Combined effects of cannabidiol and Δ9-tetrahydrocannabinol alleviate migraine-like symptoms in mice

Abstract

Background:

The therapeutic use of cannabidiol (CBD) and Δ9-tetrahydrocannabinol (THC) to treat migraine has been understudied. Using three mouse models, we examined the impact of CBD and THC on migraine-relevant behaviors triggered by: 1) calcitonin gene-related peptide (CGRP), 2) sodium nitroprusside (SNP), and 3) cortical spreading depolarization (CSD).

Methods:

Both male and female CD1 mice were treated with CBD (100 mg/kg) or THC (1 mg/kg) alone or in combinations of CBD (1, 30 or 100 mg/kg) and THC (1 mg/kg) prior to injection of CGRP or SNP. The mice were assessed for light aversion (photophobia), squint (non-evoked pain), and periorbital tactile hypersensitivity, as well as possible adverse effects. In a separate set of experiments, CSD events were optogenetically induced in familial hemiplegic migraine 1 (FHM1) mutant and wildtype littermates (WT) mice (C57BL/6 background), followed by grimace and motor assessments with and without combinations of CBD (30 or 100 mg/kg) and THC (1 mg/kg).

Results:

In CD1 mice, a 100:1 CBD:THC combination mitigated light aversion induced by CGRP and SNP in males and females. Rescue of CGRP- and SNP-induced squint was observed only in male mice with 100:1 CBD:THC. None of the treatments rescued periorbital tactile hypersensitivity in either sex. In FHM1 mutant and WT mice, the 100:1 CBD:THC ratio did not affect CSD characteristics but did reduce CSD-induced grimace features (i.e., head pain mimic). No adverse effects of any of the cannabinoid treatments were observed using cognitive, emotional, or motor tests.

Conclusions:

A 100:1 ratio of CBD:THC has a beneficial effect on some of the most bothersome migraine-related symptoms in three mouse models. Our findings support a potential therapeutic efficacy of combined CBD and THC treatments.

Introduction

Migraine is a primary headache disorder characterized by attacks of mostly unilateral recurrent throbbing headaches accompanied by neurological symptoms (1,2). Up to one-third of patients also experience auras that are characterized as cortical waves of spreading depolarization (CSD), hence the two main migraine subtypes: migraine with and migraine without aura (3). Because many patients do not benefit from current medications, new treatment options need to be evaluated for efficacy and safety. Cannabis and cannabis-derived products, such as cannabinoids, are currently being investigated (4,5). The endogenous cannabinoid system (endocannabinoid system, ECS) exerts various physiological functions, including maintaining homeostasis and modulating pain (6,7). Studies suggest that the ECS could be a good target for the treatment of migraine headache (8–10). Pretreatment with an endocannabinoid, N-arachidonoylethanolamide (AEA), was able to inhibit vasodilation caused by two potent migraine triggers, calcitonin gene-related peptide (CGRP) and sodium nitroprusside (SNP), through activation of the cannabinoid 1 (CB1) receptor in rats (11). Another study done on rats reported that the endocannabinoid arachidonyl-2’-chloroethylamide (ACEA) can inhibit CGRP release from trigeminal neurons through its actions on CB1 receptors only when TRPV1 receptors are inhibited (12). Furthermore, endocannabinoids and phytocannabinoids from the cannabis plant have been reported to act on dural immune cells in the meninges (13), which could reduce the neurogenic inflammation thought to contribute to migraine pathogenesis.

The main action of endocannabinoids is mimicked by the two most abundant components of cannabis, i.e., the phytocannabinoids cannabidiol (CBD) and Δ9-tetrahydrocannabinol (THC), which are typically obtained from Cannabis sativa L. and Cannabis indica L. and used in the majority of preclinical and clinical studies (7). Both CBD and THC bind to CB1 and cannabinoid 2 (CB2) receptors that are normally activated by endogenous cannabinoids (9). THC is a partial agonist with high affinity for CB1 and CB2 cannabinoid receptors, binding the same site as the endocannabinoids AEA and 2-arachidonoylglycerol (2-AG) (14). In contrast, CBD primarily targets CB1 as a negative allosteric modulator and activates CB2 as a partial agonist (15). It is important to note that CBD is also thought to exert its activity in vivo through the 5-HT1_A receptor, the m and d opioid receptors, as well as transient vanillin type-1 receptors (16). Low levels of CBD antagonize CB1 receptors in the presence of THC to act as a potent analgesic (17). Whereas THC has a strong psychoactive effect, CBD is considered non-psychoactive and is likely able to maximize some of the behavioral effects of THC while minimizing its less desirable effects (18,19).

Clinical studies have examined the effects of medical cannabis containing relatively high concentrations of THC in migraine patients, but thus far, studies did not take migraine subtypes into account (4,5,20,21). Several trials with medical cannabis have shown a reduction in monthly migraine attacks (8,22), however, studies remain difficult to compare due to the inclusion of different migraine subtypes, the presence of different symptoms in the patients, the use of different (or no) controls, and the use of different CBD and THC concentrations. In addition, a high concentration of THC is associated with psychoactive symptoms, anxiety, panic attacks and a high risk of addictiveness (23). Conversely, CBD-only therapies have shown contradictory results in chronic pain patients, ranging from placebo-equivalent (24) to highly effective (25,26). The use of a 1:1 ratio of CBD:THC was shown to improve or mitigate the psychoactive effects seen with THC-only administration (27–29). Despite the fact that CBD and THC act through different mechanisms, and that psychoactive effects of THC may be a limiting factor, a combination of CBD and THC could provide unique analgesic properties (14), which makes their combined use attractive as a potential therapeutic in migraine (9). The use of CBD and THC is further supported by data from experimental animal models of migraine. THC (0.32 mg/kg) applied directly after dural activation prevented migraine-like pain in a rat model (30). In mice in which headache-like symptoms were induced by infusion of CGRP, CBD (10 and 30 mg/kg) reduced migraine-like pain behavior (31), while THC suppressed features of KCl-induced CSD in rat cortical brain slices (32).

In the present study, two independent laboratories used three different preclinical migraine models to test the efficacy of CBD, THC, and CBD:THC ratios. The models were: 1) injection of the neuropeptide CGRP (33); 2) injection of the nitric oxide donor SNP (34); and 3) optogenetic induction of cortical spreading depolarization (CSD) in both wildtype (WT) and – to increase translational value of the study (35,36) – familial hemiplegic migraine type 1 (FHM1) mutant mice that carry the R192Q missense mutation in the a1 subunit of Ca_V2.1 calcium channels (37). Using the CGRP and SNP injections to challenge CD1 mice, the effects of CBD, THC, and varying combinations were tested using light aversion, periorbital tactile hypersensitivity, and automated squint assay as behavioral readouts of migraine-like symptoms (38–40). In parallel, combinations of CBD:THC were tested on the optogenetically induced CSD model, both in FHM1 mice and WT littermates, with recordings of CSD features as a measure of CSD susceptibility and using the mouse grimace scale (MGS) as a head pain mimic (35). Finally, the possible cognitive, emotional, or motor adverse effects of the different treatments were assessed. Our findings support the potential efficacy of combined CBD and THC as therapy in both migraine with and migraine without aura.

Methods

Animals

Female and male CD1 mice (Charles River, USA), 8–9 weeks of age, were acclimatized for one week, and were housed in groups of four per cage prior to experiments with injection of CGRP, SNP or vehicle phosphate-buffered saline (PBS). The estrus cycle of female mice was not assessed. Homozygous female and male FHM1 mutant mice, which had been generated by introducing the human pathogenic FHM1 R192Q missense mutation in the mouse Cacna1a gene as described before (37) (and that had been backcrossed to a C57BL/6 genetic background for 20 generations), and WT littermates, 3–6 months of age were used. For optogenetic CSD induction, these mice were crossbred with Thy1-ChR2-YFP mice (strain 7612–B6.Cg-Tg (Thy1-COP4/EYFP)18Gfng/J; Thy1/ChR2-YFP; Jackson Laboratory, ME, USA) to obtain expression of channelrhodopsin-2 (ChR2) in cortical neurons (36). After crossbreeding, Thy1-ChR2/R192Q (ChR2/FHM1) mutant mice and WT littermates (ChR2/WT) homozygous for ChR2 were obtained. All mice were housed under standard conditions (22 ± 1.5°C, humidity 45–55%, 12-h light/dark cycle) and had free access to water and food. Lights were turned on at 6 AM and turned off at 6 PM. All experiments were carried out during the light period between 8 AM and 5 PM.

For all experiments, investigators were blinded to genotype and/or drug treatment and animals were randomized to treatment groups prior to experimentation. The CSD and four-limb hanging experiments in FHM1 mutant mice and WT littermates were not powered to look at sex differences. A.K. was aware of group allocation for the CGRP and SNP experiments carried out in the University of Iowa. E.A.T. and S.H.vH. were aware of group allocation for the CSD and four-limb hanging mesh experiments carried out at the Leiden University Medical Center. Experiments involving injections of CGRP, SNP or PBS were approved by the University of Iowa Animal Care and Use Committee and carried out in accordance with the standards set by the US National Institutes of Health. Experiments involving optogenetic CSD induction were approved by the Animal Experiment Ethics Committee of the Leiden University Medical Center and the Dutch Central Authority for Scientific Procedures on Animals, in accordance with recommendations of the European Communities Council Directive (2010/63/EU). Importantly, experiments were carried out in two different laboratories, in two different countries, by different teams of investigators who were blinded to results obtained from the other team until experiments were complete. On the day of testing mice were observed for signs of distress such as failure to eat or drink, hunched posture, reduced grooming or ruffed fur, lethargy, and piloerection. During testing mice were continuously monitored for behavior, movement, and overall wellbeing. If any mouse showed signs of distress and/or was not moving, the protocol was to remove the animal from the study, however no animals were removed. All experiments were carried out in accordance with ARRIVE guidelines (https://arriveguidelines.org) and efforts were made to minimize discomfort of the animals.

Drug administration

CGRP and SNP were administered to CD1 mice by intraperitoneal (i.p.) injections at 10 μL/g body weight. Rat α-CGRP (0.1 mg/kg, Sigma-Aldrich, MO, USA) and SNP (2.5 mg/kg, Sigma-Aldrich) were diluted in Dulbecco’s PBS (Hyclone, Global Life Sciences Solutions, DE, USA). PBS was used as the vehicle control. For CGRP and SNP experiments, purified CBD (100 mg/mL) and THC (1 mg/mL) were initially obtained from Eurofins Proxy Laboratories (The Netherlands), supplied in sunflower oil. Because of issues with drug transport across continents, later experiments at the Iowa site used synthesized CBD and THC from Purisys (GA, USA), in which CBD in powder form was dissolved in sunflower oil (Sigma-Aldrich) to make a stock solution of 100 mg/mL. THC came in solid form and was melted in a water bath (85–90°C) in a glove bag under inert gas (N₂). Sunflower oil (10 mL) was added to 10 g melted THC to make a 500 mg/mL solution, then serially diluted to make a 1 mg/mL stock solution in sunflower oil. The various ratios of CBD:THC were diluted in sunflower oil and administered i.p. at 3.3 μL/g body weight, 30-min prior to CGRP, SNP or PBS injection. Dilutions were prepared the morning of the experiment. Light aversion assays were performed using both purified and synthesized agents, and because no difference was observed depending on the origin of the products (data not shown), data were pooled together for figures and analysis. The light-dark assay, rotarod, and Y-maze experiments were performed with purified cannabinoids only (Eurofins). The squint, tactile hypersensitivity, open field, and tail suspension experiments were performed with synthetic cannabinoids only (Purisys). In CSD experiments, 30:1 and 100:1 CBD:THC combinations were prepared by dilution of CBD and THC (both also from Eurofins Proxy Laboratories) in sunflower oil (Fagron, The Netherlands) and administered i.p. at 6 μL/g body weight, 30 min prior to optogenetic CSD induction. Fresh batches were prepared from stock solutions on the day before the experiment, put on a rocking table overnight at room temperature. For all experiments, vehicle controls for CBD, THC, and CBD: THC consisted of sunflower oil. All the experiments were designed as models of acute migraine.

Experimental design

A summary of the animal models, treatments, and associated tests is shown in Figure 1.

Figure 1.

Schematic overview of the animal models, treatments, and behavioral tests. The left panel illustrates a migraine without aura model in CD1 mice. Mice received intraperitoneal (i.p.) injections of cannabinoids or vehicle, followed by either CGRP or SNP to induce migraine-like symptoms. Behavioral assessments included light aversion, eye squinting, and tactile hypersensitivity. Adverse effects were evaluated using anxiety-related tests (light aversion and open field), spatial memory (Y-maze), motor coordination (rotarod), and depression-like behavior (tail suspension test) in both naïve mice and mice previously tested for migraine-like symptoms. No significant differences were observed between naïve and previously tested mice. The right panel depicts a migraine with aura model using transgenic FHM1 mutant (ChR2/FHM1) mice and wild-type littermates (ChR2/WT). Cannabinoids or vehicle were administered i.p. prior to optogenetically induced cortical spreading depolarization (CSD). Behavioral assessments included the mouse grimace scale for pain scoring. Adverse effects were evaluated using grip performance (wire grip test) before, during, and after CSD induction, as well as muscle endurance (four-limb hanging mesh test).

Migraine without aura model.

CD1 mice received intraperitoneal (i.p.) injections of either cannabinoids or vehicle prior to administration of calcitonin gene-related peptide (CGRP) or sodium nitroprusside (SNP). Following treatments, mice were assessed sequentially for light aversion, eye squinting, and tactile hypersensitivity. To evaluate adverse effects, additional behavioral assays were conducted, including anxiety-like behaviors (light aversion and open field tests), spatial memory (Y-maze), motor coordination (rotarod), and depression-like behaviors (tail suspension test). Assessments were performed on both naïve mice and mice previously subjected to migraine-like tests (light aversion, squinting, and tactile hypersensitivity), with no significant differences observed between the two groups.

Migraine with aura model.

Transgenic FHM1 mutant mice expressing channelrhodopsin-2 (ChR2/FHM1) and their wild-type littermates (ChR2/WT) were used. Mice were administered cannabinoids or vehicle i.p. prior to the induction of cortical spreading depolarization (CSD) using optogenetic stimulation. Pain-related behaviors were assessed using the mouse grimace scale. To evaluate adverse effects, grip strength was measured using the wire grip test before, during, and after CSD induction, and muscle endurance was assessed with the four-limb hanging mesh test.

The rationale for the dosing was that we started with a concentration of THC (1 mg/kg) that had a relatively low adverse effect profile and used that as an anchor for varying the CBD doses. As a lower ratio, we used 1:1 CBD:THC, which is available in the UK (Sativex) for multiple sclerosis. We then did a small pilot study with ratios up to 300:1 (data not shown since only small, single experiments) and saw the best rescue with 30:1 and 100:1, so those ratios were further examined in multiple independent experiments.

Light aversion assay

This assay is used as a surrogate for photophobia. Mice were pre-exposed to the chambers once to reduce exploratory drive and tested 48 h later. On the day of experimentation, mice were allowed to acclimatize for 1 h within the experimental room. After injection of CGRP, SNP or PBS, the CD1 mice were individually placed in light/dark boxes with two chambers; one dimly lit (55 lux) and the other not lit (0.05 lux), with the mouse being able to move freely between chambers. We found that CD1 mice respond to CGRP injections even in dim light that is approximately similar to a room lit by a computer screen. The location of the mouse in the box was determined by infrared beam tracking (Med Associates, VT, USA), as described before (38,41,42). Data were collected over a 30-min period and analyzed in sequential 5-min bins. Time in the light was reported as longitudinal data (mean ± standard error of the mean (SEM) for all mice at each 5-min interval) and as a scatterplot of individual mice (mean ± SEM for the average 5-min bin across the 30-min testing time). Motility and resting behavior in the light and dark zones were also monitored during the assay. The amount of time that the mice spent resting in either the light or dark zones was calculated as a percentage of the time spent in that specific zone, not a percentage of the total time in both zones. Mice that spent over 90% of the total time resting in the light and dark zones were excluded from the analysis and they are listed in Online Supplementary Table 1.

Automated squint assay

This assay was used to measure spontaneous, unprovoked pain that is manifested as the squint component of the grimace response. The mouse was acclimatized to a customized collar restraint for three sessions of 20 min on separate days. On the test day, the mouse was habituated to the testing room for 1 h before being placed in the restraint. Two separate 5-min recordings were performed (baseline and treatment), under normal room light condition, as described (40,41). Facial detection software (FaceX, LLC, IA, USA) was used to calculate the pixel area of the right eye every 0.1 s (10 frames per second) in the recordings. A custom MATLAB script was used to compile the values. Any frames with a tracking error rate greater than 15% were excluded from the analysis and they are listed in Online Supplementary Table 6.

Periorbital tactile hypersensitivity

This assay was used to measure facial touch sensitivity. For periorbital testing, mice were tested as described before (43). Briefly, each mouse was acclimated to its own poly-coated paper cup (Choice 4 oz. paper cups; 6.5 cm top diameter, 4.5 cm bottom diameter, 72.5 cm length) for 20 min each day for 5 to 10 days, until habituated. During each acclimation period, the D von Frey filament (Bioseb, France) was repeatedly approached to the periorbital area to lightly touch the head without applying pressure, to decrease their withdrawal reflex. Mice were considered habituated once the tip of the filament could reach the head (without any pressure) without the mouse reacting to it. A set of eight von Frey filaments was used from A (0.008 g) to H (1 g). Testing was performed according to the up and down method previously described before (43). Briefly, filaments were applied for 3 s at the periorbital area or 5 s at the plantar area. The D (0.07 g) filament was applied first. If the animal reacted (withdrew the head, wiped eyes or periorbital region, withdrew, shook, or licked hindpaw for plantar), then a lower filament was applied. If there was no reaction, then a higher filament was applied. The pattern of responses was recorded. This method was used until five applications after the first change in the pattern were assessed. The last filament applied, and the final pattern were used to calculate the 50% threshold following an established equation (44,45). Since this technique does not yield continuous thresholds and data cannot be analyzed using parametric statistics, the 50% thresholds (g) were log-transformed before being analyzed to obtain normally distributed data.

Open field test

The open field test was utilized to assess center avoidance as an indicator of anxiety-like behavior. Testing was carried out using the same equipment and light intensity (55 lux) as in the light/dark assay, but without the dark insert. Similar to the light dark assay, mice were acclimated to the room an hour before testing. They were not acclimated to the chamber. The chamber configuration is arranged to have a border measuring 3.97 cm along the periphery, resulting in a center area of 19.05 × 19.05 cm, as described before (38,41). Mice were placed in the center of the area and tested for 20 min following either vehicle or cannabinoid injection. The location of the mouse in the box was determined by infrared beam tracking (Med Associates, VT, USA), as described for light aversion. The time spent in the center area was assessed per mouse and reported as mean ± SEM of all the mice for each group.

Y maze assay

Spatial memory was measured using a three-arm Y maze assay (ANY-maze, Wood Dale, IL, USA). The test consisted of two trials separated by an intertrial period. There was no acclimatization to the maze. During the first trial, mice explored the two opened arms for 5 min while the third arm (novel arm) of the Y maze was closed. After a 2-min intertrial period in their home cage, the second trial allowed mice to explore all three arms. Mice with good spatial memory are expected to enter the previously unexplored (novel) third arm more frequently, whereas those with poor spatial memory are expected to show no preference. The number of entries into the novel arm during the second trial was assessed for an individual mouse and reported as mean ± SEM for each of treatment groups.

Rotarod assay

Impaired motor function was measured using a rotating rod suspended horizontally at a fixed height. Each animal is trained for two days and tested on the third day. The assay consisted of three trials lasting for 5 min each, separated by 10 min of resting time. On the day of testing, a mouse is placed on the rod at an initial speed of four rotations per minute (RPM). Over the next 5 min, rotation speed is gradually increased to 60 RPM. The time (in seconds) the animal can remain on the rod before falling off is recorded as a measure of motor coordination and balance. The total time the mouse stays on the rotarod was automatically reported by the apparatus (Ugo Basile, Gemonio, Italy). The three trials for each mouse were averaged. The data are reported as mean ± SEM for each mouse in a treatment group.

Tail suspension test

The tail suspension test was used to study depressive-like behavior in mice, whereby immobility indicates a state of despair under stress. During the test, the mouse is suspended by its tail using tape above an empty cage for 5 min with the entire process recorded via video, as described previously (46). There is no acclimatization. The time (in seconds) that the animal remains motionless is documented and classified as immobile only if the mouse displayed no movement for at least 2 s. The total duration of immobility was calculated and reported as mean ± SEM of all mice tested in each group.

Surgery for optogenetic cortical spreading depolarization

For CSD susceptibility experiments, optic fibers were positioned on the skull over motor (M1) cortex and recording electrodes into M1 and visual (V1) cortex, as described before (35,47). Briefly, single (M1) or paired bilateral (V1) LFP electrodes were implanted (in mm relative to bregma): 3.5 posterior/2.0 lateral/0.4 depth (V1); −0.5 posterior/2.0 lateral/0.8 depth (M1). Electrodes in cerebellum served as reference and ground. Optic fiber cannulas (400 μm) were placed on the skull over M1 cortex, bilaterally (1.5 mm anterior, 2.0 mm lateral). Post hoc histology verified electrode locations. For CSD induction followed by MGS assessment, both optic fiber and recording electrodes were positioned on the skull to minimize surgical confounds (35,47). Silver ball tip electrodes were placed over right M1, somatosensory (S1) and V1 cortex at 1.0 anterior/1.0 lateral (M1); 1.0 posterior/1.0 lateral (S1); 3.0 posterior/1.0 lateral (V1), and over cerebellum as reference and ground. A fiber optic cannula was placed on the skull over the right V1 cortex (3.5 mm posterior, 2.0 mm lateral). Optic fibers and electrodes were connected by dental cement to a 7-channel pedestal. Carprofen (5 mg/kg, s.c.) was given for post-operative pain relief and mice recovered for at least 1 week before the recordings.

Cortical spreading depolarization recordings

Mice were transferred to a home-built shielded EEG recording system with a counterbalanced swivel allowing free behavior, as described before (36,47). EEG signals were 3X pre-amplified and fed into separate amplifiers for DC-potential (500 Hz, 10X gain) and AC-potential (0.05–500 Hz, 800X gain) recordings, digitized (Power 1401 and Spike2 software, CED, UK) at 1000 Hz (DC-potential) or 5000 Hz (AC-potential) sampling rate of. Differential signals were used to detect multi-unit activity (MUA; 500–5000 Hz; 36,000X amplification; 25,000 Hz sampling).

CSD threshold assessment

CSD thresholds (a measure of CSD susceptibility as the correlate of the migraine aura) were assessed as described before (47), after 24 h of baseline EEG recordings. In brief, single 4-mW 460-nm blue light LED pulses of increasing duration were applied to the right M1 at 5-min intervals until CSD was observed. Three baseline threshold measurements were performed at intervals of at least 48 h to ensure stability. Next, the impact of acute pretreatment with a 30:1 or 100:1 CBD:THC combination was tested on separate days (in randomized order) with a day of rest in between CSD inductions. Both 60- and 30-min pretreatments were tested within the same animal, starting with a 60-min pretreatment. CSD events were identified by a DC-potential shift of ≥ 5 mV amplitude showing spread between the two recording electrodes. In case stimulation of the right M1 optic fiber did not evoke a CSD, the left M1 optic fiber was used, and recordings were obtained from the left hemisphere.

Cortical spreading depolarization induction combined with mouse grimace scale scoring

Mouse grimace scale (MGS) scores following CSD were assessed in ChR2/FHM1 or ChR2/WT mice, as described before (35). In brief, while connected to the EEG system, the mouse was placed in a glass jar that allowed free behavior, in front of a HD camera with 1920 × 1080 resolution. Video recordings were obtained before and after suprathreshold CSD induction. After a 10-min habituation period, three consecutive CSDs, separated by 5 min, were induced in V1 using a suprathreshold 30-s pulse of 460-nm blue light of 4-mW intensity. If no CSD occurred, a higher intensity was used until three CSD events were induced within a 20-min period. Videos were recorded at 25 frames/second and MGS scores were analyzed post hoc for the 10 min before the first CSD (baseline) as well as the 30- to 40-min window (30-min time point) after. CBD:THC effects were assessed at the 30-min time point, as CSD increased the MGS score at this time in both ChR2/WT and ChR2/FHM1 mutant mice (35). Each video was examined in its entirety, and MGS was evaluated from five facial action units (FAUs); orbital tightening, ear position, nose bulge, cheek bulge, and whisker positionon a three-point scale for their presence and intensity: (a value of 0 (not present), 1 (moderately visible) or 2 (severe)) (48). An MGS score was obtained for each mouse for each time point by averaging the five FAUs. The post hoc observer was blinded to genotype and experimental groups to prevent bias.

Wire grip testing after cortical spreading depolarization

In separate experiments, CSD-related motor grip function was tested. To assess grip performance, before, during and after optogenetic CSD induction, mice were allowed to grab onto a horizontal wire situated 1 cm above the floor of the electrophysiology recording cage, as described before (36,47). In brief, lifting the mouse was repeated every 10 s, starting 2 min before CSD induction, continuing for at least 12 min after the last CSD. Videos were analyzed by an observer blinded to genotype and treatment. Successful wire grabbing was scored for both forepaws at 0.5-s precision to obtain the cumulative grab duration over a 10-s bin. If a mouse grabbed the wire for a subset of the 0.5-s time window, that bin was scored as ‘grabbing’. Simultaneously recorded DC-potential was used to determine the time point of a CSD reaching the M1 cortical electrode.

Four-limb hanging test

The four-limb hanging mesh test was employed to assess effects of cannabinoid treatment on muscle endurance. After a habituation period consisting of five tests performed on separate days with at least two days in between, a baseline test was performed in the absence of treatment, followed by three experimental tests with at least two days of rest in between sessions. During each session, the hanging mesh test was performed under randomly assigned conditions (i.e., vehicle, 30:1 or 100:1 ratios of CBD:THC), administered 30 min prior to testing. On the day of the test, mice were brought into the experimental room 30 min prior to the test for acclimation. Then, mice underwent the four-limb hanging test individually, as described before (49), following a fixed sequential order (i.e., every test day the same mouse was tested first, second, third, etc.). The grid lid of a Makrolon Type II cage served as ‘mesh’, and was set tightly to a shelter at a height of 100 cm above an opened home cage filled with bedding. The mouse is positioned with its four paws on the grid, which is then inverted. The duration for which a mouse hangs on the inverted mesh is recorded, with a maximum time limit of 600 s. If the mouse falls before reaching the time limit, up to two additional attempts are allowed, each with a 120-s break in between. If the mouse jumps off the grid, it is promptly placed back on the grid without stopping the timer. If a second jump occurs, the timer is halted, and the attempt is deemed unsuccessful. Following a failed attempt, the mouse is immediately placed on the grid for a new attempt, with a maximum of six attempts per mouse. The researcher conducting the test was blinded to the mouse genotype and experimental condition. Muscular endurance was evaluated by calculating the total duration (in seconds) that a mouse could hang on the mesh multiplied with the body weight, yielding the so-called ‘holding impulse’.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software Inc, CA, USA). Given the inherent variability of mouse behavior, measurements for all behavioral tests after injections of CGRP, SNP or PBS (i.e., the light aversion assay, the squint assay, the tactile hypersensitivity test, the open field test, the Y-maze assay, the rotarod assay, and the tail suspension test), were done with at least two independent cohorts of CD1 mice. The effect sizes of experiments involving CD1 mice were based on comparable data from previous studies in the same laboratory (38,40–43). With an α of 0.05 and power at 0.80, the sample size suitable for the desired effect size was estimated at 16 per group in the light aversion assay, 13 per group in the von Frey assays, 12 per group in the squint assay, 16 per group in the adverse event assays. The effect sizes of CSD threshold and CSD-related MGS score experiments involving FHM1 mutant mice (and WT littermates) as well as ChR2-expressing FHM1 mutant mice (and WT littermates) – providing the primary readouts for effects of cannabinoids on CSD features and CSD-related head pain mimics - were based on comparable data from previous studies (35,36). With an α of 0.05 and power at 0.80, the sample size suitable for the desired effect size was estimated at five per group, which was generally used as a minimum group size for the experiments in FHM1 mutant and WT mice. Exceptions are (i) experiments assessing possible adverse effects of cannabinoids on motor dysfunction following CSD (forepaw grab duration data), and (ii) some of the 24- and 48-h time points of the time course data of post-CSD MGS scores that provided secondary readouts on possible effects of cannabinoids on longer-term recovery after CSD (data shown in Online Supplementary Figure 4), for which cases groups sizes of n = 4 were used. Prior to statistical analysis, the Shapiro–Wilk and Anderson-Darling normality tests were applied, which revealed a normal or non-normal data distribution. Based on whether the data were paired and distributed, data were compared differently. One-way ANOVA, repeated measures (RM) two-way ANOVA or Mixed-effects model (REML) followed by Tukey’s multiple comparisons test were used for the experiments where CGRP, SNP or PBS were injected in CD1 mice. For measures of the CSD experiments in FHM1 mutant mice and WT littermates (including ChR2-expressing mice), the Kruskal-Wallis with Dunnett’s multiple comparisons test was performed. Data are presented as individual values and mean ± SEM. A p-value of < 0.05 was considered significant. Significant findings are indicated by one (* = p-value 0.01–0.05), two (** = p-value 0.001–0.01), three (*** = p-value < 0.001) or four (**** = p-value < 0.0001) asterisks. Results of all statistical analyses are available in Online Supplementary Tables 1–14.

Results

Combined CBD:THC reduced CGRP- and SNP-induced light aversion

The experimental paradigm to assess the efficacy of CBD, THC, or combinations of these compounds, on light aversion is shown in Figure 2(a). CD1 mice were pretreated with cannabinoids or sunflower oil (vehicle (Veh)) 30 min prior to an i.p. injection of CGRP (0.1 mg/kg), SNP (2.5 mg/kg) or PBS. Thirty minutes later (so 60 min after the first injection), the mice were monitored using a validated light-dark assay, as a measure of photophobia, for a 30-min testing period (38,41). In the vehicle group, injection of CGRP (****p < 0.0001; Figure 2(b)) or SNP (**p = 0.0062; Figure 2(c)) led to a significant reduction in the time spent in the light chamber compared to PBS-injected mice, indicative of light aversive behavior. However, mice that had been pretreated with a 100:1 CBD:THC (100 mg/kg CBD and 1 mg/kg THC) combination exhibited a complete rescue from light aversion induced by either CGRP (Figure 2(b)) or SNP (Figure 2(c)). The rescue was most evident over the first 20 min of the testing period for both CGRP (Online Supplementary Figure 1a), and SNP (Online Supplementary Figure 1b). The steady decline in time spent in the light seen in vehicle-treated mice as well as in most other groups has been noted before and most likely reflects reduced exploratory drive as the mice become more accustomed to the testing chamber (38,42). Conversely, when CBD or THC were administered alone, CGRP- or SNP-induced light aversion was not ameliorated. Of note, in CGRP-induced light aversion, a 1:1 CBD:THC combination did not alleviate light aversion whereas a ratio of 30:1 yielded a partial rescue, with the average time spent in the light being not significantly different from the negative control (PBS) or the positive control (CGRP). Because we were not powered to visualize partial effects of drugs, and to minimize the number of animals used in this study, the 30:1 CBD:THC combination was not tested in SNP-injected mice. For the CGRP-injected mice, there were no sex differences (see open and closed symbols in the figure for males and females respectively, and statistics in Online Supplementary Table 1). There was also no difference between sexes over time in the longitudinal analysis (Online Supplementary Figure 1a, Online Supplementary Table 2). For the SNP-injected mice, while we were not powered to look at sex differences, we did not notice any trend towards sexual dimorphism in those results (Figure 2(c)), including the longitudinal analysis (Online Supplementary Figure 1b, Online Supplementary Table 2).

Figure 2.

A 100:1 CBD:THC combination effectively reduced both CGRP- and SNP-induced light aversion in CD1 mice. (a) Experimental design of the light aversion assay. (b) CGRP treatment. CD1 mice were pretreated with sunflower oil as vehicle (Veh) or cannabinoids at the indicated doses (mg/kg) and then injected with CGRP or PBS vehicle. Time spent in the light zone of individual mice is shown as the mean (± SEM) per 5-min interval. Mice given CGRP + 100:1 CBD:THC spent significantly more time in light than mice given CGRP + vehicle (*p= 0.0498) and were not significantly different from the negative control group (PBS + Veh) (p = 0.2917). For the 30:1 CBD:THC combination, there was a trend towards more time in the light, but no significant difference from either CGRP + vehicle (p = 0.3291) or from PBS + vehicle (p = 0.0962). One-way ANOVA, Tukey’s multiple comparisons, ****p < 0.0001. Groups were PBS + Veh (male n = 24, female n = 27), CGRP + Veh (male n = 24, female n = 27), CGRP + CBD 100 mg/kg (male n = 10, female n = 10), CGRP + THC 1 mg/kg (male n = 10, female n = 11), CGRP + CBD:THC 1:1 mg/kg (male n = 11, female n = 10), CGRP + CBD:THC 30:1 mg/kg (male n = 15, female n = 18), CGRP + CBD:THC 100:1 mg/kg (male n = 20, female n = 21). (c) SNP treatment. Time spent in the light of CD1 mice injected with SNP or PBS following pretreatment with cannabinoids or sunflower oil, as in panel b. Mice given SNP + 100:1 CBD:THC spent significantly more time in light than mice given SNP + vehicle (***p = 0.0005) and were not significantly different from the negative control group (PBS + Veh) (p = 0.9758). One-way ANOVA, Tukey’s multiple comparisons, ****p < 0.0001. Groups were PBS + Veh (male n = 10, female n = 12), SNP + Veh (male n = 11, female n = 10), SNP + CBD 100 mg/kg (male n = 11, female n = 12), SNP + THC 1 mg/kg (male n = 10, female n = 10), SNP + CBD:THC 1:1 mg/kg (male n = 11, female n = 9), SNP + CBD: THC 100:1 mg/kg (male n = 10, female n = 11). For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Figure 1a and b for longitudinal data over the 30-min testing period for CGRP and SNP treatments, respectively, and Online Supplementary Table 1 for full statistical data.

Combined CBD:THC rescued SNP-induced resting behavior

During the same experiments, we measured the effect of cannabinoid treatment on resting time in the dark and light. Consistent with previous studies, mice spent more time resting in the dark, but not in the light, following injection of CGRP (Figure 3(a); Online Supplementary Figure 2a) and SNP (Figure 3(b); Online Supplementary Figure 2b). For CGRP treatment, this resting behavior was not rescued by pretreatment with any combination of cannabinoids (Figure 3(a)). When data were disaggregated by sex, a rescue of the CGRP-induced resting behavior by the 100:1 CBD: THC combination only occurred in females (see statistics in Online Supplementary Table 3), while it was not rescued in males. There was no effect of cannabinoids on the time spent resting in the light. However, the 100:1 CBD:THC combination was able to rescue the increased resting in the dark induced by SNP (Figure 3(b)), and this effect was similar in males and females (see open and closed symbols in the figure for males and females respectively, and statistics in Online Supplementary Table 3). Sex statistics were not performed as this experiment was not powered to look at sex differences. Neither CBD or THC alone nor a 1:1 CBD:THC combination rescued SNP-induced resting in the dark. As with CGRP, pretreatment with cannabinoids did not affect resting behavior in the light after SNP treatment.

Figure 3.

A 100:1 CBD:THC combination rescued SNP- but not CGRP-induced resting in the dark zone in CD1 mice. Data were collected at the same time as light aversion data from the same CD1 mice shown in Figure 2. (a) CGRP treatment. Resting time (percentage ± SEM) in the dark of individual mice following CGRP treatment. CGRP significantly increased resting time under all conditions except for pretreatment with THC, which was not significantly different from CGRP + vehicle (p = 0.1983) or from PBS + vehicle (p = 0.0842). One-way ANOVA, Tukey’s multiple comparisons, ****p < 0.0001. Resting time (percentage ± SEM) in the light zone of individual mice following CGRP treatment. There was no significant effect by any of the treatments. One-way ANOVA, Tukey’s multiple comparisons, p = 0.1319. Groups were PBS + Veh (male n = 24, female n = 27), CGRP + Veh (male n = 24, female n = 27), CGRP + CBD 100 mg/kg (male n = 10, female n = 10), CGRP + THC 1 mg/kg (male n = 10, female n = 11), CGRP + CBD:THC 1:1 mg/kg (male n = 11, female n = 10), CGRP + CBD:THC 30:1 mg/kg (male n = 15, female n = 18), CGRP + CBD:THC 100:1 mg/kg (male n = 20, female n = 21). (b) SNP treatment. Resting time (percentage ± SEM) in the dark zone of individual mice following SNP treatment. SNP significantly increased resting time under all conditions except for the 100:1 CBD:THC combination, which rescued SNP-induced resting behavior against SNP + vehicle (**p = 0.0033) and PBS + Veh (p = 0.9633). One-way ANOVA, Tukey’s multiple comparisons, ****p < 0.0001. Resting time (percentage ± SEM) in the light zone of individual mice following SNP treatment. There was no significant effect by any of the treatments. One-way ANOVA, Tukey’s multiple comparisons, p = 0.1110. Groups were PBS + Veh (male n = 10, female n = 12), SNP + Veh (male n = 11, female n = 10), SNP + CBD 100 mg/kg (male n = 11, female n = 12), SNP + THC 1 mg/kg (male n = 10, female n = 10), SNP + CBD:THC 1:1 mg/kg (male n = 11, female n = 9), SNP + CBD:THC 100:1 mg/kg (male n = 10, female n = 11). For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Figure 2a and b for longitudinal data over the 30-min testing period for CGRP and SNP treatments, respectively. See Online Supplementary Table 3 for full statistical data.

A 100:1 CBD:THC combination rescued CGRP- and SNP-induced eye squint in male mice only

To assess a non-evoked pain behavior, we utilized an automated squint assessment that quantifies the eye squint aspect of the MGS response (40) (Figure 4(a)). Both male and female mice injected with CGRP or SNP exhibited a reduction in the mean eye pixel area, reflecting squint, in contrast to PBS-injected mice. Pretreatment with a 100:1 CBD:THC combination rescued both the CGRP- and SNP-induced squint, but only in male mice (Figure 4(b) and (d)). For direct comparison of the responses between sexes, the eye size for each mouse was normalized to its baseline size to obtain the change (delta) in eye size upon treatment. By this criteria, the 100:1 CBD:THC combination rescue was significantly greater in male than female mice for CGRP-induced squint (*p = 0.0472; Figure 4(c)), but while trending, did not reach significance for SNP-induced squint (ns p = 0.4074; Figure 4(e)).

Figure 4.

Cannabinoids rescued CGRP- and SNP-induced squint in male but not female CD1 mice. (a) Experimental design of the automated squint assay. (b) CGRP treatment. Average mean pixel area over 5-min (baseline and treatment). Male CD1 mice treated with a 100:1 CBD:THC combination showed a rescue of CGRP-induced squint. 100:1 was not significantly different from the baseline recording paired t-test 100:1 (p = 0.0856 against baseline 100:1 group). One-way ANOVA ****p < 0.0001. Average mean pixel area for female CD1 mice over 5 min. None of the treatment groups rescued CGRP-induced squint. One-way ANOVA ****p < 0.0001. (c) CGRP treatment delta (baseline – treatment) values of male and female CD1 mice squint values. Difference between male and female mice treated with the 100:1 ratio (*p = 0.0472, unpaired t-test). There are no other differences between sexes across all treatment groups. One-way ANOVA ****p < 0.0001. For panels b and c, the groups were PBS + Veh (male n = 19, female n = 16), CGRP + Veh (male n = 18, female n = 18), CGRP + CBD 100 mg/kg (male n = 10, female n = 11), CGRP + THC 1 mg/kg (male n = 12, female n = 12), CGRP + CBD:THC 1:1 mg/kg (male n = 9, female n = 11), CGRP + CBD:THC 30:1 mg/kg (male n = 12, female n = 12), CGRP + CBD:THC 100:1 mg/kg (male n = 15, female n = 16). (d) SNP treatment. Average mean pixel area over 5-min. Male CD1 mice treated with a 100:1 CBD:THC combination showed a rescue of SNP-induced squint. 100:1 was not significantly different from the baseline recording (p = 0.1910 paired t-test). Ordinary one-way ANOVA **p = 0.0030. For female CD1 mice, none of the treatment groups rescued SNP-induced squint. One-way ANOVA, ***p = 0.0004. (e) SNP treatment delta values of male and female CD1 mice squint values. There were no differences between sexes across all treatment groups. For the 100:1 CBD:THC combination, there was a trend towards rescue, but no significant difference between males and females (p = 0.4074 unpaired t-test). One-way ANOVA **p = 0.0026. For panels d and e, the groups were PBS + Veh (male n = 15, female n = 15), SNP + Veh (male n = 14, female n = 14), SNP + CBD:THC 100:1 mg/kg (male n = 15, female n = 14).For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Table 6 for full statistical data.

Combined CBD:THC did not rescue CGRP- or SNP-induced periorbital tactile hypersensitivity

To evaluate the influence of cannabinoid pretreatment on cutaneous tactile hypersensitivity, 50% thresholds using periorbital von Frey hair applications were assessed 30 min after CGRP, SNP or PBS injection (Figure 5(a)). CGRP and SNP induced periorbital hypersensitivity as expected (50), but none of the cannabinoid treatments rescued it (Figure 5(b) and (c)). While there appeared to be a trend towards less hypersensitivity with many of the treatment groups, there was no significant difference from the CGRP or SNP groups (Online Supplementary Table 7), and no sex differences were observed either.

Figure 5.

Cannabinoids did not rescue CGRP- or SNP-induced periorbital tactile hypersensitivity in CD1 mice. (a) Experimental design of the periorbital tactile hypersensitivity assay. (b) Sensitivity to von Frey filaments in CGRP-injected CD1 mice. The positive control CGRP was different from the vehicle group (****p < 0.0001) but there were no significant effects with any of the treatments. One-way ANOVA, Tukey’s multiple comparisons, ****p < 0.0001. Groups were PBS + Veh (male n = 16, female n = 16), CGRP + Veh (male n= 16, female n = 16), CGRP + CBD 100 mg/kg (male n = 15, female n = 14), CGRP + THC 1 mg/kg (male n = 16, female n = 16), CGRP + CBD:THC 1:1 mg/kg (male n = 16, female n = 16), CGRP + CBD:THC 30:1 mg/kg (male n = 16, female n = 15), CGRP + CBD: THC 100:1 mg/kg (male n = 16, female n = 17). (c) Sensitivity to von Frey filaments in SNP-injected CD1 mice. There was no significant effect by any of the treatments and the positive control SNP was different from the vehicle group (****p < 0.0001). One-way ANOVA, Tukey’s multiple comparisons, ****p < 0.0001. Groups were PBS + Veh (male n = 17, female n = 16), SNP + Veh (male n = 15, female n = 16), SNP + CBD 100 mg/kg (male n = 15, female n = 16), SNP + THC 1 mg/kg (male n = 14, female n = 16), SNP + CBD:THC 1:1 mg/kg (male n = 14, female n = 17), SNP + CBD:THC 100:1 mg/kg (male n = 15, female n = 17). For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Table 7 for full statistical data.

Lack of adverse effects of cannabinoid treatment

To address concerns regarding potential adverse effects of the cannabinoid treatments, light-dark assay in which light aversion was not induced (no administration of CGRP or SNP, therefore measuring anxiety-like behavior), open field (anxiety-like behavior), Y-maze (memory impairment), rotarod (motor dysfunction) and tail suspension (depression-like symptoms) tests were performed with cannabinoids administered in the absence of CGRP or SNP. None of the cannabinoid treatments affected the behavioral outcomes compared to vehicle (Figure 6(a)–(e)). As a positive control, diazepam treatment caused a significant decrease in the time to fall from the rotarod (**p = 0.0093; Figure 6(d)) and increased immobility in the tail suspension test (**** p < 0.0001; Figure 6(e)) when compared to vehicle. Hence, none of the cannabinoid treatments resulted in adverse effects on the tested behaviors in either sex.

Figure 6.

Cannabinoids did not induce adverse effects in CD1 mice. (a) Average time spent in the light during the light aversion assay in CD1 mice. None of the cannabinoid treatments changed the time spent in the light compared to vehicle. One-way ANOVA, Tukey’s multiple comparisons, p = 0.2732. Groups were Veh (male n = 11 (2 excl.), female n = 11), CBD 100 mg/kg (male n = 11, female n = 11), THC 1 mg/kg (male n = 10, female n = 12), CBD:THC 1:1 mg/kg (male n = 12, female n = 10), CBD:THC 100:1 mg/kg (male n = 11, female n = 11). (b) Average time spent in the center during the open field assay. None of the cannabinoid ratios the time spent in the center compared to vehicle. One-way ANOVA, Tukey’s multiple comparisons, p = 0.1764. Groups were Veh (male n = 8, female n = 7), CBD 100 mg/kg (male n = 8, female n = 8), THC 1 mg/kg (male n = 8, female n = 8), CBD:THC 1:1 mg/kg (male n = 8, female n = 9), CBD:THC 30:1 mg/kg (male n = 8, female n = 8), CBD:THC 100:1 mg/kg (male n = 7, female n = 8). (c) Total entries in novel arm of Y-maze assay. None of the cannabinoid ratios changed the number of entries of the novel arm compared to vehicle. One-way ANOVA, Tukey’s multiple comparisons, p = 0.5549. Groups were Veh (male n = 18, female n = 17), CBD 100 mg/kg (male n = 11, female n = 11), THC 1 mg/kg (male n = 9, female n = 11), CBD:THC 1:1 mg/kg (male n = 12, female n = 8), CBD:THC 100:1 mg/kg (male n = 11, female n = 9). (d) Average time to fall in the rotarod assay. None of the cannabinoid ratios changed the average time to fall compared to vehicle. The positive control diazepam (DZP) was significantly different from the vehicle group (**p = 0.0093). One-way ANOVA, Tukey’s multiple comparisons, ***p = 0.0001. Groups were Veh (male n = 17, female n = 15), CBD 100 mg/kg (male n = 11, female n = 11), THC 1 mg/kg (male n = 10, female n = 12), CBD:THC 1:1 mg/kg (male n = 11, female n = 10), CBD:THC 30:1 mg/kg (male n = 8, female n = 8), CBD:THC 100:1 mg/kg (male n = 11, female n = 10), DZP (male n = 7, female n = 8). (e) Average immobility time measured in the tail suspension assay. None of the cannabinoid ratios changed the average immobility time compared to vehicle. The positive control diazepam (DZP) was significantly different from the vehicle group (****p < 0.0001). One-way ANOVA, Tukey’s multiple comparisons, ****p < 0.0001. Groups were Veh (male n = 16, female n = 16), CBD 100 mg/kg (male n = 7, female n = 8), THC 1 mg/kg (male n = 8, female n = 8), CBD:THC 1:1 mg/kg (male n = 8, female n = 8), CBD:THC 30:1 mg/kg (male n = 8, female n = 8), CBD:THC 100:1 mg/kg (male n = 8, female n = 8), DZP (male n = 7, female n = 8). For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Table 8 for full statistical data.

Combined CBD:THC ameliorated acute head pain mimics following cortical spreading depolarization in ChR2/Wt mice

To evaluate the impact of combined CBD and THC treatment in the context of migraine with aura, we assessed effects of the 30:1 and 100:1 CBD:THC combinations on the characteristics of optogenetically induced CSD as well as the occurrence of a subsequent head pain mimic (facial signs of discomfort measured using the MGS score (31,44)) following CSD in freely behaving ChR2/WT mice. First, we investigated the effects of pretreatment of CBD:THC ratios or vehicle administered 60 or 30 min prior to initiation of the first CSD (Online Supplementary Figure 3a and b). Following photostimulation at the M1 cortex, CSD events (evidenced as DC shifts spreading from M1 to V1 cortex) were induced in the presence of vehicle as well as 30:1 and 100:1 CBD: THC combinations (Online Supplementary Figure 3c). Neither the threshold (Online Supplementary Figure 3d) nor the propagation rate (Online Supplementary Figure 3e) of CSD differed between vehicle and the 30:1 and 100:1 CBD:THC combinations, no matter whether cannabinoid treatment was given 60 or 30 min before inducing CSD events. Also, the half-width and amplitude of CSD was not altered by pretreatment with cannabinoids (data not shown).

Next, we assessed whether cannabinoid pretreatment affected MGS scores (as a head pain mimic) following CSD induction in a separate group of ChR2/WT mice in which photostimulation was performed through the skull at the V1 cortex and CSD events were recorded by minimally invasive skull-based DC-recordings at the S1 and M1 cortex (Figure 7(a) and (b)). Photostimulation resulted in three CSD events within a 10-min window in mice of the vehicle and cannabinoid treatment groups (examples are shown in Figure 7(c)). MGS scores, obtained 30 min after the first CSD event in the vehicle group were increased compared to baseline (assessed 30 min before CSD induction) (*p = 0.0313; Online Supplementary Figure 4a), in line with earlier findings in ChR2/WT mice (35). Pretreatment with a 100:1 CBD:THC combination showed reduced MGS scores, compared to vehicle, at 30 min after CSD (*p = 0.0158; Figure 7(d)) and compared to the level at baseline (Online Supplementary Figure 4c). Pretreatment with a 30:1 CBD:THC combination showed a non-significant trend towards reduced MGS scores, again compared to vehicle, at 30 min after CSD (Figure 7(d), see statistics in Online Supplementary Table 9). MGS scores reached baseline values within 24 h after CSD induction (Online Supplementary Figure 4). The combined results suggest that in WT mice, in particular, a 100:1 CBD:THC combination, can reduce head pain mimics following CSD without impacting basic characteristics of CSD.

Figure 7.

A 100:1 CBD:THC combination ameliorated acute head pain mimics following optogenetically induced CSD in freely behaving ChR2/WT mice. (a) Schematic representation of the locations of the recording electrodes and optic fiber used for light stimulation. CSDs were induced in the visual cortex (V1) in a minimally invasive manner by optogenetic (LED pulse) stimulation through the intact skull. Three silver ball-tip electrodes were connected to the skull overlaying the visual (V1), somatosensory (S1) and motor (M1) cortex for direct current (DC-potentials) recording of the CSD-related DC shifts. (b) Experimental design of mouse grimace scale (MGS) scoring following optogenetic induction of three CSD events. Post-CSD MGS scores were assessed using a longitudinal design, organized into four test days with at least 48 h of resting time in between. The first test day served as control without cannabinoid pretreatment. The remaining test days were randomly assigned between vehicle, 30:1 or 100:1 CBD:THC pretreatment. Injections (i.p.) were performed at 30 min prior to the start of the first CSD induction. On each test day, after a 30-min baseline recording to ensure stability of DC signals and allow for acclimatization of the mouse in the video-setup for MGS scoring, three CSDs were induced optogenetically within a 10-min timeframe. Behavioral monitoring was performed 10 min before CSD (as a baseline) and after CSD, i.e., at 30 min (1), 24 h (2) and 48 h (3) after the last CSD. (c) Representative DC-potential shifts during optogenetic induction of three CSDs in a freely behaving ChR2/WT mouse recorded 30-min after vehicle, 30:1 and 100:1 CBD:THC pretreatment. (d) MGS scores were assessed 30 min after CSD induction and presented as % calculated of pre-CSD baseline values (taken as 100%). For the raw values and time series data including the 24- and 48-h MGS scores see Online Supplementary Figure 4. Pretreatment with the 100:1 CBD:THC combination reduced MGS scores at the 30-min time post-CSD point in comparison to vehicle (Veh) pretreatment (*p = 0.0158). Effects of pretreatment with a 30:1 CBD:THC combination were less pronounced and not different against vehicle (p = 0.2542). Kruskal-Wallis test, Dunnett’s multiple comparisons, *p = 0.0123. Groups of mixed males and females were Veh (n = 6), CBD:THC 30:1 mg/kg (n = 6), CBD:THC 100:1 mg/kg (n = 6). For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Table 9 for full statistical data.

Combined CBD:THC ameliorated acute head pain mimics following cortical spreading depolarization in ChR2/FHM1 mutant mice

We assessed whether pretreatment with cannabinoids affected CSD characteristics and subsequent head pain mimics in ChR2/FHM1 mutant mice (31). To this end, we first examined effects on CSD characteristics of pretreatment of 30:1 or 100:1 as CBD:THC combinations administered 60 or 30 min before CSD induction in mice of this strain (Online Supplementary Figure 5a). For ChR2/FHM1 mutant mice, the threshold and the propagation rate of CSD were not changed by pre-administration of cannabinoids (Online Supplementary Figure 5b and c). The half-width and amplitude of CSD were also not altered (data not shown).

Next, in separate groups of ChR2/FHM1 mutant mice, photostimulation was performed through the skull at the V1 cortex to induce three CSD events for assessment of acute post-CSD head pain mimics, as described for ChR2/WT mice. Examples of CSD events in mice of the vehicle and cannabinoid treatment groups are shown in Figure 8(a). MGS scores at 30 min after the first CSD in the vehicle-injected group were increased compared to baseline (assessed 30 min before CSD induction) (*p = 0.0313; Online Supplementary Figure 4a). Pretreatment with a 100:1 CBD:THC combination administered 30 min before CSD induction showed reduced MGS scores, compared to vehicle, at 30 min after CSD (*p = 0.0103; Figure 8(b)) and compared to the level at baseline (Online Supplementary Figure 4(c)). Pretreatment with a 30:1 CBD:THC combination caused a trend towards reduction in MGS scores at 30 min after the first CSD (Figure 8(b); see statistics in Online Supplementary Table 9). Of note, the reduction in MGS scores in the presence of cannabinoid treatment, calculated as % of the mean MGS scores following vehicle treatment, was similar between ChR2/FHM1 mutant and ChR2/WT mice as for the 30:1 (p = 0.14; Mann-Whitney U test) as well as for the 100:1 (p = 0.14; Mann-Whitney U test) CBD:THC treatment groups.

Figure 8.

A 100:1 CBD:THC combination ameliorated acute head pain mimics following optogenetically induced CSD in freely behaving ChR2/FHM1 mutant mice. (a, b) Configurations of DC-potential electrodes, optic fiber placements for stimulation (using LED pulses), and experimental design for post-CSD MGS assessment were similar to those used for ChR2/WT mice (and described in Figure 7). (c) Representative DC-potential shifts during optogenetic induction of three CSDs in a freely behaving ChR2/FHM1 mutant mouse recorded 30 min after vehicle, 30:1 and 100:1 CBD:THC pretreatment. (d) MGS scores were assessed 30 min after CSD induction and presented as percentage of pre-CSD baseline values (taken as 100%). For the raw values and time series data including the 24- and 48-h MGS scores, see Online Supplementary Figure 4 a–c. Pretreatment with a 100:1 CBD:THC combination reduced MGS scores at the 30-min post-CSD time point in comparison to vehicle (Veh) pretreatment (*p = 0.0103). Effects of pretreatment with a 30:1 CBD:THC combination was less pronounced and only showed a trend-significant reduction compared to vehicle pretreatment (p = 0.06). Kruskal-Wallis test, Dunnett’s multiple comparisons, **p = 0.0037. Groups of mixed male and female males were Veh (n = 7), CBD:THC 30:1 mg/kg (n = 7), CBD:THC 100:1 mg/kg (n = 7). For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Table 9 for full statistical data.

Combined CBD:THC did not show effects on motor function in WT and FHM1 mutant mice

Next, we assessed whether cannabinoid pretreatment affected motor function following CSD, which is most relevant for ChR2/FHM1 mutant mice given that the gene mutation causes hemiplegia in patients. To this end, in mice in which CSD events were induced, a wire grip test was performed, as described before (36). In both ChR2/WT and ChR2/FHM1 mutant mice, induction of CSD reduced ability to grab the rod by the left forepaw, i.e., the paw contralateral to right hemisphere in which CSD was induced (Figure 9(a)). This coincided with the moment that the CSD wave reached the left M1 cortical electrode and lasted for ~4–5 min, in line with earlier findings (47). Pre-administration of cannabinoids did not show aggravated CSD-related impairment of grip function of the left forepaw, regardless of the genotype (Figure 9(b) and (c)). Finally, we assessed whether cannabinoid pretreatment affected muscular endurance, as measured by the four-limb hanging time (holding impulse) in mice that did not express ChR2 (Online Supplementary Figure 6a). The results showed no difference for the vehicle and cannabinoid groups in WT (Online Supplementary Figure 6b) and in FHM1 mutant (Online Supplementary Figure 6c) mice.

Figure 9.

Ratios of CBD:THC did not aggravate CSD-related transient motor dysfunction in FHM1 mutant and C57BL/6 mice. (a) Grab duration of the left and right forepaws, as assessed by wire grip testing, was assessed following optogenetic CSD induction in V1 cortex. In ChR2/WT and ChR2/FHM1 mutant mice, when a CSD reached the M1 cortex, a transient impairment of left forepaw motor function was observed that was consistent across pretreatments with vehicle, 30:1 CBD:THC or 100:1 CBD:THC. The grey area in the graphs spanning from 5 to 6 min after the end of the M1 CSD wave represents the time period used for comparison across the different treatment groups. (b) In ChR2/WT mice, none of the cannabinoid treatments aggravated motor function of the left forepaw in relation to the right forepaw against vehicle (Veh): 30:1 CBD:THC (p = 0.2264) and at 100:1 CBD:THC (p = 0.4156). Kruskal-Wallis test, Dunnett’s multiple comparisons, p = 0.1657. Groups of mixed male and female mice were Veh (n = 4), CBD:THC 30:1 mg/kg (n = 4), CBD:THC 100:1 mg/kg (n = 4). (c) In ChR2/FHM1 mutant mice, none of the cannabinoid treatments aggravated motor function of the left forepaw in relation to the right fore paw against vehicle (Veh): 30:1 CBD:THC (p > 0.9999) and 100:1 CBD:THC (p = 0.0581). Kruskal-Wallis test, Dunnett’s multiple comparisons, *p = 0.0398. Groups of mixed male and female mice were Veh (n = 4), CBD:THC 30:1 mg/kg (n = 4), CBD:THC 100:1 mg/kg (n = 4). For all panels, open and closed symbols represent males and females, respectively. See Online Supplementary Table 13 for full statistical data.

Discussion

This study investigated the effects of the cannabinoids CBD and THC on various migraine-like symptoms in three mouse models and addressed whether cannabinoids had the potential to prevent the symptoms. In the light aversion assay, pretreatment with a 100:1 CBD:THC combination, but not individual CBD or THC treatments or other CBD:THC combinations, rescued the decreased time spent in the light (light aversion) caused by CGRP and SNP. Similarly, the 100:1 CBD:THC combination rescued SNP-induced resting in the dark in both males and females, and CGRP-induced resting in the dark in females. Unexpectedly, there was no rescue of CGRP-induced resting in the dark in males. The reason for this apparent discrepancy is not known. In the spontaneous pain assay (squint), we found that only male CD1 mice pretreated with a 100:1 CBD:THC combination showed rescue from CGRP- and SNP-induced squint compared to baseline or vehicle treatments. The observed sex differences across two assays highlights that the circuitry controlling resting in the dark and squinting likely involves different pathways. This possible sexual dimorphism deserves further investigation in a study that is powered specifically to look at sex-specific actions of CBD and THC. In the periorbital tactile hypersensitivity assay, none of the cannabinoid treatments rescued CGRP- or SNP-induced responses. Following CSD in ChR2/WT and ChR2/FHM1 mutant mice, 30:1 and 100:1 CBD:THC combinations did not alter the threshold, propagation rate, half-width, or amplitude of CSD in either genotype, regardless of whether treatment was administered 60 or 30 min before CSD induction. However, a 100:1 CBD:THC combination significantly reduced MGS scores 30 min after CSD (hence it reduced the head pain mimic following CSD that had been reported before without CBD:THC administration (31)) in both ChR2/WT and ChR2/FHM1 mutant mice, while the ratio of 30:1 showed a non-significant trend towards reduction. The results obtained with the 100:1 CBD:THC ratio are especially compelling because this combination was able to reverse multiple migraine-like phenotypes (light aversion, eye squint, and grimace) in three models of migraine (CGRP-, SNP- and CSD-induced), shown in two mouse strains with a different genetic background (CD1 and C57BL/6), and in two laboratories blinded to results from the other laboratory.

The use of cannabis to alleviate pain in people has shown promise, however, a recent systematic review of the literature neither supported nor disproved the efficacy and safety of cannabis products for pain (51). With respect to migraine, several clinical trials have yielded promising results (52–54). Specifically, medicinal cannabis has been shown to reduce monthly migraine attacks and migraine severity with mild adverse effects (8,22). A recent clinical trial showed that vaporized cannabis flower containing 6% THC and 11% CBD was able to alleviate symptoms of pain, nausea, phonophobia, and photophobia in migraine patients (54). Importantly, this study showed that migraine patients that inhaled cannabis containing both THC and CBD had better pain and photophobia outcomes compared to those in the groups that received THC-dominant (6% THC, 0.03% CBD), CBD-dominant (11% CBD, 0.35% THC), or placebo doses. Thus, as seen in the present preclinical study, a combination of CBD and THC was better able to rescue pain behavior and photophobia. However, the clinical study did not specifically test touch sensitivity, and our preclinical study did not test phonophobia or nausea. In addition, differences most likely reflect different pharmacokinetics in rodents and humans and route of administration with smoke inhalation in humans versus i.p. injections in mice. Beyond pain and migraine, clinical trials have also demonstrated the efficacy and safety of purified CBD (Epidiolex) for treating epilepsy, leading to FDA approval for Lennox-Gastaut syndrome and Dravet syndrome (55,56). Importantly, various clinical trials are ongoing, and more results should soon be available.

Cannabinoids have shown promising results for treating migraine-like symptoms in preclinical studies. Anandamide helps by inhibiting endocannabinoid-hydrolyzing enzymes in the CNS and trigeminal ganglia, potentially reducing pain transmission and cortical excitability in rats (57). Another rat study showed that anandamide was able to inhibit dural blood vessel dilation caused by electrical stimulation, CGRP, capsaicin, and nitric oxide (11). In a recent study, Greco and team showed that after systemic administration, CBD reaches the brain areas involved in migraine pain and modulates migraine-related nociceptive transmission (26). Finally, THC has been shown to suppress features of KCl-induced CSD in rat cortical brain slices (32). In agreement with our results, pretreatment with CBD alone (without THC) was unable to rescue CGRP-induced photophobia in mice (31). However, that study did find that CBD alone was able to alleviate CGRP-induced periorbital tactile hypersensitivity and grimace, and noted several sex differences depending on the experimental paradigm (acute or chronic administration of CGRP, and time of administration of the treatment) (31). The reason for these differences from our results is not known but could be due, in part, to use of different mouse strains (C57BL/6 as opposed to CD1 mice) or testing paradigms. In a different mouse pain model, high concentrations of CBD or THC alone could attenuate mechanical sensitivity caused by paclitaxel, but notably there was a synergistic effect of combining subthreshold doses in a 1:1 CBD:THC combination (58). Unlike our experimental paradigm with a single administration of trigger and cannabinoids, that study used multiple days of trigger and cannabinoid administration. In this regard, it would be interesting to investigate the effectiveness of CBD and THC on chronic migraine conditions induced by repeated CGRP or SNP administration or CSD induction. Additionally, future studies with varying ratios and concentrations of CBD and THC, along with pharmacokinetic measurements should help reveal why the ratio of 100:1 is advantageous in the paradigms used here.

While both clinical and animal studies have shown that a combination of CBD and THC can be advantageous to treatments with one alone (59,60), the basis for this combinatorial effect is not well understood. Both CB1 and CB2 receptors are expressed in migraine-relevant regions of the nervous system (14,61) and meningeal immune cells (62), both of which may play a role in migraine (63). As mentioned earlier, there is a complex interplay between CBD and THC at these receptors. THC has high affinity for CB1 and CB2 receptors (14). CBD activates CB2 receptors as a partial agonist and is a negative allosteric modulator of CB1 receptors and at high concentrations it can compete with THC at the orthosteric binding site (15,64). Thus, CBD can attenuate or modulate the effects of THC (17–19), and a combination would have different effects than CBD or THC on its own. The release of CGRP in the trigeminal ganglia is influenced by the activation of cannabinoid receptors. CB1 receptors, primarily expressed in neuronal somas within the trigeminal ganglia, exhibit colocalization with CGRP and RAMP1, suggesting a regulatory role in CGRP release (12). However, while activation of CB1 may inhibit CGRP release, this effect is mitigated when TRPV1 receptors, which can stimulate CGRP release, are also activated, indicating a complex interaction between these receptor systems (12,65). Within the meninges, the presence of CB1 and CB2 receptors on a variety of immune cells that could modulate trigeminal nerve activity is especially intriguing. In support of this speculation, a study using a multiple sclerosis model showed that combining CBD and THC reduced neuroinflammation measured by cytokine levels and suppression of Th17 and Th1 cells, which was not seen when either of the cannabinoids were administered individually at the same dose (66). However, combining CBD and THC is not always positive; a rat study highlighted that co-administration of CBD and THC caused greater impairments in sustained attention compared to THC alone, emphasizing the need for more studies on the potential interactions between CBD and THC (67). Importantly, CBD and THC do not exclusively act on CB1 and CB2 receptors, but also act on other migraine relevant receptors, including 5-HT1_A and TRPV1 (16,68,69). Thus, the ability of THC and CBD to modulate each other’s effects, as well as to act on other receptors involved in migraine pathophysiology, may explain why their combination results in a greater reduction in migraine symptoms.

Adverse effects and long-term health outcomes of consuming cannabis are a concern (70). In that regard, CBD has a relatively minimal side effect profile compared to that of THC. To address the concerns, we investigated potential adverse effects of cannabinoids using the light-dark, tail suspension, rotarod, open field, and Y-maze assays. None of the cannabinoid ratios tested produced any adverse effects in these assays. Additionally, neither the 30:1 nor the 100:1 CBD ratio affected motor function or muscular endurance following CSD, in either FHM1 mutant or WT mice. Our results are similar to a study done on mice, which reported that i.p. administration with various doses of CBD and THC alone and as a 15:1 CBD:THC combination produced no adverse effects in the elevated plus maze (71). The study also reported that only higher doses of THC at 3.2 mg/kg and 6.4 mg/kg produced anxiolytic effects. Similarly, in another study, CBD did not lead to any significant effects on locomotor-exploratory-anxiety-like behaviors induced by acute nitroglycerin administration following pretreatment with CBD in rats (26). The combined findings suggest that certain cannabinoid ratios may be effective in treating some of the most bothersome migraine symptoms (light sensitivity and non-evoked pain) without producing adverse effects, providing a promising avenue for further research.

Conclusion

Our preclinical findings suggest that cannabinoids may have therapeutic potential for treating migraine symptoms without causing adverse effects. These findings are in line with previous studies that have suggested that cannabinoids may be effective in treating pain and migraine. Further studies using combinations of CBD and THC are needed to determine their clinical significance.

Supplementary Material

Supplemental material

Supplemental material for this article is available online.

Article highlights.

• A 100:1 CBD:THC combination alleviated light aversion and eye squint induced by CGRP or SNP injection in mice.

• A 100:1 CBD:THC combination reduced mouse grimace scale scores in WT and FHM1 mutant mice following CSD.

• No adverse effects of cannabinoids were observed in behavioral testing.

• Combined CBD and THC shows potential for migraine therapy.

Data availability

The datasets used or analyzed during the current study are available from the corresponding author upon reasonable request.