Protein Kinase C γ Interneurons Mediate C-fiber–induced Orofacial Secondary Static Mechanical Allodynia, but Not C-fiber–induced Nociceptive Behavior

TISSUE injury is associated with increased pain sensitivity not only at the site of tissue damage (primary hyperalgesia) but also in surrounding normal skin (secondary hyperalgesia). Whereas sensitization of nociceptive afferents largely accounts for primary hyperalgesia, secondary hyperalgesia is attributed to sensitization of central nociceptive neurons.¹  As a sign of central sensitization, secondary hyperalgesia is predictive of chronic postsurgical pain.^2–4  Chronic postsurgical pain is not only restricted to major surgeries, such as thoracic surgery and amputations, but can also be observed after simple procedures such as oral surgery, inguinal hernia repairs, and mastectomies.⁵  Identifying the mechanisms of secondary hyperalgesia is, therefore, becoming a public health priority. Interestingly, the area of secondary hyperalgesia varies between subjects but is highly reproducible within individuals, suggesting that it is a phenotypic characteristic.⁶  Magnetic resonance imaging reveals differences in brain morphology and noxious stimulation–induced neuronal activation between individuals developing large and small secondary hyperalgesia areas.⁷  This indicates that there are differences in injury-induced central sensitization according to phenotype and that a high sensitization response is a risk factor for chronic pain.

Surprisingly, little is known about such mechanisms. It is established that distinct peripheral afferents mediate the induction and expression of secondary hyperalgesia: transient receptor potential vanilloid 1 (TRPV1)-positive C-fiber input mainly contributes to induction, whereas A-fibers are involved in transmitting inputs that produce pain.^8–12  However, which A-fiber-mediated mechanical pathway is facilitated and how it is facilitated after a brief conditioning C-fiber input are still unclear.

Moreover, several pain symptoms are filed under the umbrella term of secondary hyperalgesia. Intradermal capsaicin generates a highly reproducible set of sensory changes: transient spontaneous pain followed by persisting mechanical hypersensitivity in the adjacent uninjured skin.^9,11,13  Such secondary hyperalgesia includes dynamic mechanical allodynia and hyperalgesia to punctate stimuli.^9,14,15  Moreover, because hyperalgesia to punctate stimuli is characterized by a leftward shift of the whole stimulus–response function,^9,10,16  it actually encompasses decreased pain thresholds and increased pain to suprathreshold stimuli, referred to as static mechanical allodynia and hyperalgesia, respectively. That static mechanical hypersensitivity covers a larger area^9,15  and requires smaller doses of capsaicin than dynamic mechanical allodynia⁹  suggests that the mechanical pain symptoms characterizing secondary hyperalgesia are independent. Therefore, they are likely signaled by different mechanosensitive pathways.

In this study, we addressed the mechanisms of specifically secondary static mechanical allodynia in rats, after a facial intradermal injection of capsaicin. We hypothesized that the manifestation of static mechanical allodynia is associated with the activation of a dorsally directed polysynaptic excitatory circuit within the medullary dorsal horn (MDH) that drives up the low-threshold mechanoreceptor (LTMR) inputs that transmit static mechanical information toward the lamina I nociceptive neurons.^17,18  These inputs are received by interneurons expressing the γ isoform of protein kinase C (PKCγ) within inner lamina II (II_i)¹⁷  and are normally under strong specifically γ-aminobutyric acid receptor type A–mediated (GABA_Aergic) inhibitory control. However, during the capsaicin-induced barrage from C-fibers, reactive oxygen species (ROS) are released in the MDH, leading to specifically GABA_Aergic disinhibition and, in turn, activation of the static allodynia circuit. To test this hypothesis, we combined behavioral, immunohistochemical, and Western blot analyses within the MDH.

Adult male Sprague–Dawley rats (250 to 275 g) were obtained from Charles River Laboratories (France). Rats were housed in plastic cages (three to four rats per cage) with soft bedding and free access to food and water. They were maintained in climate-controlled (23° ± 1°C) and light-controlled (12:12-h dark:light cycle), protected units (Iffa-Credo, France) for at least 1 week before experiments. All efforts were made to minimize the number of animals used. All behavioral experiments started at 10.00 am. Experiments followed the ethical guidelines of the International Association for the Study of Pain¹⁹  of the Directive 2010/63/UE of the European Parliament and of the Council on the protection of animals used for scientific purpose. Protocols applied in this study were approved by the local animal experimentation committee: Comité d’Ethique en Matière d’Expérimentation Animale Auvergne (n° CE 28-12). All experiments were conducted with the experimenters blinded to treatment conditions. Rats were randomized into treatment groups before any assessment was performed.

Capsaicin, bicuculline, muscimol, tert-butyl hydroperoxide (t-BOOH), and N-tert-butyl-α-phenylnitrone (PBN) were obtained from Sigma-Aldrich (France). KIG31-1 was obtained from Kai Pharmaceuticals (USA). It is conjugated to Tat, a peptide carrier, via a cysteine–cysteine bond at its N terminus. KIG31-1 competes with activated PKCγ for binding to the isoenzyme-specific docking proteins, receptors for activated C kinase. This strategy prevents PKCγ translocation in an isoenzyme-specific manner.^20,21  Linking of KIG31-1 to Tat enables efficient transfer of the peptide into cells.²²  Drug doses were selected based on literature^23,24  and on our preliminary experiments. No motor impairment has been reported with intrathecal or intracisternal injections of muscimol up to 1 μg,²⁵  PBN up to 100 μg,²⁰  and KIG31-1 up to 500 pmol.²⁶ 

For investigating the effects of capsaicin on rubbing nociceptive behavior and static or dynamic mechanical allodynia, rats were first habituated to stand on their hind paws on the experimenter sleeve and lean against the experimenter chest in a quiet room under red light, according to a method adapted from Ren.²⁷  The habituation required 1/2 h during which animals were tested with von Frey filaments or gentle air puffing using a calibrated pump onto a region between the right vibrissa pad and the upper right lip, with carefully avoiding touching any vibrissa. Ascending and descending series of von Frey filaments (1.0 to 12 g; Bioseb, France) were used. Each filament was tested five times at intervals of at least 5 s. The habituation session was repeated during 2 days. At the end of the second habituation session, all rats responded to 6-g von Frey filament with only a simple detection, showing a nonaversive response. The actual testing session took place on the third day. Rats were then placed in an observation field (0.6 × 0.6 m square) under red light for a 30-min habituation period during which the experimenter reached into the cage to apply 6-g von Frey filament or gentle air puffing on the face of the animal. At the end of this habituation period, rats received an intradermal injection of capsaicin (25 μg in 25 μl endolipid) or vehicle alone into the right vibrissa pad using a 30-gauge needle as previously described.²⁸  This is a relatively small dose to reduce as much as possible the area of dynamic mechanical allodynia. After 40 μg, a dose almost twice that used here, all subjects are hypersensitive to punctate stimuli, whereas only a minority exhibit dynamic mechanical allodynia.¹⁰ 

After injection, animals were placed back within the observation field for a 15-min observation period recorded by a digital camera (Sanyo, USA) followed by a 60-min mechanical testing period. Mechanical stimuli were applied with 6-g von Frey filament and gentle air puffing (1 s long) every 3 min onto the right upper lip, 0.5 to 1 cm to the injection site, that is, well outside this site, in the anticipated region of secondary hyperalgesia. We selected the 6-g von Frey filament because it was, on one side, innocuous,^17,29  but, on the other side, close to the aversive threshold.¹⁷  Each series of stimuli consisted of five stimuli, applied at an interval of at least 10 s. Stimulation was carried out when the rat was in a sniffing/no locomotion state: with four paws placed on the ground, neither moving nor freezing. The tip of the pump or von Frey filament was moved toward the target from behind the animal so that it could not see it. The 15-min recording time was divided into five 3-min blocks. For each block, a nociceptive score was determined by measuring the time (in seconds) animals spent rubbing the injected area (vibrissa pad) with the forepaw and not the hind paw. The behavioral responses to mechanical stimulations were observed and quantified according to the method developed by Vos et al.³⁰  Rat responses to mechanical stimuli consisted of one or more of the following elements: (1) detection, rats turn head toward stimulus; (2) withdrawal reaction, rats turn head away or pulls it briskly backward when stimulation is applied (a withdrawal reaction is assumed to include a detection element preceding the head withdrawal and therefore consists of two responses elements); (3) escape/attack, rats avoid further contact with the stimulus, either passively by moving their body away from the stimulus or actively by attacking the tip of the pump or the filament; and (4) asymmetric grooming, rats display an uninterrupted series of at least three wash strokes directed to the stimulated area. The following rank-ordered descriptive responses categories were formulated: no response, nonaversive response, mild aversive response, strong aversive response, and prolonged aversive behavior. Each category was given a score (0 to 4) based on the number of observed response elements. This score is assumed to reflect the magnitude of the aversiveness evoked by mechanical stimuli. Score was equal to 0 in case of no response. A mean score value was then calculated for each stimulation series. Behavior was always analyzed by a second experimenter who was blinded to animal treatment.

For experiments investigating the effects of KIG31-1 (50 pmol, in 5 μl Tat carrier), PBN (100 μg, in 5 μl artificial cerebrospinal fluid [aCSF]), or muscimol (250 ng, in 5 μl aCSF) upon capsaicin-induced rubbing nociceptive behavior or mechanical allodynia, animals were briefly (less than 3 min) anesthetized using a mask with 2% halothane and received an intracisternal injection of either drug or vehicle alone (5 μl) using a 10-μl Hamilton syringe (Dominique Dutscher, France).¹⁷  aCSF consisted of 150 mM Na⁺, 3 mM K⁺, 0.8 mM Mg²⁺, 1.4 mM Ca²⁺, 155 mM Cl⁻, pH 7.4, and 295 mosmol/kg. After recovery (less than 2 min), rats were placed in the observation field for a 30-min habituation period. Immediately after habituation, animals received an intradermal injection of capsaicin, and the rubbing nociceptive behavior and behavioral responses to mechanical stimuli were evaluated as described in Behavioral Testing section.

For experiments investigating the effects of bicuculline (0.05 μg, in 5 μl aCSF) or t-BOOH (50, 100, and 250 ng, in 5 μl aCSF) upon mechanical cutaneous sensitivity, animals were briefly (less than 3 min) anesthetized using a mask with 2% halothane and received an intracisternal injection of either drug or aCSF alone (5 μl) using a 10-μl Hamilton syringe. Immediately after recovery (less than 2 min), rats were placed into the observation field for a 45-min (bicuculline) or 125-min (t-BOOH) test period. Behavioral responses to mechanical stimuli—6-g von Frey filament and gentle air puffing (1 s long) every 3 min—onto the right upper lip were evaluated as described in Behavioral Testing section.

In a first series of experiments, we assessed extracellular signal–regulated kinase 1/2 phosphorylation (phospho-ERK1/2) under capsaicin-induced rubbing nociceptive behavior and mechanical stimulus–evoked pain. Rats were anesthetized deeply with urethane (1.5 g/kg intraperitoneal) as previously described.¹⁷  Twenty minutes after the injection, the depth of the anesthesia was assessed, and rats received an intradermal injection of capsaicin (25 μg in 25 μl endolipid) or vehicle alone into the right vibrissa pad using a 30-gauge needle. Five or 30 min later, innocuous stimulation was applied using a 6-g von Frey filament during 3 min at a frequency of 0.5 Hz to the right upper lip, 0.5 to 1 cm to the injection site, that is, well outside this site, in the anticipated region of secondary hyperalgesia. Two minutes later, rats were perfused transcardially with warm heparinized saline followed by cold 0.1 M phosphate-buffered solution pH 7.6 containing 4% paraformaldehyde and 0.03% picric acid. The brainstem was removed and postfixed for 2 h in the same fixative solution at 4°C and then cryoprotected in 30% sucrose diluted in 0.05 M Tris-buffered saline (TBS) pH 7.4 at 4°C for 24 h. Thirty micrometer coronal sections were cut on a freezing microtome and collected in TBS before being processed. Free-floating sections were placed in 2% normal goat serum (NGS) diluted in TBS containing 0.25% bovine serum albumin (BSA) and 0.3% Triton X-100 for 2 h before incubation in a polyclonal rabbit primary antibody directed against phospho-ERK1/2 (1:3,000; Cell Signaling Technology, USA) diluted in TBS–BSA–0.3% Triton X-100 overnight at room temperature. Sections were then incubated for 90 min with the secondary antibody goat anti-rabbit conjugated with peroxydase (1:400; Vector Laboratories, France). Immunoreactivity was revealed using nickel-diaminobenzidine (Vector Laboratories). In all cases, sections were rinsed in TBS several times, between and after each incubation, and finally transferred onto gelatinized slides before being coverslipped using DPX mountant (Sigma-Aldrich) for histology. Specificity controls consisted of omitting the primary antibody and incubating sections in inappropriate secondary antibodies. In all these controls, no specific staining was evident. A few selected sections were mounted separately and slightly counterstained with cresyl violet to help delineate the limits of MDH.

In a second series of experiments, we assessed ERK1/2 phosphorylation in PKCγ-immunoreactive interneurons under capsaicin-induced mechanical allodynia. Free-floating sections were first treated with 50 mM NH₄Cl diluted in phosphate-buffered saline (PBS), pH 7.4, for 30 min at room temperature. After several washes in 0.1 M PBS with 0.2% Triton (PBS-Tx), sections were blocked by preincubation in 5% NGS in PBS-Tx for 1 h at room temperature. After washes in PBS-Tx, sections were incubated with a polyclonal rabbit primary antibody directed against phospho-ERK1/2 (1:1,000; Cell Signaling Technology) and a monoclonal mouse primary antibody directed against PKCγ (1:4,000; Sigma-Aldrich), diluted in 5% NGS in PBS-Tx for 24 h at 4°C. Sections were then washed with 5% NGS followed by washes in PBS-Tx. Tissues were incubated with a Cy2-conjugated goat anti-mouse secondary antibody (1:200; Jackson ImmunoResearch, USA) and a Cy3-conjugated goat anti-rabbit secondary antibody (1:200; Jackson ImmunoResearch) diluted in 5% NGS in PBS-Tx for 1 h at room temperature. Finally, sections were washed in 5% NGS, in PBS-Tx, and then in PBS. Sections were transferred onto gelatinized slides before being coverslipped with DPX mountant for histology.

In a third series of experiments, rats were treated as in group 1 but received an intracisternal injection of KIG31-1 (500 pmol, in 5 μl Tat carrier), PBN (100 μg, in 5 μl aCSF), or vehicle alone 20 min after induction of anesthesia. Capsaicin was injected 30 min later.

In a fourth series of experiments, we assessed ERK1/2 phosphorylation under bicuculline-induced or t-BOOH-induced mechanical allodynia. Rats were anesthetized with urethane as previously described. Twenty minutes after induction of anesthesia, the depth of the anesthesia was assessed, and rats received an intracisternal injection of bicuculline (3 μg, in 5 μl aCSF) or t-BOOH (100 ng, in 5 μl aCSF). Five minutes (bicuculline) or 60 min (t-BOOH) later, innocuous stimulation of the right upper lip was performed using a 6-g von Frey filament during 3 min at a frequency of 0.5 Hz. Tissues were collected, and ERK1/2 phosphorylation or ERK1/2 phosphorylation in PKCγ-immunoreactive interneurons was assessed as described previously.¹⁷ 

Rats were deeply anesthetized with urethane as previously described. Twenty minutes after induction of anesthesia, the depth of the anesthesia was assessed, and rats received an intradermal injection of capsaicin (25 μg, in 25 μl endolipid) or vehicle alone into the right vibrissa pad using a 30-gauge needle. Five or 30 min later, innocuous stimulation was applied using a 6-g von Frey filament (at 0.5 Hz during 3 min) to the right upper lip. Immediately after, rats were decapitated for rapid tissue harvesting. Tissues containing the MDH were dissected out and cut from its rostral (0 μm; obex aperture) to its caudal end (−2,400 μm). These limits correspond to the (1) obex aperture and (2) end of the brainstem enlargement compared with the cervical spinal cord.³¹  The ventral part of the brainstem was not removed as MDH represents most of the tissue at this level. Ipsilateral and contralateral were divided by performing a cut in the midline of the tissue. Likely, ipsilateral and contralateral PKCγ interneurons were properly separated as these cells are located at least 1 mm away from the central commissure.³¹  Tissue was then rapidly frozen on dry ice and stored at −80°C for later use. Proteins were extracted and homogenized in 500 μl radio-immunoprecipitation assay buffer (Sigma-Aldrich) containing 0.5% of protease inhibitors (Sigma-Aldrich) and 0.5% of phosphatase inhibitors (Sigma-Aldrich). The homogenate was centrifuged at 4°C for 30 min at 22,000g to remove the debris. Protein concentration of the supernatants was measured with the Micro Lowry total protein kit (Sigma-Aldrich) with a spectrophotometer (Bio-Rad, France) at a wavelength of 750 nm. The supernatants (30 μg) were heated for 5 min at 95°C and loaded onto 5% stacking/12% separating sodium dodecyl sulfate (SDS)–polyacrylamide gels for the protein separation. Electrophoresis was carried out in 25 mM Tris, 192 mM glycine, and 1% SDS buffer for 2 h at 80 V, until the reference colorant reached the end of the gel. Proteins were then electrophoretically transferred onto polyvinylidene difluoride membrane (Millipore, France) in a buffer containing 25 mM Tris, 192 mM glycine, 10% SDS, and 20% methanol for 45 min at 100 V. That sample loading was equivalent in each lane was confirmed by Ponceau Red membrane staining following the blotting. The membrane was blocked with 6% nonfat dry milk and subsequently incubated overnight at 4°C with the polyclonal rabbit primary antibody directed against the phosphorylated form of PKCγ (Thr514) (1:1,000; Cell Signaling Technology) diluted in 0.05 M TBS, 5% BSA, and 0.1% Tween^® (Sigma Aldrich) 20 with moderate shaking. The membrane was then incubated 1 h at room temperature with an horseradish peroxidase-conjugated donkey anti-rabbit secondary antibody (1:10,000; GE Healthcare, France) diluted in 0.05 M TBS, 6% nonfat dry milk, and 0.1% Tween^® 20. ECL Plus solution (GE Healthcare) was used to visualize blots. Blots were captured using an automated image acquisition system (Fusion FX7; Vilber Lourmat, France) for 1 to 10 min. Blots were then incubated in a stripping buffer (62.5 mM Tris, 100 mM β-mercaptoethanol, and 2% SDS, pH 6.7) for 30 min at 50°C and probed again 1 h at room temperature with a polyclonal rabbit primary antibody directed against β-actin (1:1,000; Cell Signaling Technology) diluted in 0.05 M TBS, 3% BSA, and 0.1% Tween^® 20 as loading control.

Neurons containing phospho-ERK1/2 immunoreactivity in the MDH were photographed using a Nikon Optiphot 2 (Nikon, Japan) coupled with a 3CCD Sony DXC-950P digital camera (Sony, Japan) at ×10 and ×20 magnifications. Neurons containing phospho-ERK1/2 and/or PKCγ immunofluorescence in the MDH were photographed using a fluorescent Zeiss Axioplan 2 Imaging microscope (Zeiss, Germany) coupled with a Hamamatsu C4742–95 digital camera (Hamamatsu, Japan), by switching between fluorescein isothiocynanate and Rhodamine filter sets at ×20, ×40, and ×100 magnifications. Each image was then analyzed with Fiji-ImageJ 1.47 program (available at: http://rsbweb.nih.gov/ij³² ). Phospho-ERK1/2 and/or PKCγ-immunoreactive neurons were counted according to their location in the different laminae of the MDH from seven different sections, each taken at a given rostrocaudal plane within MDH. Intervals of 400 μm between planes ensured that cells were counted only once. The delineation of the MDH was based on the Paxinos and Watson atlas³³  and our own myeloarchitectural atlas as determined by our previous work.³¹  The data are expressed as the sum of the total number of labeled cells counted from all seven sections that were analyzed in each animal. Pictures were optimized for visual quality using Fiji-ImageJ 1.47 program at the end of the analysis.

Samples sizes were based on previous experience,^17,26,34  such numbers reflecting a balance between commonly used sample sizes in the field and a desire to reduce the use of animals in pain experiments. Data on capsaicin-induced rubbing behavior were analyzed using a Mann–Whitney test or an unpaired Student’s t test after passed the Kolmogorov–Smirnov normality test. Data on capsaicin-, bicuculline- or t-BOOH-induced allodynia were analyzed using a two- or three-way ANOVA with repeated measures followed by a Boneferroni post hoc test as indicated. Data on dose efficiency of t-BOOH-induced allodynia were analyzed by using a one-way ANOVA followed by a Dunnett post hoc test. Data on phospho-ERK1/2–immunoreactive cells were analyzed using a Mann–Whitney test or a one-way ANOVA followed by a Newman–Keuls post hoc test. Bands obtained by Western blot were quantified by densitometry using Fusion capt^® and Bio1D^® software (Vilber Lourmat). Comparison of the means of densitometric analyses was made using a one-way ANOVA followed by a Newman–Keuls post hoc test. In all cases, P value less than 0.05 was considered to be statistically significant. All quantitative analysis, graphs, and statistical tests were performed on GraphPad Prism^® 5.0 (USA) and Statistica^® 6.0 (USA). Figures were made using CorelDRAW^® 12 (Canada). All data are presented as mean ± SD.

We first assessed the behavioral changes produced by an intradermal injection of capsaicin (25 μg in 25 μl) into the vibrissa pad. Capsaicin caused a nociceptive behavior (unpaired Student’s t test, P = 0.0008; n = 5) (fig. 1A). Spontaneous nociceptive behavior was maximum within the first 5 min after capsaicin injection and had completely disappeared after 15 min. We subsequently assessed the responses to mechanical stimuli applied 0.5 to 1 cm adjacent to the capsaicin injection site, that is, within the zone of secondary mechanical hypersensitivity (see Materials and Methods), with innocuous static and dynamic mechanical stimuli: 6-g von Frey filament and air puff, respectively. Fifteen minutes after injection, mechanical stimulation with the 6-g von Frey filament evoked consistently the most aversive response categories (repeated measures [RM] two-way ANOVA with post hoc Boneferroni test, P < 0.0001, n = 5) (fig. 1B). Allodynia scores then steadily returned to preinjection levels within 45 min. Importantly, capsaicin did not change response scores to air puff stimulation (fig. 1B). Thus, rats exhibited secondary static, but not dynamic, mechanical allodynia at 0.5 to 1 cm away from the capsaicin injection site.

Extracellular signal–regulated kinase 1/2 has been reported to be quickly phosphorylated after noxious stimuli in neurons of spinal dorsal horn (DH).³⁵  But phospho-ERK1/2 signal can also be induced by tactile stimulation after inflammation or nerve injury.³⁵  Therefore, we used phospho-ERK1/2 immunoreactivity as an anatomical marker to visualize MDH neurons that are activated by the two types of stimuli, intradermal injection of capsaicin into the vibrissa pad alone or together with punctate mechanical stimulation at two delays after capsaicin injection: 5 min, when nociceptive behavior is maximum, and 30 min, when static mechanical allodynia is maximum (fig. 1, C–F). Capsaicin alone induced ERK1/2 phosphorylation mainly in ipsilateral superficial laminae I-II_o (within 5 min after injection; one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0115, n = 5) (fig. 1, C, E, and F and table 1). But ERK1/2 phosphorylation within ipsilateral MDH was further increased after stimulation applied on the area of secondary hypersensitivity with the 6-g von Frey filament (fig. 1, D–F, and table 1). Interestingly, such mechanical stimulation–evoked ERK phosphorylation was observed only when the stimulation was applied 30 min (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0118, n = 5) but not 5 min after capsaicin injection (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.8353, n = 5) (fig. 1E and table 1). Moreover, it occurred within both superficial laminae I-II_o (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0357, n = 5) and deeper laminae II_i-III_o (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0006, n = 5) (fig. 1, D and F). Importantly, there was no mechanical stimulation-induced change in phospho-ERK1/2 immunoreactivity when stimulation was applied on the area of secondary hypersensitivity with a paintbrush, 30 min after capsaicin injection (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.4190, n = 5) (table 1). These results suggest that the expression of static mechanical allodynia is associated with the activation of a neuronal pathway within MDH encompassing laminae I to III_o interneurons.

The fact that genetic impairment of PKCγ prevents nerve injury–induced cutaneous hypersensitivity to punctate mechanical stimuli^36–38  suggests that activation of PKCγ is also a prerequisite for the expression of secondary static mechanical allodynia. To test this hypothesis, we first examined the phospho-ERK1/2 signal in PKCγ-immunoreactive interneurons after mechanical stimulation applied adjacent to the capsaicin injection site with the 6-g von Frey filament. Using dual immunocytochemical labeling, we found that 30.5 ± 5.5% (n = 4) of the phospho-ERK1/2–immunoreactive cells within laminae II_i-III_o of the ipsilateral MDH were also PKCγ immunoreactive (fig. 2A).

Expression of phosphorylated PKCγ (phospho-PKCγ)—the active form of PKCγ³⁹ —in MDH was directly measured using Western blot analysis (fig. 2B; Supplemental Digital Content 1, https://links.lww.com/ALN/B238, fig. 1) at the two delays after capsaicin injection: 5 and 30 min. Whereas capsaicin alone had no effect on phospho-PKCγ, stimulation applied 0.5 to 1 cm adjacent to the capsaicin injection site with the 6-g von Frey filament led to a strong increase in phospho-PKCγ. Interestingly, phospho-PKCγ protein level was only enhanced when cutaneous stimulation occurred 30 min (451.4 ± 79.2% of baseline; one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0003, n = 4), but not 5 min after capsaicin injection (75.4 ± 25.1% of baseline; one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.5792, n = 3).

Thus, our results indicate that the activation of PKCγ interneurons and the phosphorylation of PKCγ are associated with the manifestation of secondary static mechanic allodynia. To directly test whether activation of PKCγ is necessary for the expression of this allodynia, we examined the effect of intracisternal application of the selective PKCγ antagonist, KIG31-1 (50 pmol in 5 μl), on capsaicin-induced nociceptive behavior and static mechanical allodynia. KIG31-1 had no effect on the nociceptive behavior (unpaired Student’s t test, P = 0.5379, n = 5) (fig. 3A) but completely prevented static mechanical allodynia (RM two-way ANOVA with post hoc Boneferroni test, P < 0.0001, n = 5) (fig. 3B). In addition, it strongly attenuated mechanical stimulation-evoked phospho-ERK1/2 signal in laminae I-II_o as well as II_i-III_o (30 min after capsaicin injection: Mann–Whitney test, P = 0.0286, n = 3 to 4) (fig. 3C and table 1). Altogether, these results suggest that PKCγ activation is required for the expression of secondary static mechanical allodynia.

Activation of peptidergic C-fiber nociceptor central terminals reduces selectively GABAergic inhibitory signaling within spinal DH.⁴⁰  Moreover, our group previously showed in anesthetized animals that after a microinjection of the GABA_A receptor antagonist, bicuculline, into MDH, static, but not dynamic, mechanical stimulation of the face can increase blood pressure, a reliable indicator of pain in anesthetized animals.¹⁷  This suggests that local GABA_Aergic disinhibition also produces a static, but not a dynamic, mechanical allodynia. Therefore, we hypothesized that capsaicin produces a secondary static mechanical allodynia through GABA_Aergic disinhibition.

To test this hypothesis, we first confirmed in behaving rats our previous results obtained in anesthetized animals. Intracisternal application of bicuculline (0.05 μg in 5 μl) induced a static, but not a dynamic, mechanical allodynia (fig. 4; RM two-way ANOVA with post hoc Boneferroni test, P < 0.0001, n = 5) that lasted for approximately 30 min. Then, using phopho-ERK1/2 immunoreactivity, we examined which MDH neurons are activated by mechanical stimulation under GABA_Aergic disinhibition (fig. 5). As expected, no significant phopho-ERK1/2 signal was observed after intracisternal aCSF, with or without mechanical stimulation (table 1). However, phopho-ERK1/2–immunoreactive neurons were observed 5 min after intracisternal bicuculline but only in superficial laminae I-II_o (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0423, n = 5) (fig. 5, A and C). Finally, mechanical stimulation applied to the face with the 6-g von Frey filament under GABA_Aergic disinhibition produced an even stronger phopho-ERK1/2 signal (table 1). Phospho-ERK1/2–immunoreactive neurons were now located within laminae I-II_o (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0006, n = 5), II_i-III_o (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0019, n = 5), and, to a smaller extent, III_i-V (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0139, n = 5; fig. 5, B and C). Again, 20.7 ± 3.1% (n = 4) of phospho-ERK1/2–immunoreactive neurons in laminae II_i-III_o was also PKCγ immunoreactive, indicating that PKCγ interneurons also participate in neuronal circuits activated by mechanical stimulation under GABA_Aergic disinhibition (fig. 5, D–F).

To confirm that capsaicin acts through GABA_Aergic disinhibition, we examined the effects of intracisternal application of the GABA_A receptor agonist, muscimol, on capsaicin-induced nociceptive behavior and static mechanical allodynia. Intracisternal muscimol (250 ng in 5 μl) completely prevented both spontaneous nociceptive behavior (Mann–Whitney test, P = 0.0030, n = 4 to 5; fig. 6A) and secondary static mechanical allodynia (RM two-way ANOVA with post hoc Boneferroni test, P < 0.0001, n = 4; fig. 6B).

On one hand, the secondary static mechanical allodynia is produced by the capsaicin-evoked barrage from C-fiber nociceptors.⁴¹  However, present results indicate that local circuits associated with such allodynia involve lamina II_i PKCγ interneurons. But these interneurons are targeted by myelinated, but not unmyelinated, primary afferents.^31,42  Consistently, no lamina II_i PKCγ interneurons showed phospho-ERK1/2 immunoreactivity after capsaicin injection alone. Therefore, how can the conditioning input, carried by C-fiber nociceptors, and the facilitated input, carried by A-LTMRs targeting lamina II_i PKCγ interneurons, interact with each other to induce a secondary static mechanical allodynia? One possibility is via a diffusable factor such as ROS. They are released during the capsaicin-induced barrage from peptidergic C-fiber and some Aδ-fiber nociceptors⁴³  and were shown to decrease selectively GABA release within DH.²⁴ 

Therefore, we tested whether intracisternal application of an ROS donor, t-BOOH, causes a static mechanical allodynia and unmasks the same neuronal circuit as GABA_Aergic disinhibition or intradermal capsaicin injection. Intracisternally applied t-BOOH induced a static, but not a dynamic, mechanical allodynia (RM three-way ANOVA with post hoc Boneferroni test, P < 0.0001, n = 5 per group; fig. 7). Interestingly, increasing t-BOOH concentration did not enhance the maximum score of allodynia—the highest mean score was the same for 100 ng (2.92 ± 0.7, n = 5) and 250 ng (3 ± 0.6, n = 5)—but rather the duration of t-BOOH-induced mechanical allodynia (fig. 7A). The area under the curve was thus correlated with the dose of t-BOOH (fig. 7B). At the highest t-BOOH concentration (250 ng in 5 μl), mechanical allodynia lasted for more than 2 h (fig. 7A), and a transient dynamic mechanical allodynia was present during the first minutes (fig. 7C).

We next examined, using phospho-ERK1/2 immunohistochemistry, which MDH neurons are involved in t-BOOH-induced static mechanical allodynia. The vibrissa pad was stimulated with the 6-g von Frey filaments, 60 min after intracisternal injection of t-BOOH (100 ng in 5 μl), that is, at the peak of allodynic score at this concentration (fig. 7A). t-BOOH alone had no effect on phospho-ERK1/2 levels (one-way ANOVA with post hoc Student–Newman–Keuls test, P =0.4528, n = 5; fig. 8A and table 1), but mechanical stimulation led to a strong ERK1/2 phosphorylation (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0016, n = 5; fig. 8B and table 1) within both laminae I-II_o (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0025, n = 5) and II_i-III_o (one-way ANOVA with post hoc Student–Newman–Keuls test, P = 0.0019, n = 5) (fig. 8, B and C). Moreover, again, many phospho-ERK1/2–immunoreactive neurons in laminae II_i-III_o (26.8 ± 10.3%, n = 4) were also PKCγ immunoreactive, indicating that PKCγ interneurons participate in the neuronal circuits associated with ROS-induced static mechanical allodynia (fig. 8, D–F). It is interesting to note that such ratio of dually phospho-ERK1/2–immunoreactive and PKCγ-immunoreactive neurons within laminae II_i-III_o is rather similar to those in capsaicin- or bicuculline-treated rats: 30.5 ± 5.5% and 20.7 ± 3.1%, respectively.

That intradermal capsaicin and ROS release both produce a static mechanical allodynia associated with the activation of the same MDH neuronal circuits suggests that capsaicin induces a static mechanical allodynia through ROS release. To test this hypothesis, we examined the effect of intracisternal application of the ROS scavenger, PBN, on capsaicin-induced nociceptive behavior and static mechanical allodynia as well as on mechanical stimulation–induced ERK1/2 phosphorylation within MDH. Intracisternal PBN (100 μg in 5 μl) had no effect on capsaicin-induced rubbing nociceptive behavior (unpaired Student’s t test, P = 0.6534, n = 5) (fig. 9A) but almost completely prevented static mechanical allodynia (RM two-way ANOVA with post hoc Boneferroni test, P < 0.0001, n = 5) (fig. 9B). Consistently, the level of ERK1/2 phosphorylation within MDH evoked by a stimulation applied on the area of secondary hypersensitivity with the 6-g von Frey filament was significantly reduced (Mann–Whitney test, P = 0.0286, n = 4; fig. 9C and table 1), being equivalent with that after capsaicin alone (fig. 1E).

We investigated the mechanisms of trigeminal secondary static mechanical allodynia. Monitoring neural activity within MDH with phospho-ERK1/2 immunoreactivity shows that activation of laminae I-II_o and II_i-III_o neurons, including lamina II_i PKCγ-expressing interneurons, is associated with the expression of secondary static mechanical allodynia. Western blot analysis reveals that PKCγ is concomitantly phosphorylated. Either local GABA_Aergic disinhibition or ROS release is sufficient to initiate the same static mechanical allodynia and von Frey–induced laminae I–III_o neural activation as capsaicin. Conversely, intracisternal PKCγ antagonist, GABA_A receptor agonist, or ROS scavenger prevent capsaicin-induced secondary static mechanical allodynia and associated von Frey–evoked neural activation.

Capsaicin injection into the rat vibrissa pad induced a static, but not dynamic, mechanical allodynia at 0.5 to 1 cm away from the injection site. Psychophysical studies in humans indicate that capsaicin-induced secondary hyperalgesia encompasses several pain symptoms such as dynamic mechanical allodynia and static mechanical hyperalgesia and allodynia, the two latter symptoms being filed under the umbrella term “hyperalgesia to punctate stimuli.”⁴⁴  These different pain symptoms appear to be independent from each other. And dynamic mechanical allodynia covers a much smaller area than static mechanical hypersensitivity.^9,15  Therefore, our finding that in rats mechanical allodynia is specifically static at 0.5 to 1 cm from the capsaicin injection is consistent with the cutaneous sensitivity observed in humans inside the area of hyperalgesia to punctate stimuli but outside that of dynamic mechanical allodynia.

As in DH,⁴⁵  capsaicin-induced ERK1/2 phosphorylation was restricted to superficial laminae I-II_o of MDH. But 6-g von Frey stimulation evoked ERK1/2 phosphorylation in both laminae I-II_o and II_i-III_o, indicating that facilitation of a neural circuit encompassing laminae I–III is associated with secondary static mechanical allodynia. This circuit involves lamina II_i PKCγ interneurons: (1) 20 to 30% of lamina II_i phospho-ERK1/2–immunoreactive neurons after mechanical stimulation were also PKCγ immunoreactive, (2) mechanical stimulation–induced phosphorylations of PKCγ and ERK1/2 were detected at the same delay (+30 min) after capsaicin injection, and (3) intracisternal PKCγ antagonist prevented capsaicin-induced von Frey–evoked phospho-ERK1/2 immunoreactivity and associated static mechanical allodynia. Accordingly, genetic impairment of PKCγ was shown to reduce injury-induced mechanical hypersensitivity to innocuous punctate stimuli.^36–38  Such pattern of neuronal activation is not consistent with A- or C-nociceptors transmitting inputs that produce pain: (1) in rats, TRPV1 receptors are expressed by most peptidergic and nonpeptidergic C-nociceptors^46,47  and we show that neurons activated by capsaicin are restricted to laminae I-II_o; (2) selectively stimulating A-nociceptors in the area of secondary hyperalgesia results in only lamina I neural activation.⁴⁸  The laminae I–III pattern of neural activation rather resembles that associated with dynamic mechanical allodynia where Aβ-LTMR inputs to lamina II_i PKCγ interneurons can gain access to the lamina I pain transmission circuitry.^17,26  Our results thus suggest that innocuous punctate inputs that produce secondary static mechanical allodynia are transmitted by LTMRs, which terminate within or below lamina II_i,⁴⁹  and engage lamina I neurons by way of a dorsally directed excitatory pathway including lamina II_i PKCγ interneurons. This circuit is polysynaptic as axonal terminals of lamina II_i PKCγ interneurons do not reach lamina I⁵⁰  but contact lamina II_o central neurons.⁵¹  Consistently, lamina I output neurons, which do not normally respond to innocuous stimuli, can be driven by static mechanical inputs in neuropathic rats⁵²  and by polysynaptic Aβ- and Aδ-fiber inputs in DH slices after GABA_A/glycinergic disinhibition.¹⁸  That capsaicin-induced secondary static and dynamic mechanical allodynias can manifest independently from each other^9,15 suggests that the two allodynia are signaled by different primary afferent pathways. Because dynamic mechanical allodynia is signaled by Aβ-LTMRs and PKCγ interneurons are exclusively targeted by myelinated primary afferents,^31,42  secondary static mechanical allodynia is therefore likely signaled by Aδ-LTMRs. Interestingly, mice genetically impaired for Piezo2, a major mechanical transducer in Aβ-LTMRs, exhibit no change in static mechanical allodynia in response to inflammatory mediators.⁵³ 

As previously reported,^45,54  phospho-ERK1/2 was induced 5 min after capsaicin injection. But von Frey–evoked ERK1/2 activation—as PKCγ phosphorylation—was only detected 30 min after injection. This supports the idea that the signaling cascade initiating secondary static mechanical allodynia reaches a peak more than 5 min after the capsaicin-induced barrage from C-nociceptors. Consistently, the area of secondary hyperalgesia to punctate stimuli, immediately present after injection, grows to a maximum within 15 to 30 min after capsaicin injection.⁹ 

The induction of secondary static mechanical allodynia using capsaicin requires the interaction of two primary afferent pathways: a conditioning one, carried by TRPV1-sensitive C-nociceptors terminating onto laminae I-II_o neurons, as indicated by the capsaicin-induced phospho-ERK1/2 immunoreactivity⁴⁵  (present results), and a facilitated one, signaling innocuous punctate stimuli that produce pain, carried by A-LTMRs targeting lamina II_i PKCγ interneurons. How can such two mutually exclusive afferent pathways interact with each other? One possibility is through GABA_Aergic disinhibition. First, local GABA_Aergic disinhibition in our study produced a selectively static mechanical allodynia. In a previous study in anesthetized rats, static, but not dynamic, mechanical stimulation could increase blood pressure, a reliable indicator of pain, after bicuculline microinjection into MDH.¹⁷  Second, under GABA_Aergic disinhibition, innocuous static mechanical stimulation generated a pattern of ERK1/2 activation across not only laminae I-II_o, as previously found in models of neuropathic pain,⁵⁵  but also laminae II_i-III_o. Finally, such activation involved lamina II_i PKCγ interneurons. This supports the idea that the same local circuit is associated with capsaicin-induced and GABA_Aergic disinhibition–induced static mechanical allodynia. Accordingly, we found that intracisternal administration of a GABA_A receptor agonist prevents capsaicin-induced static mechanical allodynia. This suggests that Aδ-LTMR inputs to PKCγ interneurons are under GABA_Aergic inhibition and, when this inhibition is lifted, these inputs can gain access to the lamina I pain transmission circuitry. Consistently, in DH slices, bath-applied capsaicin, which mimics C-nociceptor activation, reduces GABA_Aergic inhibition onto lamina II neurons.^40,56  And local bicuculline application enhances DH neuronal responses to Aδ-fiber–mediated, but not Aβ-fiber–mediated, mechanical and electrical stimulation.^57–59  Intracisternal muscimol also prevented spontaneous nociceptive behavior, suggesting that activation of MDH GABA_A receptors blocked the transmission of nociceptive inputs, in addition to that of Aδ-LTMR inputs, to lamina I pain circuitry. Nociceptive inputs to the superficial DH are known to be under strong presynaptic⁶⁰  as well as postsynaptic⁶¹  GABA_Aergic control. Accordingly, locally applied bicuculline also increases the response to noxious inputs.⁶²  We detected, too, an increase in laminae I-II_o p-ERK1/2 immunoreactivity after intracisternal bicuculline alone (fig. 6), suggesting an increase in resting discharge of superficial MDH neurons under GABA_Aergic disinhibition. Similarly, in humans, activation of GABA_A receptors through intrathecal application of the benzodiazepine, midazolam, can manage spontaneous pain, including perioperative,⁶³  chronic low back⁶⁴  and neuropathic pain⁶⁵  and neuropathic mechanical allodynia.⁶⁵ 

Another possibility to account for mechanical allodynia is that the conditioning input, carried by C-nociceptors, sensitizes lamina II_i PKCγ interneurons via diffusible factors such as ROS. ROS are known to be released during the capsaicin-induced barrage from C-nociceptors⁴³  and to decrease selectively GABA release within DH.²⁴  We found that the capsaicin-induced static mechanical allodynia and associated touch-evoked phospho-ERK1/2 immunoreactivity are replicated by intracisternal t-BOOH and, conversely, prevented by intracisternal PBN. Consistently, intrathecal ROS scavengers were shown to reverse the hypersensitivity to von Frey stimulation in neuropathic rats.^24,66  Although a link between ROS release in DH and secondary hypersensitivity to punctate stimuli was previously described,⁴³  the present results establish that ROS release leads to activation of the lamina I–III neuronal circuit associated with behavioral allodynia. Interestingly, alone, t-BOOH triggered no phospho-ERK1/2 immunoreactivity, indicating that ROS do not directly activate MDH neurons. On the basis of these results, we propose a model whereby ROS is released within MDH during the capsaicin-induced barrage from C-nociceptors⁴³  and produces a selective GABA_Aergic disinhibition of lamina II_i PKCγ interneurons (fig. 10).

Our results reveal a new mechanism for secondary hyperalgesia: sensitization of a polysynaptic, dorsally directed excitatory circuit whereby static innocuous inputs likely carried by Aδ-LTMRs terminating onto lamina II_i PKCγ interneurons can gain access to the pain transmission circuitry of superficial DH/MDH, producing static mechanical allodynia (fig. 10). Such circuit is unmasked under GABA_Aergic disinhibition and ROS release. Altogether, this suggests that secondary static mechanical allodynia produced by intradermal capsaicin is mediated by release of ROS in the DH/MDH during C-fiber activation, leading to the disinhibition of GABAergic circuits controlling Aδ-LTMR inputs onto PKCγ interneurons and, in turn, the sensitization of these interneurons. Thus both the primary afferents that transmit the input producing pain and the mechanisms of sensitization within MDH/DH of static mechanical allodynia are different from those of the two other mechanical pain symptoms characterizing secondary hyperalgesia: dynamic allodynia and static hyperalgesia. On one hand, secondary dynamic mechanical allodynia is signaled by Aβ-LTMR.^44,67  Aβ-LTMR inputs onto PKCγ interneurons are controlled by selectively glycinergic feed-forward inhibition.^17,26,51  Such glycinergic inhibition being lifted, Aβ-LTMR inputs can engage lamina I nociceptive neurons.²⁶  On the other hand, secondary static mechanical hyperalgesia is signaled by Aδ-nociceptors.^10,12  And pain amplification likely relies on the unmasking and/or heterosynaptic strengthening of nonactivated afferents converging onto MDH/DH neurons activated by C-nociceptors.^68–70  Together with our result that static mechanical allodynia, but not capsaicin-induced pain behavior, is sensitive to both ROS scavenger and PKC antagonist, this indicates that individualized mechanism-based treatments are required for each of these symptoms.

The authors thank Amélie Descheemaeker (Clermont Université, Université d’Auvergne, Neuro-Dol, and Inserm U1107, Clermont-Ferrand, France) for technical help during behavioral experiments and Anne-Marie Gaydier for secretarial assistance.

This work was supported by funding from Institut National de la Santé et de la Recherche Médicale (Inserm), University of Clermont-Ferrand 1 (Clermont-Ferrand, France), Fondation Gueules Cassées (Paris, France), and Région Auvergne (Clermont-Ferrand, France).

The authors declare no competing interests.