P2X7 receptor blockade reverses purinergic facilitation of neck muscle nociception in mice

Abstract

Introduction: Facilitation of neck muscle nociception mediated via purinergic signalling may play a role in the pathophysiology of tension-type headache (TTH). The present study addressed reversal of purinergic facilitation of brainstem nociception via P2X7 antagonist action in anaesthetized mice.

Methods: Following administration of α,β-meATP (i.m. 20 µL/min, 20 µL each) into semispinal neck muscles, the impact of neck muscle nociceptive input on brainstem processing was monitored by the jaw-opening reflex in anaesthetized mice (n = 20). The hypothesized involvement of the P2X7 receptor in the α,β-meATP effect was addressed with i.p. (systemic) and i.m. (semispinalis, 20 µL/min, 20 µL each) administration of P2X7 inhibitor A438079 during established facilitation; i.p. saline served as control.

Results: α,β-meATP reliably induced jaw-opening reflex facilitation (256 ± 48% (mean ± SEM), n = 20). I.p. A438079 (150, 300 µmol/kg) completely reversed this α,β-meATP effect dose-dependently. Neither saline nor intramuscular A438079 (100 µM) altered facilitated brainstem nociceptive processing.

Discussion: These data suggest that muscular structures are not directly involved in the P2X7 antagonist-mediated reversal of purinergic facilitation. Instead, involvement of neuronal structures, particularly of the central nervous system, seems more probable. The results from this animal experimental model may point to involvement of purinergic P2X7 receptors in TTH pathophysiology and may suggest potential future targets for its pharmacological treatment.

Introduction

Tension-type headache (TTH) is the most frequent primary headache with important socioeconomic impact (1). Increased pericranial muscle tenderness is a hallmark of TTH (2,3). A translational mouse model on putative pathophysiological mechanisms of TTH was developed. This model addresses the impact of nociceptive afferent input from neck muscles on central nervous system nociceptive processing monitored by the jaw-opening reflex (JOR) (4–12). The present animal model is suggested to be a translational model for the investigation of pathophysiological aspects of TTH (13). The JOR is an accepted model to investigate alterations of excitability in sensory brainstem neurons with convergent afferent input from different craniofacial tissues such as neck muscles. Noxious input into semispinal neck muscles via intramuscular infusion (i.m.) of α,β-meATP evokes sustained reflex potentiation. This indicates heterosynaptic facilitation due to access of nociceptive afferents from neck muscle to the reflex neuronal network in the brainstem. This facilitation seems to involve two stages of a peripherally triggered induction and a centrally established maintenance phase (4). Accordingly, nociceptive input from neck muscles is regarded as the main trigger for episodic TTH. With prolonged exposition, central sensitization might be induced leading to TTH chronification. Hence, aspects of dysfunctional central pain processing structures cardinal in the development of TTH can be investigated with this model.

P2X receptors (P2X1–7) are characterized as ionotropic, ATP-sensitive receptors. ATP exhibits excitatory effect on sensory neurons via activation of ionotropic P2X3 receptors which mediate nociceptive processing (14–19). α,β-methylene adenosine 5′-triphosphate (α,β-meATP) is an ATP substrate derivative binding to P2X1, P2X3 and P2X2/3 receptors (20). Intramuscular infusion of α,β-meATP into murine neck muscles facilitates brainstem nociceptive processing (4,5). Similarly, infusion of ATP into the trapezius muscle induces both pain and local tenderness in humans (21).

Besides the neuronal P2X3 receptor, P2X7 has also recently been associated with nociception (22–25). P2X7 gene disruption in mice generated a phenotype with absent hypersensitivity to mechanical and thermal stimuli in models of inflammatory and neuropathic pain (22). Additionally, up-regulation of P2X7 was demonstrated in human dorsal root ganglia and injured peripheral nerves in chronic neuropathic pain patients. This suggests a possible up-stream role of P2X7 in the development of pain conditions. In rats, selective P2X7 inhibition resulted in reduced noxious evoked activity in spinal neurons. This reduction points to modulation of central sensitization and antinociception via P2X7 antagonism (23). Furthermore, microglial P2X7 seems to be involved in long-term potentiation (LTP) of spinal nociceptive processing (25,26) and thus might mediate neuroplastic changes in chronic pain conditions. Both P2X7 antagonism and P2X7 siRNA-mediated down-regulation of P2X7 prevented the induction of spinal LTP and alleviated mechanical allodynia in rats. This implies a crucial role for neuron–microglia interaction in spinal neuroplasticity (25). A438079 is a rather new, selective, competitive P2X7 receptor antagonist (www.tocris.com) (23,24). Its antinociceptive properties have been demonstrated in different animal models of neuropathic pain including spinal nerve ligation, chronic constriction injury and vincristine administration. Depending on the model of neuropathic pain in rats, published ED50 values vary between 76 and 200 µmol/kg for exclusively systemic (i.p.) administration.

So far, P2X7 signalling in sensitization with regard to putative TTH pathophysiological mechanisms has not been investigated despite increasing evidence from different animal experimental models of P2X7-mediated nociception and neuroplastic changes. Participation of the P2X7 receptor in α,β-meATP-triggered cascades mediating facilitation of brainstem nociception seems possible. Hence, the present study addressed the effect of A438079-mediated P2X7 receptor antagonism on α,β-meATP-induced facilitation of neck muscle nociceptive processing in mice. Reversal of the α,β-meATP effect via P2X7 inhibition was hypothesized. The scientific focus was based upon systemic P2X7 antagonist administration under stable facilitation mimicking acute medication of TTH attacks.

Methods

Electrophysiological experiments were performed in adult male C57BL/6 mice (approximately 12 weeks old; Taconic, www.taconic.com). All procedures received institutional approval from the local ethics committee. The principles of laboratory animal care and use of laboratory animals (European Council Directive of November 24, 1986(86/609/EEC)) were followed. All efforts were made to minimize animal suffering and to use only the number of animals necessary to produce reliable scientific data.

Detailed descriptions of anaesthesia, surgery and electrophysiological recording have been published in detail elsewhere (27). Briefly, mice were anaesthetized by an initial intraperitoneal (i.p.) injection of a 0.5% pentobarbital sodium salt solution (Amgros I/S, Copenhagen, Denmark, www.amgros.dk) with a dose of 70 mg/kg. Depth of anaesthesia was checked by ensuring that noxious pinch stimulation (blunt forceps) of hind paw, forepaw and ear did not evoke any sensorimotor reflexes. When the animal was sufficiently deeply anaesthetized, the skin of the throat and neck were carefully shaved and lidocaine hydrochloride gel (Xylocaine® 2%, AstraZeneca A/S, Albertslund, Denmark, www.astrazeneca.com) was applied to the skin of the throat to induce local anaesthesia. Dexpanthenol eye ointment (Bepanthen®, Roche, Grenzach-Wyhlen, Germany, www.roche.com) was applied to cornea and conjunctiva of both eyes to protect them from drying. The right external jugular vein was catheterized for continuous administration of a 2% methohexital sodium salt solution (Brevimytal® Hikma, Hikma Pharmaceuticals PLC, London, UK, www.hikma.de) with a dose of 60 mg/kg per hour corresponding to a flow rate of about 0.07 mL/h for a 23 g mouse. A pair of Teflon-coated stainless steel wires (140 µm diameter) was inserted into the right anterior digastric muscle (Dig) to record electromyographic activity (EMG) and the JOR via a differential amplifier. After tracheotomy, animals were placed in a stereotactic frame and were artificially respired with a stroke volume of about 150 µL and about 200 strokes/min (MiniVent Model 845, Harvard Apparatus, Hugo Sachs Elektronik, March-Hugstetten, Germany, www.harvardapparatus.com). Body core temperature was maintained at 37.3°C with a heating blanket and a fine rectal thermal probe (FMI GmbH, Seeheim-Ober-Beerbach, Germany, www.fmigmbh.de). One platinum needle electrode each (300 µm diameter) was subcutaneously inserted into right forepaw and left hindpaw to record the electrocardiogram (ECG) via a differential amplifier. Two stainless steel needle electrodes (150 µm diameter) were longitudinally inserted into the tongue musculature (parallel, 2 mm distance) in order to apply electrical stimuli and to evoke the JOR. The oral cavity was filled up with white Vaseline (Riemser Arzneimittel AG, Greifswald-Insel Riems, Germany, www.riemser.de) to protect oral mucous membrane from drying. Neck skin was locally anaesthetized via Xylocaine®. Semispinal neck muscles on both sides were carefully exposed. One injection cannula (0.4 mm diameter) was inserted into each muscle belly of both semispinal neck muscles. Each cannula was connected via thin and short tubing to a liquid switch (CMA/110, CMA Microdialysis AB, Solna, Sweden, www.microdialysis.se). Glass microsyringes (1 mL) were connected to the liquid switch by thin tubing and were fixed in a microdialysis pump (CMA 102, www.microdialysis.se). Use of this liquid switch enabled repeated injection at the same i.m. site without changing cannula and without interruption of flow during the in vivo experiment. This procedure allowed bilateral induction of noxious input from neck muscles in order to mimic bilateral neck muscle pain in TTH patients. 100 nM α,β-meATP (Sigma-Aldrich Chemie Gmbh, Munich, Germany, www.sigmaaldrich.com) was i.m. administered with a volume of 20 µL per semispinal neck muscle during a time period of 1 minute. The same infusion site, tubing system, infusion rate (20 µL/min) and volume (20 µL per each muscle) was used for bilateral i.m. administration of A438079.

After preparation, the anaesthetized animal was rested for at least 1 hour. During this period level of anaesthesia and heart rate were routinely checked and documented, and depth of anaesthesia was maintained. All electrical signals (EMG, ECG) were recorded by bioamplifiers. EMG signals led into a data collection system (CED Micro1401, CED, Cambridge Electronic Design Ltd, Cambridge, UK, www.ced.co.uk) and a personal computer using Signal® software program (CED, Cambridge Electronic Design Ltd, Cambridge, UK, www.ced.co.uk).

The JOR was elicited by electrical stimulation of afferent nerve fibres in the tongue musculature via two needle electrodes with rectangular electrical pulses of 500 µs duration and a stimulation frequency of 0.1 Hz (Figure 1). Reflex responses were recorded in the anterior digastric muscle by electromyography (EMG). A reflex response elicited by a single stimulus is quantified by its onset latency, duration and integral (Figure 1A). The duration covers the time window between onset latency and end of the reflex response in the digastric muscle EMG. In this time window, the reflex integral (area under the curve) was calculated. Individual single-sweep analysis of the JOR was performed on a visual basis. Reflex on- and off-set were defined by the first change of basal EMG signal after each stimulus to the recovery to basal signal levels. The JOR is well quantifiable. The only source of noise is derived from the stimulus artefact due to the close proximity of electrical stimulus (tongue) and recording (digastric muscles) electrodes. Stimulus artefact and actual JOR can easily be separated (Figure 1A). The electrical threshold of the JOR was determined by applying one series of increasing and decreasing stimulus intensities from 0 to 2 mA in steps of 100 µA. The lowest stimulus intensity that just evoked a reflex response was defined as the reflex threshold. Test stimulus intensity was adjusted to approximately 150% of the reflex threshold. The JOR was evoked in a series of eight stimuli (Figure 1B). These series were repeated every 5 minutes. Drugs were A438079 (3-[[5-(2,3-dichlorophenyl)-1 H-tetrazol-1-yl]methyl]pyridine hydrochloride, Tocris Bioscience, Bristol, UK), sodium chloride solution and α,β-meATP (α,β-methylene adenosine 5′-triphosphate). α,β-meATP and A438079 were dissolved in isotonic saline solution; i.p. of saline served as a control. Former studies demonstrated stable JOR with i.m. saline infusion in this murine model (10,11). In order to reduce the number of animals used, the i.m. saline infusion was not repeated. The molarity of administered A438079 solutions for i.m. infusion was adapted to equimolar and 1000-fold of α,β-meATP concentrations of 100 nM and 100 µM corresponding to dosages of 0.16 nmol/kg and 0.16 µmol/kg, respectively. A438079 dose selection for systemic administration was based on reported antinociceptive dosages in rats (150, 300 µmol/kg) (23,24).

Figure 1.

Stimulation protocol. (A) The jaw-opening reflex (JOR) was evoked by electrical stimulation of afferent nerve fibres in the tongue via two needle electrodes. Reflex responses were recorded in the anterior digastric muscle by electromyography (EMG) (a.u.: arbitrary units). The graph shows a typical reflex response elicited by a single stimulus. Dotted lines mark onset and end latency which determine the reflex duration. In this time window, the reflex integral (area under the curve in grey) was analysed as the most prominent reflex parameter. (B) The JOR was evoked in a series of eight stimuli each (black bars). Reflex series were repeated every 5 minutes (grey bars). Drugs were applied intraperitoneally (i.p.) or intramuscularly (i.m.) into both semispinal neck muscles as marked by arrows. After three baseline series, α,β-meATP was i.m. administered and the reflex was recorded for the following 90 minutes. Subsequently, A438079 was intraperitoneally (i.p.) injected (150, 300 μmol/kg) or intramuscularly (i.m.) infused (100 µM). Isotonic saline (i.p.) served as control. The reflex was recorded for a further 60 minutes.

The experiments were conducted (Figure 1) in 20 mice (n = 5 each). The effect of subsequent i.m. or i.p. administered A438079 on α,β-meATP-induced reflex facilitation was investigated (Figure 1B). After stable baseline recording, α,β-meATP was infused i.m. Purinergic reflex facilitation was observed for the following 90 minutes. Then, A438079 was administered either i.m. (100 µM) or i.p. (150, 300 µmol/kg); i.p. saline served as control. The reflex was then monitored for further 60 minutes.

JOR integrals were analysed in each single sweep. Arithmetic mean and standard error of mean (mean ± SEM) were calculated. For the facilitative effect of α,β-meATP on JOR, percentage changes based upon average reflexes in the time window from −15 to −5 minutes. For the effect of subsequent A438079 or saline on the facilitated JOR, percentage changes based upon average reflex integrals in the time window from 80 to 90 minutes (Figure 1B). Depending on data distribution, comparison within groups was performed via paired t-test (t- and p-values), one-way repeated-measures ANOVA (one-way RM ANOVA, F- and p-values) or Friedman one-way RM ANOVA (Χ² and p-value). Differences between and within groups were analysed via two-way repeated-measures analysis of variance (two-way RM ANOVA, F- and p-values). Factors were substance and time. Data were also analysed for interaction between these factors. Post hoc comparisons were performed with the Student–Newman–Keuls test (SNK, p-values). Level of significance was set to p < 0.05. Additionally, one-way ANOVA (F- and p-values) was performed on mean JOR integrals (100 to 155 min) followed by SNK test post hoc comparisons in order to further evaluate the effect of drugs or saline on the facilitated JOR.

Results

Electrical tongue stimulation elicited the JOR with a threshold intensity of 0.7 ± 0.1 µA (mean ± SEM) in 20 healthy mice (26.8 ± 0.7 g). Test stimulus intensity was adjusted to 1.0 ± 0.3 mA corresponding to 149 ± 6% of the reflex threshold. I.m. α,β-meATP reliably induced reflex facilitation.

Effects of α,β-meATP on the basal JOR (Figure 2)

α,β-meATP reliably induced an increase of the JOR in 20 mice until 90 minutes after α,β-meATP infusion. Both single average reflex sweeps (n = 1) and grand averages (n = 5) illustrated increased brainstem nociceptive processing between groups with subsequent i.p. 300 µmol/kg A438079 and i.p. saline (Figure 2). Mean reflex values steadily increased after i.m. α,β-meATP administration and reached a stable plateau before subsequent drug administration at 256 ± 48% (mean of JOR integrals at 80, 85 and 90 minutes before subsequent administration of saline or A438079). The data from these experiments were then compared with the other groups (i.m. 100 µM A438079, i.p. 150 µmol/kg A438079) (Figure 3).

Figure 2.

Reversal of purinergic reflex facilitation with i.p. 300 µmol/kg A438079. Top: Eight rectified sweeps are averaged under baseline conditions (−5 minutes, grey), 90 minutes after i.m. α,β-meATP infusion (black) and 60 minutes after subsequent i.p. administration (155 minutes, dark grey) of either saline (left) or 300 µmol/kg A438079 (right, n = 1 each). Electromyographic activity is given in arbitrary units (a.u.). Bottom: JOR grand average (n = 5 each) are displayed for either saline (left) or 300 µmol/kg A438079 (right) under baseline conditions (−5 minutes, grey), 90 minutes after i.m. α,β-meATP infusion (black) and 60 minutes after subsequent i.p. administration (155 minutes, dark grey). JOR facilitation was completely reversed to basal levels with subsequent A438079 (300 µmol/kg) whereas facilitated digastric muscle EMG remained unchanged with subsequent saline injection (control). Single-sweep averages from the top graphs are integrated into these group averages, and went into further analysis (cf. Figure 3).

Figure 3.

Dose-dependent reversal of reflex facilitation with subsequent i.m. A438079. During established reflex facilitation 90 minutes after intramuscular α,β-meATP, i.p. saline (white), i.p. 150 µmol/kg A438079 (grey), i.p. 300 µmol/kg A438079 (dark grey) or i.m. 100 µM A438079 (bright grey) was applied (n = 5 each). (A) Reflex facilitation remained unchanged with subsequent i.p. saline and i.m. A438079. Systemically administered A438079 significantly decreased facilitated reflex integrals with dosages of 150 and 300 µmol/kg (two-way RM ANOVA). Data are presented as mean ± SEM. (B) Average reflex integral changes after i.m. injection of A438079 (100 µM), systemic saline or A438079 (150, 300 µmol/kg) are presented as box plots (dotted line: arithmetic mean). Reversal of average reflex values occurred exclusively with i.p. administered A438079. This effect was more pronounced with 300 µmol/kg dose. F- and p-values correspond to one-way ANOVA. Student–Newman–Keuls post hoc tests indicated significant differences (**p < 0.01, ***p < 0.001).

Effects of subsequent saline and A438079 (i.m., i.p.) on α,β-meATP-induced reflex facilitation (Figures 2 and 3)

After stable α,β-meATP-induced facilitation of the JOR, A438079 was administered either i.p. (150, 300 µmol/kg) or i.m. (100 µM). Saline (i.p.) served as a control (n = 5 each). Average reflex sweeps illustrated different levels of brainstem nociceptive processing between groups with subsequent i.m. A438079, systemic saline and 300 µmol/kg A438079 (Figure 2). Whereas reflexes seemed similar within baseline and facilitation time segments, recorded sweeps demonstrated clearly reduced levels of brainstem nociceptive processing with subsequent systemic A438079.

This reversal of α,β-meATP-induced reflex facilitation occurred dose dependently with systemic A438079 dosages (150, 300 µmol/kg) (Figure 3). Two-way RM ANOVA on all subsequent 5 minute interval data revealed reversal of the facilitated JOR with subsequently i.p. administered A438079 dosages of 150 and 300 µmol/kg (Figure 3A). There were significant effects of factors time (F = 4.4, p < 0.001), substance (F = 25.0, p < 0.001) and interaction (F = 2.7, p < 0.001). Reflex integrals with subsequent 300 µmol/kg A438079 were significantly lower than with all other groups (p < 0.001, vs. both saline and i.m. A438079; p < 0.01 vs. 150 µmol/kg A438079). Reflex integrals with subsequent 150 µmol/kg A438079 were significantly lower than with subsequent saline (p < 0.01) and i.m. A438079 (p < 0.001). In contrast, subsequent saline (i.p.) and i.m. A438079 (100 µM) showed no significant impact on the facilitated reflex. Differences between groups occurred from 10 minutes (corresponding to 105 minutes) after drug administration for i.p. A438079 (300 µmol/kg) and persisted until the end of the experiment. Significant differences were not present vs. 150 µmol/kg within 15 to 40 minutes after A438079 injection (corresponding to 110 to 135 minutes). Differences between groups after drug administration for A438079 occurred from 15 minutes (corresponding to 110 minutes) (i.p., 150 µmol/kg) and persisted until the end of the experiment.

For further documentation of the dose-dependent effect of systemically administered A438079, the JOR data after application of drug or saline was averaged in the time window of 5 to 60 minutes after subsequent administration (corresponding to 100 to 155 minutes) for each group (Figure 3B). The mean reflex values for i.p. saline (102 ± 5%) and i.m. A438079 (100 µM, 98 ± 9%) were significantly higher than both i.p. 150 µmol/kg (43 ± 9%) and i.p. 300 µmol/kg A438079 (−9 ± 16%).

Systemic A438079 induced persistent reversal of JOR facilitation in the recorded time window. Data support the participation of P2X7 receptor in purinergic facilitation of neck muscle nociceptive processing.

Discussion

The experiments demonstrate inhibitory effects of A438079 on α,β-meATP-induced facilitation of neck muscle nociception in anaesthetized mice. Exclusively the administration of i.p. A438079 reversed purinergic facilitation.

Electrical stimulation of afferent nerve fibres in the tongue musculature reliably evokes the JOR via a brainstem reflex network (27). Sensory neurons in the spinal trigeminal nucleus receive convergent input from primary afferents (neck muscle, tongue) and project to digastric motoneurons (28,29). Noxious input from neck muscles leads to sustained reflex potentiation evoked by α,β-meATP (4–6,11). Facilitation of brainstem nociceptive processing could be observed for at least 4 hours (11). α,β-meATP interacts with homomeric P2X3 and heteromeric P2X2/3 receptors (20), which in turn mediate nociception (14,19,30), probably via group III muscle afferents (6). This additional excitatory input from neck muscle nociceptors facilitates the tongue-evoked reflex by heterosynaptic access. The α,β-meATP effect seems independent of mechanical factors (volume) (6,9,11). This contradicts local mechanical effects of volume injection and suggests pharmacological actions of α,β-meATP. Furthermore, α,β-meATP does not induce increased semispinal or digastric EMG activity during i.m. infusion (5). This implies that basal level of muscle activity remains unaffected by i.m. α,β-meATP, i.e. it emphasizes in turn pharmacological action on sensory muscle afferents. The present animal model is suggested to be a translational model for the investigation of pathophysiological aspects of TTH (13). The experimental set-up with A438079 administration after established purinergic facilitation of brainstem nociception due to increased neck muscle nociceptive input resembles acute treatment of TTH attack.

P2X7 receptor pharmacology and its therapeutic potential gains increasing interest in inflammatory (31) and chronic pain states (32). LTP of synaptic transmission can be induced by high-frequency stimulation in the nociceptive system (33) and is thought to be involved in central sensitization, hyperalgesia, and chronic pain (34). Inhibition of P2X7 prevented induction of LTP in both in vivo and in vitro models of tetanic electrical stimulation of the sciatic nerve and alleviated mechanical allodynia in the rat (25). Accordingly, P2X7 down-regulation via P2X7 siRNA blocked LTP induction and inhibited mechanical allodynia. P2X7 gene disruption in mice resulted in abolished hypersensitivity in inflammatory and neuropathic pain models (22). Interestingly, this hypersensitivity was completely absent despite preserved normal nociceptive processing leading to a hypothesized common upstream transductional role of P2X7 in the development of pain. Similarly, selective P2X7 blockade resulted in anti-hyperalgesia in vivo in rats (35). Furthermore, the role of P2X7 in increased pain processing was demonstrated in different animal models with the antagonistic action of systemically administered A-740003 (36). In this study, different experimental pain models such as spinal nerve ligation, chronic constriction injury of the sciatic nerve, vincristine-induced neuropathy, thermal hyperalgesia induced by administration of carrageenan and Freud’s adjuvant were used. Taken together, participation of P2X7 in divergent elevated states of nociceptive transmission and pain processing seems evident.

A438079 is a selective and competitive P2X7 receptor antagonist with 19% bioavailability and a plasma elimination half-life of 1 hour (24). Applied i.p. dosages of A438079 (150 and 300 µmol/kg) in this study were adapted according to current literature (23,24). I.m. A4380789 administration was not previously described. Yet, P2X7 was recently reported to be present in skeletal muscles (37), though the majority of current literature describes the presence of this purinergic receptor in glial cells (38) and neuronal tissue (39–41). Hence, the same concentration as α,β-meATP (100 nM) was chosen in preceding A438079 pilot experiments. Because there was no effect, a concentration of 100 µM for i.m. A438079 was chosen to investigate any local effects of P2X7 inhibition. In an in vivo model of neuropathic pain via L5/6 spinal nerve ligation, systemic administration of A438079 (10 to 300 µmol/kg) resulted in a dose-dependent reversal of mechanical allodynia (ED50: 76 µmol/kg). In other rat models of pain, reported ED50 values for i.p. administered A438079 were 80 µmol/kg (chronic constriction injury), 102 µmol/kg (formalin) and 200 µmol/kg (vincristine) (23). In accordance with previously published data, in this study applied dosages suffice for antinociceptive A438079 effects. Systemic A438079 had no influence on rotarod performance tests and did not induce behavioural effects (23). Furthermore, antinociceptive effects were documented for 35 minutes. As this was the maximal observation time, a more detailed description on the duration of the A438079 effect is lacking. In the present study, the reversal of α,β-meATP-induced facilitation of brainstem nociceptive processing effect via P2X7 inhibition with 300 µmol/kg A438079 i.p. was stably documented for at least 1 hour. Beyond this regular observation time of 60 minutes (100 to 155 minutes) for putative effects of established JOR facilitation, additional data supports a lasting reversal for at least 90 minutes (300 µmol/kg).

It remains not entirely clear in which tissue the antagonistic P2X7 effect impacts in the present study. Pharmacokinetic properties of A438079 with systemic administration included a brain to plasma ratio of 2 : 1, suggesting a pronounced central nervous system A438079 distribution and action (23). The P2X7 receptors are reported to be localized in glial cells such as microglia (38), macrophages and monocytes (42,43) and neuronal tissue (18). There is evidence for presynaptic presence of P2X7 in rat cortical nerve terminals (44) and localization of P2X7 to excitatory terminals in the rat hippocampus (40), implying further P2X7 presence in the central nervous system. Focusing on P2X7 mRNA distribution in the rat brain using isotopic in situ hybridization (45), double staining revealed expression of P2X7 receptor mRNA on neurons, oligodendrocytes and microglia. However, there are implications that glial P2X7 might be crucial in signalling cascade of purinergic facilitation of neck muscle pain in the present study. α,β-meATP-induced LTP of neuronal excitation in the rat superficial spinal dorsal horn was inhibited by treatment with the glial metabolism inhibitor monofluoroacetic acid (26). These findings indicate that glial cells contribute to the α,β-meATP-induced LTP, which might be part of a cellular mechanism for the induction of persistent pain. In particular, so-called ‘primed’ spinal microglia might play a crucial role in central sensitization of nociceptive processing (46). Spinal microglia may become sensitized following their activation by, for example, disparate forms of peripheral trauma or inflammation resulting in enhanced pain intensity and duration. Furthermore, ATP release was demonstrated from nerve terminals and dorsal root ganglia neurons (47,48) enabling neuron–glia signalling (38). This ATP release was demonstrated after electrical stimulation of dorsal root ganglia neurons (49). Similarly, a P2 receptor-mediated neuronal communicative pathway has been reported in dissociated murine sensory trigeminal ganglia that probably utilizes ATP for bidirectional communication (50). Though a direct link between peripherally α,β-meATP-driven induction and a microglial involvement via P2X7 signalling in maintaining the facilitation is still missing, this synergistic cross-talk can be considered. In summary, it is not entirely clear whether reversal of purinergic facilitation via P2X7 inhibition in the present study might be purely localized centrally. Local drug administration can affect purinergic facilitation in this model (6). Both preceding and subsequent i.m. infusion of semispinal neck muscles with tetrodotoxin (TTX) prevented and reversed the α,β-meATP effect. In contrast, TTX did not alter JOR facilitation evoked by i.m. nerve growth factor infusion. TTX was locally infused only 30 minutes after α,β-meATP administration, i.e. in a time window in which the peripheral drive may not have resulted in centrally established facilitation. Consequently, it can only be hypothesized that local, subsequent TTX administration with established JOR facilitation 90 minutes after i.m. α,β-meATP would not affect facilitated brainstem nociception in this model. This would in turn confirm central sensitization uncoupled from the peripheral drive with the given timeline.

Using different experimental set-ups, the role of the P2X7 receptor in this murine model and its implication in TTH pathophysiology could be clarified further. Especially preceding systemic A438079 administration prior to purinergic facilitation could provide more insight into putative P2X7 involvement in the induction of purinergic facilitation, not only in its maintenance. Additionally, i.m. administration of other algogenic substances such as hypertonic saline or nerve growth factor in combination with A438079 would clarify further the role of P2X7 in this model. These substances provide nociceptive input into semispinal neck muscles and facilitation of brainstem nociception as well (9,10).

The demonstration that A438079 reverses purinergic facilitation of brainstem nociceptive processing in this translational murine model suggests involvement of central nervous system P2X7 receptors in the facilitation of neck muscle nociceptive processing in the brainstem. Direct functional evidence of this suggested P2X7 axis in dysfunctional central pain processing within TTH chronification is missing. These results may point to involvement of purinergic P2X7 receptors in TTH pathophysiology and may suggest potential future targets for its pharmacological treatment, particularly in regard to neuron–glia interaction.