Journal of Neurophysiology

Effects of TRPA1 Agonists Mustard Oil and Cinnamaldehyde on Lumbar Spinal Wide-Dynamic Range Neuronal Responses to Innocuous and Noxious Cutaneous Stimuli in Rats

Austin W. Merrill, Jason M. Cuellar, Justin H. Judd, Mirela Iodi Carstens, E. Carstens


Mustard oil [allyl isothiocyanate (AITC)] and cinnamaldehyde (CA), agonists of the ion channel TRPA1 expressed in sensory neurons, elicit a burning sensation and heat hyperalgesia. We tested whether these phenomena are reflected in the responses of lumbar spinal wide-dynamic range (WDR) neurons recorded in pentobarbital-anesthetized rats. Responses to electrical and graded mechanical and noxious thermal stimulation were tested before and after cutaneous application of AITC or CA. Repetitive application of AITC initially increased the firing rate of 52% of units followed by rapid desensitization that persisted when AITC was reapplied 30 min later. Responses to noxious thermal, but not mechanical, stimuli were significantly enhanced irrespective of whether the neuron was directly activated by AITC. Windup elicited by percutaneous or sciatic nerve electrical stimulation was significantly reduced post-AITC. These results indicate that AITC produced central inhibition and peripheral sensitization of heat nociceptors. CA did not directly excite WDR neurons, and significantly enhanced responses to noxious heat while not affecting windup or responses to skin cooling or mechanical stimulation, indicating a peripheral sensitization of heat nociceptors.


Mustard oil [allyl isothiocyanate (AITC)] and cinnamic aldehyde (CA) are agonists of the transient receptor potential (TRP) channel TRPA1 (Bandell et al. 2004; Bautista et al. 2006; Jordt et al. 2004). When applied to skin, they elicit burning pain, thermal hyperalgesia, and mechanical allodynia (Koltzenburg et al. 1992; Namer et al. 2005). In the oral or nasal mucosa, AITC and CA elicit burning irritation that decreases (desensitizes) across trials of repeated application (Brand and Jacquot 2002; Prescott and Swain-Campbell 2000; Simons et al. 2003), as well as heat hyperalgesia (Albin et al. 2007). Lingual application of AITC or CA excites neurons in the trigeminal subnucleus caudalis (Vc) (Carstens and Mitsuyo 2005; Simons et al. 2004; Zanotto et al. 2007). AITC excitation of Vc neurons exhibits a desensitizing temporal pattern while sensitizing responses to noxious heat (Simons et al. 2004). We presently tested whether spinal wide-dynamic range (WDR)–type dorsal horn neuronal responses to repeated cutaneous application of AITC or CA similarly exhibit a desensitizing pattern and whether their responses to mechanical and noxious thermal stimuli are enhanced after application of these chemicals, consistent with human psychophysical observations. We presently focused on WDR neurons for two reasons. First, WDR neurons respond to innocuous mechanical as well as noxious thermal stimuli, allowing assessment of the effects of AITC and CA on both types of responses within the same neuronal population, which would not be possible with nociceptive-specific (NS) neurons. Second, available evidence indicates that WDR neurons are sufficient for pain perception (Mayer et al. 1975; Price and Mayer 1975) and they are well suited to encode the intensity of noxious heat (Maixner et al. 1986). In addition, to address whether the effects of AITC and CA occur at a central (spinal) site we tested whether windup evoked by repeated C-fiber–strength electrical stimuli was affected after cutaneous application of AITC or CA.



In all, 64 adult male Sprague–Dawley rats (Harlan, San Diego, CA) weighing 430–595 g were used. The experimental procedures were approved by the UC Davis Animal Use and Care Advisory Committee. Rats were housed in a room with controlled temperature (22 ± 1°C) and lighting (lights on from 08:00 to 20:00 h), with unrestricted access to food and water.


Rats were anesthetized with sodium pentobarbital (65 mg/kg, intraperitoneal induction), and supplemented as needed so that a strong tail and paw pinch failed to evoke a withdrawal response. A tracheostomie tube was implanted, the jugular vein or lateral tail vein was cannulated with PE-50 tubing for maintenance of pentobarbital anesthesia, and wound clips were used to close the incision. Core body temperature was monitored rectally using a BAT-12 thermometer (Physitemp, Clifton, NJ) and maintained at 37 ± 0.2°C with a lamp and heating pad. During recording, anesthesia was maintained by iv pump infusion (10–20 mg kg−1 h−1) of pentobarbital.

The L6–S1 intervertebral space was identified by palpation of the spinous processes and the posterior superior iliac spine, and a midline skin incision was made from approximately L6 to T11 spinous processes. The paraspinous muscles were dissected free from the L2–T12 spinous processes on both sides, and the transverse processes were exposed by scraping off attached connective tissue. L1 and T13 spinous processes were cut and removed, and a bilateral laminectomy was performed at both levels under a dissection microscope (Wild M5) with micro-rongeurs. The dura was removed and warm agar was poured over the spinal cord. Vertebral clamps on the transverse processes of T12 and L2 were used to stabilize the animal in a stereotaxic frame (Kopf Instruments, Tujunga, CA). Needle electrodes were placed in both forelimbs to monitor electrocardiographic (ECG) activity.

In 21 rats the sciatic nerve was exposed for electrical stimulation. A 3.5-cm midline incision was made in the posterolateral upper hindlimb at the level of the biceps femoris muscle. The semitendinosus and biceps femoris muscles were separated using blunt dissection techniques and the semimembranosus muscle was reflected with fine forceps and microscissors. The left sciatic nerve was isolated above the bifurcation of the tibial and common peroneal nerve and a 3 × 15-mm strip of paraffin wax was wrapped beneath the nerve. After placement in the stereotaxic frame, the sciatic nerve was positioned onto a hooked parallel bipolar electrode (FHC, Bowdoinham, ME) and bathed in warmed mineral oil intestinal lubricant. During trials with electrical stimulation a neuromuscular blocker (pancuronium, 0.1 mg iv) was administered and the animal was ventilated using a positive-pressure pump (Harvard Apparatus, Holliston, MA). End-tidal CO2 was monitored by a Datex 254 gas analyzer (Datex-Ohmeda, Tewksbury, MA) and maintained between 3.0 and 4.0% by adjustment of tidal volume and/or respiratory rate.

Stimulation and recording

An 8- to 11-MΩ Teflon-coated tungsten microelectrode (FHC) was advanced into the dorsal horn of the spinal cord using a hydraulic microdrive (Kopf Instruments) to record single-unit activity of dorsal horn neurons. Units isolated for study were always at depths <1 mm. Action potentials were amplified and displayed by conventional means, and sent to a computer for storage and analysis using a Powerlab interface and Chart 5.0 software (AD Instruments, Grand Junction, CO) and to another running custom software (Spike; Forster and Handwerker 1990).

Single units were searched for and isolated using innocuous mechanical stimulation of the plantar surface of the ipsilateral hindpaw. Units were chosen with receptive field areas on the plantar surface of the toes, corresponding to approximately L5 spinal cord based on prior dermatomal mapping studies (Takahashi and Nakajima 1996; Takahashi et al. 1994, 1995). Of these, only units that responded to graded nonnoxious (brushing, 4–12 g von Frey) and noxious (76 g von Frey, pinch) mechanical and noxious thermal (42, 46, and 50°C) stimuli were considered for further study (WDR neurons).

For mechanical stimulation, a series of graded von Frey filaments (4, 12, and 76 g) were applied in ascending order. Each stimulus was applied for 10 s at a 1.5-min interstimulus interval to the center of the receptive field. In several experiments the receptive field was also lightly stimulated by brushing with cotton in a back-and-forth (∼1 Hz) motion for 10 s.

Thermal stimuli were delivered to the center of the receptive field using a Peltier device (13-mm diam; Physitemp NTE-2A) mounted to a micromanipulator. The thermode temperature was controlled by computer, and stimuli were delivered at a rate of about 12.5°C/s from an adapting temperature of 35°C with an accuracy of ±0.1°C. The temperature at the thermode–skin interface was continually monitored by a separate thermocouple (IT-21; Physitemp) connected to a BAT-12 thermometer (Physitemp) and displayed along with the action potential data and ECG using Powerlab. In the initial studies with AITC (n = 27 rats), 46 and 50°C stimuli were used, whereas in later studies with CA (n = 14 rats), 42, 46, and 50°C stimuli were used. In several experiments a cooling stimulus (from 35 to 10°C over a 30-s period) was also delivered.

In 14 experiments electrical stimulation of the sciatic nerve was performed. In these plus an additional 5 experiments, the hindpaw receptive field was also stimulated electrically by percutaneous needle electrodes. Constant-current stimulus trains of 16 pulses (0.7-ms duration) at 1 Hz were delivered with an S48 stimulator (Grass, West Warwick, RI). Latency ranges for C-fiber and afterdischarge activity were as described previously (Mitsuyo et al. 2006) and we presently counted all C-fiber and afterdischarge activity occurring in the 100- to 1,000-ms latency period The stimulus intensity was adjusted for each unit to be threefold that of the C-fiber threshold.

Chemical application

Ten minutes after completion of the mechanical and thermal (and electrical, if included) stimulation series, 60 s of baseline activity was recorded before application of either AITC, CA, or mineral oil (control). AITC (allyl isothiocyanate; 2 μl, 75% in mineral oil; Fluka, St. Louis, MO), cinnamaldehyde (CA; in mineral oil; Sigma–Aldrich), or mineral oil (n = 10) was then topically applied to the center of the receptive field area at 1-min intervals for 10 min using a Hamilton microsyringe attached to PE-50 tubing. This application method was chosen to match that used previously in studies of Vc neurons activated by lingual stimuli (Simons et al. 2004). Eleven minutes after the last AITC or CA droplet was applied, the von Frey and thermal stimulation series (followed by electrical stimulation, when done) were repeated. The thermal probe was replaced at the same hindpaw location using millimeter coordinates on the micromanipulator to which the thermal probe was mounted. Thermal stimulation was always initiated 2 min after replacement of the thermal probe. Application of AITC or CA was repeated after a 10-min wait period. Thus 30 min had passed between the last AITC or CA application of trial 1 and the first AITC or CA application of trial 2. On completion of the experiment the animal was killed by overdose of pentobarbital, administered intravenously.

Data analysis

The spontaneous firing rate was calculated as the sum of the total number of action potentials that occurred for 30 or 60 s before each stimulus. Responses to von Frey and thermal stimuli were quantified by summing the total number of action potentials recorded during the 10-s stimulus period, and subtracting the spontaneous firing rate per 10 s (30 s total/3). The afterdischarge was quantified as the total number of action potentials during the 30 s after the offset of the stimulus. Spontaneous firing, evoked responses, and afterdischarge to each mechanical and thermal stimulus were compared pre- versus posttreatment for each treatment group (AITC, CA, and mineral oil) using paired t-test. Responses to AITC, CA, and mineral oil were quantified by summing the total spikes during the 60-s interval after each application. Each sum was compared with the sum of the total spikes during the 60 s preceding the first application (baseline) using univariate ANOVA with post hoc Dunnett's two-sided t-test. A P value of <0.05 was taken to be significant. Statistical analyses were performed using SPSS 9.0 software. All data are means ± SE unless otherwise noted.

Data with electrical stimulation (windup) were analyzed in four ways. ANOVA was performed on all units' responses across the 16 stimuli followed by post hoc least significant differences (LSD) tests. The area under the curve (AUC) for windup was calculated by summing all action potentials in the 100–1,000 latency window after each of the 16 stimuli. Absolute windup was calculated by subtracting 16 times the initial response from AUC windup. The slope of windup was calculated by linear regression of the initial seven responses. Each measure of windup was compared across conditions (before and after the first and second applications of AITC, CA, or mineral oil) using paired tests with P < 0.05 considered to be statistically significant.


Unit sample

All units were of the WDR type and responded to noxious thermal stimulation in the 42–50°C range, as well as to innocuous mechanical stimuli. An example of a WDR unit's responses to thermal and mechanical stimuli, and AITC, is shown in Fig. 1. The mean depth was 543.2 ± 73 (SD) μm, which corresponds to the middorsal horn.

FIG. 1.

Example of allyl isothiocyanate (AITC)–sensitive lumbar spinal wide-dynamic range (WDR) neuron. A: shaded area shows extent of mechanosensitive receptive field on lateral hindpaw. Arrow: site of AITC application. B and C: raw spike traces of response to 46 and 50°C heat stimuli before mustard oil application. DF: responses to graded mechanical stimulation at indicated von Frey bending forces. G: spike trace of activity during repeated application of AITC (75%, 2-μl droplets) at 1-min intervals (arrows). Note progressive increase in firing rate with desensitization after the 5th stimulus. HL: spike traces of responses to graded noxious heat (H, I) and mechanical stimulation (JL) after sequential application of AITC. Note increased response to heat but not mechanical stimulation.

Response to AITC

In the first series of experiments, 27 units were tested for responses to repeated application of AITC and 14 (52%) responded. The example in Fig. 1G shows a buildup of firing to the initial AITC stimuli. The mean responses are shown in Fig. 2 for units excited by repeated application of AITC [Fig. 2, left, black peristimulus time histograms (PSTHs)] and for those unaffected by AITC (gray PSTH). For the responsive units, the mean firing rate during the first three stimulus applications was significantly greater compared with pre-AITC baseline but then declined to a level that was not significantly different from baseline (Fig. 2, left). After a 30-min rest period, AITC was reapplied in the same manner. Although there was a trend toward an increased firing rate, this did not reach statistical significance relative to pre-AITC baseline (Fig. 2, right) and in any event was lower compared with the first trial. Repeated application of vehicle (mineral oil) in an identical manner, did not result in any significant change in firing rate compared with premineral oil baseline in a separate group of 10 units (data not shown).

FIG. 2.

Desensitization of responses to repeated application of AITC. Shown are averaged peristimulus time histograms (PSTHs, 1-s bin width) of unit firing during repeated application of AITC (75%, 2-μl droplets) at 1-min intervals (arrows) for 10 min to the center of the receptive field area on the ipsilateral hindpaw. Black PSTHs: 14 WDR units that exhibited increased firing during the initial application of AITC. Gray PSTHs: 13 units unresponsive to AITC. Error bars are omitted for clarity. Left-hand PSTHs: responses to the first trial of sequential AITC application. Right-hand PSTHs: responses to second trial of sequential AITC application starting 30 min after the end of the first trial. *, **, ***: significantly different compared with the initial 60-s baseline period before the first application of AITC (P < 0.05, P < 0.005, P < 0.0005, respectively).

AITC sensitization of responses to noxious heat

All units were tested for responses to graded (46 and 50°C) heat before and after AITC. Figure 1 shows one unit's responses before AITC (Fig. 1, B and C) and their marked enhancement post-AITC (Fig. 1, H and I). Figure 3 A shows averaged responses to heat stimuli before (gray PSTHs) and after AITC (black PSTHs) for units that were directly activated by AITC. Responses to both 46 and 50°C were significantly enhanced post-AITC. Units unresponsive to AITC similarly exhibited a significant enhancement of heat-evoked responses post-AITC (Fig. 3B). Vehicle (mineral oil) application had no effect on heat-evoked responses in a separate group of 10 units (Fig. 3C).

FIG. 3.

AITC sensitization of WDR responses to noxious heat stimulation. A: responses to 46 and 50°C heat stimuli, before (gray PSTHs) and 11 min after the end of sequential application of AITC (black PSTHs) for AITC-sensitive WDR units. Pre- and post-AITC responses are aligned to the heat stimulus (indicated by horizontal bars). Error bars omitted for clarity. *: post-AITC response significantly different from pre-AITC (P < 0.05). B: format as in A for 13 AITC-insensitive WDR units. Note that heat-evoked responses were significantly larger post-AITC. C: format as in A for 10 separate WDR units receiving sequential application of mineral oil (vehicle control). Note the absence of difference in responses pre- vs. post–mineral oil application.

Lack of AITC effect on responses to mechanical stimulation

WDR units typically exhibited graded responses to increasing bending forces of punctate von Frey stimuli (Fig. 1, DF), which were minimally affected post-AITC (Fig. 1, JL). Figure 4 shows averaged responses to graded mechanical stimuli before (gray PSTHs) and about 11 min after AITC (black PSTHs), for AITC-sensitive (Fig. 4A) and -insensitive units (Fig. 4B). There were no significant differences in responses pre- versus post-AITC. Similarly, responses pre- and postvehicle (mineral oil) application were not significantly different (Fig. 4C). Finally, responses of units to low-threshold brushing of skin in the center of the mechanosensitive receptive field with cotton were not significantly different pre- versus post-AITC application.

FIG. 4.

Averaged responses to mechanical stimuli after application of AITC or mineral oil (control) to the receptive field area. A: averaged PSTHs of responses of AITC-sensitive neurons to stimulation with von Frey filaments with bending forces of 4, 12, and 76 g stimuli, from left to right, pre- (gray PSTHs) and 10-min postapplication of AITC (black PSTHs). Error bars omitted for clarity. B: format as in A for mechanically evoked responses of AITC-insensitive neurons pre- and post-AITC application. C: format as in A for mechanically evoked responses of 10 different WDR neurons pre- and postapplication of mineral oil (vehicle control).

AITC depression of windup

In a separate series of experiments we investigated the effect of AITC on responses of WDR units to trains of electrical stimuli delivered at C-fiber intensity and 1-Hz frequency, delivered both by percutaneous electrodes placed within the receptive field and by direct stimulation of the trunk of the common sciatic nerve. We accepted units that exhibited a progressive increase in C-fiber responses (100- to 400-ms latency range) across stimulus trials (i.e., windup; n = 7; windup slope >0.7) as well as units exhibiting constant responses across trials (flat cells; n = 8; windup slope <0.7). For windup cells, the A-fiber response (0- to 100-ms latency range) did not wind up.

For cells exhibiting windup by sciatic nerve stimulation (Fig. 5A), there was an overall treatment effect [ANOVA, F(2,333) = 33.6; P < 0.001] with post-AITC 1 and 2 conditions, both significantly different from pre-AITC (P < 0.001). All measures of windup were reduced post-AITC 1 and 2, with AUC and absolute windup significantly so (paired t-test, P < 0.05). In the same units, there was a similar significant treatment effect for windup elicited by percutaneous electrical stimulation [F(2,333) = 4.2, P < 0.05] with post-AITC 1 and 2 significantly different from pre-AITC (P < 0.01, P < 0.05, respectively]. AUC windup evoked by percutaneous stimulation was significantly reduced post-AITC 1 and 2 (Table 1).

FIG. 5.

AITC inhibits windup. A: windup of WDR neurons elicited by sciatic nerve electrical stimulation (3 × C-fiber threshold, 1 Hz). Windup responses after the first and second applications of AITC were significantly different from pre-AITC [ANOVA, F(2,333) = 33.6; P < 0.001 for both]. B: flat-responding cells. Responses after the first and second applications of AITC were significantly different from pre-AITC [F(2,381) = 44.9, P < 0.001 for both].

View this table:

Measures of windup before application of AITC (pre-AITC) and after the first and second applications of AITC (post-AITC 1 and 2)

For units exhibiting flat responses to sciatic nerve stimulation, there was also a significant treatment effect [F(2,381) = 44.9, P < 0.001] with both post-AITC 1 and 2 significantly different from pre-AITC (P < 0.001 for both). These units exhibited significant reductions in AUC and absolute windup after the first and second applications of AITC, respectively (Table 1). The same units also exhibited flat responses to percutaneous stimulation, although there was no significant treatment effect [F(2,381) = 0.62, P > 0.05] and no significant differences in any measure of windup pre- versus post-AITC (Table 1).

Of the 15 flat and windup units tested, 7 responded directly to application of AITC and 8 were unresponsive to AITC.

CA did not excite WDR units

In a separate group of 14 units, repeated application of CA did not significantly affect any unit's firing rate.

CA sensitization of responses to noxious heat

Similar to AITC, responses to graded noxious heating were significantly enhanced post-CA. Figure 6 shows an individual example (compare AC with DF), and Fig. 7 A shows averaged responses that are overlaid to emphasize the progressive increase after the first (light gray PSTHs; post-CA 1) and second applications of CA (black PSTHs; post-CA 2) compared with pre-CA (dark gray PSTH). Mean responses after both the first and second trials of CA application were significantly different from pre-CA at each stimulus temperature (paired t-test, P < 0.05).

FIG. 6.

Example of cinnamaldehyde (CA) heat sensitization of individual WDR neuron. AC: spike traces (bottom) and PSTHs (1-s bin width) above, of responses to 42°C (A), 46°C (B), and 50°C (C) before CA. Inset in A shows receptive field on toes 3 and 4. Arrow: site of CA application. DF: responses to same series of graded noxious heat stimuli post-CA. Note sensitization of responses. Firing rate did not change during application of CA. GI: responses to graded mechanical von Frey stimuli: 4 g (G), 12 g (H), and 76 g (I), respectively, pre-CA. JL: responses to graded mechanical von Frey stimuli post-CA.

FIG. 7.

CA sensitization of WDR responses to noxious heat but not mechanical stimuli. A: averaged PSTHs of responses of 13 units to 42°C (left), 46°C (middle), and 50°C (right) heat stimuli (black bars), before (dark gray PSTH) and after the first (post-CA 1; light gray PSTH) and second (post-CA 2; black PSTH) application of CA. Error bars omitted for clarity. *: significant difference between pre-CA and both post-CA 1 and post-CA 2 (P < 0.05). B: averaged PSTHs of same units in A to 4 g (left), 12 g (middle), and 76 g (right) mechanical stimuli (black bars), before and after CA (format as in A). CA did not appreciably affect the responses of 2 WDR units tested to cotton brush stimuli.

Lack of CA effect on responses to mechanical stimulation

Figure 6 shows a typical example in which responses to graded von Frey stimuli were similar before (Fig. 6, GI) and after CA (Fig. 6, JL). Figure 7B shows averaged responses to graded von Frey stimuli with no significant change after the first or second application of CA compared with pre-CA.

Lack of CA effect on responses to cooling

Thirteen units were tested for responses to cooling before and after application of CA (Fig. 8). Nine units exhibited robust responses, whereas 4 responded weakly or not at all. The mean response of all 13 units during skin cooling (Fig. 8, left) was not significantly affected (paired t-test, P > 0.05) after either of two successive trials of CA application (Fig. 8, middle and right).

FIG. 8.

CA does not affect responses to noxious cold. Shown are averaged PSTHs (error bars: SE) of 13 units to cooling receptive field, before (left) and after first (middle) and second (right) application of CA. Trace above left-hand PSTH shows temperature recorded at skin–thermode interface. Error bars: SE. There was no significant difference among conditions (P > 0.05, paired t-test).

Lack of CA effect on windup

Likewise, application of CA had no effect on windup. Figure 9 shows that mean windup curves of 5 units to sciatic nerve stimulation were similar before and after two successive trials of CA application, and not significantly different [ANOVA, F(2,237) = 0.56, P > 0.5]. No measure of windup evoked by percutaneous or sciatic nerve stimulation was significantly affected after CA (Table 2). As a vehicle control for both CA and AITC, windup elicited by percutaneous electrical stimulation was not significantly affected after two successive trials of application of mineral oil in an identical manner [Fig. 9B; ANOVA, F(2,237) = 0.08, P > 0.5] and there were no significant differences in any parameter of windup pre- versus post–mineral oil (Table 2).

FIG. 9.

CA does not affect windup. A: graph as in Fig. 6 plotting number of impulses evoked by repetitive C-fiber stimulation, before (pre-CA) and after 2 separate sequential series of applications of CA to the receptive field (post-CA1 and 2). ANOVA revealed no significant difference pre- vs. post-CA. B: graph as in A for a separate group of units in which mineral oil (vehicle control) was sequentially applied in the same manner as mustard oil or CA.

View this table:

Measures of windup before and after two successive applications of cinnamaldehyde (CA)


A main finding of this study is that both AITC and CA sensitized dorsal horn WDR neuronal responses to noxious heat while not significantly affecting their responses to other stimuli. This heat sensitization was observed irrespective of whether the AITC or CA directly excited the neuron. Furthermore, electrically evoked responses of the WDR units were either unaffected, or reduced, after application of the irritant to the skin. These data therefore support a peripheral site of heat sensitization by the TRPA1 agonists and argue against a central sensitizing action.


Roughly 50% of WDR dorsal horn units responded directly to topical application of AITC with an initial increase in firing that desensitized over a 3-min period despite continued application of AITC. A previous study reported a similar fraction (53%) of mainly WDR dorsal horn neurons to be directly activated by 4% AITC applied adjacent to the mechanosensitive receptive field in decerebrate spinalized rats (Woolf and King 1990). The desensitizing response pattern observed presently was similar to that of responses of neurons in trigeminal subnucleus caudalis (Vc) to lingual application of AITC (Simons et al. 2004; Zanotto et al. 2007), as well as the desensitizing temporal pattern of irritancy ratings elicited by lingual AITC in humans (Simons et al. 2003). The lack of response of the other half of the WDR units to topical AITC suggests that our application strategy using a high concentration (75%) and low volume (2 μl) did not have nonspecific excitatory or toxic effects. On reapplication, AITC did not evoke a significant increase in WDR neuronal firing, consistent with self-desensitization reported for AITC irritancy on the tongue that lasts >10 min in humans (Simons et al. 2003). However, Vc neurons overcome AITC self-desensitization more quickly (Simons et al. 2004), possibly due to a more rapid clearance rate in the oral mucosa compared with hindpaw skin. Consistent with this, Green (1998) showed that a 10-fold higher concentration of capsaicin is required on facial skin versus tongue to elicit equivalent burning sensations and that the time course of desensitization was much slower on the face than on the tongue.

The mechanism underlying AITC self-desensitization could involve a peripheral or central site of action. Peripherally, repeated application of AITC may lead to desensitization of TRPA1 expressed in nociceptive endings. AITC self-desensitization was recently reported to occur by a calcium- and calcineurin-independent mechanism in an in vitro assay of peptide release from skin–nerve biopsies (Ruparel et al. 2007). Alternatively (or in addition), central inhibition might contribute to the reduced response of WDR neurons to repeated application of AITC. This is supported by our observation that windup elicited by electrical stimulation of the sciatic nerve was significantly attenuated post-AITC. However, such a proposed central inhibition was insufficient to prevent AITC and CA enhancement of WDR neuronal responses to noxious heat.

CA did not directly excite the WDR units recorded presently, whereas lingual application of CA readily excited Vc neurons (Zanotto et al. 2007), presumably due to the lower diffusion barrier presented by the lingual epithelium compared with hindpaw skin. It is likely that the amount of CA that reached the sensory nerve endings in the hindpaw epithelium was insufficient to elicit action potentials, although it was sufficient to induce heat sensitization, presumably by the same peripheral mechanism as suggested for AITC (see following text). The inability of CA to excite dorsal horn cells might explain its lack of effect on windup. In contrast, AITC strongly excited about 50% of WDR cells, with a resultant depression of electrically evoked neuronal responses presumably by activation of spinal inhibitory circuits by the stronger afferent drive.

AITC and CA sensitization of heat-evoked responses

A main finding was that AITC sensitized WDR neuronal responses to noxious heat, irrespective of whether AITC directly activated the unit. CA also sensitized responses to heat even though it did not directly excite WDR units. Previous studies of WDR neurons in lamina V of mice reported significant enhancement of neuronal responses to 40°C (but not 45 or 49°C) (Eckert 3rd et al. 2006; Mazario and Basbaum 2007), and a significant enhancement of afterdischarge responses to 41 and 45°C (Martin et al. 2004) after application of AITC (10%, ∼60 μl) to the hindpaw. Lamina I NS neuronal responses to heat were also enhanced post-AITC (Eckert 3rd et al. 2006; Mazario and Basbaum 2007). Our results are generally consistent, in that AITC more strongly enhanced rat WDR neuronal responses to the lower stimulus temperatures (42 and 46°C) compared with the highest (50°C). Two previous studies reported no change in noxious heat-evoked responses of rat and mouse WDR neurons after AITC (50 or 100%) was applied adjacent to the receptive field (Pertovaara 1998; Weng et al. 2001). AITC sensitizes responses of mechanoheat-sensitive C-fiber afferents to noxious heating (Reeh et al. 1986). Because AITC did not enhance WDR neuronal responses to mechanical or electrical C-fiber stimuli, the most parsimonious explanation is that AITC applied within the cutaneous receptive field sensitized peripheral nociceptors to result in primary hyperalgesia. TRPA1 is coexpressed with the heat-sensitive channel TRPV1 in primary sensory neurons (Story et al. 2003), so AITC enhancement of nociceptor responses to noxious heat might involve a cellular mechanism by which activation of TRPA1 enhances the thermal sensitivity of TRPV1. Another possibility is that AITC causes release of inflammatory mediators that in turn lower the thermal activation threshold of TRPV1 (Chuang et al. 2001; Sugiura et al. 2002), resulting in the observed enhancement of responses, particularly to 42 and 46°C.

Mechanically and electrically evoked responses

Neither AITC nor CA affected WDR neuronal responses to graded pressure or light brushing of the mechanosensitive receptive field. This is partially consistent with recent studies showing that AITC significantly enhanced murine deep dorsal horn WDR neuronal responses to only the weakest mechanical stimulus (130 mN) while having no significant effect on responses to stronger stimuli (Mazario and Basbaum 2007). However, other studies have shown significant enhancement of neuronal responses to innocuous mechanical stimuli, and expansion of receptive fields, in rats and mice after application of AITC adjacent to the mechanosensitive receptive field of spinal WDR or NS neurons (Pertovaara 1998; Weng et al. 2001; Woolf and King 1990; Woolf et al. 1994). Such enhancement was seen in decerebrate spinalized rats (Woolf and King 1990), although another study reported the AITC-induced mechanical enhancement to be significantly attenuated in spinalized rats (Pertovaara 1998), implicating involvement of descending facilitatory pathways. The present lack of effect of AITC or CA on mechanically evoked responses of WDR neurons indicates that our method of intermittent application of small (2 μl) volumes of these agents did not produce sufficient afferent drive to engage segmental or suprasegmental pronociceptive networks.

Neither AITC nor CA enhanced the electrically evoked responses of WDR neurons, nor did CA affect responses to cooling. If anything, AITC may have activated central inhibitory mechanisms as indicated by the significant attenuation of measures of windup elicited by sciatic nerve stimulation. However, any central depressant effect was not sufficient to obscure the clear-cut enhancement of noxious heat-evoked responses observed presently.


This work was supported by State of California Tobacco-Related Disease Research Program Grant 11RT-0053 and National Institutes of Health Grants DE-13685 to E. Carstens and Initiative for Maximizing Student Diversity R25 GM-56765 to J. M. Cuellar.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract