Delta-9-tetrahydrocannabinol decreases masticatory muscle sensitization in female rats through peripheral cannabinoid receptor activation


  • Funding sources

    H. Wong is the recipient of a Mitacs Elevate Postdoctoral Fellowship. This project was jointly funded by Mitacs Canada and InMed Pharmaceuticals Inc.

  • Conflicts of interest

    S. Hossain is the chief scientific officer of InMed Pharmaceuticals Inc.



This study investigated whether intramuscular injection of delta-9-tetrahydrocannabinol (THC), by acting on peripheral cannabinoid (CB) receptors, could decrease nerve growth factor (NGF)-induced sensitization in female rat masseter muscle; a model which mimics the symptoms of myofascial temporomandibular disorders.


Immunohistochemistry was used to explore the peripheral expression of cannabinoid receptors in the masseter muscle while behavioural and electrophysiology experiments were employed to assess the functional effects of intramuscular injection of THC.


It was found that CB1 and CB2 receptors are expressed by trigeminal ganglion neurons that innervate the masseter muscle and also on their peripheral endings. Their expression was greater in TRPV1-positive ganglion neurons. Three days after intramuscular injection of NGF, ganglion neuron expression of CB1 and CB2, but not TPRV1, was decreased. In behavioural experiments, intramuscular injection (10 μL) of THC (1 mg/mL) attenuated NGF-induced mechanical sensitization. No change in mechanical threshold was observed in the contralateral masseter muscles and no impairment of motor function was found after intramuscular injections of THC. In anaesthetized rats, the same concentration of THC increased the mechanical thresholds of masseter muscle mechanoreceptors. Co-administration of the CB1 antagonist AM251 blocked the effect of THC on masseter muscle mechanoreceptors while the CB2 antagonist AM630 had no effect.


These results suggest that reduced inhibitory input from the peripheral cannabinoid system may contribute to NGF-induced local myofascial sensitization of mechanoreceptors. Peripheral application of THC may counter this effect by activating the CB1 receptors on masseter muscle mechanoreceptors to provide analgesic relief without central side effects.


Our results suggest THC could reduce masticatory muscle pain through activating peripheral CB1 receptors. Peripheral application of cannabinoids could be a novel approach to provide analgesic relief without central side effects.

1 Introduction

Recent advances in cannabinoid pharmacology have renewed hope in cannabis-based treatments for chronic pain. However, whether the peripheral endocannabinoid system can be recruited to treat specific craniofacial pain conditions remains unclear. There is a long history of the medicinal use of cannabis against pain but its use generates controversy due to the psychoactive effects of some of its components (Maione et al., 2013; Robson, 2014). The discovery of the endocannabinoid system generated new hope for cannabis-based analgesics (Maione et al., 2013; Robson, 2014). For example, inhibitors of fatty acid amide hydrolase, the primary deactivating enzyme for the endocannabinoid anandamide, have been investigated as potential therapeutic agents for chronic pain, however, they have yet to be translated to clinical application (Krustev et al., 2014; Lodola et al., 2015). Recent studies suggest that peripheral application of cannabinoids could be an effective strategy for analgesia without central side effects (Johanek et al., 2001; Yu et al., 2010; Bagues et al., 2014; Romero-Sandoval et al., 2015).

The cause of many chronic myofascial pain disorders such as myofascial temporomandibular disorders (TMD) remains unknown (Cairns, 2010). Some evidence supports the concept that it is centrally mediated, while other evidence suggests that peripheral mechanisms may also be involved; one of which is the overproduction of nerve growth factor (NGF) that leads to muscle sensitization in both humans and rats (Wong et al., 2014b). Previously, we have developed an animal model of NGF-induced muscle pain that mimics the tender points observed in the craniofacial regions of TMD patients (Wong et al., 2014a,b). Many chronic myofascial pain disorders show a marked gender-related difference with greater prevalence in women than men (Cairns, 2007). It has been reported that 2–3 times more women than men suffer from TMD-related pain (Bush et al., 1993; Cairns, 2010; Shaefer et al., 2013). Despite this gender-related difference, basic and translational pain research aimed at understanding muscle pain mechanisms using female animal has been limited (Mogil and Bailey, 2010). Intramuscular injection of NGF induced a greater and longer lasting sensitization in the female rats than the male rats, suggesting that this could be an appropriate model for chronic craniofacial muscle pain disorders (Wong et al., 2014b). We used female rats in this project with the aim to better represent the population of muscle pain patients.

In this study, we investigated whether peripheral application of tetrahydrocannabinol (THC) could relieve NGF-induced masseter muscle sensitization without central side effects and whether peripheral cannabinoid receptors play a role in the effect of THC. The results of this study indicate that peripheral cannabinoid receptors may be potential targets for treatment of muscle pain and also offer new insights into mechanism of NGF-induced muscle sensitization.

2 Methods

2.1 Animals

Female (276–322 g, n = 35) Sprague–Dawley rats were used for all experiments. Animals were housed in groups of two with a 12-h light/dark cycle. Food and water were given ad libitum. All animal procedures were reviewed and approved by the University of British Columbia Animal Care Committee.

2.2 Drugs

THC, 1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-1-piperidinyl-1H-pyrazole-3-carboxamide (AM251) and [6-iodo-2-methyl-1-[2-(4-morpholinyl)ethyl]-1H-indol-3-yl](4-methoxyphenyl)-methanone (AM630) were purchased from Cayman Chemicals (Ann Arbor, MI, USA). THC, AM251 and AM630 were dissolved in acetonitrile, all to a concentration of 10 mg/mL in a stock solution. Appropriate volumes of each solution were evaporated under nitrogen gas, and subsequently resolubilized with 4% Tween 80 in isotonic saline to final concentrations used for injection.

2.3 Immunohistochemistry

To identify trigeminal ganglion neurons that innervate the masseter muscle, the fluorescent tracer dye Fast Blue (2%; Polysciences, Warrington, PA, USA) was injected bilaterally into the masseter muscles of rats (n = 5),7 days prior to euthanization. NGF (25 μg/mL, 10 μL; Sigma, St. Louis, MO, USA) or vehicle (phosphate buffered saline, PBS, 10 μL) was injected into the left and right masseter muscles, respectively, 3 days before euthanization. Under deep isoflurane anaesthesia, animals were perfused with 120 mL cold saline followed by 120 mL of paraformaldehyde (4%). Trigeminal ganglia and masseter muscles were removed, incubated in 20% and then 40% sucrose for 48 h each to cryoprotect the tissue, and cut into 10 μm sections with a cryotome. Trigeminal ganglia sections were treated with 5% normal goat serum in PBS for 1 h and incubated overnight with commercially available primary antibodies against cannabinoid receptor type I (CB1; Santa Cruz Biotechnology, Dallas, TX, USA) and cannabinoid receptor type II (CB2; ABCAM, Cambridge, MA, USA), as well as transient receptor potential vanilloid 1 (TRPV; ABCAM). Satellite glial cells (SGCs) were identified through the use of a primary antibody against glutamine synthetase (GS; ABCAM). Masseter muscle sections were treated with primary antibodies against the pan-axonal marker PGP9.5 (ABCAM), CB1 and CB2 receptors as described above. The next morning sections were washed several times with PBS and then incubated for 1 h at room temperature in the dark in the presence of appropriate secondary antibodies with fluorescent tracers attached. After several washes in buffer, all sections were mounted on slides with covers slips and visualized with a Leica, Concord, Ontario, Canada TCS SPE confocal microscope. Fast Blue positive cells were counted and photographed for estimation of cell diameter. Images from the confocal microscope were analysed using the image processing program ImageJ (National Institute of Health, Bethesda, MD, USA). An area with nonspecific binding from the image was selected as the background. The mean and standard deviation for this area were calculated from the brightness values of the pixels by the ImageJ program. Neurons were considered positively labelled when the average intensity of the fast blue signal within the neuron was greater than 2 standard deviations above the mean background intensity. Nerve fibres were considered positive when the intensity of the fluorescent tracer signal for PGP 9.5 exceeded the 2 standard deviations of the mean background intensity. The minimum accepted length for a PGP 9.5-labeled nerve fibre was 1 μm in the masseter muscle. The specificity of the antibodies was confirmed by omission of the primary antibodies. Pre-absorption of the primary antibody with the appropriate antigen was performed for the CB1 antibody in a previous study (Sanford et al., 2008), while the specificity for the CB2 antibody was previously confirmed in knockout rats (Burston et al., 2013).

2.4 Behavioral experiments

2.4.1 Administration of NGF and THC

Three days before the behavioural experiments, rats received an injection of NGF (25 μg/mL, 10 μL; Sigma) or vehicle (PBS, 10 μL) into the left and right masseter muscles, respectively, in a randomized order under brief isoflurane anaesthesia. NGF has been shown to induce local mechanical sensitization at the site of injection which lasts from 3 h to 5 days after injection in the female rats (Wong et al., 2014b). To determine the effects of peripheral application of THC on NGF-induced sensitization, THC (1 mg/mL) or vehicle was injected into the left (NGF-injected) masseter muscle of rats under brief isoflurane anaesthesia. This concentration of THC was the lowest concentration required to reduce hypertonic saline- induced muscle pain by local injection in a previous study (Bagues et al., 2014). The masseter muscle region was shaved prior to injection and the injection sites were marked with a permanent marker for subsequent identification. The concentration of NGF was selected based on the concentration used in previous human experimental pain studies and experiments in rats (Svensson et al., 2003, 2008, 2010; Mann et al., 2006; Wong et al., 2014b). The investigator was blinded to the identity of the treatment groups until after all data were collected.

2.4.2 Mechanical threshold (MT)

MT was assessed with a rigid electronic von Frey hair (IITC Life Science, Woodland Hills, CA, USA). Before NGF injection, rats were habituated to restraint in a towel. The electronic von Frey hair was applied perpendicularly to the masseter muscle and the force was gradually increased until the animal moved its head away from the stimulus. The mechanical test stimulus was applied at 1 min intervals for 5 min and the average MT was calculated for further analysis. MT was measured daily for 5 days prior to the start of the experiment to determine that measurements were stable. MT readings recorded prior to NGF injection were used as naïve baseline. After induction of the NGF-induced sensitization, behavioural testing was performed to evaluate the antinociceptive effects of THC and the vehicle control groups (n = 6/group) at 3 days after NGF injection. On the test day, a post-NGF injection baseline (NGF baseline) was recorded before injection of the treatment groups. MT was again measured at 10, 30, 60 and 120 min after treatment injection.

2.4.3 Inverted screen test

A modified version of the inverted screen test was performed to evaluate impaired motor function in rats (Coughenour et al., 1977; Maxwell et al., 1993). Rats (n = 4/group) were placed on a screen (185 mm × 290 mm) with 5-mm-diameter holes and the screen was slowly inverted 180 degrees until the rats were suspended upside down on the bottom of the screen. The animals were observed for their ability to climb to the top of the screen in the next 60 s and were assigned to a score: (0) animals successfully climbed to the top the screen; (1) held on the screen upside down (2) or fell from the screen. The rats were trained how to perform the test daily for 3 days before they were tested. Baselines were measured at the beginning prior to treatment and the animals were retested at 10, 30, 60 and 120 min after treatment (Vehicle or THC, 1 mg/mL) by intramuscular injection.

2.5 In vivo electrophysiology

In vivo electrophysiology recordings of single ganglion neurons that innervate the craniofacial muscles were performed to investigate the mechanism of action of peripherally injected cannabinoids. Rats (n = 6/group) from the behavioural experiments were used for these subsequent in vivo electrophysiological experiments. Additional animals were, however, used for the THC/AM251 and THC/AM630 groups. Recordings were performed at least 7 days after the behavioural experiments. Experiments were conducted on the PBS-injected masseter muscles (right side) of the rats to minimize potential residual effects of the previous cannabinoid treatments.

Rats were anaesthetized with isoflurane (2–2.5% in oxygen 97–98%; AErrane; Baxter) and surgically prepared for electrophysiological recording. Heart rate, blood pressure and body temperature were monitored and a trachea tube was inserted for ventilation throughout the experiment. The hair of the face was shaved and the animal's head was positioned in a stereotaxic frame. A parylene-coated tungsten microelectrode (0.10″, 2 MΩ; A-M Systems Inc., Sequim, WA, USA) was lowered into the trigeminal ganglion through the brain via a small trephination in the skull. An incision was made over the neck to expose the brain stem, and the dura was removed to allow access for a stimulating electrode to contact the caudal brain stem. Mechanoreceptors innervating craniofacial muscles (masseter and temporalis) were identified by mechanical probing using a fine-tipped cotton swab. Antidromic collisions were performed to confirm projection of the muscle fibre to the caudal brain stem (Cairns et al., 2002; Wong et al., 2014a). A stimulating electrode (parylene-coated tungsten microelectrode, 0.10″, 2 MΩ; A-M Systems Inc.) was lowered into the ipsilateral caudal brain stem and a constant-current electrical stimulus (100 μs biphasic pulse, 10–90 μA, 0.5 Hz) was applied to evoke antidromic action potentials. Orthodromic action potentials were evoked by mechanical stimulation of the muscle. Collision was demonstrated by disappearance of the antidromic spike. The straight line distance between the stimulating and recording electrodes was divided by the latency of the antidromic action potential to estimate conduction velocity. MT was assessed with an electronic von Frey hair (IITC Life Science) by applying mechanical stimuli at 1 min intervals for 10 min to obtain baseline threshold. After baseline measurement, treatment was administered and MT was reassessed at 10, 30, 60 and 120 min thereafter. At the end of the experiments, animals were euthanized with pentobarbital (Nembutal 100 mg/kg; Abbott Laboratories, Chicago, IL, USA).

2.6 Data analysis

For the immunohistochemistry experiment, the frequency of expression of CB1, CB2 and TRPV1, respectively, in fast blue positive trigeminal ganglion neurons between NGF and vehicle-injected sides was analysed with a paired Student's t-test. For the behavioural and electrophysiological experiments, MT from NGF and vehicle-injected sides was analysed with a two-way repeated measures analysis of variance (ANOVA) with time and treatment as factors. Relative mechanical activation threshold (Rel MT) was calculated as 100 × Post-treatment MT/Baseline MT. Post hoc Holm Sidak's multiple comparison tests were used to compare postinjection mechanical thresholds between treatment groups at each time point. One-way repeated measures ANOVA on ranks was used to analyse test scores in the inverted screen tests. A probability level of 0.05 was considered significant for all tests. Error bars indicate standard error of the mean on all graphs.

3 Results

3.1 Expression of cannabinoid receptors on masseter ganglion neurons

Expression of both cannabinoid receptors was observed in the trigeminal ganglion neurons that innervate the masseter muscle 3 days after injection of saline and NGF (Fig. 1A–E). Expression of CB1 and CB2 was greater in TRPV1-positive neurons (84.3 ± 2.7% and 71.1 ± 7.3%) than TRPV1-negative neurons (CB1: 28.8 ± 2.2%, CB2: 28.9 ± 1.4%) (Fig. 1F). Injection of NGF significantly reduced the expression of both receptors (Fig. 1G). On the saline-injected side, the frequency of expression of CB1 and CB2 was 52 ± 3% and 47 ± 5% in the masseter ganglion neurons, respectively. The frequency of expression of TRPV1 was 41 ± 5%. On the NGF-injected side, the frequency of expression of CB1 and CB2 receptors was significantly reduced to 39 ± 3% and 33 ± 6%, respectively. No significant effect on TPRV1 expression was observed. Expression of CB1 and CB2 was also found in the neuronal fibres in the masseter muscles, confirming their presence in the nerve endings of trigeminal ganglion neurons (Fig. 2). In the SGCs surrounding trigeminal ganglion neurons, expression of CB1, but not CB2 was observed (Fig. 3).

Figure 1.

The photomicrographs show co-expression of CB1, CB2 and TRPV1 by masseter ganglion neurons in a female rat. Masseter ganglion neurons were identified by intramuscular injection of fast blue (A; arrows). Immunopositivity for the CB1, CB2 and TRPV1 receptors is shown in (B), (C) and (D), respectively. The composite image is shown in (E). (F) The bar graph shows the frequency of expression of CB1 and CB2 in TRPV1-positive and -negative trigeminal ganglion neurons innervating the PBS-injected masseter muscles (n = 5 rats). (G) The bar graph shows the frequency of expression of CB1, CB2 and TRPV1 in trigeminal ganglion neurons innervating the masseter muscle at 3 days after injection (n = 5 rats). Asterisks denote a significant difference by paired Student's t-tests (p < 0.05). PBS, phosphate buffered saline.

Figure 2.

The photomicrographs show co-expression of CB1 and CB2 receptors (B and C) by nerve fibres that innervate the masseter muscle in a female rat. Nerve fibres were identified immunopositivity to PGP 9.5 (A). The composite image is shown in (D). Arrows indicate labeled fibers.

Figure 3.

The photomicrographs show expression of the CB1 and glutamine synthetase, a marker for satellite glial cells, SGC) by trigeminal ganglion neurons in a female rat. Expression of CB2 was not found in the SGCs. Masseter ganglion neurons were identified by intramuscular injection of fast blue (A). Immunopositivity for CB1, CB2 and glutamate synthetase (GS) is shown in (B), (C) and (D), respectively. The composite image is shown in (E). Arrows: A. Masseter ganglion neurons, B. SGCs, D. SGCs.

3.2 Effect of intramuscular injections of THC on NGF-induced sensitization

In the behavioural studies, intramuscular injection of NGF decreased masseter muscle MT at 3 days after injection by 31%, consistent with earlier studies (Fig. 4A; Wong et al., 2014b). THC (1 mg/mL) significantly reversed NGF-induced sensitization at 10 and 30 min after injection (Fig. 4A). No effect of any injection was observed on the MT of masseter muscles contralateral to the THC injection (Fig. 4B). No impairment of motor functions was found in the inverted screen test after intramuscular injections of THC (median score for all treatments was 0).

Figure 4.

(A) The line and scatter plots show the mean (±SE) mechanical withdrawal threshold of six female rats per treatment group following intramuscular injections of tetrahydrocannabinol (THC) and vehicle in behavioural experiments. Injection of THC significantly increased the mechanical threshold (MT) in the left (nerve growth factor, NGF-injected) masseter muscle for 30 min postinjection. The *s indicate significant differences compared with the vehicle group and #s indicate a significant difference compared to the NGF baseline within the treatment group (Holm Sidak multiple comparison test, p < 0.05). (B) The line and scatter plot shows the relative mean MT of the right masseter muscles after intramuscular THC injection into the left masseter muscle. No significant differences between treatment groups were found by two-way repeated measures ANOVA.

3.3 Effect of intramuscular injections of THC on masseter muscle nociceptors

Recordings from 24 masticatory muscle mechanoreceptors were undertaken (Fig. 5). The median conduction velocity was 7.6 m/s (25th percentile: 4.5; 75th percentile: 10.3) and the median mechanical threshold was 34.3 g (25th percentile: 14.5; 75th percentile: 57.8). The population was composed of 75% Aδ fibres (2–12 m/s), 17% Aβ fibres (>12 m/s) and 8% C fibres (<2 m/s). Intramuscular injection of THC (1 mg/mL) significantly increased the MT of mechanoreceptors at 10, 30 and 60 min after injection (Fig. 6). These results from the electrophysiology experiments parallel the results obtained from the behavioural experiments. The effects of CB1 and CB2 selective antagonists on this effect of THC were examined. Co-administration of THC and the CB1 antagonist AM251 attenuated the increase in MT mediated by THC alone and the combination was not significantly different from vehicle administration (Fig. 6). Co-administration of THC and the CB2 antagonist AM630 was not different from THC alone. These results suggest intramuscular injection of THC increased MT of masseter muscle mechanoreceptors by acting through peripheral CB1 receptors.

Figure 5.

(A) Experimental setup of the electrophysiology experiments. (B) Example of a mechanical activation threshold measurement. Force was applied by an electronic Von Frey hair (bottom trace) until an action potential discharge (upper trace) was generated in the masseter muscle afferent fibre.

Figure 6.

The line and scatter plots show the mean mechanical threshold (MT) (±SE) of six mechanoreceptors per treatment group following intramuscular injections of tetrahydrocannabinol (THC), THC with AM251 or THC with AM630 into the right (untreated) masseter muscle. Asterisks indicate significant differences compared with the vehicle group and #s indicates significant differences between THC and THC/AM251 treatment groups (two-way repeated measures ANOVA, Holm Sidak multiple comparison test, p < 0.05). THC significantly increased the MT of the mechanoreceptors. This effect was attenuated by the CB1 receptor antagonist AM251 but was not affected by the CB2 antagonist AM630. These results indicate that THC was exerting its inhibitory effect through activation of the CB1 receptor.

4 Discussion

Many current drugs for the treatment of chronic pain, in particular cannabinoids, have significant central nervous system side (CNS) effects. One way to circumvent these CNS side effects is to use a low dose, local tissue administration of these drugs to target peripheral receptors. In this regard, peripherally administered cannabinoids have been proposed as one potential solution to reduce the CNS side effects associated with systemic administration of these agents (Romero-Sandoval et al., 2015). In this study, expression of both CB1 and CB2 receptors was observed in trigeminal ganglion neurons innervating the masseter muscle, providing evidence that these receptors can be targeted for masticatory muscle pain. The reduction in CB1 and CB2 receptor expression after NGF injection suggests that this mechanism may contribute to NGF-induced local myofascial sensitization. In the behavioural experiments, intramuscular injection of cannabinoids reversed NGF-induced mechanical sensitization without an effect on motor coordination. In the electrophysiology experiments, THC increased the mechanical thresholds of mechanoreceptors that innervate the masseter muscle through peripheral CB1 receptor activation. These results are consistent with earlier studies which found that local administration of THC reduced sensitization in muscle pain conditions (Johanek et al., 2001; Yu et al., 2010; Bagues et al., 2014; Romero-Sandoval et al., 2015). Our results suggest that peripheral application of cannabinoids may be effective in the treatment of muscle pain disorders and may be a more desirable strategy than systemic administration due to a lower potential for CNS side effects.

There are two known endogenous cannabinoid receptors, CB1 and CB2 (Robson, 2014). While the CB1 receptors are thought to reduce neuronal release of neurotransmitters through presynaptic inhibition, the role of CB2 receptors and how cannabinoids produce their analgesic effects remain speculative (Robson, 2014). CB1 receptors are primarily distributed throughout the central nervous system but have also been found on many peripheral tissues including trigeminal ganglion neurons, while CB2 receptors are primarily expressed by cells with an immune function (Price et al., 2003; Svízenská et al., 2008; Nadal et al., 2013; Ulugol, 2014). However, a recent study also found their expression in human myofascial tissues (Fede et al., 2016). In this study, CB1 and CB2 receptors were both observed in the trigeminal ganglion neurons innervating the masseter muscles as well as on their nerve endings in the masseter muscles. The expression of both receptors was also higher in the TRPV1-expressing neurons than in the non-TRPV1-expressing neurons. Since TRPV1 has been proposed as a nociceptor-specific transducer channel, these results suggest that CB1 and CB2 receptors may be preferentially expressed in putative muscle nociceptors, highlighting their potential as peripheral targets of analgesia (Woolf and Ma, 2007).

NGF is a neurotrophin essential for the growth and survival of sensory and sympathetic neurons (Bennett, 2001; Pezet and McMahon, 2006). In adults, it has also been found to play an important role during injury and nociception (Price et al., 2003; Watson et al., 2008). Previous studies showed that NGF may induce pain through increasing the expression of many excitatory receptors and substances associated with pain, including ion channels and receptors such as TRPV1, and voltage-gated sodium channels, such as ASIC3 and P2X3 (McMahon and Bennett, 1999; Cheng and Ji, 2008; Watson et al., 2008). NGF has also been found to increase the expression of the neuropeptides calcitonin gene-related peptide and substance P as part of the phenotypic switch that is proposed to occur in sensory neurons following injury and inflammation (Woolf and Ma, 2007; Latremoliere and Woolf, 2009). We recently found that peripheral NMDA receptors may also be a part of the mechanism whereby NGF sensitizes muscle afferent fibres (Wong et al., 2014a). In this study, we found NGF decreased the expression of CB1 and CB2, on masseter ganglion neurons, suggesting that NGF may also induce sensitization, in part, by reducing endogenous peripheral inhibitory input. This result is in contrast to an earlier study where the expression of CB1 receptors increased after induction of CFA-induced inflammatory pain (Amaya et al., 2006; Niu et al., 2012). This may be understandable as NGF does not induce inflammation (Apfel et al., 1994; Mann et al., 2006) and could suggest that in chronic pain conditions where there is no obvious inflammation, pain may be mediated, in part, by a reduction in peripheral inhibitory input, in part, through a decrease in peripheral cannabinoid receptor expression.

Peripherally administered cannabinoids may compensate for the NGF-induced decrease in CB1 and CB2 expression to provide useful analgesia. However, one could argue that the reduction in cannabinoid receptor expression may reduce the effectiveness of these compounds. However, in this study it was found that approximately 40% of masseter ganglion neurons still expressed the CB1 receptor after NGF treatment. Conditional knockout of peripheral CB1 receptors in mice demonstrated that cannabinoid-induced analgesia is mediated primarily through peripheral CB1 receptors (Agarwal et al., 2007). This is also supported by our electrophysiological results. In these experiments, THC was able to reverse the NGF-induced mechanical sensitization of muscle mechanoreceptors through activation of CB1 receptors but not CB2 receptors. Masseter muscle injection of THC had no detectable effect on the mechanical sensitivity of the contralateral masseter muscle and did not impair motor function in the inverted screen test, which indicates a local effect of the injected THC. These results suggest that although CB1 expression was reduced by NGF, activation of peripheral cannabinoid receptors was sufficient to provide effective local analgesia without systemic effects.

We did not identify a role for CB2 receptor activation in the effect of THC on NGF-induced mechanical sensitization, although it is an agonist for both receptors. Other studies, however, have suggested CB2 receptors are involved in the analgesic action of THC (Quartilho et al., 2003; Pertwee, 2008; Bagues et al., 2014). One potential explanation for the difference between this study and previous work may be that since the CB2 receptor is primarily expressed in immune cells, its effects are most readily demonstrated in inflammatory pain models. Indeed, CB2 receptors have been proposed as a target for neuroinflammation (Ashton and Glass, 2007; Adhikary et al., 2011). As NGF-induced mechanical sensitization is not associated with significant tissue inflammation, it might be expected that CB2 receptor activation would have a limited effect in this model of myofascial pain sensitivity (Mann et al., 2006).

Interaction between the neurotrophin and the endocannabinoid systems has been previously speculated (Devesa and Ferrer-Montiel, 2014). These two systems share many of the same intracellular signalling pathways that include phospholipase C, MAP kinase and phosphoinositide 3-kinase (PI3K). Activation of CB1 receptors has been shown to inhibit NGF-induced sensitization of TRPV1 receptors in mouse afferent neurons (Wang et al., 2014). Systemic administration of anandamide, an endocannabinoid, has also been found to relieve cutaneous hyperalgesia induced by intraplantar injection of NGF in rats through CB1 receptor activation (Farquhar-Smith and Rice, 2003). In this study, it was found that NGF decreased expression of peripheral CB1 and CB2 receptors without affecting the expression of TRPV1 receptors, suggesting a potential mechanism where the peripheral neurotrophic and endocannabinoid systems may influence each other directly through expression of the peripheral cannabinoid receptors. Although there are very few reports of down-regulation of receptor expression by NGF, it has been reported that NGF can decrease the expression of epidermal growth factor receptors when used to induce neuronal differentiation of rat pheochromocytoma cells (Cohen et al., 2014). This particular effect of NGF is mediated by activation of the TrkA receptor through the Ras–extracellular signal-regulated kinase (ERK) pathway. It is not known whether a similar mechanism underlies the reduced expression of CB receptors in this study. However, it is also important to recognize that our finding is based on a decreased expression of CB receptors by ganglion neurons, which may or may not, accurately reflect changes in CB receptor expression on the peripheral terminals of the afferent fibres.

Cannabis has been used to treat pain for centuries. Unfortunately, many pain sufferers who might derive benefit from its use find its psychoactive effects too unpleasant. Here, we found that local administration of THC could reverse NGF-induced mechanical sensitization in the masseter muscle through activation of the CB1 receptor without apparent CNS side effects. Intramuscular injection of NGF produces localized muscle tenderness that is reminiscent of the masticatory muscle pain reported by sufferers of TMD (Cairns, 2010). Our results suggest that peripheral cannabinoid administration may be a promising approach for the treatment of chronic masticatory muscle pain.

Author contributions

All authors contributed to the design of the study, discussed the results and commented on the manuscript. H.W. performed all the experiments and drafted the manuscript. All authors approved the final version of the manuscript.