More than 20 years ago, mechanoinsensitive (“silent”) histamine-sensitive C-nociceptors in human (2) and spino-thalamic projection neurons in the cat (12) have been identified as part of a specific pruritic pathway. More recently, molecular markers of non-histaminergic itch-specific neurons were identified in rodent, such as B-type natriuretic peptide (BNP) (13, 14) and members of the mas-related G-protein receptor family (mrgprA3, C11) (15–17) in primary afferent neurons, but also gastrin releasing peptide (GRP) (18–20) in dorsal horn neurons. Non-histaminergic itch signaling has received major interest when mas-related G-protein coupled receptors (Mrgprs) were identified on presumably itch-related neurons in the mouse, i.e., MrgprA3 (21), MrgprD (22), and MrgprC11 (23). For humans, BAM8-22, an activator of MrgrpC11 (24) and beta-alanine–an activator of MrgprD—provoke itch in humans (22, 25). Chloroquine has often been used in mice to elicit itch-behavior via activation of MrgprA3 (21, 26, 27). Thus, the plethora of new information on pathways and mediators for itch have been found and summarized in excellent reviews (8, 10, 28). All the itch-specific markers and pathways might imply that among the competing itch theories the “labeled line” theory has finally been validated.

However, not necessarily these pruritic mediators are part of a labeled line. In mice, peripheral nerve injuries provoke a broad de-novo GRP expression in DRG neurons (29). Thus, upon activation or via spontaneous activity such a “phenotypic switch” might convert nociceptors that typically inhibit itch into neurons that by releasing GRP spinally actively contribute to chronic itch. It is important to note that no direct synaptic contact might be required to get access to the spinal itch pathway, but volume transmission (30) might suffice. Single cell expression patterns have been used to define pruriceptive subpopulations of primary afferent neurons in mouse revealing the key marker proteins Mrgpr A3 (often co-expressing MrgprA1 and MrgprC11), B-type natriuretic peptide / Somatostatin, and MrgprD (31–33). In humans, classification of primary afferent nociceptors is classically based on their sensory characteristics obtained from single fiber recordings (34–38), such as mechanical thresholds and temperature sensitivity (Table 1). Currently, important work is being performed to reduce the translational gaps between rodent and human single cell expression patterns (39, 40), but also to link functional characteristics and expression pattern within a single neuron (41). This ongoing work will provide a much more detailed basis for the differentiation between itch and pain processing neurons and the extent of overlap in mediators and receptors.

Electrophysiological data from rodents and monkey did not support a “labeled line” for itch (28, 42–44) as no specific subpopulation of itch neurons was found. The results rather support the pattern theory of itch according to which nociceptors can signal itch or pain based on the combination of activated fibers resulting in a population coding (1) or based on specific discharge patterns that might differentiate between itch and pain (1).

Very focal activation of nociceptors in the skin has been suggested as a spatial discharge pattern that could explain itch without a “labeled line:” noxious stimulation that is directed only to few cells within the epidermis elicits itch in human skin (45), even when the stimulus is the algogen capsaicin known to activate most human C-nociceptors (46). It has been suggested that local activation of only few epidermal nociceptors can cause itch by a “mismatch signal” (47) or “spatial contrast” (48), provided by few activated and many non-activated nociceptive endings innervating the same skin site, for example by a minute glass wool fiber. Such a discharge pattern indicates that a noxious event is minute and localized within the epidermis. Teleologically, one might conclude that scratching off a part of the epidermis is an adequate response as it will eliminate the threat in this situation. Moreover, scratching will elicit a consistent response of all mechanosensitive polymodal nociceptors innervating this skin and might thereby terminate the “spatial contrast” itch pattern. As shown in Figure 2 such a spatial contrast can not only result from a very focal stimulation in the epidermis, but may also result from strong activation of only a subpopulation of the nociceptors at the stimulation site. Importantly, this activated subpopulation not necessarily belongs to an itch-selective class of nociceptors, it is rather the resulting “spatial contrast” of their response that is responsible for the interpretation as itch on a spinal level.

When the temporal aspects of neuronal discharge are considered, primary afferents activated by the pruritogen histamine are characterized by lower frequencies as compared to discharges to the algogen capsaicin and periods of bursting have longer intervals (49). However, this difference does not allow a clear separation. New data have found differential temporal discharge pattern to heat stimuli between potential primary afferents linked to pain or itch processing: polymodal nociceptors were differentiated according to their response to a step-like noxious heat stimulus into quick (“QC”) or delayed/slow fibers (“SC”) in the monkey (25). C-nociceptors with immediate phasic heat responses were particularly responsive to the pruritogen β-alanine, suggesting that the QC/SC classification might be helpful to characterize a pruriceptive subpopulation (25). Beyond such a characterization, it does is not obvious how this acute response pattern might be linked to a differential signal in itch or pain processing.

As noted above another important temporal aspect is duration and intensity of the discharge: longer lasting activation and higher frequency of GRP-positive neurons were required to finally allow transsynaptic activation of the tertiary GRP-receptor neurons (30). Thus, even when stimulating the itch-specific pathway via optogenetic stimulation of GRP-positive neurons single pulses were inefficient and repetitive lasting stimulation was required (30). Such a temporal pattern of delayed starting of itch after pruritogen application resonates with the psychophysics observed in humans (2). Moreover, in accordance with these considerations optogenetic studies investigating scratching upon specific stimulation generally use prolonged activation periods for example to show disinhibition via dynorphin spinal neurons (30 min at 20 Hz) (50).

According to the intensity hypothesis of itch (51) low level activation of unspecific nociceptors would induce pruritus, whereas higher discharge frequencies would provoke pain. Clinically, impairment of pain processing in neuropathy allowed to provoke itch in some patients (52): in patients with impaired nociception, but intact touch system painful stimuli were felt as itch whereas more specific stimulation using “itch powder” was ineffective (52). Thus, the authors concluded that low intensity of slow nociceptive input provokes itch, whereas increasing the intensity of noxious input by scratching would inhibit itch. Also, the observation that intradermal application of high concentrations of some pruritogens, e.g., histamine, may cause pain seems to be consistent with this hypothesis. However, application of low concentrations of algogens generally does not cause itch, just less intense pain (1). Furthermore, intraneural electrical microstimulation of human afferent C-fibers usually induces pain and, very rarely, pruritus. Increasing the stimulation frequency of intraneural microstimulation enhances the intensity of pain or itch, but no switch from pruritus to pain has been observed. Likewise, a decrease of stimulation frequency at a nerve site where pain has been elicited, decreases the magnitude of pain; but at no point does it induce itch (53).

In summary, the pure intensity theory according to which pruritus is gradually converted to itch upon increasing stimulation without change of quality can be regarded as rejected. Yet, as will be discussed below, there might also be indirect evidence for “unmasking” of itch when the pain input is impaired or treated. Thus, it might not be adequate to leave out all possible aspects of the intensity theory.

Implications for basic experimental approaches. Current experimental data primarily support the specificity and pattern theory of itch. However, recent results have shown that temporal discharge patterns have crucial effects (49) even within itch-specific pathways in the spinal cord: sustained peripheral input via pruriceptors is required to enable successful transsynaptic signal progression in the spinal cord (30). Optogenetic techniques have enabled unmatched temporal control of specific classes of primary afferents (50, 58–61). Again, neuronal specificity and temporal pattern interacted: tonic activation of pruriceptors is required to induce scratching behavior whereas short-lasting stimulation might be ineffective or may even cause withdrawal (30). Thus, the implications of temporal patterns need to be considered even when the experimental approach is focused on an itch-specific pathway, i.e., upon optogenetic stimulation of pruriceptors it is important to compare responses to acute phasic stimulation and tonic discharge. Moreover, the spatial aspect of activation might play a role even for specific activation using optogenetic stimulation in particular for transcutaneous stimulus when the limited penetration depth of blue light is taken into consideration.

When analyzing the discharge of primary afferents we usually analyze area under the curve and maximum discharge frequency. However, considering the importance of other temporal patterns such as lasting and bursting discharge we might expand our analysis accordingly as bursting discharge patterns of nociceptors with long silent periods in between can also generate some kind of spatial contrast pattern thereby merging temporal and spatial contrast theory. On the other hand, synchronous input of high discharge of many nociceptors will cause pain thereby incorporating aspects of the intensity theory to the overall outcome.

Implications for clinical approaches: The “phenotypic switch” of nociceptors that start to express GRP after axotomy (29) not only adds to the controversy about the role of this peptide in primary afferents (62, 63), but may have major implications for clinical pruritus. Beyond the direct involvement of GRP in the generation of pruritus it may render the spinal cord more susceptible for evoked pruritus: in particular acute electrical stimulation may provoke mainly pain in healthy volunteers. However, in chronic itch patients it might also evoke itch (64) suggesting that the spinal cord processing of itch has been facilitated in these patients potentially be previous release of GRP. For experimental stimulation protocols therefore phasic and tonic stimulation paradigms and possible carry-over effects by spinal GRP release need to be considered.

Animal models have been optimized to separately assess intensity of itch and pain, for example via scratching vs. wiping behavior upon injection into the cheek has been successfully established in rodents (65). However, even under these experimental conditions these models can also elicit a combination of itch- and pain-like behaviors (65, 66) suggestive of a mixed sensation. Indeed, patients with small fiber neuropathy often report concomitant itch and pain sensations (67) such as “burning itch” or “itching sting.” Rather than separating itch and pain according to the involved discipline (pain for anesthesiologists/neurologists vs. itch for dermatologists) we need to further our translational efforts not only between basic researchers and clinical scientists, but also between pain and itch specialists.