Transdermal cannabidiol reduces inflammation and pain-related behaviors in a rat model of arthritis.


Background— Current arthritis treatments often have side effects attributable to active compounds as well as the route of administration. Cannabidiol (CBD) attenuates inflammation and pain without side effects, but CBD is hydrophobic and has poor oral bioavailability. Topical drug application avoids gastrointestinal administration, and first-pass metabolism, providing more constant plasma levels.

Methods— This study examined the efficacy of transdermal CBD for a reduction in inflammation and pain, assessing any adverse effects in a rat complete Freund’s adjuvant-induced monoarthritis knee joint model. CBD gels (0.6, 3.1, 6.2 or 62.3 mg/day) were applied for 4 consecutive days after arthritis induction. Joint circumference and immune cell invasion in histological sections were measured to indicate the level of inflammation. Paw withdrawal latency (PWL) in response to noxious heat stimulation determined nociceptive sensitization, and exploratory behavior ascertained the animal’s activity level.

Results— Measurement of plasma CBD concentration provided by transdermal absorption revealed linearity with 0.6–6.2 mg/day doses. Transdermal CBD gel significantly reduced joint swelling, limb posture scores as a rating of spontaneous pain, immune cell infiltration, and thickening of the synovial membrane in a dose-dependent manner. PWL recovered to near baseline level. Immunohistochemical analysis of spinal cord (CGRP, OX42) and dorsal root ganglia (TNFα) revealed dose-dependent reductions of pro-inflammatory biomarkers. Results showed that 6.2 and 62 mg/day were effective doses. Exploratory behavior was not altered by CBD, indicating a limited effect on higher brain function. Conclusions—These data indicate that topical CBD application has therapeutic potential for relief of arthritis pain-related behaviors and inflammation without evident side effects.


Almost 50 million (22.2%) adult Americans over 18 were diagnosed with arthritis in 2007– 2009, most prominently osteoarthritis and the autoimmune disease rheumatoid arthritis. A projected increase to 67 million is anticipated by 2030 (Centers for Disease Control and Prevention (CDC), 2010). The most effective treatment for rheumatoid arthritis is injectable fusion-proteins which sequester the most prominent pro-inflammatory cytokine tumor necrosis factor α (TNFα). These chimeric antibodies may halt the progression of the disease, but side effects include immune suppression (Crawford and Curtis, 2008; Furst, 2010; Hastings et al., 2010). The neurogenic drive also contributes to the severity of arthritic inflammation (Sluka et al., 1994) and may contribute to its reoccurrence.

Cannabinoids and cannabinoid receptors are potential targets for reducing pain and inflammation (Clayton et al., 2002; Richardson et al., 2008; Zuardi, 2008). Cannabis sativa contains approximately 80 different cannabinoids, of which Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD) are primaries (Mechoulam and Shvo, 1963; Mechoulam, 1970; Turner et al., 1980). These compounds are chemically similar to endogenous endocannabinoid lipid derivatives, including anandamide (arachidonoylethanolamide) and 2-arachidonoylglycerol. At present, THC and CBD are available combined in the oral spray Sativex® (GW Pharmaceuticals plc, Cambridge, UK), which is prescribed to adult patients for neuropathic pain. Psychoactive THC side-effects are experienced, however, and long-term use of cannabis Sativa has been shown to increase the risk of developing psychosis and schizophrenia (Berman et al., 2004; Malone et al., 2010). Research is ongoing to find better, effective cannabinoids and better routes of application.

CBD, while structurally similar to THC, is a non-psychoactive cannabinoid with therapeutic potential for the treatment of neuropathic pain, cancer pain, multiple sclerosis, and inflammation (Mechoulam and Hanus, 2002; Mechoulam et al., 2002; Burstein and Zurier, 2009). The oral bioavailability of CBD is very limited due to first-pass metabolism during digestion (Mechoulam and Hanus, 2002). The THC-like cannabinoids act at CB1 or CB2 receptors, whereas CBD-like cannabinoids have a little binding affinity, leaving their role in inhibition incompletely understood. Data suggest in vitro application of CBD inhibits signaling through GPR55 and TRP channel superfamily members, and in vivo oral administration is dose-dependently reducing pro-inflammatory cytokine release (Coffey et al., 1996; Malfait et al., 2000; Akopian et al., 2008; Whyte et al., 2009). More recently, CBD was successfully delivered transdermally in different species for anti-inflammatory activity (Lodzki et al., 2003; Stinchcomb et al., 2004; Paudel et al., 2010). In this study, in vivo efficacy of transdermal CBD delivery to reduce inflammation and pain-related behaviors are tested in a rat adjuvant-induced monoarthritis model with both inflammatory and neurogenic properties (Sluka et al., 1994). Nociceptive behavior, potential adverse side effects, and inflammation-associated anatomical changes in the knee joint and neuronal tissue were assessed.

Materials and methods

2.1 Animals

All animal procedures were approved by the University of Kentucky IACUC committee and were conducted according to guidelines for the ethical treatment of experimental animals published by the Internal Association for the Study of Pain. All experiments were conducted using 260–280 g male Sprague–Dawley rats (Harlan Laboratories, Indianapolis, IN, USA) housed in individual cages on a 12 h/12 h dark/light reversed cycle and allowed access to food and water ad libitum. A total of 54 rats were used in the experiments described here, of which 21 were used as naive controls, and 23 were subjected to adjuvant-induced arthritis. To induce monoarthritis, animals were anesthetized (2–4% isoflurane) and one knee joint injected with 100 μL complete Freud’s adjuvant (CFA) (DIFCO Laboratories, Detroit, MI, USA) emulsion (2 mg/mL diluted in 1:1 normal saline: peanut oil). Rats were returned to their home cages and monitored daily. Joint circumference and pain-related behaviors were assessed prior to CFA injection and daily beginning on day three after CFA (days 3–7).

2.2 Gel preparation

All gels, including vehicle controls, were prepared by weighing the desired amount of CBD (gift from NIDA) and dissolving it in ethanol (72.5% w/w). Once dissolved, nanopore water (Barnstead NANOpure® Diamond™ ultrapure filtration system, Dubuque, IA, USA) was added, followed by isopropyl myristate (Fisher Scientific, Fairlawn, NJ, USA). Carbopol® 980 polymer (Noveon Inc., Cleveland, OH, USA) was added (0.9% w/w), and the solution was sonicated for 10 min to ensure complete incorporation of the Carbopol® 980. Polymerization of Carbopol ® 980 to form the hydroalcoholic gel was initiated by adding sodium hydroxide (0.1 N). Gels were then sonicated for 10 min, loaded into 1 mL syringes, and sealed. Gels made just prior to the initial dosing were used for the entire week since no degradation was observed, and plasma CBD concentration remained constant.

2.3 Assessment of joint inflammation

Hindlimbs were fully extended while the circumference of the knee joint was determined using a flexible tape measure wrapped around the center of the joint. Skin temperature over the joint was measured at the patella with an infrared temperature probe (Fluke, Wilmington, North Carolina, USA). Measurements were taken on days 0, 3, and 7.

2.4 Behavioural assays

2.4.1 Spontaneous pain rating— As a measure of spontaneous pain, limb posture was scored daily in the morning while animals were in their home cages by a scientist blinded to the animal’s treatment. A subjective pain-related behavioral scale was used (Sluka et al., 1993) with 0 – normal; 1 – curling of the toes, 2 – eversion of the paw; 3 – partial weight bearing; 4 – non-weight bearing and guarding, and 5 – avoidance of any contact with the hindlimb.

2.4.3 Hindpaw thermal hypersensitivity— Hindpaw heat sensitivity was assessed prior to CFA injection as well as daily starting on day three after inflammation, four h after each gel application. Paw withdrawal latency (PWL) was measured as described in our previous study (Zhang et al., 2002). Briefly, rats were acclimated for 20–30 min to individual plastic chambers (10 × 10 × 25 cm) on a glass-top table (2 mm thick). The light beam from a high-intensity projector lamp bulb (Quartzline Lamp, GE, Cleveland, OH, USA) was projected through a five × 10 mm aperture attached to an on/off switch with a digital timer. The maximal cutoff time for paw withdrawal reflex was set at 15 s to avoid skin damage. Both hind paws were tested independently for five trials at 5 min intervals by an examiner blinded to the treatment group. The mean responses ± standard error of the mean (SEM) are reported.

2.4.4 Exploratory activity— Open-field exploratory activities were assessed in a Flexfield Animal Activity System (San Diego Instruments, San Diego, CA, USA) with Photobeam Activity System software coupled to a 486 computer (Hewlett Packard, Palo Alto, CA, USA). Animal behaviors include time spent in total exploratory activity (active time, distance traveled, total beams broken, rearing events, rearing time) and resting was recorded in a 40 × 40 × 40 cm transparent plexiglass box for 45 min prior to and at the end of the experiment (Zhang et al., 2004).

2.5 Gel administration

On day three, after induction of monoarthritis, the back of each animal was shaved, and vehicle gel or gel containing 1 or 10% CBD was applied by rubbing it into the skin for 30 s. The area of gel application and volume used corresponded to the calculated final dose of CBD, i.e., 0.62 mg CBD was dosed by applying 75 μL of 1% CBD gel on 3.5 cm2 of shaved skin of the animals’ backs, 3.1 mg CBD by giving 375 μL of 1% CBD gel on 17.5 cm2, 6.2 mg equaled 750 μL of 1% CBD gel on 35.0 cm2, and 62 mg by rubbing 750 μL of 10% CBD gel onto the skin (Table 1). Joint inflammation and nociceptive behavior were assessed starting four h after gel application. On the final day of gel application, after completion of all behavioral assays, rats were killed by pentobarbital overdose, blood samples were collected for CBD plasma quantification, and transcardially perfused with 4% paraformaldehyde.

2.6 Plasma extraction

Plasma concentration of transdermally absorbed CBD was determined. As described by Paudel et al. (2010), plasma (50 μL) was combined with 500 μL of acetonitrile (ACN) and ethyl acetate (1:1, v/v; VWR, West Chester, PA, USA), vortexed for 30 s and centrifuged at 10,000 × g for 20 min. The supernatant was transferred into a 3 mL silanized glass test tube and evaporated under nitrogen in a 37 °C water bath. It was then reconstituted in 100 μL of ACN, vortexed, and sonicated for 5 min. Samples were then transferred to autosampler vials with silanized low volume inserts and 20 μL injected into the HPLC column for analysis by LC/MS.

2.7 Analytical LC/MS method

The LC/MS system used to analyze samples was comprised of a Waters Alliance 2695 pump, an autosampler, a Micromass ZQ detector, and a 996 photodiode array detector with MassLynx software (Milford, MA, USA). A Symmetry® C18 column (150 × 2.1 mm, five μm) with a Sentry Symmetry® guard column (10 × 2.1 mm, 3.5 μm) was utilized with the LC/MS system. The ZQ detector was used with an electrospray ionization probe set for single ion monitoring for CBD quantification. Analysis was performed in a negative mode for m/z 313 [CBD-H]-(dwell time: 0.3 s). Capillary and cone voltage was set at 35 kV and 40 V, respectively. Source block temperature was set at 120 °C, and desolvation temperature to 250 °C. Nitrogen was used as nebulization (flow rate: 50 l/h) and drying gas (flow rate: 450 l/h). The mobile phase was comprised of 75:25 ACN:2 mM ammonium acetate buffer w/5% ACN and used at a flow rate of 0.25 mL/min resulting in a mean CBD retention time of 5.6– 5.7 min. Standard curves were linear within the range of 2–300 ng/mL, and concentrations of samples were determined by comparison.

2.8 Immunohistochemistry

The spinal cord, dorsal root ganglia (DRG), and knee joint synovial joint capsule membranes were excised after transcardial perfusion, post-fixed with 4% buffered paraformaldehyde, cryoprotected overnight with 30% sucrose, embedded in OCT compound (Electron Microscopy Sciences, Hatfield, PA, USA) and stored at −80 °C. Tissue was sectioned (14– 20 μm) with a cryostat at −20 °C, and sections were collected on microscopic slides (Superfrost Plus, VWR, Radnor, PA) and stained. Synovial joint capsule membranes were stained with hematoxylin and eosin (HandE; Ricca Chemical Company, Arlington, TX, USA) and coverslipped using Permount (Fisher Scientific).

Spinal cord sections were stained using monoclonal mouse anti-OX-42 (1:1,000, CD11b/c; Abcam, Cambridge, MA, USA) and rabbit anti-CGRP (1:2000; Bachem, Torrance, CA, USA) antibodies. DRG sections were immunostained with mouse anti-TNFα (1:1000, #52B83; Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies. Primary antibodies were detected using the appropriate Alexa Fluor conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA). Slides were coverslipped using Vectashield Hard Set mounting medium with DAPI (VectorLabs, Burlingame, CA, USA). Controls included the absence of staining upon omission of the primary antibody and the side-to-side differences between the ipsilateral and contralateral sides as internal controls.

Images of hematoxylin and eosin (H&E) staining were acquired using a Nikon Eclipse E1000 microscope equipped with a Nikon DXM 1200F digital camera and ACT-1 software. Synovial membrane thickness was measured by drawing a perpendicular line from the outer surface to the inner margin of the intimal layer of the synovial membrane in two places within each section and averaging their length (Cunnane et al., 1999).

Immunostaining of fluorescently labeled tissue was visualized using a Nikon Eclipse E1000 microscope equipped with a Cool Snap photometric Camera ES and analyzed offline with MetaMorph software. Camera settings were kept constant within an antibody staining to quantify and compare immunoreactivity. Images of the dorsal horn were analyzed by outlining the substantial gelatinosa, determining and averaging the mean intensity of staining in this region. Background fluorescent intensity was measured outside of this area and subtracted. DRG sections were sampled by drawing regions of interest over five randomly selected sensory neuron somas, measuring the mean intensity and averaging for each animal sampled. Histological analysis was conducted in the four experimental groups: naive (uninjected control animals), CFA + VEH (monoarthritis treated with vehicle gel), CFA + low-dose CBD (monoarthritis treated with 0.62 or 3.1 mg/day CBD) and CFA + high-dose CBD (monoarthritis treated with 6.2 or 62.3 mg/day CBD). Mean intensities were averaged and compared across treatment groups.

2.9 Data analysis

GraphPad Prism version 6.1 for Windows (GraphPad Software, La Jolla, CA, USA) was used for all statistical data analyses. Data are presented as the mean ± standard error except where mentioned. Normally distributed data were analyzed with one-way ANOVA followed by Bonferroni post hoc analysis. Pain rating score data were shown as medians and were compared using Kruskal–Wallis nonparametric analyses with Dunnett’s post hoc test. For statistical analysis of the joint circumference and behavioral results, the naive and naïve rats treated with CBD were combined for comparisons to CFA + VEH (vehicle), CFA + low-dose CBD (0.62 and 3.1 mg/day), and CFA + high-dose CBD (6.2 and 62 mg/day). Values of p ≤ 0.05 were considered significant.