Physiology and Pathophysiology of Itch

Abstract

Itch is a topic to which everyone can relate. The physiological roles of itch are increasingly understood and appreciated. The pathophysiological consequences of itch impact quality of life as much as pain. These dynamics have led to increasingly deep dives into the mechanisms that underlie and contribute to the sensation of itch. When the prior review on the physiology of itching was published in this journal in 1941, itch was a black box of interest to a small number of neuroscientists and dermatologists. Itch is now appreciated as a complex and colorful Rubik’s cube. Acute and chronic itch are being carefully scratched apart and reassembled by puzzle solvers across the biomedical spectrum. New mediators are being identified. Mechanisms blur boundaries of the circuitry that blend neuroscience and immunology. Measures involve psychophysics and behavioral psychology. The efforts associated with these approaches are positively impacting the care of itchy patients. There is now the potential to markedly alleviate chronic itch, a condition that does not end life, but often ruins it. We review the itch field and provide a current understanding of the pathophysiology of itch. Itch is a disease, not only a symptom of disease.

• Scratching an itch can be intensely pleasurable, yet chronic itch is a disease that impacts quality of life to the same extent as pain.

• Signals between the environment, skin, immune cells, and afferent nerve fibers funnel into the spinal circuitry and brain to relay the sensation of itch, which promotes scratching.

• The mechanisms that underlie itch and opportunities for targeting this unmet need are reviewed.

I. INTRODUCTION

‟Itching is an unpleasant sensation which provokes the desire to scratch.ˮ So began the previous review on the physiology of itching in this journal in 1941, written by the father of investigative dermatology, Stephen Rothman. This definition was penned by Samuel Hafenreffer in 1660, with pruritus, an identical term, in place of itch. The definition continues to appear regularly at the beginning of manuscripts. Why such reinforcement occurs, unique to this topic, is indicative of the fascination that surrounds itch. Scratching an itch can also be pleasurable (269). The poet Ogden Nash wrote that happiness is to have a scratch for every itch. The sensation may be unpleasant, the scratching response rewarding, but patients who suffer from acute or chronic itch, which affects 15% of the population, are miserable.

Rothman’s review was prescient. The complete overlap between areas in which itch and pain were felt was indicative of shared neuronal pathways but could not explain the distinct motor responses of scratching an itch versus withdrawal from a painful stimulus. The triple response of Sir Thomas Lewis, manifest by vasodilatation, flare, edema and associated itch attributed to a ‟H-substanceˮ not necessarily exclusive to histamine, was consistent with mediators aside from histamine. Itch associated with systemic processes including cholestasis and hematological conditions was noted to be distinct from histamine, consistent with current understanding that clinical itches are driven primarily by mediators and pathways that are independent from histamine. The concepts of central inhibitory pathways and central sensitization were introduced.

Why do we have the sensation of itch? Conventional thinking has been that the major role of the sensation of itch and the subsequent motor response of scratching is to remove environmental insults from the skin, especially arthropods. Hence, itch is considered a protective sensation and response to the environment. More recent thinking is that in addition, scratching damages the epidermal barrier and facilitates an appropriate physiological cascade associated with sensing the insult (FIGURE 1). This cascade includes multidirectional communication between the nervous and immune systems within the skin, perhaps primarily directed against components of the microbiota, which can evoke itch and pain (FIGURE 1). The pleasure that accompanies scratching may serve to reinforce the motor activity to further ensure neuroimmune communication.

FIGURE 1.

Initiation of itch. Allergens, pruritogens, and irritants are exogenous substances which interface with the skin in acute and chronic itch. Branching terminal fibers of afferent neurons which sense these substances reach the epidermis, the uppermost viable layer of skin immediately below the stratum corneum barrier. The sensory neurons are considered histaminergic or non-histaminergic. Neural activity drives the recruitment of immune cells, including mast cells and CD4+ T cells among others. These cells release mediators that activate cognate receptors on sensory neurons to release neuropeptides to contribute to the itch-scratch cycle. Messages are relayed from the peripheral afferents to their cell bodies in dorsal root or trigeminal ganglia followed by synapsing with second-order neurons in the spinal cord. The thalamus then assists in the interpretation of messages encoding itch. IL, interleukin; STT, spinothalamic tract.

Depending on the insult, such as a contact allergen, or in association with a disease, for example atopic dermatitis or cholestasis, a pathophysiological itch may develop.

Our goal is to provide an understanding of the overall physiology and associated pathophysiology of itch. We focus on neuronal aspects, as a nervous system is needed to have the sensation of itch, as well as the multidirectional connections between the nervous and immune systems that impact itch transmission from the skin into the spinal cord. These areas are the ones most supported by objective data indicative of mechanisms, and it is appreciated that an understanding of central itch circuitry in the mouse is advancing rapidly.

Itch is rarely a pure sensation. Instead, itch typically includes degrees of burning, prickling, and stinging. This delineation is of importance because experimental stimuli as well as clinical conditions manifest such differences, which can overlap with pain. Atopic dermatitis (AD), for which itch is a sine qua non, often has a degree of pain, and Staphylococcus aureus, which is resident in AD skin, can evoke pain and itch. Our peripheral neurosensory-immune system together with central circuitry is able to differentiate between itch and pain. This observation leads to the long-standing question of whether or not there is specific neuronal coding for itch. An extension of this topic is whether the sensation of itch is restricted to skin or if other organs get itchy. We are accustomed to associating pain with visceral organs, but there is a developing consensus that processes associated with removal of material from any epithelia, be it the bladder, gut, or respiratory system, including cough, may be equivalent to itch (205).

A plethora of discoveries is shaping our understanding and helping us to understand itch. A critical caveat of ongoing research is the need to use experimental species and the challenge to translate findings from mice, canines, and non-human primates into the human condition. Scratching behavior in animals is considered to relate to human itch, although there is no objective method of asking if a non-human itches, nor is there an objective measure of human itch. If itch depends on the ability to scratch, then from an evolutionary standpoint, it can be argued that itch developed in conjunction with receptors that first appear in tetrapods (29). However, if scratching is considered a behavioral response manifest by rubbing following injection of a channel activator, then fish may have the sensation of itch (80).

In the clinic, itch is considered either acute or chronic. Acute itch is defined as an itch that lasts less than 6 wk, whereas chronic itch persists for 6 wk or longer and is experienced by 15% of the population. A variety of animal models are considered proxies for chronic itch by some investigators, but given the brevity of these models, we have difficulty making that connection. Information is lost in translation between models and human physiology, modulating the perspectives between the bench and bedside, while promising findings in mouse models (104) have often failed in the clinic (28). At the same time, appreciation of the impact of itch together with the fascinating physiology underlying this sensation is driving therapeutic success and improved quality of life for those who itch.

II. GLOSSARY

A. Types of Itches

• Acute itch: itch lasting less than 6 wk.

• Chronic itch: itch lasting 6 wk or longer (under causes, note that it can be a chronic acute mechanism).

• Neurogenic itch: induced by mediators but in the absence of neural damage. As the mediators of neurogenic itch are defined, it is likely that neurogenic itch will fall under pruritoceptive itch.

• Neuropathic itch: associated with damaged neurons, e.g., post-herpetic neuralgic itch or small fiber neuropathy.

• Pruritoceptive itch: itch associated with pruritogen activation of sensory fibers.

• Psychogenic itch: having a psychosomatic or psychiatric origin, e.g., delusions of parasitosis.

B. Selected Definitions and Terminology

• Allodynia: pain or itch from what is normally a nonpainful or itchy stimulus. It is associated with central sensitization.

• Alloknesis: triggering of itch from what is normally a nonpruritogenic stimulus includes itch.

• Atmoknesis: itch that occurs when skin is exposed to air, such as when clothing is removed.

• Central sensitization: increased responsiveness of nociceptive neurons in the central nervous system to their normal or subthreshold afferent input. It is often associated with peripheral injury or inflammation, such that persistent stimulation of nociceptors or pruriceptors leads to increased excitability of central pathways, decreased activity of inhibitory pathways, and chronic pain or itch.

• Dysesthesia: an abnormal sensation, often including some components of burning, itch, pain, pins and needles, and touch. It can occur anywhere on the skin but is often associated with the scalp.

• Neurogenic inflammation: inflammation associated with the release of mediators, particularly substance P (SP) or calcitonin gene-related peptide (CGRP), from peripheral afferent neurons, which then impact the immune system.

• Sensitive skin: a widely used term that has defied an objective definition or pathophysiology. It can occur in the setting of specific diagnoses, including atopic dermatitis, less well-defined conditions such as fibromyalgia, or it can refer to itching, burning, pricking sensations, or dysesthesias of unknown cause.

III. SENSORY NEURONS

Itch is sensed by cutaneous nerve fibers called pruriceptors. These primary afferent fibers serve as antennae and constantly sample the environment of the skin to detect and respond to cues. Signals are propagated along neural pathways to the spinal cord and brain for interpretation and response. Analogous to itch, algogenic (painful) stimuli, heat, and cold are sensed by nociceptors. It is possible that all pruriceptors function as nociceptors in humans but whether the reverse is true remains to be clarified. There are data in mice that support the concept of a small number of uniquely pruriceptive fibers. Nociceptors are categorized into two biophysically distinct major afferent classes, thinly myelinated Aδ and unmyelinated C-fibers. Aδ are 2–5 µm in diameter and have a conduction velocity of up to 8 m/s. C-fibers average 0.2–1.5 µm in diameter and have a conduction velocity less than 2 m/s. Sensory nerves do not function in isolation but are part of an interactive milieu. This milieu includes a plethora of mediators generated from the sensory neurons themselves, neighboring cells, the barrier, and the microbiome. These mediators range in size from ions to proteins, include pH, and may include localized cell or tissue tension. Each neuronal population is heterogeneous and polymodal, meaning that it has more than one detection property, often associated with specific ion channels that detect, for example, heat, cold, acid, and whether they respond to cellular stretch or a light touch from a cellular probe, being mechanically sensitive, or insensitive, and the yet ill-defined milieu that is generated by tissue inflammation. These neurons are pseudounipolar, with cell bodies residing in dorsal root (or trigeminal) ganglia. The peripheral branch of the axon extends to the skin while the central branch synapses with first-order neurons in the dorsal horn of the spinal cord. These proceed to interact with second-order neurons that comprise excitatory and inhibitory circuits that project to the parabrachial nucleus of the thalamus for processing, interpretation in additional brain regions, and a resultant motor response of scratching. C-fibers, and to a lesser extent Aδ-fibers, are responsible for itch sensing. It is not yet known if the recently described specialized glial cells that form a meshlike network near the dermal-epidermal junction and initiate the sensation of pain contribute to the sensation of itch (1).

Itch fibers comprise fewer than 10% of the C-fibers in skin. Diversity in the C-fiber population has long been recognized, is manifest in several ways, is dynamic, and evolves postnatally in mice and possibly in humans. Rather than being strictly absolute, C-fibers can be distinguished by markers and by function, often illustrated via Venn diagrams. Single-cell RNA sequencing has led to even further subdivisions, but whether these are static is not known. ‟Silentˮ nociceptors constituting almost a quarter of C-fibers in human skin have been described (246). These mechanically insensitive but heat-responsive C-afferents or silent nociceptors biophysically transition to mechanically responding fibers in the setting of injury. A conventional distinction between C-fibers, although not absolute, is by peptidergic versus nonpeptidergic markers. Peptidergic markers include SP and CGRP which are released by these fibers. They also express the nerve growth factor (NGF) receptor and tropomyosin (or tyrosine) receptor kinase A (TrkA) and somatostatin (SOM). Nonpeptidergic markers include the c-Ret neurotrophin receptor as well as artemin. In general, the c-Ret+ neurons express the isolectin IB4, a lectin that binds to nonpeptidergic neurons.

With an unbiased RNA profiling analysis that combined neuronal information with functional impact, four nonpeptidergic (NP 1–4) itch-specific classes were identified. NP1 is considered to mediate neuropathic pain and itch. NP2 and NP3 are also involved in itch sensation with NP3 responsible for the transduction of inflammatory itch (293). These data suggest that there is a neuronal segregation between chemical and acutely induced and inflammatory-mediated itch (FIGURE 2).

FIGURE 2.

Itch circuitry from the periphery to the brain. Itch-related G protein-coupled receptors (GPCRs) and cytokine receptors on sensory neurons from the skin to the spinal cord are categorized as NP1, NP2, and NP3 neurons. Some receptors are expressed in all NP1–3 neurons, whereas MrgprA3, an itch specific marker, is present only in the NP2 population. The natriuretic peptide Nppb acts on its receptor NPR1 to link transmission between dorsal root ganglion (DRG) and spinal neurons. Complex inhibitory and excitatory interneuron activity modulates the signal that will be projected to the brain. Gastrin releasing peptide (GRP) and GRP receptor (GRPR) are each key to excitatory itch circuits in the spinal cord. Inhibitory basic helix loop helix 5 interneurons (B5-I) and somatostatin (SOM) interneurons have inhibitory function. Loss of inhibition either by neuronal depletion of B5-I interneurons or the inhibitory neurotransmitters GABA, glycine, or dynorphin results in intensified itch. This pattern most likely occurs in chronic itch conditions. Enkephalinergic interneurons are important in the gate control between intense pain and itch sensation. Mechanical itch is transmitted through a different, so far unclassified, neuron population and second-order neurons in the spinal cord. The circuitry does not involve excitatory GRP but includes the inhibitory neuropeptide NPY (neuropeptide Y). Contagious itch in mice is relayed through GRPR+ neurons in the suprachiasmatic nucleus. [Modified from Dong and Dong (71), with permission from Elsevier.]

Another, and particularly useful discriminator between itch sensory neuron function, is between fibers that are either mechanosensitive or mechanoinsensitive (151). The two prototypic substances that evoke itch in human subjects are histamine and spicules of the bean plant Mucuna pruriens. The plant and its spicules are colloquially referred to as cowhage and informally known as itching powder. The sensations elicited by cowhage were known even to Rothman to be independent of histamine. Electrophysiological recordings from human and mouse primary afferents revealed that histamine acts on a subset of mechanoinsensitive C-fibers that also respond to heat and capsaicin. In contrast, cowhage activates mechanosensitive fibers in humans. Possible differences of signal intensity carried via mechano-sensitive or -insensitive neurons may be accounted for by the diameter of the respective cutaneous receptive fields. A cutaneous receptive field is defined as the area within the skin activated by a stimulus and concurrently evoking action potential (132). The activation of the receptive fields by histamine or cowhage spicules seem to differ from each other, which may be explained by distinct and overlapping C-fibers that convey different stimuli, the pruritogens themselves, or the mode of application, spicules for cowhage, and injection for histamine.

A. Transmission of Messages

All primary afferents project into the dorsal horn of the spinal cord. Peptidergic and nonpeptidergic neurons target distinct areas of the superficial laminae in the dorsal horn. The peptidergic C-afferents project into the outer region of lamina I and lamina II, whereas the nonpeptidergic subset targets the inner region of lamina II. This segregation within the spinal cord maintains functional distinction of the C-afferents. Aδ nociceptors target lamina I and deeper lamina V of the spinal cord. Low-threshold C-mechanoreceptors venture to the inner ventral part of lamina II where protein kinase C (PKC)-γ-expressing interneurons are concentrated. Deeper dorsal horn contains the wide-range dynamic projection neurons that respond to innocuous and nociceptive stimuli. Projection neurons that carry the messages beyond the spinal cord comprise 2–5% of the neuronal population, which mostly projects to the parabrachial (PB) nucleus of the dorsolateral brain stem. Within this small population, 80% express the neurokinin 1 receptor (NK1R) and likely connect with SP+/GABA– interneurons (46, 180), a neuronal population considered a fine-tuning gate of messages. This neuronal circuit is (68, 283) implicated in coping behavior associated with sustained pain and itch (120).

The above-mentioned interneurons are the gate-keepers of neuronal activity that modulate afferent input carried to the projection neurons. Morphologically, interneurons show heterogeneous features from islet, central, radial, and vertical cells (41, 99). Readers are referred to excellent articles that discuss interneuron morphology and function in detail (317, 318). Biochemical characterization somewhat simplified the interneurons into the excitatory interneurons which express vesicular glutamate transporter VGLUT2 (20, 243) and neurochemical markers including PKC-γ, neurotensin, and somatostatin, which also serve to define certain subpopulations of excitatory interneurons (228, 229, 282). The testicular orphan nuclear receptor 4 (TR4) is part of the excitatory interneuron circuitry involved in the transmission of itch and pain (298). Selective deletion of TR4 in the central nervous system resulted in loss of excitatory interneurons in the superficial dorsal horn. Phenotypically, TR4 mutant mice had increased mechanical responses but reduced responses to heat and lack of responses to various pruritogens. These findings favor the theory of converging itch and pain circuitry in the spinal cord. Gastrin releasing peptide (GRP), considered an itch-specific neurotransmitter and its corresponding receptor GRPR, are in the excitatory interneuron population (271, 272, 298) and may label an itch-specific circuit, at least in mice. Ablation of GRPR+ interneurons affected itch behavior but not pain responses, suggesting a distinction between pain and itch circuitry. Peptidergic C-fiber input to the excitatory interneurons appears to engage the neurotransmitter natriuretic polypetide b (Nppb) that is present in the TRPV1+ afferent neuronal subset (186). Initially, GRP was identified as an itch-specific neuropeptide released upon activation of peptidergic C-afferents by pruritogens which led to activation of its cognate receptor GRPR in the spinal cord (271, 272). The relative roles of Nppb and GRP in the periphery have led to a dynamic discussion in field of itch research. Currently, the integrated concept is that Nppb is released from peripheral afferents and induces GRP to be transmitted from spinal cord interneurons. In accordance with its itch-specific role, Nppb is expressed in the TRPV1+ subset that also coexpressed the members of the itch-associated mas-related G protein-coupled receptor family members MrgprA3 and MrgprC11 (170, 171). Hence, the neurons that express the receptor for Nppb, namely, natriuretic peptide receptor (Npr), are the first level of spinal cord neurons responsible for the neural itch circuitry. In summary, the excitatory itch circuit includes the peripheral release of Nppb, spinal GRP and GRPR activation in spinal interneurons that are TR4+. Complementary to the TR4+ interneurons, inhibitory interneurons express the transcription factor Bhlhb5. Loss of Bhlhb5 in basic helix loop helix 5 interneurons (B5-I) inhibitory interneurons results in excessive scratching without alteration of pain responses (229). The majority of inhibitory Bhlhb5 interneurons produce the neurotransmitter γ-aminobutyric acid (GABA), which is diminished in chronic conditions of itch and pain (FIGURE 2).

IV. LABELED LINE FOR ITCH IN MICE

In mice, acute itch is transmitted via pruritoceptive neurons that respond to chemical stimuli. These neurons express TRPV1+ and MrgprA3+ and fall predominately into the NP2 and to a small extent the NP3 populations (FIGURE 2). Acute and chronic itch are both relayed through this neuronal populations (71) consistent with a specific “itch” labeled line in this species. Whether the labeled line transmission continues in pathological conditions has not been addressed. In general, itch-specific receptors discussed above are all present in the NP2/NP3 population, including those for interleukin (IL)-31, IL-4, and IL-13 (FIGURE 2). Capsaicin activates all TRPV1+ neurons, including the itch-specific MrgprA3+/TRPV1+ neurons. The NP1 population, which is positive for MrgprD, is polymodal, responding to itch, pain, mechanical, and thermal stimuli. That this population does not show overlap with NP2 or NP3 suggests the existence of an additional line that imparts itch. In the central nervous system, as discussed above, complex circuits encode itch. Briefly, this circuitry includes Vglut2+ neurons. Itch signaling is associated with the release of Nppb, at least in the NP3 population. Since Nppb was not detected in NP2 itch neurons, NP2 and NP3 populations engage separate or overlapping circuitry for itch. In NP3 neurons, Nppb engages with its receptor NPRA followed by communication to GRPR via GRP release in the spinal cord (FIGURE 2). Mechanical itch is most likely transmitted through low-threshold mechanoreceptors in which skin structures such as Merkel cells might be involved (FIGURE 2). The mechanical itch-sensitive neurons have been shown to connect to spinal NPY+ inhibitory interneurons that receive low-threshold mechanical input. The circuitry of contagious itch requires the presence of GRPR+ neurons as discussed later (see sect. XVII).

V. NEUROANATOMY

Primary sensory neurons absorb information from the periphery to inform the spinal cord where somatosensory processing continues. In general, the circuitry is spatially organized, with nociceptive and thermosensitive afferents entering the superficial spinal cord whereas the cutaneous and proprioceptive afferents target the ventral regions of the spinal cord. Here, attention is directed towards embryological development and organization in the spinal cord as related primarily to pruritoceptive primary afferents and spinal cord segments responsible for transmission and modulation of itch.

A. Development of the Spinal Anatomy

The spinal cord is generated from the vertebrate neural plate that also gives rise to neural crest cells from which most neurons and glia originate. The neural plate thickens and invaginates, resulting in the rostro-caudal neural tube formation. Expansion of the neural tube in the head region leads to formation of the brain, whereas more caudally the same neural tube expansion forms the spinal cord. The neural tube closes along the entire body axis and neural crest cells transition from epithelial to mesenchymal cells migrating into distal parts of the embryo (159). The peripheral nervous system (PNS) is formed by cranial neural crest cells and ectodermal placodes (30). Neural crest cells give rise to dorsal root ganglia (DRG) neurons and sympathetic ganglia of the PNS. The DRG neurons form bilaterally along the developing spinal cord to innervate the skin and various organs. Remaining neural tube elements generate the truncal spinal cord which consists of only a single cell layer (10). These cells of the epithelium divide and become progenitor cells for the neural and glial cells within the spinal cord which subsequently undergo cell differentiation to form the grey and white matter of the spinal cord.

1. Laminar organization of the spinal cord

At different positions of the spinal cord, distinct neuronal types are present along the dorsal-ventral axis, which results in the laminar organization of specific neurons according to their functional and physiological properties (41, 99). Laminar segregation of the neuronal functions within the spinal cord segregate cells of motor function more ventrally while cells mediating sensory information are present in the dorsal horn. Lamina I–X in the adult spinal cord are determined by cytoarchitectonic parameters. Broadly, pain, thermosensitive afferents, so called C- and Aδ-fibers from the dorsal root ganglion, innervate the lamina I–II, touch-sensitive afferents enter the lamina II inner (i) to lamina V, and lastly, proprioceptive afferents such as Aβ and Aδ target the more ventral spinal cord including the motor neuron (MN) region. A different labeling for Rexed laminae regions are as following: marginal layer (ML) for lamina I, substantia gelatinosa (SG) for lamina II, nucleus propius (NP) for lamina III–V, and MN for lamina IX (281). Various transcription factors (TF) define during development the cytoarchitecture and diversity of neurons in the spinal cord. The itch-specific C-pruritoceptor fibers represent a subpopulation of the C-nociceptors.

2. Spinal transcription factors and neuronal cell development

The distinct combination and interplay of signaling pathways with homeodomain and basic helix-loop-helix (bHLH) transcription factors (229) regulate neurogenesis and define cellular identity in a tightly timed choreography during development of the spinal cord (150). In the ventral neural tube homeodomain, TFs are the drivers of neuronal fate mapping during development (150). The expression and presence of TFs during developmental stages is dynamic and often transient, i.e., bHLH factors ASCL1 and ATOH1 and NEUROG1 appear in proliferating progenitors but disappear during differentiation and once cells become postmitotic. The lineage of pruritoceptors in the spinal cord is presumably dependent on the proneural NEUROG1 which has been described for their DRG development (175). In the spinal cord, the same proneural lineage might represent the precursors of itch-specific neurons. For itch circuits, the bhlhb5 TF is developmentally crucial in determining the itch-inhibitory interneurons (229). The transcription factors PAX2 and TLX3 are postmitotic and remain in postnatal stages and are used as lineage markers for neuron populations. A recent publication suggested that most interneurons that receive pruritoceptive input lack PAX2 labeling (100). Sonic hedgehog is a morphogen that is crucial for the patterning of the dorsoventral axis of the spinal cord by either activating or repressing TFs (43, 97). Notch signaling maintains cell proliferation while bHLH influences cell differentiation. A concentration gradient of sonic hedgehog regulates the TF expression which is later cross-regulated by the bone morphogenetic proteins (BMP) and WNT signaling for the generation of the dorsal cell types (79, 241). BMP and WNT are instrumental in the generation of dorsal interneurons. The rostrocaudal axis is patterned and positioned via the interplay of graded concentrations of fibroblast growth factor (FGF), retinoic acid (RA), and transforming growth factor (TGF)-β member GDF11 which regulate the expression of homeobox (HOX) TFs in progenitor and postmitotic cells. Detailed discussions of the developmental fate and signaling pathways influencing the neuronal map during development are available (45, 128, 161, 187, 292). Terminal neuronal phenotypes are regulated by the expression of transcription factors such as LHX2 and LHX9 which also promote axonal guidance. Transiently expressed TFs such as bHLH members ATOH1, PTF1A, and ASCL1 directly influence gene regulation for the terminal fate of neuronal phenotypes (37, 236). GABAergic neurons, enzymes, and inhibitory neuron circuits, for example, are regulated by PTFA1 (37). These inhibitory neurons are fundamental to the processing of itch signals in the developed spinal cord. Neurons derived from the dorsal and ventral horns remain in their original segments although the laminar structure is modified as the spinal cord neuronal map is finalized.

VI. HUMAN VERSUS MOUSE NEURONS

Comparisons between mouse and human neurons have now been made with respect to the peptidergic population of nociceptors (231). Given that pruritoceptive neurons form a small subset within the nociceptors, it may be reasonable to consider that the similarities and differences are transferable to the itch-specific population. Somata size and neurochemical characterization by TrkA, a small diameter neuron marker, as well as TrkB and TrkC which label larger diameter neurons were similar between mouse and human. In contrast, 50% of human neurons express TrkA+ and Ret during adulthood, whereas the expression profile for Ret and TrkA undergoes developmental changes in mice. Additionally, human TrkA+ neurons are colabeled with TRPV1. Perhaps these double positive neurons have an evolutionary advantage, to provide protection coverage during heat or inflammation. It is speculated that the TrkA/TRPV1 neuron cluster responds to acute and highly localized painful stimuli. Perhaps a small subset of these double positive neurons is responsible for itch transmission in humans. Human and mouse DRG neurons differ in their expression profile and number of positive neurons responsible for the transmission of itch and pain.

VII. CIRCUITRY AND GATING OF ITCH

The circuitry that controls itch transmission encompasses a dense network of neurons within the superficial dorsal horn, which is innervated by C- and Aδ-fibers. Noxious and pruritic stimuli are processed mostly in the superficial dorsal horn, whereas deeper dorsal horn neurons receive nociceptive and pruritic input via polysynaptic innervation. Interplay between inhibitory and excitatory interneurons sets the tone for the spinal input/output and sensitivity of the message. The interneurons are defined by their neurochemical and neuropeptidergic signatures (281), including the expression of marker molecules SOM, neuropeptide Y (NPY), and parvalbumin (PV) (317, 318). Generally, the gross differentiation of interneurons into excitatory or inhibitory subsets is according to neurochemical and neuropeptidergic characterization. For example, SOM+ interneurons excite neurons in the mechanical pain circuit by two distinct mechanisms: 1) the superficial SOM+ interneurons receive monosynaptic Aδ- and C-fiber input to process acute mechanical pain, and 2) the second SOM+ population in the deeper layers of the spinal cord receives polysynaptic input that is modulated by intermediate inhibitory interneurons. These are engaged by Aβ afferents, which are responsible for the transmission of nonpainful mechanosensation. For this reason, activation of SOM+ interneurons directly results in spontaneous pain and decreased mechanical pain thresholds (58, 73). Additionally, SOM+ interneurons broadly overlap with the expression of excitatory mediators such as vesicular glutamate transporter 3 (VGluT3) and calretinin (CR) which are known to amplify nociceptive messaging (257). In contrast to the function of the second SOM+ population, inhibitory interneurons in the superficial dorsal horn are essential for controlling sensitivity to painful and itchy stimuli. Approximately 40% of interneurons located in lamina I–III contain the inhibitory neurotransmitter GABA, and defined as GABAergic. In a subpopulation, the GABAergic neurons coexpress, and most likely co-release glycine, the other major inhibitory neurotransmitter. Within the deeper lamina of the spinal cord, GABAergic interneurons are key factors in the Gate Control Theory (183).

The Gate Control Theory introduced by (183) explains the mechanism of control mediated by nociceptive and non-nociceptive neuronal input. Lifting of this control leads to peripheral and central sensitization processes seen in inflammation and pathological conditions of itch and pain. The theory provides that non-nociceptive large-diameter neurons “close” the gate by activating interneurons in the spinal cord which in turn “inhibit the activation of projection neurons as crucial neuronal connectors to the supraspinal nervous system” (41, 99). Noxious input is transmitted by the nociceptors which are mostly peptidergic or nonpeptidergic, unmyelinated, slow conducting C-fiber afferents and by a subset of myelinated Aδ-fibers. Activation of the nociceptors “opens” the gate and excites the projection neurons and “turns off” the inhibitory interneurons. Essential players in the Gate Control Theory are morphologically and neurochemically diverse inhibitory interneurons that release inhibitory neurotransmitters to modulate pain messages (41, 99). It has long been acknowledged that loss of inhibitory circuitry due to cellular and molecular changes results in chronic pain and amplification of pain signals to the brain. More recently, inhibitory interneurons were reported to also control itch messages (58, 73). One of the major neurotransmitters is GABA, which controls the sensory gating of itch and pain transmission in the spinal cord and marks the population of inhibitory interneurons (FIGURE 2). Besides GABA, inhibitory interneurons express NPY, dynorphin (DYN), and galanin (155, 301). A subset of DYN+ interneurons which also express glycine+ and GABA+ might modulate the signal transmission by co-release of these neurotransmitters. However, it is unclear if these neurotransmitters are differentially employed for the inhibition of itch and pain messages.

The role of GABA in itch transmission has gained attention in the last few years. Briefly, GABA signals via two distinct types of receptors. GABA-A receptors are ligand-gated ion channels that exist in pentameric structures. These also serve as receptors for benzodiazepines and ethanol. GABA-B receptors are metabotropic glutamate receptors and belong to the group C family of G protein-coupled receptors (GPCRs) (39, 40). The initial work on the role of GABA and glycine in itch circuitry suggested that strychnine, a glycine-receptor antagonist, was more effective than GABA receptor antagonists in disinhibiting the spontaneous firing in the paws of itchy mice (6). A different approach suggests GABA as a major inhibitory transmitter in a mouse model of neuropathic itch as well as inflammatory chronic itch. With a cellular approach in which medial ganglionic eminence (MGE) cells of the forebrain are transplanted into the mouse spinal cord, the group tested whether rescue of GABAergic control impacted different itch phenotypes. In bhlhb5 mutant mice, MGE transplantation of GABAergic cells resulted in improvement of neuropathic itch (42). When tested in IL-31TG mice, the MGE-GABAergic cell transplants also successfully rescued the GABAergic tone and subsequently reduced the phenotypical severe itch (54).

The above-mentioned bhlhb5+ interneurons (B5-I) are crucial in the control of itch. B5-I neurons receive input from neurons that are activated by itch counterstimuli such as menthol and capsaicin and act rather anti-pruritic. B5-I neurons inhibit the excitation of the itch specific GRPR+ interneurons (discussed further below), which are activated by pruritogen-specific circuitry. Dynorphin, a kappa-opioid agonist, marks the B5-I interneurons (139). Dynorphin might be the neurotransmitter that blocks itch mediated by chemical counterstimuli, whereas GABA might be relevant in blocking itch counterstimuli associated with scratching (mechanical input).

In terms of excitatory circuitry specific for itch, the identification of GRP has gained attention. The recent attention to GRP builds on work from Alan Cowan who demonstrated in 1983 that bombesin, a peptide with tight sequence homology to GRP, caused scratching when injected intrathecally (96). GRP has been reported to be expressed in peptidergic DRGs, released upon pruritic stimuli, and then communicates via GRPR in the spinal cord (271, 272). GRP was thus identified as the first itch-specific neuropeptide engaged in the transmission from the periphery to the central nervous system. The details of this role have been questioned (259). Newer studies describe the release of Nppb from DRG neurons which induces GRP circuitry in the spinal cord in response to multiple pruritogens (186). Moreover, the GRP+ neurons are presumably excitatory interneurons. In summary, these findings led to the current modality-specific circuit from Nppb, Nppb-receptor (Npra), GRP, and finally GRPR. In a different study, GRP-positive interneurons were found to receive noxious input and gate nociceptive information in the superficial dorsal horn. However, intense nociceptive stimulation of the GRP+ interneurons recruits enkephalinergic interneurons which in turn results in inhibition of the intense nociception (270). A “leaky gate model” was described in which GRP+ interneurons code itch but also receive nociceptive input and are thus strongly activated by intense pain, triggering enkephalin release which in turn inhibits pain.

Insight into the functional segregation of peripherally and spinally expressed somatostatin uncovered the role of SOM in itch modulation (119). It was demonstrated that optogenetic stimulation of SOM+ sensory neurons was sufficient to evoke itch. In addition, SOM in combination with Nppb or GRP potentiated the itch induced by pruritogens. Itch induced by intrathecal SOM was attenuated by the release of DYN from B5-I interneurons. This finding led to the suggestion that itch induced by SOM was mediated via inhibition of DYN+ neurons and subsequent disinhibition of GRPR+ neurons. Additionally, SOM released by primary afferents suppresses pain, leading to the possibility that at least SOM-modulated forms of itch might also inhibit pain, extending the concept of somatosensory regulation by counterstimuli.

VIII. GLIA NEURON INTERACTION

Glia are implicated in chronic itch, consistent with their involvement in inflammatory and neuropathic pain (13, 286–288). Glia, from the Greek, meaning glue, were discovered by Rudolph Virchow, the German anatomist. Virchow suggested that glia provided a matrix for neurons to be embedded in the brain. Glia have gained attention as important modulators and communicators within the nervous system in healthy and diseased tissue. In the periphery, Schwann cells, or neurilemma, produce myelin sheets around neuronal axons. Schwann glial cells (SGC) express transient receptor potential (TRP) channels and modulate neurogenic inflammation, pain, and eventually itch. Schwann cells express TRPA1, which is linked to neuroinflammation (146). Silencing of TRPA1 in nociceptors yielded mechanical allodynia attenuation without influencing microglia infiltration, whereas TRPA1 silenced in SGC resulted in reduction of allodynia and signs of neuroinflammation. The underlining mechanism of SGC-expressed TRPA1 is speculated to occur via CCL2 release, which serves as a chemoattractant for macrophages. SGC-specific TRPA1 may function separately from TRPA1 on itch-specific neurons. Further studies are required to detail the impact of SGC on inflammatory and chronic itch (FIGURE 3A). Such studies may be of particular importance in light of the identification of a mesh network comprised of specialized cutaneous Schwann-glial cells that are essential for sensing noxious (1).

FIGURE 3.

Glia-neuron interaction. A: in the periphery, Schwann cells are thought to modulate itch via TRPA1 channels and the release of CCL2. While expressed in neurons and Schwann glial cells (SGC), silencing of TRPA1 in the latter resulted in reduction of sensitization and signs of inflammation. B: spinal cord glia comprises microglia, astrocytes, and oligodendrocytes. Pruriceptors, a subpopulation of nociceptors, target lamina II inner (IIi). Activated microglia as detected by increased p38 mitogen-activated protein kinase (MAPK) signaling (323) might modulate the pruriceptors and the gastrin releasing peptide receptor (GRPR) neurons (itch signaling second-order neurons). The close proximity of GRPR and gastrin releasing peptide (GRP) in spinal cord neurons enables direct interaction. Itch signaling and amplification of the itch signal might result in a decrease of the inhibitory neurotransmitters (GABA, dynorphin, glycine) and simultaneously increase the excitatory activity via GRP. Activated astrocytes (astrogliosis), detected by increased signal transducers and activators of transcription 3 (STAT3) immunolabeling, release lipocalin 2 (LCN2) which plays a major role in itch progression and maintenance of chronic itch.

Microglia, astrocytes, and oligodendrocytes are the neuroglia in the central nervous system (232). As with pain, activation of central glial cells has been reported to contribute to itch in Nc/Nga mice, a mouse strain available in Japan that is prone to develop itch under non-pathogen-free housing conditions (252). Astrogliosis, an abundance of reactive astrocytes, was a feature involved in chronic itch. To explore the hypothesis of glial contribution to chronic itch, morphological changes of astrocytes have been observed over time. As compared with healthy, non-itchy mice, the astrocytes in the itchy mice were characterized by enlarged bodies and aberrant arborized processes, a sign of astrogliosis.

Lipocalin (LCN), akin to some other glial transmitters such as ATP, is linked to neuronal inflammation and in the modulation of neuronal excitability. In addition, intrathecal co-injection of LCN2 together with GRP enhanced GRP-elicited itch, suggesting that LCN2 plays a role in itch progression and the maintenance of chronic itch (FIGURE 3B). Aligned with these findings, serum LCN2 levels are high in atopic dermatitis and psoriasis patients (290), suggesting that LCN2 might serve as a biomarker.

Toll-like receptor-4 (TLR4) expressed in astrocytes has also been linked to chronic itch (172). TLR4 is a member of the TLR family responsible for innate immunity, recognizing exogenous ligands and pathogen-associated molecular patterns (PAMP) in the course of viral or bacterial infections (173, 174). They react also to endogenous ligands and danger-associated signals during tissue injury. Liu et al. (172) showed that TLR4 knockout (KO) mice are characterized by reduced spontaneous scratching and touch-evoked scratching in models of dry skin and contact dermatitis. The reduced behavior in the mice correlated also with less activated GFAP+ astrocytes in the spinal dorsal horn in chronic itch mice. Of note, acute itch as induced by compound 48/80 and chloroquine (CQ) were unaffected in TLR4 KO mice. Activation of TLR4 by intrathecal lipopolysaccharide indeed evoked pain and suppressed itch but still enhanced alloknesis and chronic itch, accompanied by reduced astrogliosis. Astrocyte-expressed TLR4 may thus contribute to the maintenance and sensitization of chronic itch.

Microglia constitute another type of glial cell and surpass neurons in number. They are derived from myeloid precursors, constantly screen the environment for danger, and have been implicated in maintenance of neuropathic pain (284). Scratching triggered by compound 48/80 activates microglia in the spinal cord (322). Like neuropathic pain, activation of microglia occurs within minutes. This scratch-induced microglial activation was detected with colabeling for microglia (CD11b) and phosphorylation of p38. Inhibition of the scratching by nalfurafine, a κ-opioid receptor agonist, resulted in diminished microglial p38 phosphorylation. Unlike in pain, reactive microglia are not limited to the dorsal horn of the cervical spinal cord. Glutamate or ATP released by pruritoceptors might activate microglia or itch-specific neurotransmitters such as GRP or Nppb, which may then turn on microglial activation. Furthermore, mast cells are essential to the recruitment of glial cells in the spinal cord in neuropathic and neurodegenerative diseases (95, 108).

IX. SENSITIZATION IN ITCH

Patients suffering from chronic itch, analogous to those with chronic pain conditions, perceive intensified responses to mild pruritic stimuli. This phenomenon is referred to as sensitization. Dysesthesia, the Greek term for abnormal sensation, captures the altered perception and somatosensory changes occurring in itch and pain. Alloknesis in itch and allodynia in pain are comparable terms describing the dysesthesia in patients to non-itchy/painful stimuli such as to fabrics. Exacerbated itch and pain to stimuli are categorized under the terms hyperknesis for itch and hyperalgesia for pain conditions. Some of the molecular and cellular processes described above contribute to the sensitization processes: loss of inhibitory control, activation of glial and immune cells, and increase of excitatory activity in sensory neurons. Touch-evoked pain is a concern in 50% of neuropathic pain patients (127). It is not clear if touch-sensitive neurons are responsible or whether touch-specific C-fibers are switched on in neuropathic pain and increase the sensitivity to message pain for touch.

Alloknesis has been defined as itch occurring to innocuous but dynamic tactile stimuli, which can be mildly painful (208) or above the activation threshold of mechanosensitive C-nociceptors (38). The classic concept of alloknesis by touch has been extended to itch sensitization by innocuous warmth or noxious heat (12). Another phenomenon in AD patients is cross-modality, where painful stimuli can enhance the itch sensation (114, 123). This form of dysesthesia may be termed “algoknesis” (12). Akiyama et al. (5) asked whether pruritogens cross-sensitize with respect to behavioral and neuronal responses of cultured DRG neurons. When bovine adrenal medulla 8–22 (BAM8-22) was injected, the group observed an enhancer effect on scratching behavior when followed by a second injection of BAM8–22 or the pruritogenic hexapeptide SLIGRL. DRG neurons did not respond by mirroring the behavioral responses, suggesting a different cellular mechanism of the ongoing behavioral effects (5). In contrast, SLIGRL application induced cross-sensitization for BAM8–22 which was also reflected in increased neuronal activity of DRGs, not only in the mean peak response but also the increase in the number of responsive cells. SLIGRL reduced serotonin (5-hydroxytryptamine or 5-HT) evoked scratching but did not influence histamine. Histamine cross-sensitized behavioral responses to BAM8–22 but failed to induce increased neuronal activity similar to the effects observed for BAM8-22. One might speculate that neuronal changes as observed for SLIGRL are upon effects on subsequent signaling elements employed by the subsequent pruritogen. For example, activity modulation of the respective receptor and/or the signal transduction receptors, such as TRPA1, on DRG neurons might explain the cross-sensitization triggered by SLIGRL. These findings highlight the concept of peripheral sensitization. However, these are limited to an acute setting (immediate application of pruritogens) in mice, and how these observations translate to humans is not clear.

The effects of continuous exposure to mediators on hyperknesis, alloknesis, and algoknesis remain to be determined. In a dry skin model, Akiyama et al. (3) addressed the question whether chronic itch cause sensitization of itch pathways. In this model, SLIGRL as well as 5-HT injection into the dry skin caused hyperkinetic itch responses. These behavioral responses were also reflected by enhanced neuronal activity of isolated DRG neurons from the itchy mice, suggesting a peripheral sensitization for hyperknesis in the dry skin model (3). Protease activated receptor (PAR)-2 mediated effects could result from other cellular sources than DRGs in the phenomenon of hyperknesis given that SLIRGL-mediated effects in itch versus pain have been disputed (3, 105). Of note, other mediators and neuroactive cytokines such as IL-31 or NGF might impact neuronal activity, receptor distribution, as well as nerve fiber density in an inflammatory setting such as in AD or other inflammatory conditions. For instance, in human surrogate models, intradermal injections of NGF had limited impact on chemical itch sensitization, but were reported to induce hypersensitivity to pinprick (15). Sensitivity to pinprick has been defined as hyperknesis which seems to be relayed by polymodal C-fibers, whereas when occurring secondary to an itch provocation, type I Aδ-fibers relay through central mechanisms the hyperknesis (14). In ultraviolet B-induced inflammation, it is speculated that mechanically insensitive fibers and mechano-nociceptors respond with higher intensity to stimulation, implying that heightened activity has been set through peripheral and central processes. A similar contribution of peripheral and central sensitization is assumed to occur in itchy lesions of AD patients to pinprick-evoked hyperknesis (14). MrgprD-labeled C-fibers innervating the superficial layers of the epidermis and transmitting non-histaminergic itch, in this context of more importance, are sensitized to punctate stimuli in a mouse model of contact dermatitis (215). An underlying cause for pinprick-induced hyperknesis may be the loss of polymodal C-fibers and/or the sensitization of the remaining or a certain subset of them. However, one cannot exclude the central component of sensitized pathways such as loss of descending control from supraspinal levels and enhanced excitement of spinothalamic tract (STT) neurons. For instance, STT neurons respond more intensely to “normal” input from pruritoceptive afferents (12) which results in alloknesis or hyperknesis (12, 151). The discovery that pain-related dysesthesias observed in patients are independent from capsaicin-sensitive C-fibers, segment-restricted, as well as myelinated-fiber modulated suggests a complex interplay and synaptic connections, which might be similar in itch-experienced dysesthesias. Another observation related to sensitization made in chronic itch is that innocuous thermal stimuli affect serotonin-mediated but not histaminergic itch (14). In general, there seems to be limited sensitization to histaminergic itch, but increased responses to cowhage-induced itch in chronic itch patients. Additionally, noxious heat and other algogenic stimuli, i.e., acetylcholine and bradykinin, evoke itch rather than pain in chronically suffering patients (114, 123), suggesting that tachyphylaxis does not occur with itch.

X. ITCH IN THE BRAIN

The brain represents the final station for the somatosensory processing of itch. After the spinal coding of itch under acute or sensitized conditions, the signals are transferred via projection neurons, likely NK1R+ neurons, which connect to the STT. The parabrachial nucleus (PBN) is the next supraspinal itch processing station, which branches into different brain regions. Most activated regions in the human brain are the primary and secondary somatosensory cortex (S1-S2) as detected with functional magnetic resonance imaging. In chronic itch patients, the cingulate and prefrontal cortex reveal a more robust activity as compared with healthy controls when each group is treated with histamine. Descending control from higher brain regions modulates responses in the spinal cord (6). Serotoninergic neurons from the nucleus raphe magnus (NRM) have been shown to directly connect with GRPR+ neurons and to potentiate itch. Noradrenergic neurons in the locus coeruleus modulate inhibitory interneuron activity in the spinal cord. The periaqueductal gray (PAG) is activated during scratching and was considered as an itch suppressing brain area. Tachykinin-1 (Tac1) labeled neurons in the PAG facilitate itch since their ablation significantly abolishes itch behavior in acute and chronic models. Direct activation of the Tac1+ neurons causes itch behavior in mice (92). This descending faciliatory pathway from Tac1+ PAG neurons was not linked to serotoninergic control.

The pleasure of scratching an itch is processed in the midbrain reward center. GABAergic and dopaminergic (DA) neurons in the ventral tegmental area (VTA) partake a role in the central processing of itch. In general, DA neurons are key neuronal elements in motivation and reward systems (66b), whereas GABA neurons are believed to drive aversion and disrupt reward processes (277). VTA GABA neurons are activated in acute itch and show an intensified activation when scratching is prevented, suggesting a function in processing the aversive and unpleasant components of the itch sensation. Intriguingly, DA neurons in the VTA are activated temporally delayed, and the activation is linked to the scratching behavior of the tested animals. These data support the concept that the midbrain VTA DA neurons might transfer the central pleasure component of scratching an itch. Chronic itch processing in the VTA was demonstrated to occur via DA and GABAergic neurons, suggesting a general function of the midbrain region to control acute and chronic itch. Another center in the brain is responsible for the integration of the affective component of itch. As shown for pain (325), the amygdala links anxiety and stress affections to chronic itch (242). Even though both sensory modalities converge in shared central regions, pain and itch are discernable from each other. The central pathways and circuits that differentiate itch from pain overlap anatomically, and the mechanisms by which these sensations are distinguished remain to be determined. Central mechanisms of contagious itch relay via the GRP-GRPR axis in the suprachiasmatic nucleus (SCN) in mice as discussed in section XVII.

XI. G PROTEIN-COUPLED RECEPTORS

7-Transmembrane receptors or GPCRs are exquisitely associated with itch arising in the periphery. Their ligands are direct mediators of itch when the cognate receptors are expressed on neurons and indirect when the receptors are expressed on nonneuronal cells. Extensive tables that list GPCRs associated with itch have been published, yet data supporting the importance of any particular GPCR in any acute or chronic clinical itch are remarkably limited. GPCRs that are most strongly associated with human itch are discussed, including members of the Mas-related G protein-coupled receptor (Mrgpr) family.

Itch arising in the periphery can be divided into two broad categories: histamine dependent and histamine independent. Clinical itches associated with histamine are now considered limited primarily to some patients with urticaria and drug reactions with a lesser role for histamine in AD and other conditions. This limitation accounts for the relative ineffectiveness of antihistamines in itch. It follows that histamine-independent itch is the primary driver of itch that arises in the periphery. The foundation for this understanding is built on a number of observations. These observations include the identification of mechanosensitive C-fibers responsive to cowhage but not histamine (132), that the active component of cowhage stimulates Mrgprs, and that additional substances that evoke itch in humans and scratching in mice stimulate specific GPCRs unrelated to histamine receptors.

A. Mrgprs

Mrgprs have risen to prominence with respect to itch that is independent of histamine. This prominence is manifest in a number of ways and provides a basis for the discussion here. First, Mrgprs respond to a variety of pruritogens, and these links are examined (FIGURE 4). Second, the expression of Mrgprs is limited primarily to sensory neurons implicated in itch as well as mast cells, long implicated in itch and allergy, and is also discussed. Neurons and mast cells are in close proximity and are considered to interact directly in epithelia where neuroimmune interactions are prominent and where neurogenic inflammation takes root.

FIGURE 4.

Mrgpr gene loci and receptor ligands in mice and humans. All Mrgpr genes are located on chromosome 7 in the mouse and chromosome 11 in humans. While still considered orphan receptors, substances known to activate specific mouse and human Mrgprs are indicated. Note that some substances which activate a single human receptor interact with more than one mouse Mrgpr. Similarly, some substances which activate a single mouse Mrgpr interact with more than one human receptor.

Mrgprs comprise a family of orphan GPCRs identified early in this century by an academic group led by Xinzhong Dong (72) while a postdoctoral fellow with David Anderson at Caltech and now at Johns Hopkins, and a group led by Paola Lembo and Sultan Ahmad, at Astra Zeneca (163). The respective groups referred to the receptors as Mrgprs and sensory neuron-specific receptors. The encoding of ~50 Mrgprs in the mouse indicated initially that only olfactory receptors were more numerous. A potential role in sensing was appreciated given the additional combinations of tissue distribution limited primarily to sensory neurons and mast cells, together with the appearance of Mrgprs in tetrapods. It is now recognized that the large number of Mrgpr genes in mice, due to a retrotransposon insert was the exception, as other rodents had fewer, and humans have eight. Mrgprs were divided into subclasses based on sequence homology. Mice have A, B, and C subclasses, which correspond loosely to human X1–4 while both species have D, E, F, and G (FIGURE 4). Ironically, canines, considered excellent at scratching, were found to have only a single Mrgpr, homologous to X2. These confounding observations raised the possibility that while most GPCRs were activated by a small number of ligands, perhaps Mrgprs were broad sensors, which turns out to be the case.

A role for Mrgprs in itch was first suggested in 2009 when it was found that CQ induced scratching in mice that was dependent on MrgprA3 and that CQ could also activate human MRGPRX1. The year before, Reddy, Lerner, and co-workers (224) isolated mucunain, a cysteine protease that was the active component of cowhage. While we found that mucunain, and subsequently the endogenous cysteine protease cathepsin S associated with itch and inflammation, could activate protease-activated receptors, further studies revealed that these proteases targeted Mrgprs (222, 224). The surprising finding that a protease could evoke itch via activation of a Mrgpr served to inform further studies that would explain a long-standing discrepancy between mouse and human studies involving SP. Specifically, SP is linked to itch, pain, and neurogenic inflammation. A series of publications revealed that the peripheral actions of SP are mediated via Mrgprs, throwing a wrench into the conventional view that NK1 was the primary receptor for SP. NK1 antagonists were developed in the 1990s and found to be effective in mouse models of itch and inflammation but failed in the treatment of inflammation in humans. Preliminary results from uncontrolled studies suggested that oral administration of the NK1 antagonist aprepitant might be useful for the treatment of chronic itch. These results could not be confirmed in controlled studies (285). Topical application of aprepitant was also ineffective in itch (296). Moreover, NK1 mutant mice still scratched when injected with SP (28). In contrast, Mrgpr cluster Δ−/− mice, which lack 12 Mrgprs but retain mast cell SP-responsive MrgprB2, did not scratch. SP-evoked scratching was found to be dependent on neuronally expressed murine MrgrpA1 while SP activated human MRGPRX2 (27), a finding that has recently been extended to pain (98). The explanation for the conflicting findings between mice and human turns out to be simple: The NK1 antagonists were also antagonizing mouse Mrgprs but had no impact on human Mrgprs. These observations make Mrgprs interesting therapeutic targets for the treatment of itch.

As there are more Mrgprs in mice relative to humans, mapping with respect to ligand specificity would not be expected to be one-to-one, and it is not (FIGURE 4). Several itch mediators as well as substances implicated in allergy, pseudoallergy, and inflammation activate only individual human MRGPRs, but some of these substances activate more than one mouse Mrgpr. Linking a specific endogenous substance with a pathological itch has now occurred. Bile acids activate human MRGPRX4 (182) and can account for cholestatic itch. That cholestatic itch does not occur readily in either mice or canines is consistent with the lack of a functionally homologous receptor in these other species. It is not much of a speculation to think that the itch of chronic kidney disease, which is limited to humans, may also result from an endogenous ligand activating a Mrgpr. Three receptors in humans, MRGPRX1, MRGPRX2, and MRGPRX4, are thus considered to fulfill the roles of the A, B, and C subclasses of Mrgprs while β-alanine activates MrgprD in both species.

With respect to cellular expression, in humans, MRGPRX1 and MRGPRX4 are expressed only on sensory nerves. MRGPRX2 is expressed on mast cells but has also been identified on DRG neurons (226, 311). While MRGPRX2 is not readily identified in neurons based on RNAseq of resting neurons, expression has not been considered during the process of neurogenic inflammation, a critical knowledge deficit in the context of intercellular and mediator crosstalk. With respect to expression in mice, MrgprB2 is limited to mast cells and considered generally homologous to human MRGPRX2. The Mrgprs A1, A3, and C11 are limited to sensory nerves. While MrgprA1 is highly expressed only during development, the low level of expression in adult mice retains function, as mice scratch to bilirubin.

The affinity for ligand receptor interactions is critical. Thus SP activates MrgprB2, MrgprA1, and MRGPRX2 (FIGURES 4 and 5). MrgprB2 is still present in the Mrgpr cluster Δ−/− but SP does not cause scratching in these mice. Thus, in mice, SP causes itch by activating MrgprA1 on sensory nerves rather than activating MrgprB2. In humans, if MRGPRX2 is only on mast cells, then SP evokes itch by stimulating mast cells. In this scenario, the mouse receptors MrgprA1 and MrgprB2 share functional homologies with human MRGPRX2. This concept is consistent with single human receptors sharing function with multiple mouse Mrgprs as noted in FIGURE 4. This concept is supported further by the following observations. SLIGRL, the hexapeptide derived from PAR-2, but which stimulates MrgprC11 on sensory nerves in mice to evoke itch, also activates MRGPRX2, while the potent itch inducer compound 48/80 activates Mrgprs expressed on mast cells and neurons. Returning to cysteine protease activation of Mrgprs, while cathepsin S activates MrgprC11 and MRGPRX2, the dust mite cysteine protease Der p1 associated with allergy and inflammation, but not itch, activates MrgprC11 and MRGPRX1 while cowhage activates human MRGPRX1 and MRGPRX2. Finally, vancomycin which is implicated in red man syndrome and staphylococcal delta toxin which is associated with atopic dermatitis, and thus itch, activate MRGPRX2 (213).

FIGURE 5.

Sensory neuronal receptors and signaling molecules for the transduction of itch. Sensory afferents express a multitude of receptors, providing redundancy to ensure the transmission of acute and chronic itch signaling. A number of these receptors are depicted here, including serotonin-5-hydroxytryptamine receptors (5-HTRs), promiscuous murine itch receptors MrgprA1/A3/C11, cytokine signaling receptor complexes interleukin (IL)-31RA, oncostatin M receptor β (OSMRβ), and IL-4Ra and IL-13R and voltage-gated sodium channels (Nav). A plethora of downstream signaling molecules link receptor activation with generation of action potentials, the details of which remain to be determined. In persistent itch, signaling might increase receptor expression, including the itch-related signaling transducer channels TRPV1 and TRPA1. It remains to be investigated whether downstream signaling mediators such as the mitogen-activated protein kinase (MAPK) and Janus activated kinase (JAK)/signal transducers and activators of transcription (STAT) acutely alter itch-relevant receptors via sensitization or on transcriptional levels. BAM8-22, bovine adrenal medulla 8-22; PKC, protein kinase C; PLA, phospholipase A; PLC, phospholipase C; SP, substance P.

Neuronal expression of Mrgprs has allowed these receptors to serve as markers or fluorescent tags that have been used to trace pathways of itch in mice. Of particular importance, MrgprA3 fibers were found to innervate the epidermis exclusively and respond to mechanical stimuli, noxious heat, and a number of pruritogens (103). These are features of non-histaminergic, polymodal itch fibers. Ablation of these fibers with the aid of diphtheria toxin resulted in decreased scratching behavior from a range of pruritogens while pain sensations were intact. Evaluation of dry skin and allergy in these mice, considered chronic models of itch, were also modulated. Mice were engineered further to express TRPV1 only in MrgprA3 neurons, and then treated with capsaicin. Scratching but no pain behavior was evoked. These studies established the existence of a labeled line for itch, at least in mice, in contrast to the prevailing intensity theory of itch versus pain.

A number of areas of Mrgpr biology in general and itch in particular remain to be clarified. 1) Beyond bile acids, what are the physiological endogenous or exogenous ligands for Mrgprs, and what might that imply about Mrgpr biology in humans versus mice, canines or non-human primates? 2) Do human Mrgprs of the E, F, or G subclasses have relevance to itch? 3) Since MrgprA3 neurons are considered nonpeptidergic but respond to SP, can peptidergic and nonpeptidergic neurons communicate directly? 4) Is expression of neuronal Mrgprs inducible during the course of neurogenic inflammation? 5) While CQ induces scratching in mice and can activate human MRGPRX1, hydroxychloroquine, a derivative, is used widely in the therapy of autoimmune conditions but rarely triggers itch. Therefore, might the itch associated with CQ in Africans be linked to a Mrgpr polymorphism? 6) As elevated levels of MRGPRX2 on mast cells correlate with severe urticaria (87), it may be reasonable to expect that Mrgpr polymorphisms will be associated with, or protect from, acute or chronic itches.

Taken together, Mrgprs on mast cells and neurons provide powerful links between itch, allergy, and neurogenic inflammation. The recent identification of a Mrgpr antagonist that inhibits itch in mice (28) and the demonstration of residues critical for activity via in silico modeling and in vitro activity (157, 223) are consistent with the possibility that drugs that target this class of receptors will be identified and evaluated for therapeutic benefit.

B. NK1R

Tachykinins are a family of neuropeptides. They stimulate rapid contraction of intestinal muscles as compared with the slow effects mediated by bradykinins. The best-known tachykinin with respect to itch is SP. Until recently, the cognate receptor for SP in the periphery has been considered the neurokinin-1 receptor, NK1R. It is now appreciated that while NK1R may have a role in the spinal cord with respect to itch, the relevant receptors for SP in the periphery are members of the Mrgpr family, discussed above. This shift in our understanding parallels the realization that the role of the H1 receptor in clinical itch is remarkably limited. SP is detected in primary sensory nerves that innervate the epidermis, dermis, blood vessels, and various cell types. SP exerts differentiated function on different cell types and cutaneous structures. For example, it has been suggested that SP upregulates NGF in keratinocytes which would affect neuronal regeneration (44) or lead to sensitization. A key role for SP in neurogenic inflammation is via close vicinity to cutaneous blood vessels which is aggravated through the additional feed-forward loop of NGF-dependent upregulation and release of SP in endothelial cells (185). Moreover, these tachykinin-marked sensory neurons are classified as the unmyelinated C-fibers which mark the C-nociceptors and the thinly myelinated Aδ-fibers, both of which conduct nociceptive stimuli and mediate the neurogenic inflammation. In addition to SP, CGRP is equally present and released from perivascular nerve fibers and is responsible for the flare in neurogenic inflammation (94, 211), although CGRP does not cause itch. NK1R+ neurons in the spinal cord appear to mark projection neurons for messaging itch and pain to supraspinal levels. The generation of knockin mouse with a Rosa 1s tdTomato reporter by the Ross group is enlightening (117). Within the skin, the authors detected Cre-mediated recombination in dermal fibroblasts but not in mast cells, endothelial cells, Merkel cells, or Langerhans cells as elsewhere described. The lumbar spinal cord showed tdTomato labeling within lamina I while deeper levels of the spinal cord stained for NK1R. NK1R labeling was detected in most projection neurons, important cross points for the transmission of pain and itch (8). In this context, it was reported that ablation of the NK1R+ cells in the spinal cord of the rat attenuated the scratching responses elicited by 5-HT. The study revealed that NK1R+ and GRPR+ neuron toxic ablation in an animal model of chronic itch with ovalbumin (OVA) treatment showed that SP-SAP resulted in reduction of spontaneous scratching and alloknesis scores. Toxic ablation of GRPR+ interneurons did not affect either scratching responses or alloknesis (49). This finding suggests that the spinal GRP-GRPR axis is not essential for itch sensitization. Intriguingly, the lack of the GRPR neurons as well as the NK1R+ resulted in significant reduction of enhanced responses to CQ. The segregation for this observed function of the spinal GRPR neurons may be linked to peripheral sensitization (hyperknesis) versus spinal action of sensitization (alloknesis). Nevertheless, spinal NK1R+ neurons seem to play a role in both sensitized responses as well as the itch of OVA– mice, possibly due to a selective central sensitization of these spinal neurons and their polysynaptic nature. However, there is contradictory data suggesting functional NK1R in TRPV1+ rat DRG neurons population, implying a peripheral role in nociception, particularly in heat hyperalgesia (319).

In summary, most of the data described suggest an ascending spinothalamic role for NK1R+ neurons in itch. Overall, these findings can explain the limited success of NK1R antagonists in the treatment of itch.

C. PARs

PARs are encoded by four distinct genes, PAR-1 to PAR-4 (63). PAR-1 is the receptor for thrombin. Trypsin was found to activate PAR-2 (202). PAR-3 and PAR-4 were cloned by mRNA screening of rat platelets (310). PAR-1, -3, and -4 are expressed on neuronal, epithelial, astrocytes, and immune cells (204). Protease cleavage of PARs unmasks a sequence which acts as a tethered ligand and binds to conserved sequences to induce transmembrane signaling (204). The sequence of the tethered ligand is characteristic and distinct for each PAR, with SLIGRL derived from mouse PAR-2 being the most studied. PAR-induced cellular signaling engages either G proteins or arrestins. For example, PAR-1, -2, and -4 interact with G proteins, which results in activation of different kinases such as mitogen-activated protein kinase (MAPK) or phosphatidylinositol 3-kinase (PI3K) or protein kinases as well as tyrosine kinases depending on the activating ligand and residual cell type (266). The activation of PARs increases transcription of cytokines, chemokines, and growth factors which regulate different cellular processes.

Proteases have long been recognized as mediators of non-histaminergic itch. Itch induced by mucunain, the active ingredient of cowhage, was first described by Arthur and Shelley in 1955 (22). Mucunain induces long-lasting, histamine-independent itch without pain. Arthur and Shelley (22) speculated that mucunain might be a protease and induce itch directly via action on sensory nerves or via an indirect mechanism through the release of pruritic compounds. Protease-induced non-histaminergic itch was investigated, identifying trypsin and chymotrypsin as pruritogens (140). Overexpression of proteases or PAR-2 in mice results in a severe eczematous phenotype (86). Moreover, in a severely itchy mouse strain overexpressing serine protease channel activating protease-1 (CAP-1/Prss8), the phenotype could be rescued by backcrossing into PAR-2 KO animals. The concept of PAR-2-mediated itch was challenged by the discovery that the synthetic peptide SLIGRL also activates MrgprC11. It was found that SLIGRL-induced itch was lost in Mrgpr cluster Δ−/− mice. MrgprC11 relies on the presence of the RL-NH₂ sequence of the synthetic peptide. The controversy has not been fully clarified, but the possibility remains that the naturally occurring proteases might trigger itch via PAR-2. Steinhoff and colleagues (140)suggested that trypsin and tryptase activate PAR-2 on DRG neurons to trigger histamine-independent itch and both proteases induce neurogenic inflammation and activate PAR-1 and PAR-4. PAR-2 is also involved in pain, which complicates the picture. Effects were reduced in PAR-2-deficient mice, suggesting that the major neuronal function of PAR-2 is to induce hyperalgesia. Trypsin-mediated itch in mice did not require the presence of either MrgprC11 or PAR-2, implying the involvement of an additional receptor, perhaps PAR-4, which is expressed in rodent DRG neurons (23, 25). Regarding signaling, Patricio et al. (209) employed the PAR-4-specific activating hexapeptide AYPGKF-NH2 (AYP) to show that AYP activated the TRPV1+ neuron population and that AYP-elicited itch was diminished in TRPV1-deficient mice. The TRPA1 antagonist HC-0300310 blocked AYP-elicited itch, suggesting the involvement of this channel. However, TRPA1-deficient mice showed augmented scratching to AYP that was mirrored in calcium studies of TRPA1−/− DRG neurons when stimulated with AYP. A compensatory mechanism might play a role in the heightened responses to AYP in TRPA−/− mice and neurons. Another controversial finding of the study is the colocalization of PAR-4 with GRP in peripheral neurons (209). The GRPR antagonist RC-3095 blocked itch induced by AYP in mice, suggesting the involvement of GRP-GRPR circuitry. Due to its ability to cross the blood-brain barrier (9, 19), RC-3095 might exert its effect centrally to block GRPR consistent with prior findings by Mishra and Hoon (186). Future studies with more specific genetic knockout animals will be needed to fully explain this complex picture and its relevance to humans.

D. Histamine

Histamine is one of the most studied substances in preclinical and clinical research. However, its importance in clinical itch outside of urticaria and drug reactions is now considered limited, as that of histamine-independent itch has gained favor. Histamine is a short-acting, endogenously produced, widely distributed biogenic amine (253). Playing a major role in allergic associated inflammation, histamine regulates the maturation, activation, and chemotaxis of immune cells (278) besides exerting immune regulatory functions on monocytes, T cells, macrophages, and various other immune cells (158). Histamine activates, with different affinities, four distinct GPCRs, H1–4, named chronologically in order of their discovery (2). Histamine, the gold standard for histaminergic itch mechanisms studies (FIGURES 1 and 5), causes intense itch accompanied by vasodilation, redness, flare, and swelling in human when applied to the skin. In mice, histamine is often used as a positive control in the context of investigations that involve other pruritogens. The major source of histamine is differential granule release, together with other bioactive substances, from mast cells, following IgE- or Mrgpr-dependent stimulation. Histamine may be released from additional types of cells, including neurons, basophils, and keratinocytes. H1 and H4 receptors are expressed on DRGs (FIGURE 5), although the expression on the latter is low (293). H1R couples to Gα_q/11 G protein and induces signaling via phospholipases A₂ (PLA₂) and phospholipase C (PLC)-β3 as well as protein kinase C-δ (124). H4R couples to Gα_i/o proteins and is also expressed on a variety of immune cells and other cell types from the intestine and lungs (66a). Neuronal activation of itch via histamine is linked to TRPV1 (124) as well as TRPV4 (144).

The first antihistamine directed against the H1R, phenbenzamine, was marketed in 1942. H1 antihistamines act to downregulate constitutive receptor activity and are thus inverse agonists rather than H1-receptor antagonists (164). First-generation antihistamines are known for their sedative side effects. It was not until three decades later that the H2R antagonist cimetidine was put into clinical use for the treatment of gastric acid disorders. The first H3 inverse agonist is in use for narcolepsy (147). Cetirizine is a second-generation H1R antagonist. It is a metabolite of hydroxyzine, a widely used first-generation antihistamine. JNJ 7777120 is the first highly selective H4R antagonist, but its development has been curtailed due to adverse effects. Other H4R antagonists are under clinical evaluation for inflammatory disorders such as AD (300), while serliforant is a H4R antagonist undergoing evaluation for vestibulopathy (24).

Urticaria or hives are recurrent, pruritic, and characterized by central swelling with surrounding erythema in any body region. In contrast to most itches, the urge to scratch may be replaced with an urge to rub. What underlies the response of rub versus scratch remains a mystery. Urticaria lasts less than 6 wk in the acute form and may be associated with medications (nonsteroidal anti-inflammatory drugs, opiates, and narcotics), foods, viral infections, and contact allergens. Fifteen percent of such patients also have angioedema. Chronic urticaria continues for more than 6 wk and may be associated with autoimmune disorders such as systemic lupus erythematous. H1R antagonists have been a mainstay of treatment, although many are now treated with biologics directed to IgE and additional targets. It is increasingly appreciated that SP may be a major player in chronic urticaria as well as chronic spontaneous urticaria (CSU) which is characterized by spontaneous lesion reappearance. High levels of SP have been detected in CSU patients. Emerging concepts suggest that SP activates MRGPRX2, an emerging therapeutic target on mast cells, and participates in urticaria and the release of multiple allergens, including histamine. Antihistamines have been widely used in the treatment of allergic disorders for decades based on findings that histamine promotes inflammation and modulates immune responses. Beyond itch, HR are associated with tumorigenesis, manifest as either promoting or inhibiting tumor growth (181). Histamine was believed to promote neuroinflammation in diseases such as multiple sclerosis; however, this finding has been questioned as blocking H4R enhances inflammation (31). There have been discussions about histamine in the role of airway inflammation based on its activity on lung macrophages and the subsequent induction of pro-inflammatory cytokines such as IL-6. In an animal model of allergic airway inflammation, H4R KO animals showed less pulmonary infiltration of eosinophils and lymphocytes with a concomitant reduction of TH2 immune activation (74). Blocking H4R in a model of pulmonary fibrosis showed beneficial effects in reducing inflammation (176, 177). Histamine activates also H1R on pulmonary epithelial cells and induced TLR3 expression leading to IL-8 secretion to accentuate inflammation. Depending on the cell population that is driving the disease pathology, targeting either HR might be beneficial. In an AD model, H4R blocking attenuated itch responses in mice, presumably through reduction of IL-31 secretion from TH2 cells (230). Blocking H4R showed promising results in the psoriasis-like model, induced with imiquimod, through reduction of TH1 responses. These findings underline the pleiotropic effects of histamine. Of note, histamine exerts a protective role in some other preclinical models such as in the trinitrobenzene sulfonic acid model of colonic inflammation in mice (309). In contrast, when colitis was induced with dextran sodium sulfate, targeting H4R improved the condition in this model (244).

E. Serotonin

5-HT is a monoamine neurotransmitter. Injection of serotonin into skin causes itch in humans (34, 299) and scratching in mice (190). Despite much research, including data linking serotonin to allergic skin disease and to pruritus associated with both uremia and cholestasis (279), the importance of serotonin as either a peripheral or central mediator of itch in humans is unknown. This uncertainly is likely a reflection of the complexity of the serotonergic system. This complexity is exemplified by the existence of 14 serotonin receptors from 7 receptor families (106). All are GPCRs except for 5-HT₃, which is a ligand-gated ion channel. The diversity of receptors makes it challenging to know exactly which are important in itch. Most of these receptors are expressed in spinal neurons involved in ascending transmission. In the periphery, serotonin can be released from mast cells, platelets, and endothelial cells during injury and inflammation and acts on small- to large-diameter DRG neurons (307). Serotonin is a component of the inflammatory soup known to excite Aβ- and C-fibers (156). Treatment of pruritus by targeting serotonin receptors adds complexity rather than clarification. SSRIs can help treat pruritus, but they increase serotonin (centrally) by preventing reuptake. Ondansetron, an antagonist at the 5-HT₃ channel, anecdotally can be of benefit in the treatment of cholestatic itch, uremic itch, and opioid itch. However, such anecdotes provide limited useful information.

F. Additional GPCRs

Endothelin-1 (ET-1) is a 21-amino acid peptide and potent vasoconstrictor implicated in pruritus (142). Produced by endothelial, immune, and neuronal cells, ET-1 engages with two GPCRs, endothelin-A receptor (ETAR) and endothelin-B receptor (ETBR). In mice, ET-1 mediates itch behavior via ETAR and ERK1/2. ETAR+ neurons are of peptidergic origin and overlap with most of the itch receptors. ET-1 induces robust itch responses and is often used in rodent research as a non-histaminergic itch agent and has also been administered to humans via iontophoresis. Intriguingly, the responses in humans seem to be partially linked to histamine-induced itch, most likely via mast cell release based on the findings that a H1R blocker partially blocked the itch responses (142). It would be of interest to test other HR blockers to delineate whether ET-1 is histamine-related itch inducer in humans.

Prostanoids are derived from membrane lipids. Prostanoids are formed via the cyclooxygenase (COX) pathway from arachidonic acid (AA). Prostaglandins are the unstable forms of prostanoids and have been implicated in itch. Prostaglandin E₂ (PGE₂) is itch-inducing in AD patients, but this seems to rely on histaminergic itch (113). Despite increased levels of prostanoids or other AA derivatives such as thromboxane in AD, it is not clear whether itch is induced directly by these molecules or whether they function as inflammatory sensitizers in itch. The transmembrane protein TMEM79, which is expressed in keratinocytes and neurons, was reported to be crucial in pathological responses to oxidative stress (78), most likely in AD. TMEM79 loss in keratinocytes triggers mast-cell degranulation and subsequently histaminergic itch with PGE₂ and H4R/H1R interdependency. Since histamine is not considered a major pruritic contributor to AD, the relevance of this finding is not clear.

Leukotrienes are derivatives of AA and have been long implicated with itch. In particular, leukotriene (LT) B₄, a lipid chemoattractant and activator of granulocytes (238), has gained attention as relevant to itch signaling when released from keratinocytes upon itch stimulation by SP and SLIGRL (18, 139a). IL-31-mediated itch may employ keratinocyte derived LTB₄ (16). LTB₄ inhibition was achieved either by blocking the synthesis of the chemoattractant or by blocking the chemoattractant with pharmacological antagonists. These data support the notion that itch is a complex interplay of neuronal and immune components within the skin and very likely other organs.

Sphingosine 1-phosphate (S1P) is a bioactive signaling lipid associated with a variety of inflammatory diseases. These include not only AD and psoriasis (126), but also multiple sclerosis, which is often accompanied by itch and pain (206) that may be neuropathic in nature (267).The interaction of S1P with the S1PR3 receptor leads to activation of TRPA1 pruriceptors and TRPV1 nociceptors (110). These data reveal that distinct molecular mechanisms may underpin the discrimination of itch versus pain in the periphery.

That cannabinoids are used recreationally led to the identification of endogenous lipid-based compounds and receptors. Use of cannabinoids to treat inflammatory skin disease has been suggested and reported anecdotally (35). The increased legalization of marijuana is likely to drive therapeutic use, especially for itch and AD. Two cannabinoid GPCRs have been identified (191, 214). In mice, both receptors have been detected in sensory neurons, whereas in humans, only CB2R has been so localized (11, 263). Activation of CBR in sensory neurons may decrease neuronal activity and modulate the axon-flare response. Tetrahydrocannabinol, a bioactive component of marijuana, blocked scratching behavior elicited by compound 48/80 (245). These antipruritic effects were CB1R dependent as revealed by use of antagonists. Upon stimulation with hapten, CB1R mutants were susceptible to the development of features of AD (88). N-acetylethanolamines are beneficial for the treatment of itch through activation of the endocannabinoid system, and topical N-palmotoylethanolamine (PEA) has been reported to decrease uremic itch scores to baseline (273). Topical PEA cream also reduced the itch in a cohort of mild to moderate AD patients when applied twice daily for 4–6 wk (75).

XII. TRP CHANNELS

TRP channels constitute a family of ion channels expressed widely across species. They have been strongly implicated in itch, although their necessity has been questioned (234). TRP channels were first identified in Drosophila in 1969 (62). Their importance in nociception came almost three decades later, when it was determined that capsaicin activated a channel that became known as TRPV1. More than 25 distinct TRP channels have been described in mammals. TRPs respond to various stimuli via an influx of cations. TRPs are expressed in diverse cell types, including immune cells, neurons, and keratinocytes, and participate as sensors to maintain skin homeostasis. Based on their amino acid sequences, TRP channels form six subfamilies. These include TRPA (ankyrin), TRPC (canonical), TRPM (melastatin), TRPML (mucolipin), TRPP (polycystin), and TRPV (vanilloid). TRP channels are gated by stimuli ranging from small molecules and ions to thermal as well as mechanical inputs, underlying their role in tissue homeostasis and sensation. Their characteristic structure contains a central channel pore and four subunits built around the pore. TRPs feature six transmembrane domains with short interconnections to each other and their long NH₂-terminal domains are paired with short COOH-terminal domains, both of which extend into the cytoplasm. TRPs are generally nonselective mono- or divalent cation-gated channels with no anion permeability. Only a few TRPs have demonstrated specific selectivity towards calcium or magnesium, while most are nonspecific and gated by both cations. The association between TRPs and sensing has made them targets of drug development, but side effects, such as altered thermoregulation with respect to antagonists of TRPV1, have confounded drug approval. Furthermore, polymorphisms in TRP channels that manifest with alterations in pruritus, and which could thus guide drug development, have not been reported.

A. TRPV1

TRPV1 is gated by vanilloids, comprising compounds that have a vanillyl group, of which capsaicin and its analog resiniferatoxin, which is 1,000 times more potent, are members. Additional TRPV1 activators include thermal stimulus more than (50) 42°C (49), extracellular protons and anandamide, a lipid derived from cannabinoids. TRPV1 is found mostly on peptidergic nociceptor neurons and is a marker for the sensory neurons mediating pain and itch (50, 196, 197). TRPV1 is also sensitized by molecules released during inflammation, including bradykinin, prostaglandin, and retinoids (52, 313). Activation via inflammatory mediators results in the downstream signaling by protein kinases C and A as well as phosphatidylinositol bisphosphate (PIP₂) (227). In inflammatory pain, TRPV1 expression is upregulated, which may account for the increased heat perception (121, 294, 295, 308). Targeting TRPV1 to treat inflammatory pain seemed like a promising treatment strategy. Unfortunately, antagonizing TRPV1 has side effects such as hyperthermia and loss of noxious heat perception (138).

The role of TRPV1 in the mediation of itch associated with histamine, but not other tested pruritic mediators, was uncovered when it was found that the injection of histamine was attenuated in TRPV1-deficient mice. In contrast, itch elicited from the injection of ET-1 and serotonergic itch from α-Me-5-H were not altered. Seemingly, most pruritogens do not engage the TRPV1 channel for the transduction of the itch signal. IL-31 is one of the pruritogens that mediates itch via both TRPV1 and TRPA1 (FIGURE 5).

Efforts are ongoing to tease apart, if possible, the relative contributions of TRPV1 as compared with TRPA1 with respect to itch versus inflammation. In the squaric acid dibutylester (SADBE) model of allergic contact dermatitis, both channels were required for scratching behavior while genetic or pharmacological ablation of the TRPV1 population promoted inflammation (84). In this model, C-fiber neuronal activity precedes immune responses and regulates to some extent immune functions as suggested elsewhere (33). In contrast, TRPV1 was necessary for inflammation in the imiquimod-induced psoriasis mouse model of inflammation (225).

B. TRPA1

The ankyrin family possesses multiple copies of the ankyrin motif residing in the NH₂-terminal domain and is required for their assembly and gating of the channel (50, 51, 188, 221). TRPA1 was found initially to have a role in sensing pain and cold and participating in inflammation, and was activated by a range of noxious or pungent substances, including mustard oil, garlic oil, allyl isothiocyanate, and pollutants as well as ∆(9)-tetrahydrocannabinol (THC) and metabolites of acetaminophen (135, 210, 268). TRPA1 has emerged as the channel most tightly associated with the transduction of acute and chronic itch, although that role has now been questioned (234). TRPA1 is expressed by a subpopulation of small-diameter C-fiber polymodal afferent in the dorsal root, trigeminal, and nodose ganglia marked by the coexpression of the NGF receptor TrkA and TRPV1 (135, 268). TRPA1+ fibers project to peripheral tissues including the skin and the viscera. Their depolarization prompts the release of SP, neurokinins, and CGRP which promote extravasation, vasodilation, and neurogenic inflammation as well as hypersensitization to algesic stimuli. TRPA1 has been suggested to function as a gatekeeper for the neuroinflammatory release of neuropeptides. TRPA1 participates in non-histaminergic acute scratching in mice from CQ and the endogenous enkephalin BAM8–22 (168, 302, 303). CQ directly activates MrgprA3 while BAM8–22 activates MrgprC11, each of which couples to TRPA1 (FIGURE 5). With respect to dry skin induced by acetone-ether-water (AEW), and considered by some to recapitulate dry skin in humans and thus be a model of chronic itch, TRPA1 was necessary as TRPA1 KO mice genes scratched less as compared with their wild-type littermates and phenotypic changes were less impressive (303). TRPA1 was similarly necessary for the full manifestation of the effects of oxazolone and urushiol allergic contact dermatitis in mice (168). However, the requirement for TRPA1 in these processes has been questioned (234).

C. TRPV4

TRPV4 is expressed in keratinocytes, macrophages, and nociceptive neurons (145, 178). It has been considered a polymodal sensory modulator for thermal and chemical stimuli as well as osmotic and mechanical sensory inputs. It is also implicated in serotonin-induced itch (7). Serotonin levels have been associated with different pruritic diseases (141, 258), and there are many GPCRs in the serotonin receptor family. In TRPV4 KO mice, scratching behavior is attenuated in response to 5-HT, but not CQ, SLIGRL, or histamine (7). Single-cell calcium imaging was employed to highlight serotonin activation of HTR2A and coupling to TRPV4. The signals were extinguished in TRPV4 KO DRG neurons but remained unaltered in DRG neurons isolated from TRPV1 or TRPA1 KO mice. A link to the HTR7 receptor and TRPA1 has also been reported (190). By sensory neuron gene expression studies across genetically distinct mouse strains, the authors connected expression levels of HTR7 with non-histaminergic itch in the DBA mouse line. HTR7 was labeled in sensory neurons and colocalized with TRPA1 in hairy skin. Both receptors were required to induce neuronal responses by the HTR7 agonist LP-44. HTR7 was also implicated in persistent itch induced by topical MC903. Integrating both serotonin receptors into setting of allergic and nonallergic chronic itch, it has reported that lineage specific ablation of TRPV4 in macrophages and keratinocytes reduces itch in the AEW-dry skin or SADBE-induced ACD model.

D. TRPM8

The normal function of TRPM8 is to sense cold. In contrast to other TRP channels, TRPM8 is activated by menthol or icilin and inhibits or masks the sensation of itch due to the cooling sensation (275). Menthol-containing emollients provide an approach for temporary relief for itch. Skin cooling attenuates spinal neuron responses to histamine injections (131a). Deficiency of TRPM8 in sensory neurons results in attenuated responses to its activators (59), whereas more intense cold, <10°C, was still detected, most likely by other cold-sensitive channels. TRMP8 agonists have now been evaluated in clinical studies. A pilot study tested two TRPM8 agonists, (1R,2S,5R)-N-[2-(2-pyridinyl)ethyl]-2-ispropyl-5-methylcyclohexancarboxamide and menthoxypropanediol, in a double-blind randomized study with dry skin chronic itch patients (262), with promising reduction of itch. In a study of pedal pruritus in atopic dogs, the TRPM8 agonist 1-diisopropylphosphorylheptane was evaluated but did not improve pruritus (276). The effectiveness of targeting TRPM8 remains to be clarified as do the mechanisms that connect TRPM8+ neurons to spinal itch circuitry.

E. TMEM16a Versus TRP Channels in Itch

The concept of TRP channels as major signal transducers for itch has been challenged, as noted at the beginning of this section. Undem’s group used preparations that included skin with intact DRGs and spinal nerves and demonstrated that the calcium-activated chloride channel TMEM16a, rather than TRPs, are activated following stimulation of histamine or CQ (234). It was suggested that technical differences between the neurophysiology methods and the site of nerve recordings could explain discrepancies with prior findings. How this fascinating study impacts the reported supportive role of TRPs in inflammation will be of interest to the field.

XIII. VOLTAGE-GATED SODIUM CHANNELS

The contribution of voltage-gated sodium (Nav) channels to itch is garnering attention (160, 240). Nav channels play a central role in the transmission of sensory information. Peripheral sensory neurons express Nav1.7, Nav1.8, and Nav1.9 (154). Conventionally associated with transmitting pain signals, it is increasingly recognized that Nav channels contribute to itch (FIGURE 5). Loss-of-function mutations in SCN9A, which encodes Nav1.7, results in a congenital deficiency to pain, although itch has not been addressed (64, 198). Gain-of-function mutations cause painful neuropathies (111, 118). Erythromelalgia is an autosomal dominant condition caused by a gain-of-function alteration in Nav1.7 (65) that leads to burning pain and erythema in the extremities (293a) as well as a small fiber neuropathy (82). Gain-of-function changes in Nav1.8 and Nav1.9 also result in small fiber neuropathies (82) and familial episodic pain (321). Impairment of Nav channels impacts depolarization, inactivation, or recovery of neuronal firing and the net hyperexcitability of DRG neurons (280). Nav antagonists are in preclinical and clinical studies (280). With respect to Nav1.9, a woman with a gain-of-function mutation expressed in a heterozygous fashion presented with severe itch and concomitant partial loss of pain sensation (240). To better understand the impact on itch, Nav1.9 and the homologous mutation was studied in mice. Expression was primarily detected in unmyelinated, nonpeptidergic small-diameter DRG neurons, which expressed MrgprA3 or MrgprC11, a subpopulation linked to non-histaminergic itch. Loss of Nav1.9 led to reduction in histamine, BAM8–22, and CQ elicited scratching. Peak neuronal responses for histamine and CQ were similar between KO and wild-type mice, but the percentage of responding neurons was reduced. Like the human Nav1.9 mutation, heterozygous Nav1.9L799p/wt mice scratched more frequently than their wild-type littermates, thought to be due to deactivation or hyperpolarization changes in the channel resulting in excess Na⁺ influx (162). In line with these findings, a small population of MrgprA3+ neurons from Nav1.9L799p/wt mice was found to be hyperexcitable and may contribute to spontaneous itch activity (240). With respect to Nav1.7, a monoclonal antibody directed against the paddle region of the channel showed efficacy in mouse models of inflammatory and neuropathic pain as well as models for acute and chronic itch (160). Intrathecal injection suppressed itch induced by peripheral administration of compound 48/80, CQ, and GRP while systemic and intrathecal delivery suppressed chronic itch in the dinitrofluorobenzene (DNFB) hapten model of contact sensitivity. Like other Nav channels, Nav1.7 contributes to tetrodotoxin-sensitive sodium channel-mediated excitatory synaptic transmission. The primary afferents expressing Nav channels synapse with lamina IIo neurons and connect to lamina I neurons as part of the circuitry involved in chronic itch and pain.

XIV. CYTOKINES AND ITCH

The immune system, nervous system, and epithelia have evolved to balance homeostasis while protecting against infections, injuries, and threats. The complex interactions between these systems are multifactorial and multidirectional. Danger signals can result in dysregulated responses that are detrimental, with pathophysiological consequences. Direct communication between cells and mediators was implied by the close vicinity of Langerhans cells and mast cells to sensory neurons (21, 115, 265). The simplified version of neuroimmune interaction has been modulated and has gained multiple layers of complexity in which neurons participate actively in the recruitment of immune cells. It is likely that each cell type in skin can communicate directly or indirectly. Since Rothman’s review, the view of a neuroimmune interplay has been widened by understanding the cross-communication between the nervous system and the immune system via small molecules and cytokines. These messenger molecules are released by immune cells and activate receptors on sensory neurons. Key cytokines involved in itch and nociception include IL-2, IL-4, IL-13, and IL-31. While many cytokines are involved in inflammation, none has been implicated directly in pruritus, and they are not discussed further. A variety of small molecules, including histamine and serotonin, are released by controlled and differential granule release from mast cells and interact with neuronal receptors while sensory nerves release the neuropeptides SP and CGRP to activate immune and non-neuronal cells. We focus on the cytokines that interact with cognate receptors on neurons to drive itch and inflammation.

A. IL-31

IL-31 is the cytokine most directly implicated in mediating itch (FIGURE 5). Blocking antibodies to IL-31 or its receptor rapidly decrease itch in mice, canines, and humans (195, 237). The anti-inflammatory effect of such blockade is in flux (200). IL-31 was discovered in 2004. It is a four-helix bundle cytokine with minimal homology to IL-6 but classified as a member of the IL-6 family (109). IL-31 was found initially to be produced by activated CD4+ T lymphocytes but is also made by mast cells, macrophages, basophils, eosinophils, keratinocytes, and dendritic cells (69, 125, 199, 235). Transgenic mice that overexpress IL-31, driven by the Lck promoter in lymphocytes, developed eczematous skin lesions and immune-triggered hair loss. The promoter region includes binding sites for several transcription factors, including NFAT1, STAT6, JunB, AP-1, and NFκB which likely enhance IL-31 expression in TH2 lymphocytes and mast cells (109, 320). IL-31 is not considered a classic TH2 cytokine as the latter are associated with the induction of allergy and eosinophilia, whereas IL-31 tends to be downstream or separate from TH2 processes (109, 136, 207). IL-31 signals through a receptor heterodimer complex consisting of IL-31RΑ and oncostatin-M receptor β (OSMRβ). Both subunits are detected in the skin, brain, lung, trachea, skeletal muscle, testis and ovary, prostate, placenta, spleen, thymus, bone marrow, and blood leukocytes with OSMRβ being the more constitutively expressed subunit (69, 70). Our research has detailed the cellular as well as the molecular mechanism by which IL-31 exerts it effects. IL-31RA has been localized on sensory neurons in mice and humans (32, 55). We analyzed the receptor distribution on sensory neurons and found expression in 3–5% of the total neuron population. MrgprA3, another marker of itch sensory neuron, is expressed in a similar percentage of DRGs. Neurons that respond to histamine, SLIGKV/SLIGRL, and CQ partially overlap with the IL-31RA+ neuron population. There was also overlap of TRPV1 and TRPA1 expression. IL-31 directly induces itch when injected into the murine skin, although questions remain as to whether IL-31 may have nociceptive effects. The expression and distribution of IL-31 receptors on sensory neurons indicate a direct messenger function in sensory aspects of inflammation including itch and possibly pain.

1. IL-31 receptors and downstream signaling

In contrast to other IL-6 cytokines, IL-31 does not interact with the gp130 receptor subunit but binds with low affinity first to IL-31Rα. Following receptor engagement, a plethora of downstream signaling pathways are activated, including Janus activated kinase (JAK)-Stat, Ras/Erk, and the PI3K/AKT (60). Activated TH2 cells are the primary but not exclusive source of IL-31 secretion. Typical AD triggers including Staphylococcus aureus and OVA induce the secretion of IL-31 from TH2 cells (55) while human beta defensin (hBDN) and the antimicrobial peptide cathelicidin-37 (LL-37) induce release from mast cells (199), now recognized to be mediated via Mrgprs. IL-31 is thus ideally suited for the neuroimmune interplay in innate immunity and cutaneous inflammation. Human basophils, which are effectors of allergy, produce IL-31 and are also activated by IL-31 (217). Activation of IL-31RA on basophils provokes the release of IL-4 and IL-13, key cytokine drivers of AD, suggesting a multicellular role for IL-31 to coordinate inflammation and drive itch. Additional insights into the downstream effects of IL-31 have been gained by in vitro studies in which recombinant cytokine is injected into mouse skin. IL-31 influences differentiation and expression of the structural protein filaggrin (61). In AD, impairment of filaggrin results in skin barrier dysfunction (76, 77). These data imply that IL-31 is involved in structural processes in addition to inflammatory ones. The injection of IL-31 daily for 2 wk resulted in keratinocyte proliferation, skin thickening, associated impaired barrier function manifest by increased transepidermal water loss (TEWL), which is a feature of AD, and increased transcription of TRPV2, a signal transducer channel localized on sensory neurons, but not other TRP channels.

2. IL-31 and itch

IL-31 is one of the first cytokines that has been described to directly act on sensory neurons to trigger itch. The initial itch studies were done with osmotic pumps that release IL-31 into wild-type and c-Kit mast cell-deficient mice (69) and transgenic mice overexpressed IL-31. The overexpression resulted in these mice in acanthosis, parakeratosis, and hyperplasia, and an intense immune phenotype infiltrate was present in all and resembled the human disease. Increased numbers of mast cells were noted in the skin of IL-31 transgenic animals, a finding also in human AD. The most direct explanation that accounts for these observations is that while mast cells are not necessary for aspects of the inflammation associated with IL-31, by interacting with its receptor on various cell types, including sensory neurons, this cytokine influences feedback and feed-forward networks between the nervous and immune systems and keratinocytes. In addition, the differential granule release from mast cells, together with pruritogens, exacerbates the itch-scratch cycle. Single injections of IL-31 caused dose-dependent itch in mice that was reduced in mice lacking TRPV1 and TRPA1. These data are consistent with IL-31 acting as a direct pruritogen (55).

It has been suggested that IL-31 elicits acute itch, whereas IL-4 and IL-13 are drivers of some cases of chronic itch. However, due to a constant inflammatory release of IL-31, this cytokine is suitable to target for both acute and chronic itch (FIGURE 6).

FIGURE 6.

Acute and chronic itch in humans. In acute itch, histaminergic itch is induced by histamine via activation of its receptors on sensory neurons. Non-histaminergic itch can be induced by a plethora of molecules, of which some are identified. One such molecule is interleukin (IL)-31 which is released from immune cells including CD4⁺ TH2 cells and mast cells. Upon release, IL-31 activates a heterodimeric receptor complex consisting of IL-31Rα and oncostatin M receptor β (OSMRβ) on sensory neurons to elicit itch directly. It is possible that this process induces the release of neuropeptides such as substance P (SP) from sensory neurons. SP activates MRGPRX2 on mast cells and may be induced on sensory neurons. Itch triggers degranulation of mast cells and the differential and controlled liberation of various neuropeptides, proteases, and cytokines. A continuous inflammatory milieu may result in chronic itch, which is debilitating in many diseases. IL-31 has been linked to chronic itch skin conditions. During inflammation, immune cells release other cytokines including IL-4 and IL-13 which signal via the heterodimer IL-4Rα and IL-13Rα1 and are reported to function as neuronal enhancers for different itch pathways. Bile acids have recently been reported to activate MRGPRX4 and may thus contribute to cholestatic itch.

The IL-31 responsive neuron population comprises 3–5% of DRG neurons, reflecting transmission via a specific population. The IL-31 population is comprised in the NP3– inflammatory itch sensing neurons which are also activated by histamine (FIGURE 2). The highest overlap of IL-31 responsive neurons was found to be with TRPA1+, consistent with the reduced itch in responses TRPA1 KO mice. IL-31 mediated itch signals through an ERK1/2-dependent mechanism rather than the JAK-STAT pathway. Since ablation of JAK1 in sensory neurons inhibited IL-31 itch (203), the details of signaling in acute versus chronic itch associated with IL-4 and IL-13 require additional clarity. Targeting IL-31 in AD results in significant reduction of itch with no improvement of the inflammation, which raises questions about the connection between itch and inflammation (137). A possible explanation for the missing efficacy in the inflammatory axis might be that a complete depletion of IL-31 is required to block itch and inflammation simultaneously as seen with phototoxic ablation of sensory itch (200) Another neuroimmune-regulatory role of IL-31 is the neuropoietic stimulation on sensory neurons similar to NGF (83) which is consistent with increased nerve fiber density in lesional skin of IL-31 transgenic mice and human AD. Transcriptional profiling of mouse DRGs revealed overlap but also differences between stimulation with IL-31 as compared with NGF. These data should be interpreted with care as NGF is added to the cultured DRG neurons that are stimulated with IL-31. IL-31 stimulation resulted in enhanced expression of neuropoietic genes as compared with NGF (83). Intriguing findings were that the neuropoietic functions exerted by IL-31 were linked to STAT-3 and were TRPV1 independent, which contrasts with the ERK1/2 phosphorylation and presence of TRPV1 associated with itch. IL-31, like other representatives of the IL-6 cytokine family, employs multiple signaling pathways during inflammation, neuropoiesis, and itch.

3. IL-31 in human pruritic diseases

IL-31 levels correlate with the severity of pruritus in AD, prurigo nodularis, and CTCL but not in nonpruritic psoriasis patient samples, consistent with transcriptional upregulation in pruritic AD (53, 194, 256). Expression was induced in human peripheral blood mononuclear cells by staphylococcal superantigen, which is associated with atopy, but not influenced by either influenza or HSV. IL-31 is produced by CTCL T cells, leading to increased IL-31 levels in serum and peripheral blood mononuclear cells, and is correlated with pruritus severity (256). Localization studies with immunofluorescence labeling for IL-31, OSMRβ, and IL-31RA revealed that moderately and severely CTCL skin had higher expression of these components and correlated with visual analogue scale ratings for pruritus. It is possible that IL-31 is the mediator of itch in CTCL patients. The relevance of IL-31 in chronic spontaneous urticaria has been examined (219). IL-31 levels were higher than in negative controls but less than in patients with AD. It has been suggested that in addition to mast cells, IL-31 induces basophil migration and release of IL-4 and IL-13 to contribute to the pathophysiology (218). Overall, the relationship of levels of IL-31, OSMRβ, and IL-31RA in pruritic states, as determined by immunological and molecular approaches, is an active area of investigation. Current observations emphasize the pleiotropic effects of IL-31 and its receptors in inflammatory diseases associated with pruritus.

B. IL-4 and IL-13

The TH2 cytokines IL-4 and IL-13 are major drivers in the proinflammatory orchestration in AD. Both cytokines have been detected at elevated levels in the skin and serum of AD patients. IL-4 is apparent at the onset of AD, whereas IL-13 is associated with chronicity (FIGURE 6). A role for IL-4 and IL-13 in chronic itch, not just AD, has been reported (203). These cytokines activate cultured DRG neurons in mice but do not elicit itch behavior. Rather, IL-4 increases the responsiveness of sensory neurons to other itch mediators, whereas IL-31 activates such neurons and evokes scratching behavior (55). It remains unclear whether IL-13 exerts similar sensitization effects to IL-4. IL-4Rα appears fundamental to chronic itch as conditional knockout mice, lacking IL-4Rα gene in nociceptive neurons, do not develop either chronic itch or histopathological features in the MC903 model of AD-like model. Neither IL-5 nor its receptor had a role in this model. These findings support the cellular and molecular basis for the therapeutic effects of dupilumab, an antibody directed towards the IL-4Rα subunit and the first biologic to treat moderate to severe AD (FIGURES 6 and 7).

FIGURE 7.

Peripheral and central targets of itch. Peripheral itch targets are receptors and mediators of chronic itch including the interleukin (IL)-31 pathway. Targeting IL-4 and IL-13 with dupilumab has been successful in atopic dermatitis. Antagonists of MRGPRX2 might be beneficial in the treatment of inflammation, itch, and urticaria. Antagonism of MRGPX4 has the potential to benefit the treatment of cholestatic itch when driven by bile acids. In neuropathic pain, targeting peripheral GABA-R has proven analgesic properties, which could be translated to itch. Central targets of itch act in the central nervous system and include the use of κ-opioid receptor (KOR) agonists and μ-opioid receptor (MOR) antagonists. Targeting spinal serotonin-pathways via 5-hydroxytryptamine (5-HT) 3 and 7 receptors may achieve clinical success in chronic itch management. Antagonists directed against the NK1R most likely work spinally. Targeting itch-specific circuits such as gastrin releasing peptide receptor (GRPR) might represent a therapeutic avenue. Increasing inhibitory tone via replenishing GABA reduces acute and chronic itch in mice. Gabapentin and pregabalin, widely used to treat neuropathic pain, are prescribed in the clinic to treat chronic itch patients. DRG, dorsal root ganglion; GRP, gastrin releasing peptide; SP, substance P.

XV. OTHER CYTOKINES AND THEIR RELATIONSHIP TO ITCH

IL-33, IL-25, and thymic stromal lymphopoietin (TSLP) are cytokines that function upstream of IL-4, IL-13, and IL-31. They are released from keratinocytes as well as from epithelial cells from other tissues and contribute to the initiation of type 2 inflammation. IL-33 and TSLP evoke scratching behavior in mice and activate mouse sensory neurons (169, 304). None of these cytokines has been shown to evoke itch directly in humans. However, their pleiotropic effects and complex signaling are consistent with at least an indirect impact on itch.

The proinflammatory type 1 cytokine tumor necrosis factor (TNF)-α and its receptor TNFR1 are implicated in acute and chronic itch via peripheral and central sensitization mechanisms in mice (184). Known for its role in chronic pain, cancer, and neuropathy, TNF-α excites and sensitizes primary afferents (130, 233), eventually via TRPV1 upregulation. To evaluate the role of TNF-α in acute itch, the authors employed compound 48/80 and CQ and demonstrated that TNFR1 KO mice showed fewer scratching responses from these itch-inducing agents. Additionally, the dry skin model revealed that TNF-α expression is upregulated in the skin; DRG and the cognate receptor expression is elevated in spinal cord tissue. The findings in mice may correspond to observations in human itch. Thus thalidomide has been effective in many dermatologic disorders, including inflammatory and noninflammatory itches (249). These broad effects likely follow from the impact that thalidomide has on protein ubiquitination (85) and its capacity to decrease TNF-α production (189). Similarly, by regulating neuronal plasticity in pain, TNF-α increases N-methyl-d-aspartate currents in lamina II neurons in the spinal cord and sensitizes spinal cord neurons in chronic itch (180).

A different class of cytokines, namely, chemokines, are regulators of multiple cellular processes (56). Chemokines are expressed in different tissue types, including the central nervous system, and have chemotactic function with implications in neuroinflammation and chronic pain (89–91). CXCL10 signals via CXCR3, a GPCR, both of which have been associated with maintenance of chronic pain post peripheral nerve and chronic constriction injury (57). CXCR3 was recently reported to mediate itch induced by an allergic contact dermatitis mouse model (216). Neuronal responses to CXCL10 were intensified in cultured DRG neurons dissected from mice treated with SADBE as compared with vehicle-treated ones. In a recent study, CXCR3 and its ligand CXCL10 were reported to modulate chronic itch and alloknesis phenomenon induced by the dry skin model and 48/80 injections (131). CXCR3 KO mice exhibited less signs of alloknesis with lower signs of astrocyte activation.

A. JAK Inhibitors

Oclacitinib is a JAK inhibitor (JAKi) approved by the United States Food and Drug Administration in 2013 for the treatment of itch and allergic diseases in canines. The drug was launched in early 2014. The supply was exhausted within 2 mo. This situation exemplified the pent-up demand for effective therapeutics for itch. It raised the profile of the unmet need for both dogs and humans (167). JAKi are emerging as treatments for inflammation in general, including allergic and autoimmune disorders. JAKs were known to be expressed in sensory neurons, which led to the evaluation of a JAKi as a potential therapy for chronic itch (203). The canonical cytokines IL-4 and IL-13 signal via JAKs as do upstream cytokines IL-33 and TSLP. IL-31 transmits itch via MAPK but utilizes JAK signaling to stimulate growth of small fiber neurons. JAK1-JAK3 are also activated by IL-2, IL-7, IL-9, IL-15, and IL-21, whereas IL-6, the prototypic proinflammatory cytokine, signals through JAK1, JAK2, and TYK2 (248).

In the clinic, topical application of tofacitinib, a JAK1/3i, led to itch improvement starting at day 2. Baricitinib, an oral JAK1/2 selective inhibitor, reduced chronic itch and skin inflammation in a phase 2 clinical trial with moderate to severe AD. Genetic depletion of JAK1 in sensory neurons resulted in diminished itch in mice treated with the inflammation inducing vitamin D analogue MC903 (203). Consistent with this finding, a JAK1 gain-of-function point mutation in mice resulted in a pruritic dermatitis phenotype with scratching and stepwise progression into immunological abnormalities (312). The application of petrolatum to the ears masked the damage to the barrier and delayed the immune dysregulation of elevated TH2 cytokines and increased expression of serine proteases, particularly kalikrein-6 and marapsin which are skin barrier regulators. A JAK1 gain-of-function mutation in humans caused severe pruritic dermatitis (67). The JAK inhibitor ruxolitinib led to improvement. Collectively, these data emphasize the role of JAKs in inflammatory conditions and chronic itch. There is a growing interest in developing JAKi as topical treatments for mild-moderate AD to circumvent systemic adverse side effects. Tofacitinib, the first commercially available JAKi, was approved originally for the treatment of rheumatoid arthritis. The ointment formulation of tofacitinib tested in AD patients successfully improved the Eczema Area and Severity Index (EASI) scores to 80% after 4 wk of treatment while showing antipruritic action as early as 24–48 h (36). Similarly, JTE-052 provided itch relief in moderate-to-severe AD patients (192). The relative contributions of neuronal versus immune JAKs in pruritus are being clarified.

XVI. MODELS AND TECHNIQUES

Histamine-induced itch is the most studied and well-understood acute itch despite the findings that most chronic itch forms are of non-histaminergic nature. Preclinical models of non-histaminergic itch are the chemically as well as mechanically induced itch forms broadened with models of dry skin, allergic, atopic, cholestatic, and genetic. Moreover, there has been growing effort to assess itch and the associated dysesthesia in humans with surrogate models. The realization that scratching in animals occurs after algesic and pruritic agents has led to the search for different technical assessments of itch.

A. Measuring Itch in Mice

Behavioral assessment of itch in mice was initially studied by few researchers and was overshadowed by pain research. Behavioral testing to measure itch was undertaken in different ways. Until 2008, pruritogens were assayed by injection into the nape followed by recording and analysis of scratching by bouts over time (149). This approach could not necessarily distinguish pruritic from nociceptive behaviors. An important advance was introduced by Shimada and LaMotte (250). Their cheek model discriminates between itch- and pain-associated behavior (250). Algogens such as capsaicin trigger wiping of the injected ipsilateral cheek with the forelimbs, whereas pruritogens such as histamine evoke scratching via the hindlimbs. This approach is now the standard. Akiyama et al. (4) reported that when mice are injected with algogens such as capsaicin, the mice behavioral responses of wiping were reduced by morphine. In contrast, histamine-evoked scratching was diminished by naltrexone, a μ-opioid antagonist. The authors additionally tested whether spicules, native with cowhage or inactivated, when loaded with histamine or capsaicin would result in wiping or scratching. Cowhage-evoked itch is accompanied by nociceptive sensations such as a stinging, burning sensation (152). Cowhage-evoked scratching peaked between 10 and 15 min while wiping behavior remained through the recording time of 30 min. Even inactivated spicules loaded with histamine elicited more scratch and wiping behavior as compared with the cowhage-loaded ones. Similarly, capsaicin-induced behavior post spicule application was more enhanced than cowhage-elicited responses with more prominent wiping behavior suggesting that the behavioral assessment is more objective than the classical assay. Of note, while cowhage spicules are excellent tools for evaluating itch in humans, mice rapidly remove the spicules, and they are not generally used in mice. By means of the cheek model, the authors uncovered that formalin-induced behavior is more reflective of itch rather than pain or in addition to dose-dependent nociception. Most of the other classical algogens, such as mustard oil and bradykinin, evoked predominantly wiping behavior. LaMotte et al. (153) also introduced another model that provides behavioral differentiation between itch and pain, namely, the calf assay for itch. Processing of somatosensory information for the calf model happens in the lumbar spinal cord region which has been more extensively studied as compared with the trigeminal processing when signals are conveyed from the cheek. An advantage is the better characterized neuronal population in the lumbar spinal cord responding to chemical and mechanical stimuli. Injections of capsaicin evoke licking behavior, whereas histamine injections result in biting responses like paw injections, a behavioral assessment widely used in pain research (316). Profiling neurons from different dermatomes may contribute to more fully understanding similarities and differences between the inputs and the processing of itch and pain.

B. Genetic Models

Another axis through which to study itch is via genetically modified strains of mice in which the mediation, transmission, or coding of itch has been modified (271, 272). Genetic manipulation of the molecules involved in the transmission of itch either heightens or reduces itch or protects from itch. A variety of models have been introduced in preceding sections. Examples include overexpression of the cytokine IL-31, which results in excessive scratching and self-inflicted lesions and transcription factor bhlhb5 KO animals which have a severe neuropathic itch phenotype. One of the early mouse models for itch is the NC/Nga mouse (129, 251). Spontaneous signs and symptoms include itching, erythema, scaling, dryness, and alopecia at 8 wk of age. However, the NC/Nga mice pathophysiology is not exactly predictable and hence bears a poor rate of reproducibility. Models that have been considered to be proxies for AD and chronic itch include the overexpression of cytokines IL-4, and TSLP, separate from IL-31 (314). A mouse model of CTCL was developed by inoculation of human Myla cells in which levels of microRNA (miRNA) were increased in mouse serum (104). This study uncovered an unusual role for miRNA in chronic itch triggered by the pathology of CTCL. Patients with CTCL suffer from chronic itch for which no treatment paradigms have been established. This model provides a molecular basis to understand the origin of chronic itch in CTCL which was linked to levels of circulating miRNA that were also induced in human patients (220).

C. Itch with Sensitization Models

Most commonly used models for itch are sensitization of the skin by application of either haptens or allergens such as ozaxolone, vitamin D analogues, or OVA (179, 289). In correlation with human AD, the oxazolone model presents an epidermal serine protease signature and a concomitant decrease of antimicrobial peptides such as the β-defensin-3. This model, as well as that using the hapten trinitrochlorobenzene applied to the skin of the hairless mouse strain, shifts the TH1 responses towards TH2 as seen in AD (179). Vitamin D [1α,25-(OH)₂D₃] or MC903 can be used to induce AD-like skin inflammation (165) MC903 sensitization was linked to the overexpression of TSLP. Epicutaneous sensitization with OVA on shaved skin (102) of different strains of mice induces an eruption that mimics the features of human AD (261). Erotic or dry skin itch can be achieved through repeated application of an AEW solution at any skin location with any mouse strain (201). For psoriatic itch, the topical application of imiquimod, a topical anti-cancer drug, results in spontaneous scratching bouts accompanied by alloknesis (239).

D. Human Itch and Surrogate Models

In humans, itch is often accompanied by nociceptive sensations of burning, pricking, and stinging. Psychophysics is used to measure these sensations using the visual analogue scale while the numerical rating scale is used in the setting of clinical trials to measure itch. Latency, peak, and duration can be incorporated into the itch measurements (254, 255). Cowhage is an ideal tool for evaluating itch in humans. Native cowhage elicits itch, burning, pricking, and stinging. In addition, spicules can be readily inactivated in an autoclave and then ‟reconstitutedˮ with other test agents. The spicules function as microneedles and do not generate nociceptive responses when inactivated. Like murine itch models, pruritogens such as histamine, BAM8-22, and serotonin can be tested in humans via injection, but alternatively by reconstituting spicules. Histaminergic itch is conveyed by the mechanoinsensitive C-fiber population, whereas non-histaminergic itch such as cowhage-elicited itch employs polymodal C-fibers (193). Histamine and mast cell degranulators cause the typical wheal, a manifestation of acute protein extravasation, and flare reaction, caused by neuropeptides.

XVII. CONTAGIOUS ITCH

Contagious or social itch is triggered by visual or other cues in the absence of a conventional pruritogen. In contrast to itch from exposure and infection from a contagious organism, such as the scabies mite, contagious itch is triggered by visual exposure to somebody else scratching or the auditory cues, as in a lecture about pruritic dermatological skin conditions (112). Contagious itch via audiovisual transmission is experienced by humans and non-human primates and has been reported in rodents (112, 247), although the significance of this itch is not clear. The initial description was of a human audience exposed to presentations with pictures of insects, scratch marks, and allergic reactions followed by a presentation with soothing cues with pictures of babies. A significant number of scratching movements was measured during the “itchy” presentation, with increased itch perception reported by the audience, which contrasted with the relaxing portion of the presentation. Mirror neurons were identified initially within the ventral premotor area when monkeys conducted a certain movement or in observation of this movement (122).

An effort has been made to associate the cellular and molecular mechanism of contagious itch with the GRP-GRPR axis (315). Here, naive mice were reported to exhibit contagious itch when observing videos of BRAFNaV1.8 mice, which exhibit exaggerated scratching behavior (324), in the absence of olfactory or auditory stimulation. Dependence on SCN was suggested as this area receives direct visual input; contagious itch was absent in GRP and GRPR mutant mice and mice in which GRPR was conditionally depleted in the SCN, yet was restored by delivery of GRP or by optogenetic stimulation (315). The importance of these findings has been questioned, given that contagious itch was not observed in wild-type mice injected with histamine (166).

XVIII. CLINICAL ITCHES

Itch can arise anywhere along the pathways between the skin and brain. This anatomy encouraged itch to be classified as pruritoceptive when arising in the skin, neuropathic when associated with neuronal pathology, and psychogenic when associated with a psychological condition (291). This approach frames a brief discussion, for which multi-authored reviews are available. A recent report that suggests a descriptive classification of chronic itch (264) has the benefit of additional detail incorporating skin findings together with underlying systemic conditions to assist in establishing itch as a disease while potentially being less informative with respect to mechanism and pathophysiology.

Most itches are pruritoceptive. They arise either from exogenous environmental substances or from the persistent presence of endogenous mediators that are produced in or make their way to the skin. Itches associated with insect bites or poison ivy are examples of acute itches, resulting from allergic responses in each case. Chronic itches are often associated with inflammatory skin disease, such as AD. The co-dependency of these itches and connection with neurogenic inflammation are exemplified by the observation that neural lesions that impact the areas of inflammatory skin disease eliminate the sensation of itch while resulting in resolution of the inflammation (26). Chronic itch can also be a part of noninflammatory systemic diseases, including cholestasis, chronic kidney disease, or even pregnancy. These should be viewed as chronic acute itches as the itch rapidly resolves in most instances shortly after organ transplant or birth. The mediators driving these itches are being determined. With respect to cholestatic itch, bile acids interact with MRGPRX4, while lysophosphatidic acid (LPA) produced by autotaxin (148) interacts with LPA receptors on sensory neurons. With respect to the itch of chronic kidney disease, we have hints that there may be parallels with that of cholestatic itch. As an extension, it is tempting to speculate that the severe itches of primary biliary cirrhosis and polycythemia vera may relate to heme metabolism.

There is no shortage of other challenging itches. The advent of targeted anticancer immunotherapies is frequently associated with therapy-limiting pruritus, for which the mechanism is unknown and no animal models are available (306). Neuropathic and psychogenic itches are chronic and frequently intense. Examples of the former include post-herpetic itch, notalgia paresthetica, and brachioradial pruritus which can be associated with nerve compression. A number of severe itches, including genital itch, which likely include degrees of central sensitization, are not readily categorized. Ocular itch, thought to be of an allergic nature, calls for a topical approach for treatment. Itch from parasites, particularly helminths, including nematodes of the genus Strongyloides, causes misery worldwide, regardless of economic background, and often presents as a diagnostic dilemma and marked delay in diagnosis (274). Again, the mediators are not known.

Not only is the treatment of chronic itch a challenge, but also careful evaluation leads to the determination of an underlying etiology in a fraction of patients (306). When a cause cannot be determined, these patients are considered to have chronic pruritus of unexplained origin or CPUO (143), an acknowledgment of the classic description of fever of unexplained origin or FUO by Petersdorf and Beeson (212). While patients with chronic itch frequently note that they would prefer to commit suicide than continue to suffer (101), documented reports of suicide are rare.

We are often asked if itch differs between the sexes. This question, as well as potential differences in itch that could be associated with mental health, economic differences, or race, which assumes that races can be distinguished, has been considered (66). There are no objective data to support any such differences based on physiology or pathophysiology. There are data to suggest that the processing of itch and pain-related sensations can differ between men and women exposed to native cowhage spicules or cowhage spicules prepared with histamine or capsaicin (107). Differences in the presentation of pruritus based on race and sex have been suggested (133). No difference in the prevalence of itch between sexes has been identified when each has the same disease, in this case, chronic kidney disease (47, 116). In our itch clinic, we sometimes have the impression that on average, woman tolerate severe itch better than men do, but we recognize the subjective nature of this comment. Interpretation of any data that suggest differences must take into account potential differences in interview style. For example, asking if a patient itches or has other nociceptive concerns versus whether a patient spontaneously reports itch can impact findings.

Identical questions have been raised with respect to pain with similar findings. In mice, it has been suggested that male mice utilize microglia in the spinal cord to mediate pain, whereas female mice use T cells, a finding that may be influenced by hormonal differences (260). It is thus possible that therapeutic approaches could be targeted to patient physiology.

XIX. TARGETING ITCH

Just as clinical itches can arise anywhere along the pathways between the skin and brain, interruption of a pathway has the potential to modulate itch (FIGURE 7).

Approaches to treating itch are represented in FIGURES 7 and 8. FIGURE 8 is an acknowledgment of the style of Dr. Seuss.

FIGURE 8.

Itch-scratch cycle. The many areas where therapeutic approaches can be directed are depicted. These areas range from the skin barrier, depicted as a funnel with associated immune cells together with afferent fibers that flow into the spinal circuitry and on to the brain.

The skin is viewed as a funnel topped with its barrier and filled with keratinocytes that support a web of sensory fibers. Side chambers topped by inflammatory cells feed into the funnel. The neck of the funnel consists of neuronal fibers that coalesce into the spinal cord. These axons continue to the brain where the signals are interpreted as itch followed by generation of the motor response of scratching which impacts the skin barrier to drive the itch-scratch cycle.

Therapy can thus be directed to protecting or repairing the barrier with topical approaches; blocking sensory neuronal activity or neuropeptide release, immune cell activity, or mediators; or modulating spinal or brain activation or inhibitory pathways. Approaches directed to each of these areas are in clinical use. The combination of fundamental research together with recognition of genetic polymorphisms that impact itch and inflammation may lead to targeted therapeutic approaches with limited off-target effects.

XX. FUTURE DIRECTIONS

Remarkable progress is being made in understanding the physiology of itch. At the same time, we are at the proverbial end of the beginning with respect to having a deep understanding of itch. Questions related to details of mechanisms remain at every level of the itch-scratch cycle, and addressing these questions will provide ongoing opportunities for investigators. The development of humanized mice has the potential to partially overcome the translational gap between preclinical and clinical findings. Why is there so much redundancy with respect to itch signaling pathways? Further to this question, with the recent implication of neutrophils in innervation and itch (297), essentially all cells that populate the skin can contribute to itch. Does itch relate to tickle and touch? Is there communication between free nerve endings, the glial network at the epidermal-dermal junction, Meissner corpuscles, Ruffini endings, and Merkel cells or are these structures independent one from the other? What are the relative and specific contributions of the various receptor and channels involved in itch, including the Piezo channels that are important in touch? It is likely that polymorphisms in receptors and channels, beyond that found already for Nav1.9, will be identified in genetic databases that lead to increased itch. The possibility of finding variants that protect from itch may be elusive as it involves proving a negative.

Imaging techniques are always advancing. The generation of mice in which neurons are functionally labeled with genetically encoded calcium indicators, together with the capacity of genetically labeling cell types in the skin, will allow for detailed visualization of neuroimmune interactions and potentially their bidirectional communication. Can imaging techniques be developed that allow for the visualization of C-fibers in humans in vivo? Will brain and spinal cord imaging in mice and humans eventually lead to resolution on a cellular level and contribute to understanding details of central itch processing and differences with pain?

With respect to itch therapy, as with pain, the placebo effect for drugs or vehicles can exceed 50%. The psychological and psychosomatic factors that contribute in a positive way to the placebo effect are thus important. The inverse, manifest as a negative expectation, or nocebo effect, can be detrimental to therapy (305). How might the learning processes associated with these effects be harnessed to enhance therapeutic benefit? Can we take advantage of inhibitory circuitry to modulate itch? What is the mechanism of action of natural products that may help itch, such as flavonoids? Why are current therapies that target the IL-31 pathway so effective at treating itch but not concurrent inflammation? Besides drugs, physical modalities, such as therapeutic approaches using lasers, already help to relieve itch associated with burn wounds and scars, although the mechanism of efficacy is uncertain. Is such monochromatic light targeting mast cells or something else? Heat and cold can impact neuronal function. Can they be applied to the skin or by point sources in the DRG to target itch? Can objective measures of itch, or other senses, be developed so as to improve upon psychophysical measures and the numerical and visual analog rating scales? Whether itch can be alleviated completely while maintaining other sensory modalities remains to be determined, but we are entering an age in which the suffering from itch may be replaced by the pleasure of scratching.

GRANTS

This work was supported by Pfizer Aspire Program Grant 233429, Dermira, Inc. Grant 231828, and National Institutes of Health Grant 1R21AR067399.

DISCLOSURES

F. Cevikbas holds equity in Dermira, Inc. E. Lerner is on the Scientific Advisory Board of Escient Pharmaceuticals.