Animal Models of Itch: Scratching Away at the Problem

36



Carstens and Kuraishi



potential benefit in studies of normal itch sensation and its physiological

mechanisms, as discussed extensively below.

There is also a pressing need for the development of animal models of

chronic itch that mimic the pathophysiology of atopic dermatitis and other

dermatological and systemic conditions associated with severe chronic itching in humans. Currently, there are few such animal models, and these are

also discussed.

A variety of systemic disorders have itching as a symptom (4), in

particular liver dysfunctions such as cholestasis or biliary cirrhosis (5), and

renal failure (6). The pathophysiological mechanisms for these types of itch

are unknown. The possibility of pathophysiological alterations in central

itch signaling mechanisms, with emphasis on opioid involvement, is also

considered below.



II.



HINDLIMB SCRATCHING AND RELATED BEHAVIORAL

MODELS OF PERIPHERALLY EVOKED ITCH



The association between itch sensation and scratching has led to the use of

scratching behavior in animals as an assay for itch. Most studies involve i.d.

injection of a pruritogen into the nape of mice or rats, and count the number

of bouts of hindlimb scratching directed toward the stimulus (2,3). To what

extent is this a viable itch model?

To be a selective model of itch, the behavioral scratching response

should be consistent with the following properties of itch:

 selectively induced by pruritic but not algesic stimuli

 alloknesis (‘‘itchy skin’’) (i.e., manifestations of itch sensation

elicited by innocuous mechanical stimulation of skin surrounding

the pruritic area)

 induction or exacerbation of itch by opioids

 suppression of itch by



–opioid antagonist, naloxone

–noxious stimuli (scratching, heat)

–innocuous stimuli (cooling, rubbing)

–distraction

–mast cell degranulation (e.g., compound 48/80)

–pretreatment of skin with capsaicin.

Regarding the first point, there is evidence both for and against a specific

association between itch and scratching. Spontaneous scratching behavior

occurs in arthritic rats (7); the scratching was reduced by morphine in a



Animal Models of Itch



37



naloxone-reversible manner but was not reduced by an antihistamine

(astemizone), suggesting that the scratching may reflect pain rather than

itch. Hindlimb scratching is also elicited by intrathecal microinjection of a

variety of chemicals, including neurokinins, capsaicin, morphine, and many

others (8–10). In some instances, the intrathecally induced scratching was

reduced by morphine (10), again suggesting that it was pain-related. However, scratching is evoked by intrathecal drugs in spinalized rats (8), raising

the possibility that it reflects the activation of motor scratch reflex circuits

rather than sensory systems. Facial scratching is also elicited by intracerebroventricular injection of morphine (11,12) and other substances (see

below). Scratching is also a component of normal grooming behavior. These

data therefore indicate that scratching behavior per se does not necessarily

reflect itch sensation.

However, recent behavioral data suggest that directed hindlimb scratching may distinguish between itch and pain (Table 1). Intradermal injection of

substances that induce itch in humans, including substance P (13), the mast

cell degranulator compound 48/80 (14), serotonin (5-HT; 15), and plateletactivating factor (PAF; 16), elicit dose-related scratching in rodents (2,3,17–

19) (Figs. 1a, b and 3b). In contrast, i.d. injection of algesic agents, such as

capsaicin and formalin, elicits either no or very little scratching, which is not

dose-related (3,18,19). Curiously, i.d. histamine does not elicit scratching in

ddY mice (3,18) (Fig. 1a and b) or Sprague–Dawley rats (19) even though it

is the definitive pruritogen in humans (20,21). This discrepancy may be

explained by the fact that cutaneous mast cells in rodents contain little

histamine but high concentrations of 5-HT (22–24), which may be pruritic in

certain rodent strains. Other rodent strains, including the ICR mouse (25)

and hairless guinea pig (17), exhibit scratching dose-related to i.d. histamine

(Table 1). In the latter study, PAF also elicited significant scratching. The



Table 1 Chemicals That Do (+) and Do Not (À) Induce Directed Scratching

When Given i.d. in Three Rodent Species

Chemical

Histamine

Substance P

Compound 48/80

5-HT

PAF

Leukotriene B4

Capsaicin

Formalin



Mouse



Hairless guinea pig



Rat



À (ddY); + (ICR)

+

+

+



+



À



+

+



+



+

À

À



À

À



38



Carstens and Kuraishi



Figure 1 Hindlimb scratching elicited by various chemicals injected i.d. in mice. (a)

Graph plots mean number of scratching bouts per hour (error bars: S.E.M.) in mice

receiving i.d. 5-HT (filled circles), with little scratching following i.d. histamine (filled

diamonds). Open circle: vehicle (saline). (From Ref. 18.) (b) Graph plotting mean

scratching bouts per hour elicited by i.d. injection of compound 48/80 (filled circles)

and substance P (filled triangles), with little scratching following i.d. histamine

(diamonds). (From Ref. 3.)



available behavioral data thus indicate that directed hindlimb scratching

behavior may distinguish between pruritic and algesic stimuli, supporting it

as a viable animal model of itch.

In a recent study of scratching induced by i.d. 5-HT in rats, head or

whole body shaking was correlated with scratching (19), suggesting that this

type of grooming behavior may also be a useful parameter to assess itch.

The possibility of alloknesis, to our knowledge, has not yet been addressed using animal scratching models.

The third point, that opioids should evoke the itch-related behavior,

is supported by several studies showing that facial scratching is elicited

by opioids delivered by intracerebroventricular, intramedullary, or intrathecal

routes (9,11,12,26–30). The opioid-evoked scratching was shown in some of

these studies to be reversed by naloxone. Naloxone significantly attenuated

directed hindlimb scratching elicited by i.d. injection of 5-HT (18) or substance

P (31). Substance P–induced hindlimb scratching was also significantly

reduced by pretreating the injected skin area with capsaicin or compound

48/80 (31). The capacity for noxious or innocuous counterstimulation to

reduce hindlimb scratching in animals has not yet been investigated. Interestingly, the degree of scratching elicited by i.d. 5-HT was markedly less when the

investigator was present, rather than absent during data collection (18),

suggesting that scratching is susceptible to distraction.

In summary, many of the expected properties of itch are fulfilled in

studies of directed hindlimb scratching by rodents, supporting its utility as



Animal Models of Itch



39



an animal behavioral model of itch. Furthermore, the data suggest that 5HT, substance P, mast cell degranulators, and PAF are pruritogenic in rodents, whereas histamine appears not to be pruritic in some common rodent

strains.

Hindlimb scratching directed toward the eye after topical application of

chemicals to the ocular surface has been used to assess conjunctival itch in

hairless guinea pigs (17). In this model, histamine, 5-HT, PAF, and prostaglandin E2 elicited significant dose-related ocular scratching, whereas even

high concentrations of the algesic agents bradykinin, acetic acid, or saline did

not.

Biting has also been suggested to reflect itch, based on a recent study

assessing licking and biting directed toward the site of i.d. 5-HT injected into

the hindpaw of mice (32). 5-HT elicited approximately equal numbers of

licks and bites over a 60–100 nmol dose range. Biting, but not licking, was

significantly attenuated by both naloxone and the 5-HT antagonist, methysergide. In contrast, formalin elicited a characteristic biphasic temporal

pattern of hindpaw licking but no biting. These results suggest that hindpaw

biting is analogous to scratching and may represent a means of noxious

counterstimulation of the paw to relieve itch sensation.



III.



MULTIPLE ITCH MECHANISMS



The pharmacology of peripherally evoked scratching has recently come under

study (Table 2). In humans, itch elicited by histamine, substance P, and PAF is

reduced by H1 receptor antagonists as well as mast cell degranulation,

implicating histamine liberated from mast cells as the final effector (21).

Consistent with this, the H1 antagonist, pyrilamine, substantially reduced

histamine-evoked ocular scratching in guinea pigs (17). However, recent data

from mice indicate that substance P-evoked scratching may partly involve a

separate NK-1 receptor-mediated mechanism (31) (Table 2). Moreover, PAFinduced scratching in guinea pigs was reduced by PAF antagonists (WEB

2086 and CV-6209) but not by H1 antagonists (17). Leukotriene B4 (LTB4)

elicits scratching in mice that is attenuated by an LTB4 antagonist (33). LTB4

may play a role in substance P-induced scratching (34). Emedastine, which

blocks histamine H1 receptors as well as histamine release from mast cells,

significantly attenuated scratching induced by i.d. LTB4, substance P, and

histamine, but not 5-HT, in ICR mice (35). Interestingly, the blocking effect

of emedastine was much greater for scratching elicited by LTB4 and

substance P compared to histamine, suggesting that part of emedastine’s

antipruritic effect may be via a blockade of LTB4 activity. The other H1

antagonist, azelastine, which reduces pruritus in chronic hemodialysis

patients (36), may involve the action and production of LTB4 in its



40



Carstens and Kuraishi



Table 2 Multiple Itch Mechanisms (See Text for Explanation)

Chemical

Histamine

Substance P/

NK-A

PAF

5-HT



Leukotriene B4



Animals

scratching # by

H1 receptor antagonists

NK-1 antagonists; naloxone;

pretreatment of skin with

capsaicin; compound 48/80

PAF antagonists

5-HT1/2 antagonists, not 5-HT3;

naloxone; pretreatment of skin

with capsaicin; compound 48/80

LTB4 antagonist (ONO-4057);

emedastine; azelastine



Humans

itch # by

H1 antagonists

H1 antagonists;

compound 48/80

H1 antagonists;

compound 48/80

5-HT3 antagonists

(ondansetron)

Azelastine (hemodialysis

patients)



antipruritic effect (37). Collectively, the data presented above suggest that

there may be multiple mechanisms involved in itch, in addition to the

liberation of 5-HT or histamine from cutaneous mast cells.

Scratching was elicited in mice by 5-HT2, but not 5-HT1A or 5-HT3,

agonists (Fig. 1a), and was reduced by 5-HT1/2 antagonists (methysergide and

cyproheptadine) but not by 5-HT3 antagonists including ondansetron (18),

indicating that 5-HT-induced scratching in mice is at least partly mediated via a

peripheral 5-HT2 receptor. These differences might be explained by species

differences in 5-HT receptor mechanisms involved in itch, and/or pathophysiological changes in 5-HT signaling due to the dermatitis.

The currently available data therefore suggest the existence of multiple

pharmacologically distinct peripheral itch mechanisms. It will be interesting

to determine if a common population of peripheral ‘‘itch’’ receptors expresses

a multiplicity of molecular receptors for each of the suspected pruritic agents,

or if there is some degree of chemical selectivity in responses of peripheral

chemonociceptors that project to central itch signaling pathways.



IV.



MODELS OF ALLERGIC ITCH



When the ICR mouse was given an i.d. injection of antigen-specific immunoglobulin E and then an intravenous injection of antigen, it showed scratching directed toward the immunoglobulin-injected site (38). The passive

cutaneous anaphylaxis-induced scratching was suppressed by H1 receptor

antagonists (cetirizine and terfenadine), but the inhibition of scratching was



Animal Models of Itch



41



incomplete at a dose that almost completely abolished plasma extravasation.

H2 receptor antagonists (famotidine and ranitidine) inhibited the scratching

without affecting plasma extravasation. These findings suggest that histamine is a key mediator in itch associated with immunoglobulin E-dependent

allergic reactions.

Mosquito bites frequently cause allergic skin reactions and itching in

humans. When ICR mice were exposed to mosquitos (Aedes albopictus), the

initial mosquito bites did not elicit scratching but repeated bites (twiceweekly exposure) led to a gradual increase in scratching (39). In mice

receiving repeated injections of an extract from the mosquito salivary gland,

the first mosquito bites elicited marked scratching and plasma extravasation, suggesting that mosquito bite–induced scratching resulted from an

immediate allergic reaction. Terfenadine did not affect mosquito biteinduced scratching at a dose that almost completely abolished plasma

extravasation and markedly suppressed histamine-induced scratching (39).

The reason why histamine has a larger role in itch-associated responses

following passive cutaneous anaphylaxis compared to mosquito bites is

unclear. The amount of histamine released by mosquito bites might be too

small to cause itching. Taken together, these findings suggest the presence of

histamine-mediated and histamine-independent mechanisms of immediate

allergic itching.



V.



MODEL OF DRY SKIN–ASSOCIATED ITCH



Skin dryness is apparent in several pruritic skin diseases, such as senile

xerosis, seasonal xerosis in winter, and atopic dermatitis. It may be also

associated with itch in patients with renal failure or cholestasis (5,6). Daily

treatment of mouse skin with a mixture of acetone and ether followed by tape

stripping led to an increase in spontaneous scratching (40). This treatment

disrupts the cutaneous barrier and decreases the hydration of the stratum

corneum during the initial 2 days, and resulted in a gradual increase in spontaneous scratching from days 3 to 5. The treatment did not affect the number

of total and degranulated mast cells in the skin and increased spontaneous

scratching in mast cell-deficient mice, suggesting that scratching was independent of cutaneous mast cells.



VI.



MODELS OF CHRONIC ITCH



Potentially promising models include the ‘‘itchy’’ (NC/jic) mouse (41,42),

the hairless guinea pig (17), and neonatal capsaicin treatment in rats (43).



42



Carstens and Kuraishi



A potential animal model of chronic itch is neonatal treatment of rats

with capsaicin, which destroys a substantial fraction of unmyelinated primary

afferent fibers and leads to a significant elevation in spontaneous scratching

and associated skin damage (43). Administration of morphine resulted in a

significant delayed increase in scratching in these animals, whereas naloxone

significantly decreased scratching. Control animals receiving vehicle as neonates showed little spontaneous scratching that was not influenced by

morphine.

Another potential chronic itch model is the NC/jic mouse (41,42).

Within 2–6 months after birth, the majority of these animals spontaneously

develop skin lesions (eczema, bleeding, and alopecia) associated with

excessive scratching. They scratched the face and the rostral part of the

body all day long. Under 12-hr light/dark conditions, there was a greater

number of scratching bouts during darkness than during light, but a

circadian rhythm was not apparent (Fig. 2). Using differential gene display

analysis, the animals that developed skin lesions expressed an increased level

of mRNA for myocyte-specific enhancer-binding factor compared to animals that did not develop skin lesions and did not show increased scratching

(41). A subsequent study of these animals revealed that the spontaneous

scratching was reduced by distraction and naloxone, suggesting that it was

itch-related (42). 5-HT, but not histamine or substance P, elicited scratching

when given i.d. (42). 5-HT antagonists reduced the 5-HT-evoked, but not

spontaneous, scratching, indicating that 5-HT may not play a significant role

in mediating spontaneous scratching. It was also shown that the animals did

not develop skin lesions, scratching, or increased plasma immonoglobulin G

levels when reared in a specific pathogen-free environment. However, when

transferred to a conventional environment, the animals developed all of

these symptoms within about 4 weeks. These results suggest that the NC/jic

mouse may represent an animal model of itch associated with chronic

dermatitis, affording the possibility to develop novel antipruritic treatment

strategies for clinical cases of itch that are poorly treated by antihistamines

or other drugs.



VII.



MODELS OF CENTRALLY MEDIATED ITCH



As mentioned above, the microinjection of opioids and other substances via

the intracerebroventricular (11,12,28), intramedullary (26,27), or intrathecal

route (8,10,29) elicits facial scratching behavior. It has been suggested that

scratching induced by central opioid administration may represent central

itch, presumably via activation of central itch signaling pathways as dis-



Animal Models of Itch



43



Figure 2 Spontaneous scratching of NC mice. Small magnets were worn just above

the ankle of both hindlimbs of NC mice (20–23 weeks old) and the limb movement

was automatically monitored using a scratching counting system (NS-SC01; Neuroscience, Inc., Osaka, Japan). The height of the columns represents the number of

scratch bouts per hour. Results are expressed as the mean F S.E.M. of four animals.



cussed further below. In mice, facial scratching was elicited by intracerebroventricular administration of morphine and the A-receptor agonist,

DAMGO (but not y-receptor or n-receptor agonists), in a dose-dependent

manner, and morphine-evoked scratching was reduced or abolished by distraction (experimenter present during data collection) and naloxone (12).

These data suggest that scratching may be itch-related and that it is mediated by a A-opiate receptor.

It should be noted, however, that a variety of other substances given via

the intracerebroventricular route elicit scratching and other grooming

behaviors, including substance P (44,45), ACTH (46), and neuropeptides

such as TRH, bombesin, neurotensin, and neuromedin (28,47–49). However,

the neuropeptide-induced scratching was antagonized by naloxone (47) in

support of a role for central opioid receptors.

The idea that opioids activate a central itch signaling pathway is

relevant to the pruritus of cholestasis, which is associated with an increase in

circulating opioids and can be relieved, to some degree, by naloxone (5).

Rats with experimental biliary stenosis exhibited naloxone-sensitive analgesia (50) and increased hepatic concentrations of opioids and mRNA for

preproenkephalin (51,52), although they did not show increased scratching



44



Carstens and Kuraishi



(5). Moreover, intramedullary microinjection of plasma extracts from

cholestasis patients with pruritus elicited a naloxone-reversible facial

scratching in monkeys (53). These data suggest that chronic itch suffered

under certain systemic diseases may involve the activation of central itch

pathways by circulating pruritic agents such as opioids.



VIII.



NEURAL MECHANISMS OF ITCH



A major benefit of an animal model of itch is that it affords the opportunity

to investigate the underlying neural mechanisms. Two main theories have

been proposed to explain itch. Specificity theory proposes the existence of

specific ‘‘itch’’ receptors linked to an itch signaling sensory pathway. Intensity theory postulates that itch and pain are signaled by a common population of neurons responsive to both pruritic and algesic stimuli. Itch would

be signaled by a low firing rate and pain by a higher firing rate in such a

nonspecific sensory pathway.

Recent studies have provided evidence favoring itch specificity. Human

microneurographic studies have uncovered a class of mechanically insensitive cutaneous receptors with slowly conducting unmyelinated afferent fibers

that respond to cutaneous histamine over a time course that closely matches

that of concomitant itch sensation (54). Subsequently, a subpopulation of

spinothalamic tract neurons in the superficial dorsal horn (lamina I) with

similar response characteristics was identified in the cat (55). A small number

of the latter neurons responded selectively to histamine but not the algesic

agent, mustard oil, whereas others responded nonselectively to both histamine and mustard oil.

The existence of an itch-specific sensory pathway is further supported

by intraneural microstimulation experiments. Stimulation near the axon of

one or a few polymodal nociceptors elicits a sensation of pain that increases

with stimulus frequency, but does not become itch at low frequencies (56).

Conversely, microstimulation at some intraneural sites elicits itch that

increases in intensity with stimulus frequency, but never becomes painful

at high stimulus frequencies (57).

Although there may exist a small population of itch-selective spinal

cord neurons, superficial dorsal horn neurons more commonly respond to

both algesic and pruritic stimuli. Using i.d. histamine as a search stimulus,

Jinks and Carstens (19,58) identified nociceptive-specific and wide dynamic

range (WDR)-type neurons in the rat superficial dorsal horn that responded

to subsequent i.d. histamine, as well as a variety of other suspected pruritic

(5-HT) and algesic chemicals (capsaicin, mustard oil, and nicotine). Some of



Animal Models of Itch



45



these neurons were insensitive to mechanical stimuli, but still responded to

pruritic as well as algesic chemical stimulation. Of note, the responses of such

neurons to i.d. 5-HT were prolonged, often lasting 20–40 min (Fig. 3a). This

matches the time course of scratching elicited by i.d. 5-HT in behavioral

studies (Fig. 3b) (19). Furthermore, both behavioral scratching and neuronal

responses showed significant tachyphylaxis to repeated i.d. 5-HT injections.

These correlations suggest that the chemically nonselective superficial dorsal

horn neurons might carry information relevant to itch. How the nervous

system discriminates between itch and pain based on input from such

nonselective neurons is not clear, but might conceivably depend on firing

rate. In this regard, superficial dorsal horn neuronal responses to noxious

stimuli (e.g., capsaicin) exhibited a higher-frequency firing rate compared to

responses to 5-HT (19).

Although the role that WDR and nociceptive-specific dorsal horn

neurons play in itch is far from clear, they exhibit some properties consistent

with itch sensation. The histamine-evoked responses of dorsal horn neurons

are suppressed by mechanical (rub and scratch) and noxious heat stimuli as



Figure 3 Prolonged responses of superficial dorsal horn neurons to i.d. 5-HT compared with scratching. (a) Averaged peristimulus–time histogram (PSTH; bin width:

1 sec) of neuronal firing in response to i.d. microinjection of 5-HT (60 mM; 1 Al) at

arrow. Error bars on PSTH are omitted for clarity. Single units were recorded in

pentobarbital-anesthesized Sprague–Dawley rats. 5-HT was microinjected i.d. into

neuronal receptive field on ipsilateral hindpaw. Inset shows recordings sites (dots)

compiled on section through the L5 dorsal horn. (b) Graph plots mean number of

hindlimb scratching bouts (error bars: S.E.M.) directed toward the site of injection of

5-HT in the nape of Sprague–Dawley rats. Scratching bouts were counted in 2-min

intervals. (From Ref. 19.)



46



Carstens and Kuraishi



well as cooling (58–60), consistent with the antipruritic effects of these counterstimuli (61–65). The histamine-evoked responses of superficial dorsal horn

neurons were often facilitated by low doses of morphine given intrathecally,

although higher morphine doses uniformly depressed neuronal responses to

histamine as well as noxious heat (58). It should be borne in mind, however,

that histamine may not be pruritic in Sprague–Dawley rats and future studies

of itch mechanisms should focus instead on 5-HT, which may be pruritic in

this species.



REFERENCES

1.



Koltzenburg M, Handwerker HO, Torebjork HE. The ability of humans to lo¨

calise noxious stimuli. Neurosci Lett 1993; 150:219–222.

2. Berendsen HHG, Broekkamp CLE. A peripheral 5-HT1D-like receptor involved in serotonergic induced hindlimb scratching in rats. Eur J Pharmacol

1991; 194:201–208.

3. Kuraishi Y, Nagasawa T, Hayashi K, Satoh M. Scratching behavior induced

by pruritogenic but not algesiogenic agents in mice. Eur J Pharmacol 1995; 275:

229–233.

4. Krajnik M, Zbigniew Z. Understanding pruritus in systemic disease. J Pain

Symptom Manage 2001; 21:151–168.

5. Bergasa NV, Mehlman JK, Jones EA. Pruritus and fatigue in primary biliary

cirrhosis. Bailliere’s Best Pract Res Clin Gastroenterol 2000; 14:643–655.

6. Murphy M, Carmichael AJ. Renal itch. Clin Exp Dermatol 2000; 25(2):103–

106.

7. DeCastro-Costa M, Gybels J, Kupers R, Van Hees J. Scratching behaviour in

arthritic rats: a sign of chronic pain or itch? Pain 1987; 29:123–131.

8. Bossut D, Frenk H, Mayer DJ. Is substance P a primary afferent neurotransmitter for nociceptive input? II. Spinalization does not reduce and intrathecal morphine potentiates behavioral responses to P substance. Brain Res

1988; 445:232–239.

9. Frenk H, Bossut D, Urca G, Mayer DJ. Is substance P a primary afferent

neurotransmitter for nociceptive input? I. Analysis of pain-related behaviors

resulting from intrathecal administration of substance P and 6 excitatory compounds. Brain Res 1988; 455:223–231.

10. Wilcox GL. Pharmacological studies of grooming and scratching behavior

elicited by spinal substance P and excitatory amino acids. Ann NY Acad Sci

1988; 525:228–236.

11. Konigstein H. Experimental study of itch stimuli in animals. Arch Dermatol

¨

Syphilol 1948; 57:829–849.

12. Tohda C, Yamaguchi T, Kuraishi Y. Intracisternal injection of opioids induces

itch-associated response through mu-opioid receptors in mice. Jpn J Pharmacol

1997; 74:77–82.