Activation of medullary dorsal horn γ isoform of protein kinase C interneurons is essential to the development of both static and dynamic facial mechanical allodynia

The γ isoform of protein kinase C (PKCγ), which is concentrated in a specific class of interneurons within inner lamina II (IIi) of the spinal dorsal horn and medullary dorsal horn (MDH), is known to be involved in the development of mechanical allodynia, a widespread and intractable symptom of inflammatory or neuropathic pain. However, although genetic and pharmacological impairment of PKCγ were shown to prevent mechanical allodynia in animal models of pain, after nerve injury or reduced inhibition, the functional consequences of PKCγ activation alone on mechanical sensitivity are still unknown. Using behavioural and anatomical approaches in the rat MDH, we tested whether PKCγ activation in naive animals is sufficient for the establishment of mechanical allodynia. Intracisternal injection of the phorbol ester, 12,13‐dibutyrate concomitantly induced static as well as dynamic facial mechanical allodynia. Monitoring neuronal activity within the MDH with phospho‐extracellular signal‐regulated kinases 1 and 2 immunoreactivity revealed that activation of both lamina I–outer lamina II and IIi–outer lamina III neurons, including lamina IIiPKCγ‐expressing interneurons, was associated with the manifestation of mechanical allodynia. Phorbol ester, 12,13‐dibutyrate‐induced mechanical allodynia and associated neuronal activations were all prevented by inhibiting selectively segmental PKCγ with KIG31‐1. Our findings suggest that PKCγ activation, without any other experimental manipulation, is sufficient for the development of static and dynamic mechanical allodynia. Lamina IIiPKCγ interneurons have been shown to be directly activated by low‐threshold mechanical inputs carried by myelinated afferents. Thus, the level of PKCγ activation within PKCγ interneurons might gate the transmission of innocuous mechanical inputs to lamina I, nociceptive output neurons, thus turning touch into pain.

Combining intracisternal injection of both phorbol ester and a specific antagonist to the γ isoform of PKC reveals that PKCγ activation within the medullary dorsal horn leads to facial static and dynamic mechanical allodynia in naïve rats. Thus, the level of PKCγ activation within the inner lamina II PKCγ interneurons that receive innocuous mechanical inputs, might gate the transmission of such inputs to lamina I, nociceptive output neurons, and thus the transformation of touch into pain.

allodynia; orofacial; pain; phorbol ester; 12; 13‐dibutyrate; trigeminal

Chronic pain syndromes initiated by tissue damage and inflammation (inflammatory pain) or lesions to the nervous system (neuropathic pain) are characterized by pain hypersensitivity. Key symptoms of persistent pain syndromes include mechanical allodynia, painful sensation caused by innocuous mechanical stimuli like light touch. Moreover, two types of mechanical allodynia can be clinically discriminated in patients: a static mechanical allodynia evoked by applying a gentle pressure on the skin and a dynamic mechanical allodynia elicited by lightly stroking it.

Central mechanisms have been implicated in the development of mechanical allodynia. Touch‐sensing, low‐threshold mechanoreceptors (LTMRs), including all heavily myelinated Aβ‐fibre as well as Aδ‐fibre and C‐fibre LTMRs, terminate in the deep spinal (SDH) and medullary (MDH) dorsal horns, between inner lamina II (IIi) and lamina V (Seal et al., [39] ; Todd, [42] ; Li et al., [15] ). Thus, lamina I neurons, one of the main nociceptive output pathways from the SDH/MDH to the brain, do not receive direct input from LTMRs. However, anatomical markers of neuronal activation, Fos protein or phospho‐extracellular signal‐regulated kinases 1 and 2 (ERK1/2), have revealed that an activation of interneurons within the superficial SDH/MDH, encompassing lamina I to outer lamina III (IIIo), is associated with mechanical allodynia (Bester et al., [2] ; Miraucourt et al., [21] , [22] ). This indicates that a crosstalk between laminae can develop; LTMR inputs may then gain access to the pain transmission circuitry of the superficial SDH/MDH via polysynaptic, ventral‐to‐dorsal excitatory pathways and elicit pain.

Among the laminae I–IIIo neurons that are activated by mechanical stimulation under mechanical allodynia are interneurons within lamina IIi that specifically express the γ isoform of protein kinase C (PKCγ) (Miraucourt et al., [21] , [22] ). Interestingly, genetic and pharmacological impairment of PKCγ has been shown to prevent static as well as dynamic mechanical allodynia and the associated stimulus‐evoked neuronal activations (Malmberg et al., [19] ; Martin et al., [20] ; Miraucourt et al., [21] , [22] ; Zou et al., [48] ). Together with the evidence that such lamina IIi PKCγ interneurons receive inputs from myelinated LTMRs (Neumann et al., [25] ; Lu et al., [18] ; Peirs et al., [31] ), this suggests that these neurons are key elements for dorsally‐directed circuits driving tactile inputs to lamina I nociceptive‐specific neurons.

Importantly, however, evidence for the involvement of PKCγ activation in the unmasking of dorsally‐directed local circuits onto lamina I neurons and associated development of static and dynamic mechanical allodynia was only obtained in animal models of pain, under conditions of disinhibition (Miraucourt et al., [21] , [22] ) or nerve injury (Malmberg et al., [19] ). Therefore, the selective role of PKCγ activation in mechanical allodynia is still unknown. Here, using behavioural and immunohistochemical techniques, we assessed the effects on the cutaneous mechanical sensitivity of naive animals of the selective activation of PKCγ by combining the local application of a phorbol ester with that of a selective PKCγ antagonist, KIG31‐1. These experiments were performed within the trigeminal system of rats.

Male Sprague‐Dawley rats (250–275 g) were obtained from Charles River (L'Arbresle, 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‐h/12‐h dark/light cycle) protected units (Iffa‐Credo) for at least 1 week before experiments. All efforts were made to minimize the number of animals used. All behavioural experiments started at 10.00 h. Experiments followed the ethical guidelines of the International Association for the Study of Pain (Zimmermann, [47] ), of the Directive 2010/63/UE of the European Parliament and of the Council on the protection of animals used for scientific purpose. The protocols applied in this study were approved by the local animal experimentation committee (Comité d'Ethique en Matière d'Expérimentation Animale Auvergne, no. 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.

Phorbol ester, 12,13‐dibutyrate (PDBu) was purchased from Sigma (France) and dissolved into 0.3% ethanol in artificial cerebrospinal fluid. KIG31‐1 was obtained from Kai Pharmaceuticals (San Francisco, CA, 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 isozyme‐specific manner (Mochly‐Rosen & Gordon, [23] ; Churchill et al., [6] ). Linking of KIG31‐1 to Tat enables efficient transfer of the peptide into cells (Chen et al., [5] ).

For investigating the effects of PDBu on cutaneous mechanical sensitivity, rats were first habituated to stand on their hindpaws on the experimenter's sleeve and to lean against the experimenter's chest in a quiet room under red light, according to a method adapted from Ren ([35] ). The habituation required 0.5 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 right upper lip, carefully avoiding touching any vibrissa. Ascending and descending series of von Frey filaments (1.0–12 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 a 6‐g von Frey filament with only a simple detection, showing a non‐aversive 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 a 6‐g von Frey filament or gentle air puffing on the face of the animal. At the end of this habituation period, animals were briefly (<3 min) anaesthetized using a mask with 2% halothane and received an intracisternal injection (into the cisterna magna) of either PDBu (0.03, 0.3 and 3 nmol/5 μL saline with 10% dimethyl sulphoxide) or vehicle alone (5 μL) using a 10‐μL Hamilton syringe. The artificial cerebrospinal fluid consisted of: 150 mm Na+, 3 mm K+, 0.8 mm Mg2+, 1.4 mm Ca2+, 155 mm Cl−, pH 7.4, 295 mosmol/kg. After injection, animals were placed back within the observation field for a 10‐min observation period recorded by a digital camera (Sanyo, USA) followed by a 120‐min mechanical‐testing period. Mechanical stimuli were alternately applied with a 6‐g von Frey filament and gentle air puffing (1 s) every 3 min onto the right upper lip. 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 behavioural responses to mechanical stimulations were observed and quantified according to the method developed by Vos et al. ([46] ). The rat responses to mechanical stimuli consisted of one or more of the following elements: (i) detection, i.e. rat turns head toward the stimulus; (ii) withdrawal reaction, i.e. rat turns 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 response elements); (iii) escape/attack, i.e. rat avoids further contact with the stimulus, either passively by moving its body away from the stimulus, or actively by attacking the tip of the pump or the filament; and (iv) asymmetric grooming, i.e. rat displays an uninterrupted series of at least three wash strokes directed to the stimulated area. Based on the number of observed response elements, a score was given (allodynia score: 0–4) that is assumed to reflect the magnitude of aversiveness evoked by mechanical stimulation. Five rank‐ordered descriptive response categories were formulated according to the study of Vos et al. ([46] ): no response (score 0), non‐aversive/detection response (score 1), mild aversive response (score 2), strong aversive response (score 3), and prolonged aversive behaviour (score 4). A mean score value was then calculated for each stimulation series. Behaviour was always analysed by a second experimenter who was blind to animal treatment.

For experiments investigating the effects of KIG31‐1 on PDBu‐induced mechanical allodynia, animals were briefly (<3 min) anaesthetized using a mask with 2% halothane and received an intracisternal injection of either KIG31‐1 (50 and 100 pmol/5 μL) or Tat carrier alone (5 μL) using a 10‐μL Hamilton syringe (Miraucourt et al., [21] ), 30 min before the intracisternal injection of PDBu. The effect of a single dose of PDBu or PDBu + KIG31‐1 was investigated in each animal.

Changes in motor performance were assessed using the accelerating rotarod (8500; Ugo Basile, Comerio, Italy), in which rats were required to walk against the motion of a rotating drum, with the speed increasing from 4 to 40 rpm over 10 min. The time taken to fall off the rotarod was recorded as the latency (in s). Rats were tested before and at 30 and 40 min after the PDBu (3 nmol/5 μL) or vehicle that was injected intracisternally under brief halothane anaesthesia.

We assessed ERK1/2 phosphorylation under PDBu‐induced mechanical allodynia. Rats were anaesthetized deeply with urethane (1.5 g/kg, i.p.) as previously described (Miraucourt et al., [22] ). At 20 min after the induction of anaesthesia, the depth of the anaesthesia was assessed, and rats received an intracisternal injection of PDBu (3 nmol/5 μL) or vehicle (5 μL). Rats were stimulated 30 min later with a paintbrush (dynamic mechanical stimulus: for 10 min at 0.5 Hz) or 6‐g von Frey filament (static mechanical stimulus: 60 times at 1 Hz) onto a region between the right vibrissa pad and upper lip. The dynamic mechanical stimulus used here was a paintbrush because it produces a stronger activation than air puff and thus probably activates more neurons. Rats were perfused transcardially 2 min later with warm heparinized saline followed by cold 0.1 m phosphate‐buffered solution, pH 7.6, containing 4% paraformaldehyde and 0.03% picric acid. This 2‐min delay was selected because it corresponded to the peak of pain‐induced phospho‐ERK1/2 immunoreactivity in the SDH (Ji et al., [11] ). 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, pH 7.4, at 4 °C for 24 h. Coronal sections (30 μm) were cut on a freezing microtome and collected in Tris‐buffered saline before being processed. Free‐floating sections were placed in 2% normal goat serum (NGS) diluted in Tris‐buffered saline containing 0.25% bovine serum albumin and 0.3% Triton X‐100 for 2 h before incubation in a polyclonal rabbit primary antibody directed against phospho‐ERK1/2 (1 : 3000; 9101L; Cell Signaling Technology, USA) diluted in Tris‐buffered saline containing 0.25% bovine serum albumin and 0.3% Triton X‐100 overnight at room temperature (20–22 °C). Sections were then incubated for 90 min with the secondary antibody goat anti‐rabbit conjugated with peroxidase (1 : 400; Vector Laboratories, Les Ulis, France). Immunoreactivity was revealed using nickel‐diaminobenzidine (Vector Laboratories). In all cases, sections were rinsed in Tris‐buffered saline several times, between and after each incubation, and finally transferred onto gelatinized slides before being coverslipped using DPX mountant for histology. Specificity controls consisted of omitting the primary antibody and incubating sections in inappropriate secondary antibodies. In all of 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 the MDH.

In a second series of experiments, we assessed ERK1/2 phosphorylation in PKCγ‐immunoreactive interneurons under PDBu‐induced dynamic mechanical allodynia. Free‐floating sections were first treated with 50 mm NH4Cl diluted in phosphate‐buffered saline, pH 7.4, for 30 min at room temperature. After several washes in 0.1 m phosphate‐buffered saline 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 : 1000; Cell Signaling Technology) and a monoclonal mouse primary antibody directed against PKCγ (1 : 4000; P 8083; 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, West Grove, PA, 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 phosphate‐buffered saline. Sections were transferred onto gelatinized slides before being coverslipped with DPX mountant for histology.

For experiments investigating the effects of KIG31‐1 on ERK1/2 phosphorylation (1 : 3000; Cell Signaling Technology) under PDBu‐induced mechanical allodynia, animals were anaesthetized deeply with urethane. At 20 min after induction of anaesthesia, rats received an intracisternal injection of either KIG31‐1 (100 pmol/5 μL) or Tat carrier alone (5 μL), 30 min before the intracisternal injection of PDBu (3 nmol/5 μL). At 30 min after PDBu injection, rats were stimulated with a paintbrush (dynamic mechanical stimulus: for 10 min at 0.5 Hz). Rats were then processed as above.

Neurons containing phospho‐ERK1/2 immunoreactivity in the MDH were photographed using a Nikon Optiphot two coupled with a 3CCD Sony DXC‐950P digital camera at ×10 and ×20 magnifications. Neurons containing phospho‐ERK1/2 and/or PKCγ immunofluorescence in the MDH were photographed using a fluorescent Zeiss Axioplan two Imaging microscope coupled with a Hamamatsu C4742–95 digital camera, by switching between fluorescein isothiocynanate (FITC) and rhodamine filter sets at ×20, ×40 and ×100 magnifications. Each image was then analysed with the fiji‐imagej 1.47 program (http://rsbweb.nih.gov/ij;Schindelinet al.,[38]). Phospho‐ERK1/2‐immunoreactive 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 the MDH. Intervals of 400 μm between planes ensured that cells were counted only once. The delineation of the MDH was based on the atlas of Paxinos & Watson ([30] ) and our own myeloarchitectural atlas as determined by our previous work (Peirs et al., [31] ). The data are expressed as the sum of the total number of labelled cells counted from all seven sections that were analysed in each animal. Pictures were optimized for visual quality using the fiji‐imagej 1.47 program at the end of the analysis.

The sample sizes were based on previous experience (Miraucourt et al., [21] , [22] ), 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. Five rats per group were used in each experiment. Results are presented as mean ± SEM. Data were analysed using the statistical software sigmaplot (Systat Software Inc., San Jose, CA, USA). Firstly, descriptive statistics for all groups were calculated. The normality of data distribution was tested using the Shapiro–Wilk test. One‐way or two‐way anova followed by a Bonferroni post hoc test (normally distributed data) or a Duncan post hoc test (non‐normally distributed data) was used to analyse behavioural and phospho‐ERK1/2 immunoreactivity data. The level of significance was set at P < 0.05. Figures were made using either SigmaPlot or CorelDRAW® 12.

The intracisternal injection of PDBu produced a dose‐dependent facial static as well as dynamic mechanical allodynia (Fig. [NaN] ). Interestingly, neither static nor dynamic mechanical allodynia was present at the first measurement (9–12 min after injection). However, they both progressively developed over the first 20–30 min after injection to reach a peak (25–40 min after injection) and subsequently decrease (Fig. [NaN] A and C). Increasing the dose of intracisternal PDBu (from 0.03 to 3 nmol) led to (i) an increase in the peak score, (ii) a decrease in the time to peak (static: from 30–35 to 25 min; dynamic: from 40 to 28 min) and (iii) an increase in the duration of allodynia. Accordingly, there was a dose‐dependent increase in the areas under the curve of static (Fig. [NaN] B; ‘F3,16 = 52.79’, P < 0.0001) and dynamic (Fig. [NaN] D; ‘F3,16 = 47.79’, P < 0.0001) mechanical allodynia. We found no motor impairment in these PDBu‐injected rats (Table [NaN] ).

Assessment of the psychomotor effects of PDBu by the rotarod test

1 Values are given as mean ± SEM (in s).

It is of note that PDBu‐induced static and dynamic mechanical allodynia developed concomitantly, e.g. following 3 nmol PDBu, they reached a peak at the same delay after the injection (25–28 min) and lasted about 100 min (compare Fig. [NaN] A and C). Moreover, they exhibited similar dose–response curves (compare Fig. [NaN] B and D).

We used phospho‐ERK1/2 immunoreactivity as an anatomical marker (Ji et al., [12] ) to visualize MDH neurons that were activated by static (von Frey filament) and dynamic (light stroking with a paintbrush; see ) mechanical stimuli in rats preemptively treated with intracisternal PDBu (3 nmol). Compared with vehicle‐treated rats, static as well as dynamic mechanical stimulation of the face produced a strong ERK1/2 phosphorylation within mostly the ipsilateral MDH of PDBu‐treated rats (Fig. [NaN] ). We monitored phospho‐ERK1/2 immunoreactivity in the different laminae of the ipsilateral and contralateral MDH of vehicle‐treated and PDBu‐treated rats, with and without mechanical stimulation of the face (Fig. [NaN] ). A strong ERK1/2 phosphorylation appeared only after static (Fig. [NaN] A) and dynamic (Fig. [NaN] B) mechanical stimulation in PDBu‐treated rats. Interestingly, these two stimulation‐induced phospho‐ERK1/2 immunoreactivities displayed the same lamina distribution, both predominating in lamina I, the most superficial outer part of lamina II (laminae I–IIo), IIi and IIIo. Within the ipsilateral MDH, there were significant main effects of treatment (static stimulation: ‘F3,16 = 8.51’, P < 0.001; dynamic stimulation: ‘F3,16 = 15.78’, P < 0.001) and laminae (static stimulation: F2,16 = 20.73, P < 0.001; dynamic stimulation: ‘F2,16 = 11.72’, P < 0.001). There was also an effect of interaction between treatment and laminae (static stimulation: ‘F6,16 = 2.84’, P = 0.019; dynamic stimulation: ‘F6,16 = 3.94’, P = 0.004). It is noteworthy that, compared with static stimulation‐induced ERK1/2 phosphorylation, the dynamic stimulation‐induced ERK1/2 phosphorylation was larger and, although predominantly ipsilateral, bilateral (compare Fig. [NaN] A and B). This is probably because stroking the skin with a paintbrush produces a much larger cutaneous stimulation than a von Frey filament.

Among the laminae I–IIIo interneurons, the activation of which is associated with mechanical allodynia, are many lamina IIi PKCγ interneurons (Miraucourt et al., [21] , [22] ). To investigate if the activation of PKCγ interneurons was also associated with PDBu‐induced mechanical allodynia, we examined the phospho‐ERK1/2 signal in PKCγ‐immunoreactive interneurons after mechanical stimulation with a paintbrush in PDBu‐treated rats. Using dual immunocytochemistry, we found that 26.5 ± 5.6% (n = 5) of phospho‐ERK1/2‐immunoreactive cells within laminae IIi–IIIo of the ipsilateral MDH were also PKCγ immunoreactive in rats under PDBu‐induced dynamic mechanical allodynia (Fig. [NaN] ). It is noteworthy that, in control rats (intracisternal injection of vehicle), only a very few phospho‐ERK1/2‐immunoreactive cells (4.0 ± 4.0%; n = 5) were also PKCγ immunoreactive.

Phorbol ester, 12,13‐dibutyrate is known to activate all protein kinase C (PKC) isoforms (for review see Mochly‐Rosen et al., [24] ). Therefore, to test whether activation of PKCγ is involved in PDBu‐induced cutaneous mechanical hypersensitivity, we examined the effect of intracisternal injection of the selective PKCγ antagonist KIG31‐1 (50 and 100 pmol, 30 min before 3 nmol PDBu intracisternal injection) on PDBu‐induced static and dynamic mechanical allodynia. KIG31‐1 dose‐dependently attenuated static (Fig. [NaN] A; ‘F2,16 = 14.89’, P < 0.001) as well as dynamic (Fig. [NaN] B; ‘F2,16 = 22.22’, P < 0.001) mechanical allodynia. The effect of KIG31‐1 on mechanical allodynia was actually due to its antagonism of PKCγ activation as KIG31‐1 alone has no effect on mechanical cutaneous sensitivity (Miraucourt et al., [21] ; Petitjean et al., [33] ). Interestingly, comparison of the dose–response curves of the inhibition of PDBu‐induced static and dynamic mechanical allodynia by KIG31‐1 (Fig. [NaN] A and B, insets) reveals that the effect of KIG31‐1 was equivalent on both mechanical allodynia; the largest dose of KIG31‐1 produced a 84.4 ± 9.5% and 87.7 ± 6.6% (n = 5/group) reduction in the areas under the curve of PDBu‐induced static and dynamic mechanical allodynia, respectively.

We also monitored the effect of the intracisternal injection of KIG31‐1 (100 nmol) on dynamic mechanical stimulation‐induced phospho‐ERK1/2 immunoreactivity in PDBu‐treated (3 nmol) rats. Intracisternally injected KIG31‐1 significantly reduced the number of phospho‐ERK1/2‐immunoreactive neurons in both the ipsilateral and contralateral MDH (Fig. [NaN] C). Within the ipsilateral MDH, there were significant main effects of treatment (‘F1,8 = 8.60’, P = 0.006) and laminae (‘F2,8 = 54.31’, P < 0.001). There was also an effect of interaction between treatment and laminae (‘F2,8 = 5.13’, P = 0.011). Thus, selectively inhibiting PKCγ activity almost completely suppressed both PDBu‐induced static and dynamic mechanical allodynia and, accordingly, reduced the associated neuronal activation within the superficial MDH.

These experiments were designed to investigate the effect of selectively activating PKCγ within the MDH on the trigeminal mechanical sensitivity. We found that intracisternal injection of PDBu concomitantly induced static as well as dynamic mechanical allodynia. Monitoring neuronal activity within the MDH with phospho‐ERK1/2 immunoreactivity revealed that the activation of both laminae I–IIo and IIi–IIIo neurons, including lamina IIi PKCγ‐expressing interneurons, was associated with the expression of mechanical allodynia. PDBu‐induced mechanical allodynia and associated MDH neuronal activation were prevented by selectively inhibiting segmental PKCγ.

It is now well established that PKC activation within the SDH evokes pain hypersensitivity. Locally applied phorbol esters have been shown to produce punctate (measured with von Frey filaments) mechanical allodynia and hyperalgesia (Palecek et al., [29] ; Sluka & Audette, [41] ) and heat hyperalgesia (Kamei & Zushida, [13] ; Niikura et al., [26] ). Such mechanical and heat pain hypersensitivity is associated with the enhanced activity of nociceptive neurons, both background and responses to cutaneous mechanical stimuli (Palecek et al., [28] ; Lin et al., [16] ; Peng et al., [32] ). Interestingly, however, only neuronal responses to brush and punctate stimuli, but not to pinch, appear to be enhanced (Palecek et al., [28] ; Lin et al., [16] ; Peng et al., [32] ). Finally, phorbol esters facilitate excitatory synaptic transmission within the SDH, including that of the first sensory synapse (Gerber et al., [9] ; Palecek et al., [29] ; Sikand & Premkumar, [40] ), and concomitantly reduce inhibitory synaptic transmission (Lin et al., [16] ; Peng et al., [32] ). It is of note, however, that, whereas phorbol esters were shown to reduce the postsynaptic responses to glycine and GABA (Lin et al., [16] ), they can conversely facilitate presynaptic GABA release (Vergnano et al., [45] ). Altogether, these previous results clearly indicate that the activation of PKC within the SDH leads to the development of central sensitization resulting in pain hypersensitivity. Unfortunately, phorbol esters activate the eight PKC isozymes (for review see Mochly‐Rosen et al., [24] ). It is therefore not possible to assign a given effect to the activation of a selective PKC isozyme. By combining the intracisternal injection of PDBu with that of the selective PKCγ antagonist KIG31‐1, the present approach circumvents the lack of selectivity of phorbol esters; effects that are elicited by PDBu but suppressed by KIG31‐1 involve the selective activation of PKCγ, the PKC isozyme that is restricted to a population of lamina IIi SDH/MDH interneurons (Polgar et al., [34] ; Peirs et al., [31] ).

That locally applied KIG31‐1 prevents PDBu‐induced static as well as dynamic mechanical allodynia supports the conclusion that the activation of PKCγ is a key molecular event in the production of both types of mechanical allodynia. Previous results from our laboratory (Miraucourt et al., [21] , [22] ) and other laboratories (Malmberg et al., [19] ; Martin et al., [20] ; Zou et al., [48] ) indicate that the activation of the PKCγ isozyme within the SDH/MDH is necessary for the induction of either form of mechanical allodynia. The present report also suggests that PKCγ activation is not only necessary but also sufficient. Activation of PKCγ within lamina IIi PKCγ interneurons appears to determine whether LTMR inputs will gain access to the pain transmission circuitry of the superficial SDH/MDH and elicit pain.

In this study, we used ERK1/2 phosphorylation to assess MDH circuits underlying PDBu‐induced static and dynamic mechanical allodynia. Stimulation of the skin with an innocuous von Frey filament or light brush under intracisternal PDBu activated ERK1/2 within I–IIo and IIi–IIIo neurons, including many lamina IIi PKCγ‐expressing cells. Such a pattern of neuronal activation is similar to that reported for the dynamic mechanical allodynia produced by intracisternal injection of the glycine receptor antagonist strychnine (Miraucourt et al., [21] , [22] ). It is now clear that polysynaptic ventral‐to‐dorsal excitatory pathways, through which LTMR inputs can gain access to the pain transmission circuitry of the superficial SDH/MDH, are associated with mechanical allodynia. Thus, LTMR inputs were shown to engage lamina I nociceptive neurons: (i) in neuropathic rats expressing static mechanical allodynia (Keller et al., [14] ; Lu et al., [18] ); (ii) in rats expressing dynamic mechanical allodynia after local glycinergic disinhibition (Miraucourt et al., [21] ); and (iii) in SDH slices after combined GABAAergic and glycinergic disinhibition (Torsney & MacDermott, [43] ). Our results are thus consistent with the conclusion that dorsally‐directed local circuits are associated with mechanical allodynia. Such circuits probably begin in lamina IIi PKCγ interneurons that are known to receive inputs from myelinated LTMRs (Neumann et al., [25] ; Lu et al., [18] ; Peirs et al., [31] ). LTMRs signalling mechanical sensitivity to innocuous punctate (static mechanical allodynia) or light brushing (dynamic mechanical allodynia) stimuli would thus directly activate lamina IIi PKCγ interneurons and these interneurons in turn engage lamina I nociceptive neurons through a series of synaptically connected neurons, including lamina IIo central interneurons (Lu et al., [18] ).

Our results also suggest that the local circuits associated with static and dynamic mechanical allodynia are very similar, if not the same. Indeed, static and dynamic mechanical allodynia were concomitantly induced after PDBu and exhibited similar time‐courses (at all tested doses of PDBu) and dose–response curves (for both PDBu and KIG31‐1). Such a conclusion is surprising as there is evidence for the mechanisms underlying static and dynamic mechanical allodynia being different. Indeed, static and dynamic mechanical allodynia can manifest alone in patients (Ochoa & Yarnitsky, [27] ) or develop independently in animal models of pain (Field et al., [7] ; Sasaki et al., [37] ). Moreover, the static form of mechanical allodynia is selectively revealed under conditions of GABAAergic disinhibition (Roberts et al., [36] ; Miraucourt et al., [22] ) and the dynamic form of mechanical allodynia under conditions of glycinergic disinhibition (Miraucourt et al., [21] , [22] ), indicating that SDH/MDH circuits associated with static and dynamic mechanical allodynia are under selective GABAAergic and glycinergic inhibitory controls, respectively. Finally, static mechanical allodynia is selectively sensitive to morphine (Field et al., [7] ,[8] ; Gonzalez et al., [10] ; Catheline et al., [4] ; Miraucourt et al., [22] ) and amitriptyline (Field et al., [8] ), and the dynamic mechanical allodynia to amimexiletine and ketamine hydrochloride (Sasaki et al., [37] ). One possibility is that the two circuits underlying static and dynamic mechanical allodynia are only different in part. Static and dynamic mechanical allodynia are known to be signalled by different primary afferents, i.e. static mechanical allodynia by Aδ‐fibre (Ochoa & Yarnitsky, [27] ; Treede & Cole, [44] ; Field et al., [7] ; Seal et al., [39] ) and dynamic mechanical allodynia by Aβ‐fibre (Campbell et al., [3] ) LTMRs. It is therefore possible that the pathways signalling static and dynamic mechanical allodynia are different until the first synapse onto lamina IIi PKCγ interneurons but common from PKCγ interneurons onto lamina I nociceptive output neurons. The different modulations of static and dynamic mechanical allodynia might thus be accounted for by the different properties of Aβ‐fibre or Aδ‐fibre LTMRs and their synapses onto PKCγ interneurons. For instance, excitatory Aβ‐fibre inputs onto PKCγ interneurons have recently been shown to be controlled by selectively glycinergic feed‐forward inhibitory circuits (Lu et al., [18] ).

It is of note that only a small part of lamina IIi PKCγ interneurons were activated by mechanical stimulation, von Frey filaments or a paintbrush under PDBu application. Obviously, it is possible that Aβ‐LTMR and Aδ‐LTMR are received by two different sets of PKCγ interneurons, which would then initiate the activation of two parallel series of synaptically connected interneurons, transmitting Aβ‐LTMR and Aδ‐LTMR information to lamina I nociceptive output neurons. However, a more likely explanation is that only part of the lamina IIi PKCγ interneurons are directly involved in allodynia circuits. Indeed, our group (Alba‐Delgado et al., [1] ) recently provided evidence for lamina IIi PKCγ‐expressing interneurons being a heterogeneous class of interneurons, with at least two morphologically and functionally different subpopulations, i.e. central and radial PKCγ interneurons.

In summary, the present results provide further evidence of the essential contribution of PKCγ activation to the manifestation of both static and dynamic mechanical allodynia. They suggest that, once PKCγ interneurons are sensitized, LTMR inputs that synapse onto these PKCγ interneurons can gain access to the pain transmission circuitry of the superficial SDH/MDH and elicit pain. Consistent with this, ablating parvalbumin‐expressing inhibitory interneurons in naive mice was recently also shown to produce mechanical allodynia via the disinhibition of PKCγ interneurons (Petitjean et al., [33] ). Altogether, our results point to PKCγ and the PKCγ interneurons as potential therapeutic targets for the treatment of mechanical allodynia.

The authors declare no competing financial interest.

Study concept and design: N.P.D. and R.D. Acquisition of data: N.P.D. and A.D. Analysis and interpretation of data: N.P.D., A.D., R.D. and A.A. Drafting of the manuscript: N.P.D., R.D. and A.A. Obtained funding: R.D. Study supervision: A.A. and R.D.

This work was supported by funding from the Institut National de la Santé et de la Recherche Médicale (Inserm), Université Clermont‐Ferrand 1 (France), and Région Auvergne (France). We thank Anne‐Marie Gaydier for secretarial assistance.

Abbreviations

II i inner lamina II

II o outer lamina II

III o outer lamina III

ERK1⁄2 extracellular signal‐regulated kinases 1 and 2

LTMR low‐threshold mechanoreceptor

MDH medullary dorsal horn

NGS normal goat serum

PBS‐Tx phosphate‐buffered saline with 0.2% Triton

PDBu phorbol ester, 12,13‐dibutyrate

PKC protein kinase C

PKCγ γ isoform of protein kinase C

SDH spinal dorsal horn

Graph: Intracisternal injection of PDB u dose‐dependently induces a facial static as well as dynamic mechanical allodynia. Time‐courses of the changes in behavioural responses (A and C) and bar histograms of the corresponding areas under the curve ( AUC s) (B and D) evoked by static (6‐g von Frey filament) (A and B) and dynamic (air puff) (C and D) mechanical stimuli applied on the face of rats intracisternally injected with vehicle (5 μL) or PDB u (0.03, 0.3 and 3 nmol in 5 μL). Data are represented as mean ±  SEM. Statistical analysis in B and D was performed using one‐way anova followed by Bonferroni post hoc test ( n  = 5 per group). ** P  <   0.01, *** P  < 0.001.

Graph: Neural circuits associated with PDB u‐induced static and dynamic mechanical allodynia both encompass laminae I– IIIo. Images of phospho‐ ERK 1/2 immunolabelling in the MDH of rats at 30 min after intracisternal injection of vehicle (5 μL) (A) or PDB u (3 nmol in 5 μL) with static (6‐g von Frey filament) (B) or dynamic (light brushing) (C) mechanical stimulation applied on the face. From left to right: contralateral MDH , ipsilateral MDH and a higher magnification of the box in the ipsilateral MDH of the same animals. Dotted lines on the right indicate, from dorsal to ventral, the limits of laminae I, IIo and IIi. There are no or very few phospho‐ ERK 1/2‐immunoreactive cells following intracisternal vehicle but many within laminae I– IIo as well as within laminae IIi – IIIo following intracisternal PDB u plus static or dynamic mechanical stimulation. It is of note that phospho‐ ERK 1/2‐immunoreactive cells are much more numerous after dynamic (C) than static (B) mechanical stimulation.

Graph: Neural circuits associated with PDB u‐induced static and dynamic mechanical allodynia both encompass laminae I– IIIo. (A and B) Bar histograms of the number of phospho‐ ERK 1/2‐immunoreactive cells in the different laminae of the ipsilateral (left) and contralateral (right) MDH in rats after intracisternal injection of PDB u (3 nmol in 5 μL) plus mechanical stimulation ( PDB u‐S), vehicle alone (5 μL), PDB u alone and vehicle plus mechanical stimulation (Vehicle‐S). Mechanical stimulation was applied on the face with either a 6‐g von Frey filament (60 times at 1 Hz) (A) or light brushing (0.5 Hz for 10 min) (B). IIIi , inner lamina III. Data are represented as mean +  SEM. Both the von Frey filament and light brush induce a strong phospho‐ ERK 1/2 immunoreactivity, similarly located in the superficial MDH , within laminae I– IIo and IIi – IIIo. It is of note that a paintbrush induces a much stronger phospho‐ ERK 1/2 immunoreactivity than a von Frey filament (compare ordinates). Statistical analysis was performed using a two‐way anova followed by a Duncan post hoc test ( n  = 5 per group). * P  <   0.05, ** P  <   0.01, *** P  <   0.001.

Graph: PKC γ‐immunoreactive interneurons in lamina IIi are activated by dynamic mechanical stimuli in rats under PDB u‐induced mechanical allodynia. Fluorescence images of phospho‐ ERK 1/2‐immunoreactive cells (red), PKC γ‐immunoreactive interneurons (green), and double‐labelled neurons (white arrows) within lamina IIi of the MDH in rats after intracisternal injection of PDB u (3 nmol in 5 μL) and dynamic mechanical stimuli (lightly stroking the skin with a paintbrush) applied on the face. Insets show the double‐labelled neurons.

Graph: Intracisternal injection of KIG 31‐1 dose‐dependently prevents both the cutaneous mechanical hypersensitivity and associated neuronal activation within the superficial MDH induced by intracisternal PDB u. Time‐courses of changes in behavioural responses evoked by static (6‐g von Frey filament) (A) and dynamic (air puff) (B) mechanical stimuli applied on the face of rats intracisternally injected with PDB u (3 nmol in 5 μL) at t  =   0. Rats were preemptively intracisternally injected with either Tat carrier (5 μL) or KIG 31‐1 (50 or 100 pmol in 5 μL). Insets : Bar histograms showing the areas under the curve ( AUC s) of the corresponding time‐courses. (C) Bar histograms of the number of phospho‐ ERK 1/2‐immunoreactive cells in the different laminae of the ipsilateral (left) and contralateral (right) MDH in rats after intracisternal injection of PDB u (3 nmol in 5 μL) plus dynamic mechanical stimulation in rats preemptively intracisternally injected with either Tat carrier (5 μL) or KIG 31‐1 (100 pmol in 5 μL). Data are represented as mean ±  SEM. Statistical analysis was performed using one‐way anova followed by a Bonferroni post hoc test (A and B) and a two‐way anova followed by a Duncan post hoc test (C) ( n  = 5 per group). * P  <   0.05, *** P  <   0.001. III i, inner lamina III.