Allosteric Cannabinoid Receptor 1 (CB1) Ligands Reduce Ocular Pain and Inflammation

Cannabinoid receptor 1 (CB1) activation has been reported to reduce transient receptor potential cation channel subfamily V member 1 (TRPV1)-induced inflammatory responses and is anti-nociceptive and anti-inflammatory in corneal injury. We examined whether allosteric ligands, can modulate CB1 signaling to reduce pain and inflammation in corneal hyperalgesia. Corneal hyperalgesia was generated by chemical cauterization of cornea in wildtype and CB2 knockout (CB2−/−) mice. The novel racemic CB1 allosteric ligand GAT211 and its enantiomers GAT228 and GAT229 were examined alone or in combination with the orthosteric CB1 agonist Δ8-tetrahydrocannabinol (Δ8-THC). Pain responses were assessed following capsaicin (1 µM) stimulation of injured corneas at 6 h post-cauterization. Corneal neutrophil infiltration was also analyzed. GAT228, but not GAT229 or GAT211, reduced pain scores in response to capsaicin stimulation. Combination treatments of 0.5% GAT229 or 1% GAT211 with subthreshold Δ8-THC (0.4%) significantly reduced pain scores following capsaicin stimulation. The anti-nociceptive effects of both GAT229 and GAT228 were blocked with CB1 antagonist AM251, but remained unaffected in CB2−/− mice. Two percent GAT228, or the combination of 0.2% Δ8-THC with 0.5% GAT229 also significantly reduced corneal inflammation. CB1 allosteric ligands could offer a novel approach for treating corneal pain and inflammation.


Introduction
The cornea has one of the densest concentrations of unmyelinated sensory nerve endings in the body [1,2], which are highly sensitive to mechanical stimulation, temperature, and various chemicals mediators, through receptors such as the transient receptor potential family [3]. Damage or irritation to these nerve endings resulting from ocular surface manipulations such as cataract surgery [4], long-term and improper use of contact lenses [5][6][7], and frequent exposure to irritating environmental and chemical stimuli (infection, air pollutants, hazardous chemicals, air pressure etc.) [1,8,9], can lead to local release of inflammatory mediators, including calcitonin gene-related peptide (CGRP), causing corneal inflammation [10]. Stimulation of corneal nerves following damage can result in ocular pain, either to normally non-noxious stimuli (allodynia) and/or as a heightened pain response to noxious stimuli (hyperalgesia) [9,[11][12][13], which can result in sensitization and neuropathic pain over time [14].
Current pharmacotherapies for corneal pain and inflammation include topical steroids, non-steroidal anti-inflammatory drugs, antibiotics, as well as other agents for neuropathic pain including, tricyclic antidepressants, GABAergic drugs (e.g., gabapentin), opioids, etc. [12,15,16]. However, these treatments are not always effective enough to produce adequate pain relief, especially where both pain and inflammation may need to be controlled [13,15,17].
Modulation of the endocannabinoid system (ECS) has emerged as a novel approach to treat pain and inflammation, among other conditions [18][19][20]. The ECS is comprised of G protein-coupled receptors (cannabinoid receptor 1, CB1; and cannabinoid receptor 2, CB2), endocannabinoids, and the enzymes responsible for their synthesis and degradation [21][22][23][24][25][26][27]. Support for a functional role for the ECS, and specifically CB1 activation in modulating TRPV1-mediated corneal pain, was provided by a recent paper demonstrating that cannabinoids that act at CB1, such as tetrahydrocannabinol (THC), are able to reduce corneal hyperalgesia to a capsaicin challenge and, additionally, also produce a reduction in corneal injury-induced inflammation [28].

GAT211 and GAT229
Potentiated the Anti-Nociceptive Effects of ∆ 8 -THC, Whereas GAT228 Directly Reduced Corneal Pain Different concentrations of the racemic compound GAT211, and the resolved enantiomers GAT229 and GAT228, were applied topically in WT mice to establish the effective concentrations required to reduce the corneal pain score compared to the vehicle-treated group (27 ± 7, n = 8; Figure 1D). Some compounds were then tested in combination with subthreshold concentrations of ∆ 8 -THC. Administration of 0.4% ∆ 8 -THC did not reduce the pain score in capsaicin-challenged corneas (p > 0.05, n = 6) as previously reported [28], nor did administration of 0.2% ∆ 8 -THC (p > 0.05, n = 6).

Discussion
Hyperalgesia is a well-documented symptom of inflammatory and/or neuropathic pain [13,[37][38][39]. In our mouse model of corneal injury, we observed a heightened pain response (hyperalgesia) to chemical stimuli. Previous studies have shown that activation of CB1 reduces TRPV1-mediated pain and inflammation in other models of pain, including nerve growth factor-sensitized pain [40,41], as well as inflammation-induced pain in urinary bladder [42]. In the cornea, we have previously reported that activation of CB1 by ∆ 8 -THC reduces a TRPV1-induced corneal pain response, provoked through a capsaicin challenge, and reduces neutrophil infiltration [28]. Consistent with our findings, CB1 activation in cornea has been implicated in TRPV1 desensitization and a decrease in pro-inflammatory mediators after corneal injury [20]. CB1 has also been reported to be important for the normal course of corneal wound healing [20,43]. Therefore, together with previous data [28], we have further provided evidence to support that activation CB1 may be a good target for corneal neuropathic pain management by direct modulation of the sensation of pain, and the inflammatory response which may lead to sensitization over time.  . Values represent mean ± SD. Arrow in (A) points to one of many infiltrating neutrophils. Scale bar: 50µm. For statistical analysis one-way ANOVA with Dunnett's post hoc test (compared to vehicle) was used. * p < 0.05, ** p < 0.01, **** p < 0.0001. However, therapeutic use of CB1 orthosteric agonists may be limited due to side-effects such as dose-dependent receptor desensitization, and off-target effects [30][31][32]44,45]. CB1 positive allosteric modulators may be advantageous in that they may provide an alternate means to modulate CB1, but with fewer of these limitations. PAMs may stabilize receptor conformations in such a way that they can fine-tune the effects of orthosteric ligands [46,47], increasing affinity and/or efficacy of binding, and ultimately resulting in changes in downstream signaling [33,44,48,49]. Previous studies have reported that administration of GAT211 and ZCZ011, a related CB1 PAM, reduced mechanical and cold hyperalgesia in mouse models of neuropathic and inflammatory pain [35,36]. GAT211 treatment did not show evidence of anti-nociceptive tolerance throughout the entire 19 day dosing period, or physical dependence (measured by paw tremors) at 20 days following the treatment [35]. Similarly, there was no difference in the anti-nociceptive effect of ZCZ011 following 6 days of chronic dosing (40 mg/kg, i.p. b.i.d) compared to acute dosing at day 1 (40 mg/kg, i.p.) [36]. Administration of GAT211 alone also did not produce side-effects associated with orthosteric activation of CB1, such as those produced by ∆ 9 -THC, or WIN55,212-2 [35].
In this paper, we have shown that topical application of the racemic ago-PAM GAT211, or the PAM GAT229, potentiated the corneal anti-nociceptive effects of a subthreshold dose of the orthosteric agonist ∆ 8 -THC. Additionally, GAT229 also potentiated the anti-inflammatory effects of a subthreshold dose ∆ 8 -THC measured at 6 h following capsaicin stimulation. Dose-dependent potentiation of cannabimimetic side-effects with the combination of CB1 PAMs with orthosteric agonists, however, was reported by both Slivicki [35] and Ignatoawaska-Jankowska [35] and their colleagues. Therefore, avoiding administration of the combination of CB1 PAMs with exogenous orthosteric agonists may be advisable. Local increases of endocannabinoids at the site of pathology that are enough to potentate the actions of CB1 PAMs may be more desirable as a treatment paradigm; it would enable PAMs to be administered without requiring the addition of orthosteric agonists. Such was the case in models of neuropathic and inflammatory pain treated with GAT211 and ZCZ011 [35,36], and for intraocular pressure lowing in a mouse model of ocular hypertension, an effect which was absent in normotensive mice [50]. In our corneal hyperalgesia model, administration of GAT211 or GAT229 reduced corneal pain when combined with subthreshold ∆ 8 -THC, but not on their own. This may be due to lack of, or insufficient, local endocannabinoid production to allow PAMs to potentiate activity at CB1, at least at the time point used to measure the pain response in this study. Unlike previous studies investigating the in vivo effects of CB1 PAMs, our model is relatively acute, with drug administration occurring shortly after cauterization, and capsaicin challenge occurring only 6 h later. In contrast, AEA and 2-AG were significantly increased in the brain and spinal cord 3 and 7 days following injury [51], and at least 14 days in the dorsal root ganglia [52], consistent with the timeline of effects observed with PAM administration [36]. Further studies using a more chronic model of corneal pain would therefore be useful to investigate if increases in endocannabinoid levels following injury in the cornea are sufficient to permit PAM potentiation of CB1 activation by endocannabinoids. Although avoiding PAM-potentiated psychoactivity may not be an issue with respect to topical administration to the eye, it has been reported that following ocular topical administration, WIN 55,212-2 can be detected in the urine [53]. While psychoactivity from this route of administration has not been reported, chronic application over time of exogenous potent CB1 agonists may still have the potential to result in systemic drug levels. Therefore, use of lower doses of CB1 agonists together with a PAM for topical use may also have advantages to avoid systemic side-effects.
Finally, unlike GAT211 or GAT229, the allosteric agonist GAT228, significantly reduced corneal pain and inflammation on its own, consistent with actions of an allosteric agonist, as previously reported in vitro [34]. While the actions of allosteric agonists at CB1 have yet to be fully characterized, it is possible that activation through the allosteric site may retain some advantages of allosteric modulators over orthosteric agonists [54,55]. However, it was recently suggested that in absence of a ligand at the orthosteric site, GAT228 may bind with equal affinity to either the allosteric or orthosteric site [56]. While we cannot exclude that the actions of GAT228 are due to orthosteric rather than allosteric agonism, in radioligand binding experiments GAT228 (up to 1 µM) did not displace the orthosteric agonist, implying allosteric interactions [34]. Further investigation of chronic dosing of GAT228 could be worthwhile and may identify if use of this compound could provide benefit over more traditional orthosteric agonists.
This paper provides further evidence supporting the role of CB1 in modulating capsaicin-evoked corneal pain responses and inflammation. Here, we present for the first-time evidence that allosteric activation of CB1 using the PAMs GAT211 or GAT229, in combination with subthreshold dose of CB1 orthosteric agonist ∆ 8 -THC, or the CB1 ligand GAT228 alone, reduced both corneal pain and inflammation. CB1 PAMs in combination with subtherapeutic dose of orthosteric agonists, or CB 1 allosteric agonists alone, could be a novel approach for the treatment of corneal pain and inflammation.

Experimental Animals
All animal experiments and care complied with the Canadian Council for Animal Care guidelines (http://www.ccac.ca/) and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Protocols were approved by the Dalhousie University Committee on Laboratory Animals or by the Indiana University Animal Care Committee.

Induction of Corneal Injury
Corneal injury in mice was induced as previously described in Thapa et al. (2018), using a protocol adapted from a rat model of corneal hyperalgesia [57]. Briefly, mice were anesthetized using 2-3% isoflurane and corneal injury was induced in both eyes using a silver nitrate-coated (MedPro ® , 75% silver nitrate, 25% potassium nitrate; AMG Medical Inc., Montreal, QC, Canada) micro-applicator brush (Centrix Inc., Shelton, CT, USA). The micro-brush was held in contact with the cornea for 2 s, producing a distinct superficial white lesion (~1 mm diameter) in the epithelial cell layer. Cauterized eyes were then rinsed several times with room temperature saline. An ocular lubricant (Systane ® , Alcon Canada Inc., Dorval, QC, Canada) was applied to the corneal epithelial surface to reduce corneal drying. Mice recovered fully from anesthesia within 3-5 min post-cauterization.

Assessment of Behavioral Pain Response
Corneal sensitivity to an acute capsaicin stimulus was evaluated in animals at six hours following cauterization injury, as previously described [28]. Mice were lightly restrained and injured eyes were challenged with capsaicin (1 µM) stimulation applied topically (5 µL). Each animal was given a single dose of capsaicin per eye (right eye first, followed by left eye) to elicit a pain response and to avoid the desensitization which may occur from repeated application. Topical application of capsaicin was associated with rapid blinking that occurred over a 30 s period and was accompanied by occasional eye wiping. The number of blinks and eye wipes were used as an indicator of ocular pain, as previously described [28]. Pain behaviors were recorded in the tested eyes using an iPhone 5S (8 megapixel). Offline analysis was carried out by an experimenter blinded to the treatments given. Videos were analyzed offline in slow motion (play speed 0.5, Windows Media Player version 10) and the pain response was scored by adding the total number of blinks and eye wipes recorded in each eye over the entire 30 s recording period following capsaicin application to give a composite pain score.
Neutrophil migration was quantified in corneal sections using an Axiovert 200M microscope with a Hamamatsu Orca R2 Camera (Zeiss, Thornwood, NY, USA). Three representative images (imaged at 20 × magnification) were taken from each section, corresponding to the right and left corneal peripheries and from the center of the cornea. Neutrophils from these three images were counted and summed to represent the total neutrophil per section. A total of 6-8 sections with 120 µm intervals were analyzed from each sample and were averaged. For each experimental group, 4-7 eyes were analyzed.

Data Analysis
Statistical analysis was performed in GraphPad Prism version 6. Unless otherwise indicated, one-way analysis of variance (ANOVA) with Dunnett's post hoc was used to compare data between groups of three or more, while analysis between two groups was performed using t-test. All data are represented as group mean ± standard deviation and were considered significant at p < 0.05.