Itch

Peculiarity of Itch Sensations

Itch (pruritus) is a peculiar modality in the realm of somatic sensations. Obviously, it serves nociceptive functions, but it is clearly distinct from pain as a sensation and also with respect to inducing stimuli. It is restricted to the skin and some adjoining mucosae. For the neurophysiologist, the most striking difference applies to the connected reflex organization: whereas the application of painful stimuli to the skin—at the extremities in particular—provokes withdrawal reflexes, itching stimuli provoke the very characteristic scratching reflex. The connection between itch and scratching is so close that itching stimuli are also called “scratchy” in English and similarly “kratzig” in German. This close connection indicates that the neuronal apparatus for itch has developed as a nocifensive system for removal of irritating objects and agents affecting the skin. One might also describe scratching as a reflex pattern that is used in situations in which the noxious stimulus has already invaded the skin. In this situation, withdrawal would be useless; instead, it appears to be more appropriate to scratch the injured site. On the other hand, pathological pruritus raises a major therapeutic problem in a number of diseases. In some cases the itching may be so severe that it heavily impedes the patient’s quality of life.

Classification of Itch

Based on the underlying induction mechanisms, pruritus has been classified as pruriceptive, neurogenic, neuropathic, and psychogenic ( ).

• Pruriceptive itch refers to types of itch in which pruritic mediators, such as histamine, cause itch via activation of peripheral pruriceptors.

• Neurogenic itch is generated centrally by stimulation of itch-mediating pathways in a non-diseased central nervous system; pruritus on spinal application of opioids would be a common example.

• Neuropathic itch denotes types of itch that are caused by diseases of the nervous system, such as post-herpetic itch ( ) or brachioradial pruritus ( ).

• Psychogenic itch refers to types related to illusional states, seen for example in parasitophobia.

This nomenclature is based on mechanism and should therefore also be useful for therapeutic decisions. Because the exact pathophysiology of many clinical forms of itch is yet unclear and combinations of peripheral and central mechanisms are likely to occur, integration of pathophysiology, classification, and therapy will require further effort. A clinically oriented classification has been proposed that mainly differentiates between itch initiated by skin lesions and itch with a presumed systemic cause without evidence of primary skin involvement ( ).

The classifications proposed do not take into account different “flavors” of itch; questionnaires on itch based on the McGill pain questionnaire have been developed and indicate that the itch sensations may be further differentiated. Although this approach has been used successfully in pain research, its impact on itch research remains to be clarified. However, there is no doubt that the type of itch elicited, such as by histamine, can be differentiated from the burning itch induced by endothelin ( ). In addition, differences between histamine- and cowhage-induced itch have been reported ( ). Moreover, differential behavioral responses to itch (scratching with atopic dermatitis versus rubbing with nephrogenic or hepatic itch) are in line with the existence of different itch pathways.

Primary Afferent Pruriceptive Neurons

According to the intensity hypothesis of itch ( ), low-level activation of nociceptors would induce pruritus, whereas higher discharge frequencies would provoke pain ( ). Although the observation that intradermal application of high concentrations of some pruritogens (e.g., histamine) may cause pain seems to be consistent with this hypothesis, many other observations do not support it. Application of low concentrations of algogens does not generally cause itch, just less intense pain. 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 in stimulation frequency at a nerve site where pain has been elicited decreases the magnitude of the pain, but at no point does it induce itch ( ).

Thus, results from microstimulation experiments involving cutaneous C fibers best match the hypothesis that there is a small group of itch-mediating C fibers among a much larger population of pain-mediating units. Indeed, C fibers responding to the application of histamine in parallel with the itch ratings of subjects have been discovered among the group of mechano-insensitive C afferents ( ), thus suggesting that there is a specific pathway for itch ( Fig. 14-1 ). In contrast, the most common type of C fibers, mechanoheat–sensitive nociceptors (CMH or polymodal nociceptors), are either insensitive to histamine or only weakly activated by it ( ). Hence, this fiber type cannot account for the prolonged itch induced by the intradermal application of histamine.

The histamine-sensitive or “itch” fibers (i.e., pruriceptors) are characterized by a particular low conduction velocity, large innervation territories, mechanical unresponsiveness, and high transcutaneous electrical thresholds ( ; ). In line with the large innervation territories of these fibers, two-point discrimination for histamine-induced itch is poor (15 cm in the upper part of the arm) ( ). The excellent locognosia for histamine-induced itch in the hand ( ) might therefore be based on central processing compensating for low spatial resolution in the periphery.

The relative prevalence of the different C-fiber types in human skin nerves has been estimated from recordings in the superficial peroneal nerve ( ). Polymodal nociceptors, which respond to mechanical, heat, and chemical stimuli, are about four times as abundant as mechano-insensitive nociceptors in young healthy volunteers, but their proportion decreases in the elderly (2.5 times) ( ). Mechano-insensitive nociceptors ( ) are activated by chemical stimuli ( ) and can be sensitized to mechanical stimulation in the presence of inflammation ( , ). Among the mechano-insensitive afferent C fibers is a subset of units that have a strong and sustained response to histamine. They account for about 20% of the mechano-insensitive class of C fibers (i.e., about 5% of all C fibers in the superficial peroneal nerve). Specific activation of histamine-positive chemo-nociceptors by prostaglandin E ₂ ( ) in combination with the pruritogenic effects of prostaglandins ( ) provides a strong argument for a specific neuronal system for itch processing that is separate from the pain pathway ( ).

Non-Histaminergic Pruriceptive Neurons

The axon reflex flare is a neurogenic vasodilatation that characteristically surrounds a histamine stimulation site and is induced by release of neuropeptides from mechano-insensitive C fibers ( ). Absence of an axon reflex flare therefore suggests that the itch is independent of histamine-sensitive C fibers. Indeed, itch was induced by pain in an early study in the absence of a flare response, thus indicating a histamine-independent action ( ). Itch without an axon reflex flare can also be elicited by weak electrical stimulation ( , ), further evidence that the sensation of itch can be dissociated from cutaneous vasodilatation.

Cowhage spicules inserted into human skin produce itch at an intensity comparable to that following the application of histamine ( , ). However, mechano-responsive “polymodal” C-fiber afferents, the most common type of afferent C fibers in human skin ( ), can be activated by cowhage in the cat ( ) and, according to recent studies, also in non-human primates ( ) and in human volunteers ( ) ( Fig. 14-2 ). These fibers are unresponsive to histamine and not involved in sustained axon reflex flare reactions ( ). This is consistent with the observation that cowhage-induced itch is not accompanied by a widespread axon reflex flare ( ; ). Although in humans the segregation between histamine-positive, mechano-insensitive fibers and cowhage-positive mechanosensitive fibers is clear-cut, in monkey, mechanosensitive C fibers also responded to histamine ( ). The different histamine response might be explained by higher histamine concentrations with intradermal injection than with iontophoresis.

Aδ fibers responding to the insertion of cowhage for several minutes ( ) suggest an additional role of afferent input from myelinated fibers. Differential block of myelinated afferents does not reduce capsaicin-induced pain and only slightly reduces histamine-induced itch; however, it massively reduces cowhage-induced itch, at least in part of the subjects ( ). The exact role of Aδ-fiber input for cowhage-induced itch is unclear because the reduced skin temperature induced by the nerve blocking maneuver in these experiments might also reduce cowhage-induced activation.

The active compound cysteine protease mucunain has been identified lately and been shown to activate proteinase-activated receptor 2 (PAR-2) and even more potently PAR-4 ( ). Given that cowhage spicules can activate a large proportion of polymodal nociceptors, we face a major problem in explaining why activation of these fibers by heat or by scratching actually inhibits itch whereas activation by cowhage produces it. This problem will be discussed further in connection with central itch pathways.

Spinal Pruriceptive Neurons

The concept of dedicated pruriceptive neurons has been extended by the results obtained from cat spinal cord recordings. A specific class of dorsal horn neurons projecting to the thalamus that respond strongly to histamine introduced into the skin by iontophoresis has been demonstrated ( ). The time course of these responses was similar to that of itch in humans and matched the responses of peripheral C itch fibers (see Fig. 14-1 ). These units were also unresponsive to mechanical stimulation and differed from the histamine-insensitive nociceptive units in lamina I of the spinal cord. In addition, their axons had lower conduction velocity and anatomically distinct projections to the thalamus. The itch-selective units in lamina I of the spinal cord form a distinct pathway projecting to the posterior part of the ventromedial thalamic nucleus, which projects to the dorsal insular cortex ( ), a region that has been shown to be involved in a variety of interoceptive modalities such as thermoception, visceral sensations, thirst, and hunger.

Thus, the combination of dedicated peripheral and central neurons with a unique response pattern to pruritogenic mediators and anatomically distinct projections to the thalamus provides the basis for a specific neuronal pathway for itch.

This is also supported by studies performed in rodents. Dorsal horn neurons bearing the receptor for gastrin-releasing peptide (GRPR) have been identified as being crucial for the scratch behavior in a variety of itch models ( ). There was some reduction in scratching by constitutively inactivating the gene encoding the GRPR gene or by pharmacologically blocking the receptor. However, selective deletion of GRPR-bearing cells by a toxin linked to the GRPR ligand bombesin (bombesin-saporin) completely abolished scratching behavior, whereas nociceptive behavior was virtually unchanged ( ). This indicates that GRPR-expressing dorsal horn neurons may be indispensable for the itch response in this species, although not necessarily the GRPR receptor alone might be responsible. However, recent data on bombesin-induced itch that could not be blocked via GRPR agonist have shed some doubt on the GRPR specificity of the bombesin results ( ).

In contrast to the above evidence for a specific pathway for itch, histamine-sensitive projection neurons in the monkey were found to also respond to mechanical stimuli and to capsaicin ( , ), and in rodents, overlapping between nociceptive and pruriceptive neuronal responses was found ( ). This does not necessarily contradict the concept of a “specific pathway.” One has to distinguish between “selectivity” (i.e., only a subgroup of neurons respond to a particular pruritogenic substance) and “membrane specificity” (a subgroup of neurons responds only to a group of pruritogenic agents). Membrane specificity is not necessarily required for a “specific,” or better, “selective” pathway. A “selectivity hypothesis” for itch processing has been discussed before by several authors ( , , , , , , ).

Interestingly, not only is histamine- and cowhage-induced itch processed separately in primary afferent neurons, but the separation is also maintained on the spinal level. Spinothalamic projection neurons in the dorsal horn could be separated into a histamine- and cowhage-responsive population without overlap ( ). Moreover, the thalamic projections of the two subgroups also differ. The histamine-sensitive pathway can be assumed to be selective for itch, albeit recordings in monkey spinothalamic tract neurons also suggest some mechanical and capsaicin sensitivity (see above). In contrast, the cowhage-sensitive pathway may be regarded as unspecific in that activation of mechano-heat–responsive polymodal nociceptors is probably underlying the generation of cowhage-induced itch ( ). There may, however, be a kind of “spatial selectivity” for this form of itch if this pathway contains neurons that are exclusively activated by polymodal nociceptors with terminals in the superficial epidermal layers of the skin since it has been shown that cowhage exerts its pruritogenic action only in the epidermis and not in deeper layers ( ).

It is interesting to note that “specificity” is discussed not only for neurons but also for mediators ( ); for example, the classic algogen capsaicin generally provokes pain when applied to human skin, but it induces itch when applied to the tip of an inactivated cowhage spicule ( ). This important finding indicates that the spatial characteristics of the application may be crucial and may functionally convert an algogenic mediator to a pruritic mediator. The highly localized stimulation in the epidermis strongly activates some of the local nociceptors while their immediate neighbors remain silent, thereby resulting in a mismatch signal of activation and absence of activation from this site. It has thus been hypothesized that this mismatch might be perceived by the central nervous system as itch ( ). Teleologically, it is obvious that scratching behavior in the case of a very localized superficial noxious focus is an adequate response because it can eliminate the presumed cause. Moreover, scratching activates all the mechanosensitive nociceptors in the stimulated area, and thus the mismatch signal of activated and non-activated nociceptors at this site is terminated.

Supraspinal Itch Processing

The development of refined methods of functional imaging in the past 2 decades made it possible to study supraspinal itch processing in awake humans. However, a serious obstacle is the slow onset and “waxing and waning” nature of experimental itch induced by one of the aforementioned itch mediators. Imaging methods depending on fast input signal variation, such as electroencephalography (evoked cortical potentials) or magnetoencephalography, are less suitable for these studies. This obstacle least affects the regional cerebral blood flow (rCBF)-based imaging method of ¹⁵ O-H ₂ O positron emission tomography (PET) because of its low time resolution. One study ( ) used intracutaneous injection of histamine. Activations were found mainly in the motor and premotor areas of the cortex, in the cerebellum, in the anterior and posterior cingulated gyrus, and in frontal areas, predominantly in the left hemisphere ( , , ). It was concluded that the significant co-activation of motor areas supports the notion that itch is inherently linked to a desire to scratch. Later PET studies using histamine prick ( ) or histamine iontophoresis ( ) found a similar activation matrix but with some differences regarding activation of a thalamic site. This activation matrix seemed to be similar for pain and itch ( , ). Modulation of itch by painful cold stimuli was investigated with ¹⁵ O-H ₂ O PET ( ). The periaqueductal gray matter (PAG) was activated only when painful and itching stimuli were applied simultaneously. This activation was combined with reduced activity in the anterior cingulate, dorsolateral prefrontal cortex, and parietal cortex, thus suggesting that the PAG might be involved in the central inhibition of itch by pain. In a recent PET study using histamine iontophoresis on the back of the left hand, brain activations of healthy controls and patients suffering from atopic dermatitis were compared ( ). Although the itch sensations induced by the stimuli were not different in both groups, more cerebral areas were activated in patients than in healthy controls. Activation in the patients was significantly higher in the contralateral thalamus and ipsilateral caudate and pallidum.

Whereas rCBF-related PET studies typically have a time resolution of 50–70 seconds, functional magnetic resonance imaging (fMRI) allows scanning of the brain at a much higher time resolution, typically 2 to 3 seconds ( Fig. 14-3 ). These studies should be performed in a “percept-related” manner; that is, changes in local oxygenation (blood oxygenation level–dependent [BOLD] effect) should be correlated with variations in itch intensity. This was not always regarded in earlier fMRI studies on itch, where itch ratings were performed at 30- or 60-second intervals ( , ). One group modulated the time course of the itch sensation by interposing cooling stimuli in a regular manner to overcome the problem of slowly changing time courses of itch sensations and to allow a more distinguished correlation with the BOLD changes ( , ). To achieve regular modulation of the histamine-induced itch response without interference from another stimulus modality, we used intracutaneous microdialysis probes perfused with histamine solution or saline for control. Histamine applied with this method induced a strong itch response that could quickly be terminated by flashing the probes with local anesthetics ( ). With these types of stimulation, roughly the same regions were activated in fMRI investigations as described in the previous PET studies, but in addition, more subcortical regions were encountered (ipsilateral caudate nucleus, contralateral claustrum, bilateral putamen, and bilateral thalamus). Most strikingly, we observed stronger activation of the cerebellum than with heat pain stimulation. These findings again demonstrate the close relationship between itch and a motor action, namely, scratching.

Itch Modulation by Painful and Non-Painful Stimuli

Inhibition of itch by painful stimuli has been demonstrated experimentally with the use of various painful thermal, mechanical, and chemical stimuli. Electrical stimulation via an array of pointed electrodes (“cutaneous field stimulation”) has also been used successfully to inhibit histamine-induced itch for several hours in an area around a stimulated site 20 cm in diameter. The large area of inhibition suggests a central mode of action ( ). Consistent with these results, itch is suppressed inside the secondary zone of capsaicin-induced mechanical hyperalgesia ( ). This central effect of nociceptor excitation by capsaicin should be clearly distinguished from the neurotoxic effect of higher concentrations of capsaicin, which destroy most C-fiber terminals, including fibers that mediate itch ( ). The latter mechanism, therefore, also abolishes pruritus locally until the nerve terminals are regenerated.

Not only is itch inhibited by the enhanced input of pain stimuli, but vice versa, inhibition of pain processing may also reduce its inhibitory effect and thus enhance itch ( ). This phenomenon is particularly relevant to spinally administered µ-opioid receptor agonists, which induce segmental analgesia often combined with segmental pruritus ( ), and has also been confirmed in animal experiments ( ). Very recent results suggest that the analgesic and pruritic effects of µ-opioids might be mediated by different isoforms (MOR1 versus MOR1D), which would have major therapeutic implications ( ). Conversely, κ-opioid antagonists have been found to enhance itch ( ). In line with these results, the κ-opioid agonist nalbuphine has been shown to reduce μ-opioid–induced pruritus ( ), and the concept has already been tested successfully in chronic itch patients with a newly developed κ-opioid agonist ( ).

Central inhibition of itch can also be achieved by cold stimulation. In addition, cooling has a peripheral inhibitory effect: histamine-induced activation of nociceptors can be reduced by cooling ( ). Also in humans, cooling of a histamine-treated skin site reduced the activity of the primary afferents and decreased the area of “itchy skin” or “hyperkinesis” around the application site ( ). Conversely, warming the skin would lead to an exacerbation of itch. However, as soon as the heating becomes painful, central inhibition of pruritus will counteract this effect ( ).

Very recent work on the antipruritic effects of subpopulations of primary nociceptive afferents indicates that input from the vesicular glutamate transporter 2 (VGLUT2)-positive subpopulation is especially crucial for the inhibition of itch behavior by painful stimuli ( , ). When VGLUT2 and thereby glutamate release was selectively eliminated in Na _v 1.8-positive nociceptors by conditional genetic knockout techniques, inflammatory and neuropathic pain responses were grossly abolished, but spontaneous scratching behavior and increased experimental itch were massively enhanced ( ). Most interestingly, capsaicin-induced pain behavior was changed to scratching behavior in these mice, thus suggesting that the lack of noxious input via VGLUT2-positive nociceptors disinhibited itch ( ). The exact nature of the crucial nociceptor class is still unclear inasmuch as another group did not find increased scratching when VGLUT2 was selectively eliminated in transient receptor potential vanilloid 1 (TRPV1)-positive primary afferent neurons ( ).