Itchy channels and where to find them

Itch (pruritus) is an unpleasant sensory experience that triggers a desire to scratch. It is induced by activation of a subset of cutaneous C fibres in response to environmental irritants and some endogenous mediators (such as histamine). As a somatic sensation itch is closely related to pain. Indeed, itch‐mediated afferents are mostly classed as nociceptors; moreover, high concentrations of pruritogenic substances (e.g. histamine) produce pain, while some pain‐inducing agents (e.g. capsaicin) can produce itch when applied topically to the skin (reviewed in Lee et al. 2016). As with pain, acute physiological itch is a healthy reaction as it informs the organism of harmful conditions and helps to eliminate them (e.g. by removing parasites or potentially damaging agents from the skin). Yet, chronic itch developing in some pathologies (e.g. atopic dermatitis) can be an excruciating disease.

Recent years saw swift progress in the understanding of itch circuits and deciphering molecular mechanisms of itch signalling. Thus, histaminergic and non‐histaminergic itch have been identified as distinct phenomena and the receptors of several non‐histaminergic pruritogens have been discovered. In particular, members of Mas‐related G protein‐coupled receptors (Mrgprs) MrgprA3, MrgprC11 and MrgprD were identified as itch‐mediating receptors for chloroquine (Liu et al. 2009), pruritic peptide BAM8‐22 (Liu et al. 2009) and β‐alanine (Shinohara et al. 2004), respectively. The Mrgpr‐expressing primary afferents were, thus, identified as ‘itch’ afferents (Liu et al. 2009; Liu & Dong, 2015).

Intense current research is focused on the elucidation of how G protein‐coupled receptors (GPCRs) of itch generate receptor potential in the peripheral endings of itch afferents. The major emphasis thus far was on the members of the transient receptor potential (TRP) channel family TRPV1 and TRPA1 (see excellent review by Kittaka & Tominaga, 2017). Histamine receptors expressed in sensory neurons (H1 and H4) and some Mrgprs couple to the G_q/11 signalling cascade involving activation of phospholipase C (PLC) and hydrolysis of phosphatidylinositol 4,5‐bisphosphate (PIP₂) into diacylglycerol (DAG) and inositol trisphosphate (IP₃). The DAG‐mediated branch of this signalling cascade may involve activation of protein kinase C and phospholipase A₂, while the IP₃ branch results in release of Ca²⁺ from the IP₃‐sensitive intracellular stores and modulation of a plenitude of downstream Ca²⁺‐sensitive targets (Gamper & Rohacs, 2012). A number of these pathways were implicated in pruritogen‐induced activation of TRPV1 and TRPA1 with the assumption that these, in turn, produce depolarization of itch nerve endings sufficient to trigger action potentials (Kittaka & Tominaga, 2017). However, due to inherent difficulties with directly assessing and controlling the excitability of nerve endings, the majority of these studies used indirect approaches, such as electrophysiological and imaging experiments with the dissociated sensory neurons in culture in combination with genetic models and behavioural tests.

In the new study published in the current issue of The Journal of Physiology, Ru and colleagues developed an improved skin–nerve preparation in which electrophysiological recordings from intact itch afferents are combined with fast, reliable and repeatable subcutaneous application of drugs (Ru et al. 2017). In this preparation the dorsal side of the mouse skin is dissected out along with the blood vessels (subscapular artery and its branches) and the nerves (with their associated T7–T10 dorsal root ganglia). The presence of the subscapular artery made it possible to overcome the low diffusion of the externally applied drugs through the skin using intra‐arterial injections. The accuracy of the intra‐arterial drug delivery has been carefully verified with dye injections.

Using this preparation it became possible to directly test the effect of genetic deletion or pharmacological inhibition of TRPV1, TRPA1 and other candidate ion channels on the excitation of peripheral endings of itch afferents in response to pruritogen application to the skin. This powerful approach delivered a number of clear‐cut but rather unexpected results. Thus, the authors showed that histamine and non‐histaminergic agonists (chloroquine, BAM8‐22 and ovalbumin) excite largely the same subpopulation of TRPA1‐ and TRPV1‐expressing C fibres (∼25% of total cutaneous C fibres). This is at odds with the previous observation that selective silencing of trigeminal histaminergic fibres does not significantly affect non‐histaminergic itch and vice versa (Roberson et al. 2013). This discrepancy may reflect differences between spinal and trigeminal systems and requires further investigation.

But a real surprise came when the responses of itch C fibres from trpa1 ^−/− or trpa1 ^−/−/trpv1 ^−/− double‐knockout mice were investigated. Astoundingly, deletion of TRPA1, or TRPA1 together with TRPV1, had no effect on the itch afferent excitation by chloroquine, histamine or ovalbumin. Moreover, regarding the excitation of itch afferents by chloroquine, the following conditions were also without an effect: (i) the double deletion of TRPC3 and TRPC6; (ii) the pharmacological inhibition of TRPA1 on the trpc3 ^−/−/trpc6 ^−/− background; and (iii) the pharmacological inhibition of TRPC3/TRPC6 on the trpa1 ^−/−/trpv1 ^−/− background. Even a broad‐spectrum TRP channel blocker ruthenium red, which inhibits most of the sensory TRP channels, was without an effect. In these experiments the absence of TRPA1 and/or TRPV1 channel activity in the fibre has been confirmed with allyl isothiocyanate (AITC) and capsaicin, respectively. Evidence presented herein argues strongly against the important role of TRP channels in the itch signal initiation at the level of free nerve endings of the cutaneous itch fibres, which is in stark contrast to the body of previous literature on the role of TRP channels in itch (see Kittaka & Tominaga, 2017 for review).

The authors suggested that TRPA1/TRPV1 channels expressed somewhere else within the itch circuitry (but not at the itch nerve endings) might be important for the scratch reflex. This could account for the reduced scratching seen in some studies on TRPV1 and TRPA1 knockout animals (Shim et al. 2007; Wilson et al. 2011). Additionally, it is possible that some degree of TRP channel inhibition/desensitization could have been produced by the surgical procedure (e.g. due to excessive activation of mechanosensitive fibres or due to the tissue damage‐induced release of some excitatory compounds). In this case the TRP channel contribution to the pruritogen‐induced itch fibre excitation could have been underestimated. It has to be noted though that some measures to prevent desensitization have been taken by Ru and colleagues (e.g. a cyclo‐oxygenase inhibitor has been included in the media throughout). In addition, most recorded itch afferents responded robustly to AITC and capsaicin, indicating that active TRPA1 and TRPV1 channels were present.

Notably, in contrast to the previous study (Wilson et al. 2011), Ru and colleagues found no reduction in behavioural itch response to chloroquine in trpa1 ^−/− mice. Moreover, chloroquine‐induced inward current and excitability in the DRG neuron somata were also unaffected by TRPA1 deletion (again, cf. Wilson et al. 2011). Thus, the discrepancy with earlier studies cannot at present be explained solely by the ‘upstream’ effects of TRPA1/TRPV1 in CNS or by some easily tractable technical issues. Clearly, further research will be required to resolve this controversy.

If not TRP, what is the mechanism for the pruritogen‐induced activation of itch fibre endings? A somewhat similar uncertainty about the role of TRPV1 channels surrounded the mechanisms of nociceptive fibre excitation by the inflammatory mediator bradykinin (reviewed by Petho & Reeh, 2012). Search for additional players resulted in the identification of a TRPV1‐independent signalling cascade whereby bradykinin B₂ receptors, acting via the G_q/11–PLC pathway, produced simultaneous activation of depolarizing Ca²⁺‐activated Cl⁻ channel TMEM16A and inhibition of hyperpolarizing M‐type K⁺ (KCNQ) channels (Liu et al. 2010). Ru and colleagues tested if a similar mechanism could be at play in the GPCR‐induced activation of itch fibres. Consistent with the involvement of the G_q/11–PLC pathway, chloroquine‐induced C fibre excitation was almost abolished in the PLCβ3 knockout mice. Furthermore, pharmacological inhibition of TMEM16A with the selective inhibitor MONNA reduced both the chloroquine‐induced action potential discharge from the itch fibres in vitro and the chloroquine‐induced scratching behaviour in vivo.

Although these pharmacological experiments do not offer the same level of conclusiveness as the data obtained in knockout mice (a limitation acknowledged by the authors), they indeed suggest that at least some chloroquine‐induced itch can be attributed to the GPCR‐induced activation of Ca²⁺‐activated Cl⁻ channels in the itch fibre endings. Interestingly, M channel inhibition with the selective blocker XE991 did not induce itch fibre activity in vitro; however, the in vivo effect of XE991 has not been tested.

The excitatory action of the Cl⁻ channel TMEM16A in sensory afferents, suggested in this (Ru et al. 2017) and previous (Liu et al. 2010; Cho et al. 2012; Jin et al. 2013; Takayama et al. 2015) studies is at odds with known inhibitory action of another Cl⁻ channel abundantly expressed in sensory afferents, the GABA_A channel (see e.g. Takkala et al. 2016; Du et al. 2017). The reasons for this apparent discrepancy are currently unknown but could involve distinct densities and localization (e.g. peripheral endings vs. more proximal parts of the fibre) of these channels.

Addressing the conundrums highlighted by the interesting study of Ru and colleagues will require a concerted effort of researchers studying the somatosensory system and is likely to bring an important next step to our understanding of peripheral mechanisms of itch and pain.

Additional information

Competing interests

None declared.