ATP Induces Sustained Facilitation of Craniofacial Nociception Through p2x Receptors on Neck Muscle Nociceptors in Mice

Abstract

Noxious input from neck muscles probably plays a key role in tension-type headache pathophysiology. ATP selectively excites group III and IV muscle afferents in vitro. Accordingly, ATP infusion into trapezius muscle induces strong pain and local tenderness in healthy man. The present study addresses the impact of ATP on neck muscle nociception in anaesthetized mice. Craniofacial nociceptive processing was tested by the jaw-opening reflex via noxious electrical tongue stimulation. Within 2 h after injection of 100 nmol/l or 1 μmol/l ATP into semispinal neck muscles, reflex integrals significantly increased by 114% or 328%, respectively. Preceding intramuscular administration of the P2X receptor antagonist PPADS (3–100 nmol/l) suppressed the ATP effect. Subsequent application of PPADS (100 nmol/l) caused a total recovery of facilitated reflex to baseline values. ATP induces sustained facilitation of craniofacial nociception by prolonged excitation of P2X receptors in neck muscles.

Introduction

Tension-type headache (TTH) is the most frequent primary headache (1, 2). Psychophysical studies in patients have indicated an important role of neck muscle nociception in the pathophysiology of TTH (2, 3). Pericranial tenderness and hardness of head and neck muscles are significantly more marked in TTH patients than in healthy volunteers (3–7). Pressure pain detection and tolerance thresholds of head and neck regions are significantly decreased in TTH patients compared with healthy controls (8). Pericranial tenderness is positively associated with both the intensity and frequency of TTH (3, 4). Recently, the impact of static contraction of the shoulder and neck muscles on muscle tenderness and headache in TTH patients has been addressed (9). Patients are more liable to develop shoulder and neck pain in response to static exercise than are healthy controls.

Muscle contraction causes increase of ATP concentration and muscle tension (10, 11). Concentrations of ATP and its analogues in human vastus lateralis muscle increase in the exercising muscle interstitium, at a rate associated with intensity of muscle contraction and the magnitude of muscle blood flow (10). Electrical stimulation of ventral roots in decerebrate cats evokes an increase of interstitial ATP concentration in triceps surae muscles by 200% (11). ATP administration into skeletal muscles selectively excites group III and IV muscle afferents (12–14). Accordingly, ATP injection into trapezius muscle induces moderate to strong pain and produces local tenderness in humans (15). The excitatory effects of ATP on sensory neurons are known to be mediated by ionotropic P2X receptors (16–18). The P2X3 receptor, especially, seems to be significantly involved in nociception (16).

Nociceptive processing in the brainstem can be assessed by the jaw-opening reflex (JOR). This brainstem reflex can be elicited by electrical, thermal and mechanical stimulation of the craniofacial region (19–24). Primary trigeminal afferents synapse on excitatory sensory neurons of the spinal trigeminal complex. These neurons project ipsi- and contralaterally to excite digastric motoneurons (25, 26). The reflex seems to be a suitable model to investigate alterations of excitability in sensory brainstem neurons with convergent afferent input from different craniofacial tissues such as the neck muscles.

The present study addressed the effect of topical ATP administration into neck muscles on nociceptive brainstem processing in anaesthetized mice assessed by the JOR.

Methods

The electrophysiological experiments were performed in 44 adult male C57BL/6 mice (about 12 weeks old; 20–28 g; Charles River Laboratories, Sulzfeld, Germany). All procedures received institutional approval from the local ethics committee of the University of Aachen (ref. no. 50.203.2-AC 15, 16/03). The principles of laboratory animal care and use of laboratory animals [European Communities 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.

The detailed description of anaesthesia, surgery and electrophysiological recording in mice has recently been published (27). Mice were anaesthetized by an initial i.p. injection of a 0.5% pentobarbital sodium salt solution (Sigma-Aldrich, Taufkirchen, Germany) with a dose of 70 mg/kg. Depth of anaesthesia was checked by ensuring that noxious pinch stimulation (blunt forceps) of the hind paw, the forepaw and the ear did not evoke any sensorimotor reflexes. When the mouse was sufficiently deeply anaesthetized, the skin of the throat was carefully shaved and Lidocaine gel (AstraZeneca, Wedel, Germany) was applied to induce local anaesthesia. Dexpanthenol eye ointment (Roche, Grenzach-Wyhlen, Germany) was applied to the cornea of both eyes to protect it from drying. The right external jugular vein was catheterized for continuous administration of a 2% methohexital sodium salt solution (Lilly, Bad Homburg, Germany) 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) via a differential amplifier. After tracheotomy, animals were placed in a stereotaxic frame and were artificially respired with a stroke volume of about 150 μl and about 200 strokes per minute for the duration of the experiment (MiniVent Model 845; Harvard Apparatus, Holliston, MA, USA). The percentage expiratory carbon dioxide concentration was continuously monitored (Capstar-100; CWE Inc., Ardmore, PA, USA) and adjusted to a physiological level that prevented spontaneous breathing. The body core temperature was maintained at 37.5°C with a heating blanket and a fine rectal thermal probe (FMI, Seeheim-Ober Beerbach, Germany). One platinum needle electrode each (300 μm diameter) was subcutaneously inserted into the left forepaw and the right hind paw 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. The oral cavity was filled up with white Vaseline (Riemser, Greifswald-Insel Riems, Germany) to protect the oral mucous membrane from drying. Semispinal neck muscles on both sides were carefully exposed. A pair of Teflon-coated stainless steel wires (140 μm diameter) was inserted into the right semispinal neck muscle to record EMG activity from 5 min before to 5 min after intramuscular injections. One injection canula each (0.4 mm diameter) was inserted into the muscle belly of both semispinal neck muscles. Each canula was connected via thin tubing to a glass microsyringe (1 ml) that was fixed into a CMA/102 microdialysis pump (CMA Microdialysis, Solna, Sweden). This procedure allowed bilateral induction of noxious input from neck muscles in order to mimic bilateral neck muscle pain in TTH patients. Via these canulas intramuscular injections of 20 μl α,β-methylene ATP (ATP; 100 nmol/l and 1 μmol/l; Sigma-Aldrich) or pyridoxal phosphate-6-azophenyl-2′,4′-disulphonic acid (PPADS; 3, 10, 30, 100 nmol/l; Sigma-Aldrich) were performed with a flow rate of 20 μl/min. After surgery and placement of all electrodes, the anaesthetized animal was rested for at least 1 h. In this time period level of anaesthesia (reflexes) and heart rate were routinely checked and documented, and depth of anaesthesia was maintained.

All electrical signals (EMG, ECG) were recorded by bioamplifiers and led into a data collection system (CED micro 1401) and a personal computer using the Signal and Spike2 software programs (CED, Cambridge, UK).

The JOR was elicited by rectangular electrical pulses of 500 μs duration with a stimulation frequency of 0.1 Hz. The electrical threshold of the JOR was determined by applying 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 JOR threshold (I_JOR). The test stimulus intensity was adjusted to about 125% of the I_JOR. The JOR was evoked in blocks of eight stimuli each. These blocks were repeated every 5 min. After three stable baseline JOR blocks, ATP was administered into both semispinal neck muscles. JOR blocks were repeated every 5 min for up to 2 h after ATP injection (Fig. 1a). In order to investigate the role of P2X receptors in the induction of ATP effects, the P2X receptor antagonist PPADS (3, 10, 30 or 100 nmol/l) was intramuscularly applied 20 min before ATP injection. After ATP administration JOR blocks were repeated for 1 h (Fig. 1b). In order to characterize the role of P2X receptors in the maintenance of ATP effects, PPADS was applied 35 min after ATP injection in a second experimental series. JOR blocks were repeated for 1 h after intramuscular PPADS administration (Fig. 1c).

Figure 1

Stimulation protocol. The jaw-opening reflex was evoked in blocks (vertical black bars) of eight electrical stimuli each. The reflex blocks were repeated every 5 min. In three different experimental groups drugs were injected into both semispinal neck muscles (marked by arrows). (a) Sole injection of ATP. (b) Combined preceding PPADS and subsequent ATP administration. (c) Combined preceding ATP and subsequent PPADS application.

A former experimental study has shown that isotonic saline (0.9%, 20 μl) injection into semispinal neck muscles does not induce any alterations of the jaw-opening reflex in anaesthetized mice (28). Thus, the facilitatory effect of other substances such as hypertonic saline, nerve growth factor or ATP injection may not be due to a simple volume effect in the muscle (28, 29).

Onset latency, duration and integral of the JOR were analysed in each single sweep. Arithmetic mean and standard error were calculated (mean ± SEM). One-way repeated measures analysis of variance (ANOVA) or Friedman repeated measures ANOVA on ranks (Friedman ANOVA) were applied in order to analyse differences within groups. Multiple comparison procedures within groups were performed by the Holm–Sidak post hoc test, the Student–Newman–Keuls post hoc test or the Dunn’s post hoc test. Differences between groups were analysed by one-way ANOVA (normality) or Kruskal–Wallis one-way ANOVA on ranks (Kruskal–Wallis ANOVA). Multiple comparison procedures between groups were performed by the Holm–Sidak post hoc test or the Dunn’s post hoc test, respectively. Furthermore, JOR thresholds and mean integrals before and after intramuscular injection were analysed by paired t-test.

Results

The JOR was elicited in 44 experiments in mice by electrical tongue stimulation with a threshold stimulus intensity of I_JOR = 687 ± 26 μA (mean ± SEM). The test stimulus intensity was adjusted to 852 ± 32 μA corresponding to 124.0 ± 0.8% of the I_JOR (n = 44).

In nine mice 100 nmol/l ATP solution was injected into semispinal neck muscles (Fig. 2). Under baseline conditions I_JOR was 740 ± 45 μA, JOR latency and duration were 5.2 ± 0.3 ms and 4.0 ± 0.2 ms, respectively (mean ± SEM; n = 9). Within 1 h after ATP administration JOR significantly altered (Friedman ANOVA): reflex integral (χ² = 91.7, P < 0.001) and duration (χ² = 43.0, P < 0.001) increased by 79.4 ± 16.3% and 15.0 ± 5.0%, respectively, onset latency significantly decreased by 7.8 ± 1.2% (χ² = 100.6, P < 0.001). Within 2 h after intramuscular ATP injection, integral and duration significantly increased by 114.0 ± 20.9% (χ² = 143.7, P < 0.001) and 15.8 ± 5.4% (χ² = 44.1, P < 0.05), onset latency significantly decreased by 10.5 ± 1.5% (χ² = 158.7, P < 0.001). The Student–Newman–Keuls test explained these results by significant changes of the JOR between time points 10 and 120 min after ATP injection in comparison with the three baseline JOR blocks. The I_JOR significantly decreased to 510 ± 50 μA after injection corresponding to 58.3 ± 8.4% of baseline threshold (paired t-test: t = 5.12, P < 0.001).

Figure 2

Reflex facilitation after intramuscular 100 nmol/l ATP injection. (a) Grey and black graphs represent averages of eight rectified reflex sweeps 5 min before (light grey; pre, −5 min), 30 min after (dark grey; post, 30 min), and 120 min after (black; post, 120 min) ATP injection (100 nmol/l, 20 μl) into both semispinal neck muscles in one mouse. Electromyographic (EMG) activity is given in arbitrary units. (b) Reflex integral (○ ), duration (▴), and latency (▪) from 15 min before to 120 min after injection of ATP (100 nmol/l, 20 μl) into both semispinal neck muscles in nine mice. The changes from baseline (time points −15, −10, −5) are expressed as arithmetic mean and standard error. The start of injection is marked by an arrow. Asterisks represent the level of significance in comparison with the baseline blocks as calculated by the Friedman ANOVA (∗P < 0.05; ∗∗∗P < 0.001).

In five experiments 1 μmol/l ATP solution was injected into semispinal neck muscles (Fig. 3). Before injection I_JOR was 600 ± 63 μA, JOR latency and duration were 5.0 ± 0.2 ms and 3.8 ± 0.2 ms (mean ± SEM; n = 5). One hour after ATP administration, JOR integral (χ² = 59.3, P < 0.001) and duration (χ² = 45.3, P < 0.001) increased significantly by 187.7 ± 64.6% and 21.8 ± 5.5%, respectively, and latency decreased significantly (χ² = 63.6, P < 0.001) by 10.9 ± 2.0%. Within 2 h after ATP injection, integral (328.3 ± 90.8%; χ² = 114.8, P < 0.001), duration (33.1 ± 18.6%; χ² = 90.8, P < 0.001), and latency (−14.0 ± 2.4%; χ² = 114.3, P < 0.001) altered significantly. The I_JOR decreased significantly to 380 ± 58 μA after injection, corresponding to 62.4 ± 3.2% of baseline value (paired t-test: t = 2.557, P < 0.05).

Figure 3

Reflex facilitation after intramuscular 1 μmol/l ATP injection. (a) Grey and black graphs represent averages of eight rectified reflex sweeps 5 min before (light grey; pre, −5 min), 30 min after (dark grey; post, 30 min), and 120 min after (black; post, 120 min) ATP injection (1 μmol/l, 20 μl) into both semispinal neck muscles in one mouse. Electromyographic (EMG) activity is given in arbitrary units. (b) Reflex integral (○ ), duration (▴), and latency (▪) from 15 min before to 120 min after injection of ATP (1 μmol/l, 20 μl) into both semispinal neck muscles in five mice. The changes from baseline (time points −15, −10, −5) are expressed as arithmetic mean and standard error. The start of injection is marked by an arrow. Asterisks represent the level of significance in comparison with the baseline blocks as calculated by the Friedman ANOVA (∗∗∗P < 0.001).

Within 2 h after intramuscular injection, Kruskal–Wallis ANOVA showed significant differences between 100 nmol/l and 1 μmol/l ATP groups. Effects on JOR latency (H = 9.0, P < 0.01) and JOR integral (H = 5.5, P < 0.05) were significantly stronger with a higher dosage of ATP. Baseline (preinjection) JOR parameters (threshold, stimulus intensity, JOR latency and duration) were not statistically different between both groups. Intramuscular injection of ATP did not evoke EMG activity in semispinal neck muscles (Fig. 4).

Figure 4

Neck muscle activity during ATP injection. Electromyographic (EMG) recording of right semispinal neck muscle and right digastric muscle activity, and heart rate recording (1/min) in one specimen before, during and after intramuscular injection of ATP (100 nmol/l, 20 μl) into both semispinal neck muscles. EMG recordings were performed by intramuscular wire electrodes. EMG activity is given in arbitrary units.

In order to investigate the role of P2X receptors in the induction of ATP-evoked JOR facilitation, the P2X receptor antagonist PPADS was applied intramuscularly 20 min before ATP administration. After PPADS (3, 10, 30, 100 nmol/l) application into neck muscles, JOR integral and duration did not change in comparison with baseline reflex blocks for all concentrations according to ANOVA. JOR latency decreased significantly after 3 nmol/l PPADS (ANOVA: F = 7.209, P < 0.001) and 100 nmol/l PPADS (Friedman ANOVA: χ² = 12.4, P < 0.05).

Preceding administration of PPADS solution altered the effect of ATP on the reflex (Fig. 5). With 100 nmol/l PPADS and 100 nmol/l ATP (n = 10), integral (χ² = 27.7, P < 0.05) and latency (χ² = 29.9, P < 0.01) decreased significantly by 15.0 ± 9.3% and 3.7 ± 2.1%, respectively (Friedman ANOVA), and duration remained unchanged (ANOVA). After 30 nmol/l PPADS administration in five mice, ATP did not induce changes in integral (Friedman ANOVA) and duration (ANOVA), but induced reduction of latency by 4.0 ± 1.3% (ANOVA: F = 2.308, P < 0.05). With preceding injection of 10 nmol/l PPADS (n = 5), JOR integral increased significantly by 37.2 ± 18.4% (Friedman ANOVA: χ² = 33.4, P < 0.01), whereas latency and duration remained unaffected after ATP (ANOVA). With the lowest dosage of PPADS (3 nmol/l; n = 5) with subsequent ATP injection, JOR integral (71.5 ± 12.9%; Friedman ANOVA: χ² = 33.0, P < 0.01) and duration (26.9 ± 16.5%; χ² = 25.0, P < 0.05) increased significantly, and onset latency decreased by 8.1 ± 2.0% to 4.8 ± 0.3 ms (ANOVA: F = 4.9, P < 0.001).

Figure 5

Dose-dependent suppression of reflex facilitation by preceding PPADS. (a) Modulation of reflex integral after sole intramuscular administration of 100 nmol/l ATP (no PPADS, s, n = 9) and after combined preceding PPADS (3 (j, n = 5), 10 (▴, n = 5), 30 (h, n = 5), 100 nmol/l (•, n = 10)) and subsequent ATP (100 nmol/l) injection into both semispinal neck muscles. Data are expressed as percentage changes from baseline (time points −15, −10, −5 min). The changes from baseline are expressed as arithmetic mean and standard error. Asterisks represent the level of significance in comparison with the baseline blocks as calculated by Friedman ANOVA (∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001). (b) Top: grey and black graphs represent specimen averages of eight rectified reflex sweeps 5 min before (grey; pre, −5 min) and 60 min after (black; post, 60 min) ATP injection (100 nmol/l, 20 μl) into both semispinal neck muscles in one mouse each. Electromyographic (EMG) activity is given in arbitrary units. Bottom: corresponding averages (mean + SEM) of reflex integrals between time points 5 and 60 min after intramuscular ATP administration without (no PPADS; n = 9) or with preceding PPADS injection (3 nmol/l n = 5; 10 nmol/l n = 5; 30 nmol/l n = 5; 100 nmol/l n = 10). Average integral under 100 nmol/l PPADS significantly differed from sole ATP application (Dunn’s post hoc test, ∗P < 0.05).

Effects of ATP on reflex integral (Kruskal–Wallis ANOVA: H = 22.2, P < 0.001) and duration (H = 20.268, P < 0.001) differed significantly depending on preceding administration of various PPADS doses (3, 10, 30, 100 nmol/l) and sole application of ATP (Fig. 5). The Dunn’s post hoc test explained these differences by significantly stronger changes in JOR integrals with sole application of 100 nmol/l ATP in comparison with the group with preceding administration of 100 nmol/l PPADS (Q = 4.1, P < 0.05). Reflex integrals also differed significantly between 3 nmol/l PPADS and 100 nmol/l PPADS groups (Q = 3.4, P < 0.05). Regarding reflex duration, the Kruskal–Wallis ANOVA revealed significant differences between sole ATP application and combined injection of PPADS and ATP (H = 20.3, P < 0.001). Reflex effects after combined administration of 100 nmol/l PPADS and 100 nmol/l ATP differed significantly from sole ATP injection (Q = 3.3, P < 0.05) and from combined 10 nmol/l PPADS and ATP injection (Q = 4.0, P < 0.05). Reflex onset latencies did not differ between the various groups (ANOVA).

In order to characterize the role of P2X receptors in the maintenance of ATP-evoked reflex facilitation, PPADS (100 nmol/l; n = 5) was applied 35 min after 100 nmol/l ATP, at a time point when JOR was clearly facilitated (see Figures 2 and 6). After intramuscular administration of ATP, reflex integral and duration increased significantly by 55.9 ± 11.6% (Friedman ANOVA: χ² = 23.8, P < 0.01) and 12.9 ± 1.8% (ANOVA: F = 5.1, P < 0.001), respectively, and latency decreased significantly by 3.3 ± 1.6% (ANOVA: F = 4.0, P < 0.01) (mean ± SEM). This facilitatory effect under ATP did not differ statistically from the results in the above experiments (ANOVA; see Figure 2). Within 1 h after subsequent intramuscular administration of 100 nmol/l PPADS, facilitated JOR totally recovered to baseline values (before ATP injection; Fig. 6). Regarding three (facilitated) reflex blocks before PPADS injection (time points 20, 25, 30 min), JOR integral decreased significantly within 1 h after PPADS injection (ANOVA: F = 4.7, P < 0.0001) and JOR integral increased significantly under sole ATP administration (Friedman ANOVA: χ² = 61.7, P < 0.000001). From time points 85–95 min reflex integrals accounted for 155.2 ± 30.0% (mean ± SEM) after sole ATP injection and 11.3 ± 14.3% after combined ATP and PPADS administration. These integrals were significantly different (t-test: t = 3.4, P < 0.01). The I_JOR of 620 ± 49 μA under baseline conditions did not differ statistically from the I_JOR of 580 ± 37 μA after ATP and subsequent PPADS administration (n = 5).

Figure 6

Neutralization of reflex facilitation by subsequent PPADS injection. (a) Grey and black graphs represent averages of eight rectified reflex sweeps 5 min before (light grey; pre), 30 min after 100 nmol/l ATP injection (black; post ATP), and 60 min after 100 nmol/l PPADS injection (dark grey; post PPADS) into both semispinal neck muscles in one mouse. Electromyographic (EMG) activity is given in arbitrary units. (b) Reflex integral (○ ), duration (▴) and latency (▪) under ATP (100 nmol/l, 20 μl) and subsequent PPADS (100 nmol/l, 20 μl) injection into both semispinal neck muscles in five mice. The changes from baseline (time points −15, −10, −5 min) are expressed as arithmetic mean and standard error. (c) Comparison of reflex modulation under sole ATP administration (100 nmol/l; s, n = 9) and combined administration of preceding ATP (100 nmol/l) and subsequent PPADS (100 nmol/l; •, n = 5). The changes from baseline (time points −15, −10, −5) are expressed as arithmetic mean and standard error. Based on the blocks at time points 20–30 min, reflex significantly increased under ATP condition (Friedman ANOVA, ∗∗∗P < 0.001) and significantly decreased under combined ATP and PPADS administration (ANOVA; ∗∗∗P < 0.001).

Discussion

Injection of ATP into semispinal neck muscles induced sustained facilitation of the JOR (latency, duration, integral, threshold) for at least 2 h in a dose-dependent manner in anaesthetized mice. The P2X receptor antagonist PPADS suppressed the induction of reflex facilitation in a dose-dependent manner and neutralized an established facilitation as shown by subsequent administration.

The pharmacological properties of α,β-meATP and PPADS and the dose-dependent effect on neck muscles indicate a P2X receptor-mediated mechanism in the present study. The dose-dependent blockade of reflex facilitation by preceding PPADS is compatible with an α,β-meATP-induced receptor activation. The neutralization of established reflex facilitation by subsequent PPADS administration points to sustained activation of P2X receptors in order to maintain reflex effects. α,β-meATP selectively excites homomeric P2X1, P2X3 and heteromeric P2X2/3 receptors (30). Whereas P2X1 receptors are not localized in skeletal muscles, P2X3 receptors are characterized by fast desensitization (18). Thus, there is some evidence that the described effect on brainstem nociception may be mediated by heteromeric P2X2/3 receptors that desensitizes very slowly (16, 31).

Recent studies suggest a strong involvement of neck muscle nociception in the pathophysiology of TTH corresponding to the previously used term ‘muscle contraction headache’ (2, 3). TTH patients suffer not only from headache attacks but also from pericranial tenderness and hardness of head and neck muscles (3–7). Static exercise induces shoulder and neck pain in TTH patients more frequently than in healthy controls (9). Thus (tonic) muscle activation may be connected to the development of TTH. It is well known that contraction of extremity muscles causes a significant increase of interstitial ATP concentration (10, 11). Raised concentrations of ATP and its metabolites may initiate different purinergic signal pathways within the neuromuscular system. One of these pathways may be a long-term excitation of P2X receptors inducing central sensitization of brainstem nociceptive processing. This neurobiological phenomenon possibly prepares the ground for pericranial tenderness and the development of episodic and chronic TTH.

According to a recent study applying the same animal experimental model, intramuscular administration of an identical volume of isotonic saline changed neither the reflex integral nor the reflex threshold (28). Thus, the facilitation was not induced by a simple mechanical volume effect on muscle receptors. Injections of hypertonic saline (5%) or recombinant human β-nerve growth factor in mice neck muscles induced similar effects on brainstem nociceptive processing (28, 29). Different algogenic substances such as capsaicin, bradykinin, nerve growth factor and ATP can induce muscle nociceptor activation by local intramuscular administration (12, 13, 32). ATP administration into human trapezius muscle produced a moderate and prolonged pain sensation and tenderness. ATP in combination with other active substances (bradykinin, serotonin, histamine and prostaglandin E₂) produced unacceptable side-effects in relation to pain within a few minutes after start of infusion into trapezius muscle (15, 33). In decerebrate cats, α,β-meATP was injected into the popliteal artery in order to test its effects on triceps surae muscle afferents (14). With single fibre recordings, excitation of groups III and IV muscle afferents was documented. PPADS blocked this activation in both fibre classes. In anaesthetized rats, Reinöhl and colleagues investigated the responsiveness of 46 group IV receptors of the gastrocnemius-soleus muscle in the sciatic nerve to intramuscular ATP administration (12). In about 67% of high- and low-threshold mechanosensitive units, action potentials appeared with a delay of up to 28 s after completion of the injection and lasted between 11 and 555 s. In a similar study of the same group, response duration to ATP application ranged between 11 and 781 s (13). Thus, ATP seems to be an appropriate chemical stimulator in order to induce nociceptive input from skeletal muscles. Exogenous ATP injection can also activate C and Aδ afferent fibres in the skin. In healthy volunteers, an intracutaneous ATP injection at high concentration activated C nociceptors of the peroneal nerve, without preference for mechano-responsive or mechano-insensitive units (34). Using a rat in vitro skin-nerve preparation, ATP selectively activated the peripheral terminals of nociceptive sensory neurons in the skin (35). The present study provides some indirect evidence of the afferent fibre classes excited by ATP administration. According to continuous electromyographic recordings of semispinal and digastric muscles, no muscle activity was evoked during ATP injection (see Fig. 4). An appreciable excitation of Aα efferents and/or group Ia and/or group II afferents would have induced muscle activity at least in the semispinal muscle via a direct or a monosynaptic spinal pathway. Thus, an appreciable activation of thick myelinated fibres can be excluded.

Recent studies have demonstrated that ATP is not only the most important chemical energy source of the organism but also a significant intercellular messenger molecule that interacts with membrane receptors (purinoceptors) on a broad range of cell types (18, 31, 36). ATP receptors are classified into G-protein coupled receptors, so called P2Y, and ATP-gated cation channels, so called P2X receptors (18, 30, 37, 38). P2X receptors, in particular, play an important role in excitatory nociceptive processing (16, 18, 31, 39). ATP binds to P2X receptors which are localized in the central and peripheral nervous system. The P2X receptors are found in rat in the dorsal horn of the spinal cord and in substantia gelatinosa where nociceptive neurons synapse (40). P2X receptors expressing sensory fibres innervate skin, muscles and viscera (41). ATP is physiologically released from neuronal and non-neuronal cells in the periphery and central nervous system and activates target cells via purinergic receptors (42). ATP-induced postsynaptic depolarization may open voltage-operated Ca²⁺ channels. Postsynaptic Ca²⁺ influx itself can evoke Ca²⁺-dependent release of ATP into extracellular space, which can induce depolarization and Ca²⁺ influx in the sense of a positive feedback loop (30, 39). ATP released into the spinal cord is thought to play a role in central sensitization.

P2X receptors can be activated by ATP and various ATP analogues (18, 31, 36). ATP is the most potent activator of P2X receptors, but its rapid degradation to many analogues aggravates pharmacological characterization of special receptor mechanisms. Degradation studies have shown dephosphorylation of ATP, ADP, and AMP with half-lives of about 15–20 min (43, 44). The same studies proved that α,β-meATP is very slowly degraded to α,β-meADP with about 80% remaining after 1 h. Thus, unlike ATP, α,β-meATP is resistant to enzymatic degradation and can be used to identify P2X receptors in multicellular preparations including whole animals (45, 46). The stable P2X receptor agonist α,β-meATP shows high sensitivity to homomeric P2X1, P2X3 and heteromeric P2X2/3 receptors (30). Chemical stability and a narrow receptor profile were the main arguments for selecting the P2X receptor agonist α,βme-ATP in our experiments. Animal experimental studies have documented the algogenic character of α,β-meATP (47–51).

For pharmacological characterization of P2X-mediated ATP effects in the neck muscles, we decided to apply PPADS. Two principal alternative P2X receptor antagonists are available, suramin and PPADS. Recently, it was indicated that suramin not only inhibits P2X receptors but also inhibits binding to N-methyl-D-aspartate (NMDA) receptors in the brain at concentrations comparable to those required to inhibit responses in P2X receptors. Since both NMDA and P2X receptors form Ca²⁺-selective channels, suramin may block Ca²⁺ entry into neurons depolarized by either ATP, glutamate or any other neurotransmitter or drug acting via P2X or NMDA receptors (52, 53). By contrast, PPADS at concentrations up to 100 μmol/l does not affect NMDA-activated current (54). Suramin, but not PPADS, inhibits GABA-activated ion current in rat hippocampal neurons (55, 56). Consequently, we selected PPADS as a potent selective antagonist for P2X receptors.

The JOR is a markedly convergent brainstem reflex that can be elicited by electrical, thermal and mechanical stimulation of the orofacial region such as the lips, tongue, tooth pulp, orofacial skin, cornea and meninges (19–24, 28, 29). Electrical tongue stimulation was performed at an intensity of 852 ± 32 μA, corresponding to 124.0 ± 0.8% of the reflex threshold. The applied electrical test stimuli caused a pricking painful sensation in man, indicating an excitation of thin-myelinated, nociceptive group III muscle afferents corresponding to Aδ-fibres (57, 58). In muscle nerves of the cat, group IV afferents were assumed to be activated by electrical stimuli with intensities of at least 10 times the nerve threshold (59, 60). Electrical stimulation of dorsal roots in spinal cord slices evoked C-fibre potentials only with intensities corresponding to about nine-fold the Aδ fibre threshold (61). In microelectrode studies in human nerves, C-fibre activity could be evoked only by electrical stimuli with intensities of 15–20 times the detection threshold (62). Thus, the applied stimulus intensity in the present study probably excites group III but not group IV muscle afferents. The JOR seems to be an adequate model to investigate long-term modulatory effects on craniofacial nociceptive processing in living animals.