Meningeal transient receptor potential channel M8 activation causes cutaneous facial and hindpaw allodynia in a preclinical rodent model of headache

Background

Migraine is a common, often debilitating, neurological disorder that can last from hours to several days. The recent Global Burden of Disease Study found migraine headache to be the third most prevalent disease on the planet, following dental caries and tension type headache (1). Despite the magnitude and impact of migraine, the mechanisms that lead to this disorder remain unclear. Trigeminal nociceptors innervate the meninges and are sensitive to both chemical and mechanical stimulation (1,2). These nociceptors may play a role in the development of migraine pain, but the mechanisms by which they are activated remain to be fully elucidated.

The transient receptor potential (TRP) family of non-selective cation channels is involved in several physiological and pathological processes (3). TRP channels are sensitive to both thermal and chemical stimuli and have been proposed to act in mammals as thermosensors as well as detectors of endogenous inflammatory states and external irritants (4,5). In support of this hypothesis, pharmacological as well as genetic evidence has clearly implicated certain TRP channels in the detection or transduction of sensory stimuli. There are three super-families of TRP channels, TRPV, TRPC and TRPM (6), as well as several other smaller TRP families, some with only a single member (e.g. TRPA). The TRPM subfamily has eight members: TRPM1–M8. The TRPM8 channel was formerly known as the cold and menthol receptor (CMR1) due to its responsiveness to cold temperature as well as to cooling agents such as menthol (7–9). TRPM8 is activated by noxious and non-noxious cold ranging from ∼28℃, down to 8℃ (10). TRPM8 can also be activated by cooling agents such as icilin (as well as menthol mentioned above), behaving like many ligand-gated channels in response to these agents (11). Transcripts of TRPM8 are found in a subset (<15%) of small diameter sensory neurons (7,9). As alluded to above, TRPM8 expression in the trigeminal and dorsal root ganglia is thought to confer innocuous cold sensitivity to the somatosensory system (12,13), particularly in sensory neurons innervating cutaneous tissues. However, TRPM8 is also expressed in neurons innervating deep tissues, such as the bladder and colon (14,15). As these neurons are not exposed to cold temperatures, cold may not be an activating stimulus in these tissues and the channel may respond to other endogenous activators (16,17). Possible endogenous activators/sensitizers in deep-tissue afferents include stimuli such as lysophospholipids, cyclopentenone prostaglandins and phosphatidylinositol biphosphate among others (18). Thus, TRPM8 may be a sensor of a variety of internal and external stimuli.

One of the most consistent genetic findings in migraine patients comes from several recent genome-wide association studies (GWAS) that revealed single nucleotide polymorphisms (SNPs) in and around the TRPM8 locus (19–21). The mutations can lie in coding regions of the gene, but are often within the 5’ untranslated region. It remains unclear how these mutations influence TRPM8, e.g. by altering channel expression or function, but these mutations have been verified across several populations of migraineurs (22–25) and suggest that TRPM8 may play a role in migraine. The purpose of this study was to investigate whether activation of TRPM8 in the meninges produces behaviors in rodents consistent with migraine pain using a preclinical model of headache.

Methods

Results

Dural application of 1 nmol icilin produced robust facial (Figure 1(a)) and hindpaw (Figure 1(b)) allodynia that peaked 3–4 hours later. This allodynia was dose-dependent as shown in the time courses and by area-over-curve plots in Figure 1(c,d). Allodynia was not observed in response to dural application of vehicle (PEG 300). Animals treated with 1 nmol icilin displayed facial withdrawal thresholds significantly different from controls from 2–5 hours, thresholds were trending towards baseline at five hours and had completely returned to baseline by 24 hours. Animals treated with 100 pmol or 10 pmol icilin did not produce facial or hindpaw responses significantly different than controls. OPEN IN VIEWER

Although icilin is commonly used as an activator of TRPM8, blockade of this channel with a selective antagonist would further support the conclusion that headache-like responses following dural icilin are mediated by TRPM8. First, an in vitro calcium influx assay was used to examine blockade of TRPM8 by the antagonist AMG1161. In cultured CHO cells stably transfected with the rat TRPM8 channel, AMG 1161 (Figure 2) displayed concentration-dependent inhibition of TRPM8 activation by icilin and cold with IC₅₀ values of 23 ± 0.9 nM (n = 6 independent experiments with three replicates for each concentration) and 11 ± 0.4 nM (n = 2 independent experiments with three replicates for each concentration), respectively. In order to address selectivity, similar studies were performed against related TRP channels including TRPV1 (activated with capsaicin), TRPV3 (activated with 2-APB), TRPV4 (activated with 4-alpha PDD), TRPC5 (activated with cold temperature) and TRPA1 (activated with allyl isothipcyanate; AITC). Corresponding IC₅₀ values were >10 µm, >10 µm, >20 µm, >40 µm and >40 µm, respectively (data not shown). These data indicate that AMG1161 has functional selectivity over similarly related TRP channels. OPEN IN VIEWER

In order to determine whether icilin produces its behavioral effects via activation of TRPM8, AMG1161 was given to rats prior to dural stimulation. Oral pretreatment with AMG1161 (10 mg/kg) 30 min prior to application of 1 nmol icilin on the dura prevented the reduction in facial and paw withdrawal thresholds due to icilin administration. There was a significant decrease in facial and hindpaw allodynia at the three hour and four hour time points compared to icilin alone and also for facial allodynia at the five hour time point (hindpaw allodynia was not significant at five hours) (Figure 3(a,b)). Oral pretreatment with vehicle or AMG1161 had no effect on animals given vehicle on the dura (Figure 3(a,b)) indicating that AMG1161 alone had no effect on withdrawal thresholds. OPEN IN VIEWER

Sumatriptan (as well as other triptans) is considered to be a standard first-line abortive agent in the treatment of human migraine patients but is not given for other forms of pain (34). Prior animal studies have used sumatriptan as a probe to determine whether behavioral responses are headache-like (27,30). Thus, efficacy of sumatriptan against icilin-induced behavioral responses was examined to determine whether these behaviors are also consistent with headache. Simultaneous treatment of rats with sumatriptan (0.6 mg/kg, s.c.) and dural icilin led to partial attenuation of the decrease in both facial and paw withdrawal thresholds observed with icilin alone (Figure 4). Sumatriptan treatment did not cause any changes in withdrawal thresholds in animals given vehicle (PEG-300) onto the dura (Figure 4). OPEN IN VIEWER

NOS inhibitors have also been found to be efficacious in humans with migraine as pharmacological blockade of NOS with the non-selective inhibitor L-NG-Monomethylarginine (L-NMMA) significantly reduced migraine pain in a small clinical trial (35,36). Similar to the experiments described above with sumatriptan, we also assessed whether the icilin-induced behaviors in this preclinical model were sensitive to NOS inhibitors. Animals were pretreated with another non-selective NOS inhibitor L-NAME (20 mg/kg, s.c.) or vehicle 15 min prior to icilin administration (Figure 5(a,b)). Animals that received L-NAME/icilin displayed significantly higher thresholds (i.e. reduced allodynia) both in the face and hindpaw compared to those of the vehicle/icilin group. No significant differences were found between the L-NAME/icilin group and vehicle groups. Together with the efficacy of sumatriptan, these data indicate that the behavioral response in rats due to dural TRPM8 activation is consistent with headache. OPEN IN VIEWER

Discussion

Migraine is a debilitating disorder that affects a population of otherwise healthy individuals and for which current treatments are often ineffective. Despite the prevalence of migraine, the underlying pathophysiology is not well understood and thus identification of novel targets for migraine therapies is challenging. Although pain signaling from the meninges has been proposed to play a role in the development of the pain of migraine, little is known about the mechanisms by which afferents are activated during a migraine event. We and others have described numerous mechanisms capable of initiating nociceptive signaling from the meninges in animals (26,27,37) but human data supporting a role for many of these mechanisms are largely absent. In contrast, human data implicating TRPM8 in migraine has been reported in several recent GWAS of migraine patients (21–25,38) but preclinical data supporting a role for TRPM8 in migraine do not yet exist.

Using a preclinical model of headache, we now demonstrate that stimulation of the dura with the TRPM8 agonist icilin evokes cutaneous allodynia, a common feature of migraine.

Previous preclinical work has shown cutaneous allodynia after dural afferent activation with numerous compounds (39–41) and thus these data are consistent with other pro-nociceptive stimuli applied to the dura. An interesting observation from these studies is the development of mechanical allodynia following TRPM8 activation, a type of hypersensitivity not typically observed after activation of this channel (cold allodynia is usually the phenotype associated with TRPM8). However, the current studies are not testing the primary site of activation (i.e. the dura mater) and are dependent on referred allodynia of the facial and hindpaw regions. It is thus not known whether thermal thresholds are altered in the dura mater where TRPM8 is activated. Allodynia was nonetheless attenuated by pretreatment with the TRPM8 antagonist AMG1161, a compound that shows several hundred-fold selectivity over related TRP channels, indicating this affect was due to specific activation of TRPM8 in the dura. Additionally, both sumatriptan and L-NAME attenuated the cutaneous facial and hindpaw allodynia due to dural application of icilin. This is presumably due to attenuation of afferent input from the meninges and inhibition of subsequent central sensitization necessary to establish referred facial and hindpaw allodynia (27), although our data do not prove this mechanism. The ability of sumatriptan and L-NAME to attenuate icilin-induced allodynia in this model suggests that the behavioral responses due to activation of TRPM8 within the dura are consistent with headache. Triptans are one of the most common abortive agents for migraine and NOS inhibitors have also been shown to be efficacious in inhibiting migraine pain in humans (35,36). Taken together, these data suggest that activation of TRPM8 on meningeal afferents contributes to headache.

Although several GWAS now exist implicating SNPs in TRPM8 in migraine, the actual effects of the mutations on channel expression and function are not yet known. Consequently, it is difficult to determine whether patients with these mutations have increased or decreased TRPM8 function. Activation of TRPM8 on sensory neurons can initiate afferent signaling but, interestingly in the case of this channel, the sensation that is ultimately perceived may be pro-nociceptive or anti-nociceptive, possibly depending on the context. In animals, TRPM8 is essential for both neural and behavioral responses to noxious cold, as well as cold mimetics (42). Deletion or antagonism of TRPM8 reduces inflammation-induced cold hypersensitivity (43–45), indicating that this channel is necessary for cold allodynia in pain states. But TRPM8 can also produce analgesia in states of inflammation (8,46) and menthol is widely used as a topical analgesic. Ultimately, it may depend on whether TRPM8 is activated alone or in the presence of other stimuli. In support of this concept, it was recently shown that, while activation of TRPM8 alone is nociceptive, TRPM8 activation can decrease nociception due to stimulation of other TRP channels (47). Our data demonstrate that activation of TRPM8 alone in the meninges is pro-nociceptive but whether TRPM8 is activated alone during migraine in humans is not clear. Future studies will ideally determine how migraine-associated TRPM8 SNPs impact channel function. These studies will also shed light on whether patients are more likely to have increased or decreased channel expression and function and will help further uncover how TRPM8 may contribute to migraine (i.e. in a protective or causative).

It should be noted that prior studies on TRPM8 expression in meningeal afferents found little to no channel expression on these neurons (48). This is seemingly at odds with our current findings that activation of TRPM8 in the dura is pro-nociceptive. Differences in species could account for the discrepancy in findings as our studies were performed in rats while the prior work was in mice and expression of TRPM8 on dural afferents may be species dependent. Additionally, other studies using mice have found TRPM8 expression in the meninges but found that the expression is restricted to specific regions (49). TRPM8 may thus be expressed in the meninges in mice, but not equally across the tissue. Another possibility is that TRPM8 is not expressed on meningeal afferents but is found on other cell types in the dura. Mast cells have been implicated in migraine (50) and TRPM8 is expressed on this cell type (51). Additionally, we recently reported that dural fibroblasts may also contribute to headache pathology (52) and TRPM8 is expressed on several types of fibroblasts (53). Thus, activation of the channel on non-neuronal cells could indirectly initiate afferent signaling. Future work is necessary to more conclusively determine the expression pattern of the channel in the meninges and how its activation may ultimately contribute to headache.

These studies provide the first preclinical evidence that TRPM8 can play a role in headache disorders such as migraine and they provide a potential mechanism for how the channel may contribute to migraine pain, i.e. activation within the meninges and subsequent initiation of afferent signaling. Although it still remains to be determined how TRPM8 is impacted by the mutations in human migraine patients, these data suggest that meningeal TRPM8 is capable of contributing to headache. Importantly, these findings imply that TRPM8 antagonists may have efficacy as novel migraine therapeutics.

Article highlights

Activation of TRPM8 in the dura mater produces behavior in rats that is consistent with headache.