated, sensory neurons, in turn, shape the immune response through the release of various neuropeptides. Collectively, advances in neuroimmunology demonstrate that the nervous system is an integral part of the overall immune response in both health and disease.
Three broad areas of investigation remain open to major advancement. First, the specific cellular sources of neuromodulatory cytokines during the course of an immune response and how these cells home toward and interact with specific sensory neurons remains poorly defined. Second, which chronic inflammatory and sensory disorders can be modulated by targeting specific cytokines and/or neuropeptides in human disease is an exciting area of therapeutic inquiry. Third, what specific mediators and signaling pathways drive the release of neuropeptides from neurons in the skin as well as mucosal surfaces remains largely unknown. Ultimately, additional insights into these processes will likely open new avenues of therapeutic intervention for sensory, inflammatory, and infectious disorders of the skin and beyond.
We thank members of the Kim Laboratory for reviewing the manuscript and Dr. Isaac Chiu for insightful discussions.
This work was supported by grants from the National Institute of Arthritis and Musculoskeletal and Skin Diseases at the National Institutes of Health (NIH; K08AR065577 and R01AR070116), the American Skin Association, a Doris Duke Charitable Foundation Clinical Scientist Development Award, and LEO Pharma. A.M.T. is supported by a grant from the National Institute of Allergy and Infectious Diseases at the NIH (T32AI716340).
Abbreviations used in this article:
calcitonin gene–related peptide
dorsal root ganglion
group 2 innate lymphoid cell
transient receptor potential A1
thymic stromal lymphopoietin
vasoactive intestinal peptide.
B.S.K. has served as a consultant for AbbVie, Inc., Menlo Therapeutics, and Pfizer, Inc., and on advisory boards for Cara Therapeutics, Incyte Corporation, Kiniksa Pharmaceuticals, Menlo Therapeutics, and Regeneron Pharmaceuticals, Inc. B.S.K. is a stockholder of Locus Biosciences and is a founder and chief scientific officer of Nuogen Pharma, Inc. The other authors have no financial conflicts of interest.