Pruritus (itch) is a sharp tingling sensation resulting from a potentially harmful stimulus, chronic and neurogenic itch are persistent bouts of itching with (chronic) and without stimulus (neurogenic). Although scratching is a reflex response to itching, if the irritation is extensive there is more likely to be a pain reflex before scratching which potentially means that scratch is to alleviate irritation rather than to protect against what has caused it. Due to the overlap with pain, historically it was thought that itch was the same mechanism and pathway as nociception but interpreted differently as it is a lower intensity (Frey, 1922), in effect a warning that pain may occur if the body stays exposed to the stimulus but is not immediately dangerous. This has largely been rejected in favour of the suggestion that there are specific neurons controlling itch (Schmelz et al., 1997). However, the overlap between itch and pain mean a specific itch system may be too simple and the neurons may conduct nociceptive signals as well (selective theory). The following review will tackle the evidence for the selectivity theory with regards to peripheral neuron types and discuss the potential mediators and corresponding receptor subtypes responsible for itch signalling at the peripheral terminals of dorsal root neurons.
Identification of itch signalling neurons
It has been hypothesised that pruritus is transmitted along primary afferents specific to itch (Schmelz et al., 1997). Yosipovitch contested specificity theory with claims that noxious heat and scratching inhibit itch as it changes ones perception from itch to pain. This claim was supported by their results identifying the intensity of itch decreasing by 0.9cm (P<0.05) (Yosipovitch et al., 2005). However, the scale used was a visual analogue scale which although they showed to be reliable in repetition, is not the most accurate way to measure itch intensity because it is designed to measure subjective characteristics. Yet itch can be quantifiably measured in the frequency and duration of scratch methods which have high construct validity, as many species including humans scratch to alleviate itch. Their claim is supported more conclusively in 2015` with the co-expression of itch neuronal markers and TRPV1 (table1) suggesting noxious heat would have an effect on itch signalling (Usoskin et al., 2015). Usoskin’s team identified neurons responding to itch by using RNA-sequencing to group dorsal root neurons into sub-populations based on RNA expression. They identified 13 sub-groups of sensory neurons, with non-peptidergic unmyelinated C-fibres having 3 sub-groups identified (NP1-3)(Usoskin et al., 2015). These groups contained populations of neurons expressing itch receptors. Different populations expressed different itch receptors at varying expression levels (figure1A-D) suggesting that the different populations are involved in different types of itch. Their results identify selectively high expression of somatostatin in NP3 neurons (0.83)(Table1), this was therefore used as a marker for NP3 neurons.
RNA-sequencing can produce false positives due to artefacts (Ozsolak and Milos, 2011). This can lead to mean data that is not representative of the expression of that RNA. Usoskin’s team made efforts to minimise this limitation by setting a threshold for expression for each gene and only considering the genes that exceeded that expression when grouping populations ((Usoskin et al., 2015) supplementary methods).
Itch receptors and itch mediators
Histamine
Histamine was the first identified itch mediator which might be expected due to the alleviation of acute itching by antihistamines, importantly histamine helped explain the antagonistic relationship between itch and pain (scratch alleviates itch). Nilsson’s team used cutaneous field stimulation (nociceptive electrical impulses) to abolish histamine induced itch in an area of skin, 4 hours after, itch intensity was still 32% lower than control (Nilsson and Schouenborg, 1999). Although histamine is used experimentally to induce itch, pathological itch is unaffected by histamine (Klein and Clark, 1999) and expression maps identify only low levels of the histamine receptor Hrh1 in NP2 and NP3 (table1). The low receptor expression and lack of effect in pathological pruritus may explain why acute itch is only a low level irritation. High receptor occupancy would be required to evoke action potentials which may also explain why high local histamine concentrations are required for itch such as in inflammation.
Interleukin-31
Previous studies identified that interleukin-31 (IL-31) was implicated in chronic disease (Dillon et al., 2004; Takaoka et al., 2006). Since then research identified IL-31 as a key mediator of atopic-dermatitis in mice (Grimstad et al., 2009). It took until 2013 to demonstrate that IL-31 induced scratching with a single acute dose (Arai et al., 2013) see in figure2A. Arai observed that lengthy scratching to IL-31 was higher if applied to lesioned skin, supporting IL-31s role in chronic itch from previous studies (figure2B) but no receptor had been conclusively identified. Usoskin et al then uncovered that NP3 neurons expressed the IL31ra receptor (figure1G) in equal abundance to TRPV1 (0.58)(table1). They investigated this further by treating mice with IL-31 and observing the frequency of scratching increase from 10 to 50 (figure2D). Usoskin also states that NP3 neurons are involved in chronic itch, therefore so is IL-31, which supports the work done by Dillon and Takaoka’s teams (Dillon et al., 2004; Takaoka et al., 2006). However, Dillon et al had to overexpress IL-31 and used transgenic mice to introduce IL-31ra into the epithelial cells. These are not usual causes of atopic-dermatitis so construct validity of the model is limited. In addition, although the model showed good face validity as the mice had similar symptoms to the human disease, their results could not be translated into wider context because no empiric evidence was provided for increased scratching, despite it being stated in the text. Takaoka later produced a similar study but including results showing increased frequency of scratch (figure2C), implicating IL-31 in physiological as well as pathological itch.
Serotonin
Serotonin was found to be involved in itch when Weisshaar applied it to human skin (Weisshaar et al., 2004). Serotonin was also shown to increase scratching in vivo (figure1H) and 5HT1f was highly expressed (0.83) in NP3 neurons and scratching increased from 10 to 80 in response to 5HT (figure2D) which supports the claim that serotonin receptors have a functional role in itch (Usoskin et al., 2015).
Morita’s team showed that 5HT7 is coupled to TRPA1 cation channels by expressing 5HT7 or TRPA1 alone in human embryonic kidney cells and showing no inward calcium current, then co-expressing them and observing an inward current (Morita et al., 2015) also see figure3A. This suggests they are coupled but does not confirm that they cause itch or that other ion channels are not involved. So the team produced knockout mice, to show that there was a significant reduction in scratching if either 5HT7 (65s reduced to 25s) or TRPA1 (35s reduced to 5s) was ablated, suggesting both are required for itch (figureB/C). In addition, the TRPA1 knockout suggests that without TRPA1, 5HT7 does not lead to itch therefore is not coupled to other ion channels involved in itch.
The role of another TRP channel, TRPV4, has been implicated in 5-HT mediated itch (Akiyama et al., 2016), yet in vivo, Morita showed reduced scratching time from 35 to 5 seconds in response to 5-HT without TRPA1 suggesting only TRPA1 is sufficient for itch. However, Akiyama’s results show significantly reduced scratching in TRPV4 knockouts (100 bouts reduced to 20) but no difference between wildtype and TRPA1 knockouts (figure3D). This discrepancy (compare figure 3C and D) is not explained in the literature so a follow-up study should be undertaken to identify 5-HT response in vivo and in voltage-clamp experiments of isolated neurons with TRPV4, TRPA1 or both ablated to identify the contribution of each.
Histamine and IL-31 induce itch along with inflammation and serotonin induces itch at micromolar concentration but pain at millimolar concentration (Morita et al., 2015). This could be evidence that itch neurons also respond to pain, however the response could be from separate classes of neurons with the same receptors.
Conclusion
In summary, it is clear that a sub-population of non-peptidergic neurons (NP3) transmit pruriceptive signals. It is still not completely clear whether these neurons are specific to itch signalling or selective for itch. The lack of co-localisation between itch markers and IB4 suggest the neurons are itch specific but involvement of TRPA1, TRPV1 and TRPV4 in pruritus and nociception suggest that the neurons could be prompted to respond to nociceptive stimuli. To conclude, the evidence from these mediators and receptors strongly suggests that itch neurons can conduct nociceptive signals but it is still possible that the neurons are itch specific. A future approach may be to use optogenetics to induce action potentials specifically in itch neurons by introducing photoactivated ion channels under the somatostatin promoter. Then record the in vivo response to identify whether when scratching is induced these neurons become silent or nociceptive.
