Abstract
Patients with arthritis report using cannabis for pain management, and the major cannabinoid delta-9-tetrahydrocannabinol (Δ9-THC) has anti-inflammatory properties, yet the effects of minor cannabinoids on arthritis are largely unknown. The goal of the present study was to determine the antiarthritic potential of the minor cannabinoid delta-8-tetrahydrocannabinol (Δ8-THC) using the collagen-induced arthritis (CIA) mouse model. Adult male DBA/1J mice were immunized and boosted 21 days later with an emulsion of collagen and complete Freund’s adjuvant. Beginning on the day of the booster, mice were administered twice-daily injections of Δ8-THC (3 or 30 mg/kg), the steroid dexamethasone (2 mg/kg), or vehicle for two weeks. Dorsal-ventral paw thickness and qualitative measures of arthritis were recorded daily, and latency to fall from an inverted grid was measured on alternating days, to determine arthritis severity and functional impairment. On the final day of testing, spontaneous wire-climbing behavior and temperature preference in a thermal gradient ring were measured to assess CIA-depressed behavior. The Δ8-THC treatment (30 mg/kg) reduced paw swelling and qualitative signs of arthritis. Δ8-THC also blocked CIA-depressed climbing and CIA-induced preference for a heated floor without producing locomotor effects but did not affect latency to fall from a wire grid. In alignment with the morphologic and behavioral assessments in vivo, histology revealed that Δ8-THC reduced synovial inflammation, proteoglycan loss and cartilage and bone erosion in the foot joints in a dose-dependent manner. Together, these findings suggest that Δ8-THC not only blocked morphologic changes but also prevented functional loss caused by collagen-induced arthritis.
SIGNIFICANCE STATEMENT Despite increasing use of cannabis products, the potential effects of minor cannabinoids are largely unknown. Here, the minor cannabinoid delta-8-tetrahydrocannabinol blocked the development of experimentally induced arthritis by preventing both pathophysiological as well as functional effects of the disease model. These data support the development of novel cannabinoid treatments for inflammatory arthritis.
Introduction
Inflammatory arthritis is an immune disorder that affects one or more joints, causing pain, swelling, reduced range of motion, and stiffness that frequently worsens with age (Berman et al., 2018; Krenn et al., 2018). The most common type of inflammatory arthritis, rheumatoid arthritis, globally afflicts over 18 million people (roughly the population of New York) (Global Burden of Disease, 2021). Current frontline treatments for inflammatory arthritis include disease-modifying antirheumatic drugs (DMARDs), glucocorticoids, and nonsteroidal anti-inflammatory drugs. While these treatments may manage symptoms and slow disease progression in some patients, their efficacy is limited, and adverse effects are increased with prolonged use. For instance, long-term use of glucocorticoids can induce hyperglycemia, osteoporosis, and myopathy (McMahon et al., 1988; Schakman et al., 2008; Wang et al., 2020), whereas chronic nonsteroidal anti-inflammatory drug use increases risk of thrombotic events and gastrointestinal ulceration or bleeding (Fosbøl et al., 2010; Sostres et al., 2013). Additionally, the immunosuppressant properties of biologic DMARDs help control inflammation, but also increase susceptibility to viral and bacterial infection (Singh et al., 2015). Therefore, alternative and complementary treatments for inflammatory arthritis are needed.
Cannabinoids are used increasingly as alternative treatments for arthritis, in part due to the rapidly shifting, although not empirically grounded, public perception regarding the therapeutic effects of cannabis-based products. While there is a dearth of clinical data to ascertain the efficacy of cannabinoid drugs (Sarzi-Puttini et al., 2019), some reports have highlighted their anti-arthritic potential. Namely, one clinical trial found that patients who were given Sativex, a 1:1 combination of the major cannabinoids delta-9-tetrahydrocannabinol (Δ9-THC) and cannabidiol (CBD), exhibited pain relief and significant improvements in their arthritis severity (Blake et al., 2006). Likewise, preclinical data using experimental animal models suggest that cannabinoids may have anti-arthritic properties. For example, orally administered CBD reduced arthritis severity, joint damage, and cytokine release in the mouse collagen-induced arthritis (CIA) model (Malfait et al., 2000). In addition, endocannabinoid-modulating drugs such as JZL184 and URB597 block the development of CIA and resulting pain behaviors in mice (Kinsey et al., 2011; Nass et al., 2021). Similarly, the major and minor cannabinoids Δ9-THC and cannabigerol decrease proinflammatory cytokine release in rheumatoid synovial fibroblasts (Lowin et al., 2022, 2023). Thus, it is plausible that cannabinoid-based treatment could yield true benefits for arthritic patients, but more investigation is necessary to determine the efficacy of lesser-studied minor cannabinoids.
The potential clinical use of Δ9-THC is hampered by its well-known psychotropic effects, producing somnolence, motor impairment, and cognitive side effects (Bosker et al., 2012; Grotenhermen and Müller-Vahl, 2016; Hudson et al., 2022). However, the less potent minor cannabinoid delta-8-tetrahydrocannabinol (Δ8-THC) (Compton et al., 1993; Govaerts et al., 2004) produces qualitatively similar but milder cognitive effects in humans (Hollister and Gillespie, 1973; Kruger and Kruger, 2023) and mice (Ten Ham and De Jong, 1974; Compton et al., 1991; Vanegas et al., 2022), consequently presenting a potentially lower risk profile of adverse effects. Indeed, Δ8-THC has been well-tolerated in human studies (Hollister and Gillespie, 1973; Abrahamov et al., 1995), and its immunosuppressive effects have been documented in several animal models (Loveless and Munson, 1976; Thapa et al., 2018; Kakar et al., 2022). However, Δ8-THC has yet to be evaluated using a model of inflammatory arthritis.
Thus, the current study was designed to probe the anti-arthritic potential of Δ8-THC in comparison with a clinically relevant treatment (i.e., dexamethasone (DEX)) in a mouse model of autoimmune arthritis (i.e., collagen-induced arthritis). One of the many limitations of studying clinical populations or preclinical models of inflammatory arthritis is a lack of predictive biomarkers for pain or loss of function (Wang et al., 2013; Zhao and Moots, 2020; Nass et al., 2021) Thus, in addition to assessing arthritic severity via semiquantitative scoring, paw thickness, histology, and tissue cytokine levels, the present study also assessed paw functionality, including grip strength, climbing, and thermal preference, to determine the potential use of these assays as behavioral markers of arthritic severity.
Materials and Methods
Animals.
Male DBA/1J mice (Jackson Laboratories, Bar Harbor, ME) aged 9–12 weeks at the start of the experiment were randomly assigned to all treatment conditions. Females were excluded from study because it has been previously reported that females resist developing CIA (Holmdahl et al., 1986; Jansson and Holmdahl, 2001). Mice were housed 3–5 per cage in an Optimice 100-cage IVC rack system (Animal Care Systems, Centennial, CO) on Teklad Sani-Chips bedding (Inotiv, West Lafayette, IN) in an Association for Assessment and Accreditation of Laboratory Animal Care International accredited facility. Mice were maintained on a 12/12-hour light/dark cycle (lights on at 0600) with free access to food and bottled water throughout. The Institutional Animal Care and Use Committee at the University of Connecticut approved all experimental protocols prior to testing. All testing was conducted during the light phase, by an experienced experimenter blinded to the treatment conditions.
Drugs.
The minor phytocannabinoid delta-8-tetrahydrocannabinol was generously provided by the NIDA drug supply program (Bethesda, MD). The synthetic glucocorticoid DEX was purchased from Sigma-Aldrich (St Louis, MO). All drugs were dissolved in a vehicle solution of 5% ethanol, 5% Kolliphor EL (Sigma-Aldrich), and 90% normal saline (i.e., 1:1:18 parts by volume; Kinsey et al., 2011; Nass et al., 2021). All solutions were injected subcutaneously at a volume of 10 μl/g body mass. All compounds were injected starting the day of the collagen booster immunization (i.e., 21 days after the initial immunization), and continuing daily until the end of the study.
CIA Model.
Under general isoflurane anesthesia (Covetrus, Dublin, OH), 100 μl of an emulsion of bovine type II collagen (2 mg/ml) dissolved in 0.05M acetic acid (Chondrex, Inc., Redmond, WA; #20022), in an equal volume of complete Freund’s adjuvant was injected subcutaneously, ∼1 cm from the distal end of the tail. The emulsion, which was a similar color and consistency to mayonnaise, was prepared on ice using a Tissue Tearor (Bartlesville, OK, USA) just prior to administration and preloaded into 10 ml syringes; otherwise, the preparation would separate if left too long. Twenty-one days later (Brand et al., 2007; Nass et al., 2021), mice received a second ‘‘booster’’ exposure to the collagen, this time emulsified in an equal volume of incomplete Freund’s adjuvant (IFA), injected midway between the tail base and first exposure site. Control mice received two injections of an emulsion of IFA and 0.05M acetic acid without collagen. Complete Freund’s adjuvant was prepared in-house and consisted of heat-inactivated Mycobacterium tuberculosis H37Ra (4 mg/ml) (BD Biosciences, Franklin Lakes, NJ; # BD 231141), 85% heavy mineral oil (Fisher Scientific, Pittsburgh, PA; #O122–1), and 15% mannide monooleate (Fisher Scientific; #AC332130250). The IFA was prepared identically but without M. tuberculosis (Nass et al., 2021).
Measuring Paw Arthritic Morphology.
Each morning after the ‘‘booster’’ injection of IFA with (CIA) or without (control) collagen, forepaws and hind paws were examined using a clinical scoring system commonly used in the CIA model (Brand et al., 2007; Miyoshi and Liu, 2018; Zhu et al., 2019). The magnitude of paw inflammation was quantified by assigning the numbers 0–4 as follows: 0, normal or no obvious differences in appearance versus healthy mice; 1, erythema and swelling in one or two toes, but no apparent swelling of paw or ankle; 2, erythema and swelling in three or more toes or mild swelling of entire paw; 3, erythema and moderate swelling extending from the ankle to the metacarpal/metatarsal joints; and 4, erythema and severe swelling encompassing the ankle, foot, and all digits or ankylosed paw and toes. The scores for each limb were summed for each mouse, resulting in a composite arthritis score with a maximum of 16 total points. In addition to clinical scoring, hind paw thickness (i.e., the distance between the plantar surface and dorsum, as measured between the walking pads) was measured daily to the nearest 0.01 mm using a digital micrometer (Mitutoyo, Japan) to quantify paw edema, as previously described (Kinsey et al., 2011; Nass et al., 2021). The mean of the hind paw measures for each mouse was used for analysis. Forepaw thickness was not measured because this measure was highly variable in pilot studies.
Inverted Grid Grip Strength.
Individual mice were placed on a stainless-steel wire bar perforated floor cage insert (Allentown, LLC) that was suspended 18 cm over a padded surface (Crawley, 2008). Adhesive tape was applied to the outer 3 cm of the grid to prevent mice from climbing over the perimeter edge. The grid was gently inverted for up to 60 seconds, and the latency to fall was measured to the nearest 0.01 second using a stopwatch. To reduce the probability of the mouse falling while inverting the grid, the mouse was placed vertically on the grid so that its head was oriented toward the ceiling, and then the rostral side of the grid was gently lowered as if hinged on the side caudal to the mouse. The task was performed every two days after the booster injection.
Automated Wire-Climbing.
Climbing is an ethologically important behavior in mice that is depressed by different pain states (Santos et al., 2023). To assess the effects of CIA on spontaneous climbing behavior, mice were placed individually into clear, plastic cylinders (18 cm H x 7.5 cm d) and allowed to freely explore for 10 minutes. Each cylinder was lined from bottom to top with 1/4 in wire mesh around 75% of the inner perimeter (no mesh in front 25% to permit unobscured video recording). Additionally, the top of the cylinder was covered by a perforated plastic lid to permit airflow but prevent the mouse from climbing out or escaping. Behavior was video-recorded and measured in real-time using ANY-maze software (Stoelting, Wood Dale, IL, USA), by separating the top and bottom halves of the cylinder into ‘climbing’ and ‘not climbing’ zones. The mouse was scored as being in either zone if its centroid (i.e., midpoint) crossed into the zone. To validate this automated measure, a subset of videos (n = 20) was deidentified, and climbing was hand-scored by a trained and blinded observer. For hand-scoring, climbing was operationalized as all four paws of the mouse being off the floor of the apparatus. There was no difference in ANY-maze scored and hand-scored climbing (t(18) = 2.047; P = 0.0555) (Supplemental Fig. 1), so ANY-maze scored climbing is reported here. Climbing was tested once, on day 14 after the booster injection.
Thermal Preference.
The thermal gradient ring (TGR) (Stoelting Co, Wood Dale, IL) is an automated test apparatus comprised of an enclosed, conductive ring connecting two hot/cold plates that create a temperature differential across the ring (i.e., 20–40°C). The ring is lit from above by arrays of visible and infrared LEDs. Individual mice were placed in the 23 ± 1°C zones (i.e., about room temperature) on alternating sides of the ring to start and allowed to freely explore for 16 minutes. Mouse location and movement were tracked using an overhead camera and analyzed in real time using ANY-maze software. Dependent variables include the weighted average of temperature/zone occupancy (i.e., ‘temperature preference’) and distance traveled in meters. Typically, locomotor activity is increased and temperature preference is lower in the first 8 minutes of testing, consistent with exploratory behavior rather than true thermal preference (Supplemental Fig. 2). Therefore, only data from minutes 9–16, where locomotor activity was more stable, were used for analyses. Temperature preference was tested once, on day 14 post-booster injection.
Histologic Analysis of the Ankle Joint.
On day 15 post-booster, hind paws were obtained by transecting the right hind leg from each mouse at the mid-tibia level. Skin was removed from the ankle and mid-foot but left in place around the digits. A slit was made in the skin dorsal to the digits to allow penetration of fixative, and the digit tips (claws) were removed. Hind paws were fixed in 4% paraformaldehyde for 6 days at 4°C on a rocking platform with changes every other day, then washed in phosphate-buffered saline and decalcified in 14% EDTA until no trace of mineralized tissue remained, as verified by X-ray (Faxitron, Tucson, AZ). Hind paws were dehydrated through a graded ethanol/xylenes series and infiltrated with paraffin under vacuum with four changes. Hind paws were embedded in sagittal orientation and sectioned at 12 microns, and sections mounted in alternate pairs on frosted slides. Equal numbers of slides were stained with H&E or with Safranin-O/Fast Green. The “standardized microscopic arthritis scoring of histologic sections (SMASH) method was used for scoring (Hayer et al., 2021). Slides were examined and scored using a Nikon E800 light microscope by a single experienced reviewer who was blinded to the identity of the samples. Scores were performed on a minimum of three sections at least 48 microns apart per hind paw from a minimum of three different slides. Every joint present in the section was scored for four separate parameters as described in Hayer et al. (2021): synovial infiltration, cartilage proteoglycan loss (from Safranin-O stained sections only), cartilage erosion, and bone erosion. Using the criteria described in detail in Hayer et al. (2021), scores were assigned for each joint in 0.5 increments from 0.0 to 3.0 (with 3.0 being most severe). Multiple scores of the same joint from different sections of the same hind paw were averaged together so that each joint from each paw was assigned one single score value for statistical comparison. The mean score from all joints was used to obtain a single SMASH score value for each hind paw.
Paw Cytokine Quantification.
On day 15 post-booster, left hind paws were transected just above the ankle joint, snap-frozen in liquid nitrogen, and stored at -80°C until assay. Samples were homogenized using a Tissue Tearor (Bartlesville, OK, USA) in ice-cold phosphate buffered saline and centrifuged at 4°C and 3220g for 16 minutes. The supernatant was collected and used for quantification of cytokines (interleukin (IL)-1β, IL-6, and vascular endothelial growth factor-A (VEGF-A)), in triplicate, via Ella Multiplex Panels (ProteinSimple, Minneapolis, MN, USA). Per manufacturer specifications, samples were diluted 1:1 prior to loading into the cartridge using the sample diluent included in the Ella kit.
Statistical Analyses.
All data were analyzed by ANOVA, followed by Bonferroni post hoc comparisons. Data for clinical scores, edema, and the inverted grid test were compared using two-way mixed ANOVA, with experimental treatment as a between-subjects variable and days post-booster as a within-subject variable. Daily effects in repeated measures (i.e., clinical score, edema, and inverted grid test), climbing, temperature preference, histology, and paw cytokine data were analyzed using one-way ANOVA with treatment as a between-subjects variable. All data are represented as mean ± S.E.M. with an alpha level of α = 0.05 to determine statistical significance.
Results
Δ8-THC Attenuates CIA-Induced Paw Swelling.
As previously reported, CIA caused significant morphologic changes in the mice. Specifically, a repeated-measures ANOVA revealed a main effect of treatment on clinical arthritis score [F(4,43) = 27.973; p < 0.0001; Fig. 1A] and paw thickness [F(4,43) = 18.422; p < 0.0001; Fig. 1B] and significant interactions between treatment and experiment day for each measure, respectively [F(56,588) = 23.020; p< 0.0001; F(56,602) = 11.069; p< 0.0001]. Bonferroni post hoc comparisons revealed that mice in the CIA group had significantly higher clinical scores (M = 4.73; S.E.M. = 0.5) [mean difference = 4.68, p< 0.0001] and thicker paws (M = 3.46 mm; S.E.M. = 0.06 mm) [mean difference = 0.56 mm, p< 0.0001] than controls (M = 0.1 & 2.91 mm; S.E.M. = 0.06 & 0.01 mm) starting on days 7–8. There was a delayed onset of CIA-induced morphologic changes (day 10 versus day 7) in mice treated twice daily with Δ8-THC (30 mg/kg). Furthermore, Δ8-THC-treated mice had significantly lower clinical scores (M = 2.77; S.E.M. = 0.31) [mean difference = −1.97, p = 0.009] and less paw swelling than CIA mice (M = 3.17 mm; S.E.M. = 0.04 mm) [mean difference = −0.29 mm, p=0.0067].
Δ8-THC attenuates CIA-induced effects on paw morphology. Mice subjected to CIA and treated twice daily with Δ8-THC (30 mg/kg) for two weeks have significantly lower (A) clinical arthritis scores and (B) less hind paw swelling than vehicle-treated arthritic controls. Data are expressed as mean ± SEM (n = 9-10). *P < 0.05 versus (non-CIA) control, #P < 0.05 versus CIA.
Δ8-THC Attenuates CIA-Induced Behavioral, Functional Deficits.
The effects of Δ8-THC on CIA-induced functional deficits (i.e., latency to fall from an inverted grid) were assessed every two days. There was a main effect of treatment on latency to fall [F(4,43) = 14.461; p< 0.0001; Fig. 2A] and a significant interaction between treatment and experiment day [F(24,258) = 5.578; p< 0.0001]. CIA decreased latency to fall from the grid (M = 23.42 seconds; S.E.M. = 2.89 seconds) versus controls (M = 42.67 seconds; S.E.M. = 2.3 seconds) [mean difference = −20.42 s, p = 0.0007]. Mice treated repeatedly with DEX (M = 51.64 seconds; S.E.M. = 1.98 seconds) did not differ from controls [mean difference = 10.81 s, p = 0.0602] indicating that the steroid treatment blocked this functional impairment of CIA. However, mice treated with either the 3 mg/kg (M = 19.3 seconds; S.E.M. = 2.56 seconds) [mean difference = −26.29 s, p< 0.0001] or 30 mg/kg dose of Δ8-THC (M = 30.89 seconds; S.E.M. = 2.58 seconds) [mean difference = −12.4 3 s, p= 0.0275] had a significantly shorter latency to fall than controls, suggesting that Δ8-THC produced a transient delay but ultimately had no effect on CIA-impaired grip function in the inverted grid test.
Δ8-THC prevents CIA-induced pain behaviors without producing locomotor deficits. Mice were tested every two days in the inverted grid test to assess CIA-depressed paw function. (A) Δ8-THC delayed CIA-decreased latency to fall from the inverted grid. On the last day of the experiment, spontaneous wire-climbing behavior and thermal preference in the TGR were measured. (B) Δ8-THC (30 mg/kg) blocked CIA-depressed climbing and (D) CIA-induced preference for warmth (C,E) without locomotor deficits. Data are expressed as mean ± S.E.M. (n = 9–10). *-***P < 0.05–0.0005 versus (non-CIA) control, #-###P < 0.05–0.0005 versus CIA, †P < 0.05 versus DEX.
Pain-depressed behaviors were quantified 14 days after the booster treatment via the wire-climbing and thermal gradient ring tests, respectively. There was a main effect of treatment on wire-climbing behavior [F(4,43) = 14.083; p < 0.0001; Fig. 2B], immobility during the climbing test [F(4,43) = 10.684; P < 0.0001; Fig. 2C], temperature preference [F(4,43) = 4.694; p =0.0031; Fig. 2D], and distance traveled in the TGR [F(4,43) = 8.536; p< 0.0001; Fig. 2E]. Explicitly, mice in the CIA group climbed less (M = 23.61 s; S.E.M. = 16.63s) [mean difference = -378.18 s, p< 0.0001], were more immobile (M = 351.4 seconds; S.E.M. = 43.19 seconds) [mean difference = 266.09 seconds, p <0.0001], preferred warmer temperatures (M = 33.44°C; S.E.M. = 1.79°C) [mean difference = 4.2°C, p =0.0136], and traveled less in the TGR (M = 3.23 m; S.E.M. = 1.37 m) [mean difference = -10.95 m, p< 0.0001] than controls (M = 401.79 s, 85.31 seconds, 29.24°C, 14.18 m; S.E.M. = 74.37 s, 37.29 seconds, 1.94°C, 1.28 m). In contrast, mice subjected to CIA and treated with the higher dose of Δ8-THC (30 mg/kg) climbed significantly more (M = 213.26 s; S.E.M. = 79.57 s) than vehicle-treated CIA mice [mean difference = 189.65 s, p= 0.0265]. Similarly, CIA mice treated with Δ8-THC (30 mg/kg), preferred the same temperature (M = 28.18°C; S.E.M. = 1.08°C) as non-CIA controls [mean difference = 1.06°C, p= 0.5085] and were more mobile than untreated CIA mice in the climbing (M = 213.6 seconds; S.E.M. = 44.15 seconds) [mean difference = −137.8 seconds, p = 0.0173] and TGR tests (M = 10.62 m; S.E.M. = 2.32 m) [mean difference = −3.56 m, p = 0.1085], indicating that Δ8-THC prevented pain-depressed behaviors caused by CIA. It is worth noting that the positive control drug DEX was more effective than Δ8-THC (30 mg/kg) at preventing CIA-depressed climbing (M = 493.18 s; S.E.M. = 64.42s) [mean difference = 279.92 s, p = 0.0015] and CIA-induced immobility in the climbing test (M = 51.77 s; S.E.M. = 38.99 s) [mean difference = −161.83 s, p = 0.0057]. However, Δ8-THC (30 mg/kg) produced antiarthritic effects comparable to DEX in the TGR test, including attenuation of CIA-induced thermal preference (M = 27.56°C; S.E.M. = 0.57°C) [mean difference = -0.62°C, p = 0.7042] and locomotor deficits (M = 12.05 m; S.E.M. = 1.23 m) [mean difference = 1.43 m, p = 0.5257].
Δ8-THC prevents synovial inflammation and bone erosion in the ankle joint.
To further assess the physiopathological effects of CIA, histology was performed on the ankle joints of a subset of individuals (n = 2–4) from each group. Fig. 3 shows representative histologic examples of each group. As expected, the feet of control (non-CIA) mice and CIA mice treated with dexamethasone were similar and did not display signs of arthritis. However, in the feet of mice with CIA there was massive macrophage infiltration into the synovium and connective tissue, along with cartilage and bone erosion. Remarkably, the feet of CIA mice treated with 30 mg/kg Δ8-THC were similar to the feet of control mice and lacked major synovial macrophage infiltration. Proteoglycan loss was also ameliorated, although bone erosion still occurred. Little response was observed with low-dose Δ8-THC (3 mg/kg) treatment.
Histology of Δ8-THC effects in the mouse CIA model. (A,B) Metatarsal-phalangeal joint from a control (no CIA) mouse stained with (A) H&E and (B) Safranin-O/FG showing normal thin appearance of synovium (*) and robust proteoglycan-positive chondrocytes in the articular cartilage (arrowheads in B). (C) Tibia-talus-calcaneus joints from a positive control mouse (CIA + dex); the synovium is multi-layered but free of macrophage infiltration, and the articular cartilage stains robustly with Safranin-O (arrowheads). (D) Metatarsal-phalangeal and (E) tibia-talus joints from two different CIA mice stained with H&E; note extensive macrophage synovial infiltration (m/si) and cartilage/bone erosion (c/be). (F) Tibia-talus and (G) metatarsal-phalangeal joints from two different CIA mice treated with 30 mg/kg Δ8-THC (high dose); in (F), only scattered macrophages are present in the synovium (*) and the articular cartilage stains robustly with Safranin-O (arrowheads): compare with the control tibia-talus joint in (C); in (G) there is moderate loss of periosteal/trabecular bone (long arrows) but no macrophages are present in the synovium (*): compare with the control joint in (A). (H,I) Metatarsal-phalangeal and calcaneus-cuboid joints from two different CIA mice treated with 3 mg/kg Δ8-THC (low dose); in both feet, note the extensive macrophage synovial infiltration (m/si) and cartilage/bone erosion (c/be); and in (I) note the extensive loss of proteoglycan staining from the articular cartilage (arrowheads). Magnification ×100. Scale bar = 1.0 mm.
Quantitative analysis of SMASH scores revealed a main effect of treatment on synovial infiltration [F(4,13) = 10.604; p = 0.0005; Fig. 4A], cartilage proteoglycan depletion (i.e., safranin-O staining intensity) [F(4,13) = 16.214; p = 0.0051; Fig. 4B], cartilage erosion [F(4,13) = 3.790; p = 0.0296; Fig. 4C], and bone erosion [F(4,13) = 6.214 p =0.0051; Fig. 4D]. CIA mice had significantly higher scores in all categories (M = 1.75–2.56; S.E.M. = 0.085–0.37) compared with non-CIA controls (M = 0.37–0.79; S.E.M. = 0.08–0.25) [mean difference = 0.86–2.06, p =0.0005–0.0220], except for cartilage erosion (M = 1.23; S.E.M. = 0.38) [mean difference = 0.858, p = 0.1031]. Compared with the CIA group, Δ8-THC-treated (30 mg/kg) mice had significantly reduced synovial inflammation (M = 0.56; S.E.M. = 0.15) [mean difference = 2, p = 0.0007], proteoglycan loss (M = 0.56; S.E.M. = 0.15) [mean difference = −1.91, p = 0.0017], cartilage erosion (M = 0.17; S.E.M. = 0.08) [mean difference = −1.06, p =0.0494], and bone erosion (M = 0.35; S.E.M = 0.15) [mean difference = −1.4, p = 0.0141], indicating that Δ8-THC is protective against the pathologic changes associated with inflammatory arthritis. Notably, there were no differences between DEX and high-dose Δ8-THC treatment in any SMASH category (M = 0.05–0.52; S.E.M. = 0.05–0.15) [mean difference = 0.03–0.49, p = 0.393–0.9636], although, this comparison should be interpreted cautiously due to the small sample size of the DEX group (n = 2).
Δ8-THC prevents CIA-induced joint damage and inflammation. Quantitative analyses were performed on feet collected from mice at the end of the experiment. Δ8-THC blocked CIA-induced increases in (A) synovial inflammation, (B) proteoglycan loss, (C) cartilage erosion, and (D) bone erosion. Data are expressed as mean ± S.E.M. (n = 4). *-**P < 0.05–0.005 versus (non-CIA) control, #-##P < 0.05–0.005 versus CIA, †P < 0.05 versus DEX.
Δ8-THC Attenuates CIA-Induced Influx of Proinflammatory Cytokines.
At the end of behavioral testing, paw cytokines were assessed in whole hind paw homogenates. There was a main effect of treatment on levels of the cytokines IL-1β [F(4,41) = 16.525; p< 0.0001; Fig. 5A], IL-6 [F(4,41) = 5.72; p =0.0009; Fig. 5B], and VEGF-A [F(4,41) = 11.4; p< 0.0001; Fig. 5C]. The CIA mice had increased levels of IL-1β (M = 9290.5 pg/g; S.E.M. = 680.55 pg/g) [mean difference = 8303.12 pg/g, p< 0.0001] and IL-6 (M = 730.99 pg/g; S.E.M. = 199.99 pg/g) [mean difference = 672.69 pg/g, p = 0.0002] and decreased VEGF-A (M = 851.63 pg/g; S.E.M. = 102.4 pg/g) [mean difference = −2157.72 pg/g, p< 0.0001] compared with non-CIA controls (M = 987.37, 58.31, 3009.34 pg/g; S.E.M = 109.51, 7.3, 207.53 pg/g). Treatment with Δ8-THC (30 mg/kg) effectively blocked CIA-induced alterations of IL-1β (M = 5388.47 pg/g; S.E.M. = 1609.27 pg/g) [mean difference = −3902.03 pg/g, p= 0.0096], IL-6 (M = 407.19 pg/g; S.E.M. = 155.01 pg/g) [mean difference = -323.8 pg/g, p= 0.0595], and VEGF-A (M = 2161.02 pg/g; S.E.M. = 487.11 pg/g) [mean difference = 1309.4 pg/g, p= 0.0019]. The positive control DEX was more effective than high-dose Δ8-THC at reducing CIA-increased levels of IL-1β (M = 133.09 pg/g; S.E.M. = 20.22 pg/g) [mean difference = 5255.39 pg/g, p =0.001] and IL-6 (M = 49.32 pg/g; S.E.M. = 5.78 pg/g) [mean difference = 357.87 pg/g, p= 0.044], but Δ8-THC (30 mg/kg) attenuated CIA-depressed VEGF-A to the same extent as DEX (M = 2501.88 pg/g; S.E.M. = 160.64 pg/g) [mean difference = 1207.28 pg/g, p = 0.4071].
Δ8-THC blocks effects of CIA on paw cytokine levels. Whole hind paws were collected at the end of the experiment and homogenized to quantify IL-1β, IL-6, and VEGF-A via multiplex ELISA. Levels of (A) IL-1β and (B) IL-6 were increased, and (C) VEGF-A was decreased in the CIA condition. These CIA-induced changes were blocked by Δ8-THC (30 mg/kg) or DEX treatment (twice daily, 14 days). Data are expressed as mean ± S.E.M. (n = 8–10). *-***P < 0.05-0.0005 versus (non-CIA) control, #-###P < 0.05-0.0005 versus CIA, †P < 0.05 versus DEX.
Discussion
The present study provides novel evidence that the minor phytocannabinoid Δ8-THC possesses analgesic, anti-arthritic, and immunomodulatory properties. Moreover, Δ8-THC was effective at reducing both the morphologic and functional behavioral aspects of arthritis in mice. That is, Δ8-THC (30 mg/kg) reduced paw swelling and clinical arthritis signs present in CIA-treated mice. Correspondingly, Δ8-THC attenuated both CIA-depressed wire-climbing and preference for warmth in a thermal gradient ring test, without producing locomotor effects in either test. Finally, Δ8-THC prevented joint destruction and altered inflammatory cytokine levels resulting from CIA. Together, these data support the development of Δ8-THC as a novel therapeutic treatment of inflammatory arthritis.
The present work sheds light on a relatively understudied branch of cannabinoid research. Traditionally, research efforts and public attention have focused on the major phytocannabinoids, Δ9-THC and CBD. However, there are over 100 additional minor cannabinoids produced in the cannabis plant that have been generally overlooked until very recently. These minor cannabinoids display promising therapeutic effects for a variety of human health conditions including chronic pain, inflammation, cancer, chemotherapy-induced emesis and hypophagia, diabetes, insomnia, and emotional disorders (Baek et al., 1998; Izzo et al., 2009; Brierley et al., 2016, 2017; Li et al., 2016; Segat et al., 2017; Gallily et al., 2018; Pellati et al., 2018; Smeriglio et al., 2018; Wong and Cairns, 2019; Zagzoog et al., 2020; Anand et al., 2021; Ferrarini et al., 2022). Limited evidence also suggests that the gut microbiota may mediate some of the anti-inflammatory effects of minor cannabinoids (Forte et al., 2021; Kalkan et al., 2023), although the degree to which such mechanism(s) may affect inflammatory arthritis was beyond the scope of the present study.
Traditionally, preclinical pain assessment has focused on pain-stimulated behavior, such as allodynia and hyperalgesia elicited by touch or thermal stimuli. However, many compounds that yield promising analgesic effects in pain-stimulated assessments have failed to produce results in clinical trials (Negus, 2019). In humans, chronic pain often results in a reduced ability to function and perform activities that were once enjoyable (Hadi et al., 2019). Indeed, restoration of these types of pain-depressed behaviors remains a priority of clinical treatment of chronic pain (Dworkin et al., 2008). Climbing is an innovative measure of pain-depressed and spontaneous behavior, and the wire-climbing test was designed to exploit this ecologically relevant exploratory behavior (Santos et al., 2023). The present finding that wire-climbing is significantly depressed by CIA treatment supports the expanded use of this inexpensive, yet powerful paradigm that is not subject to locomotor confounds.
Similarly, the TGR assesses pain-depressed (thermal preference) and pain-depressed (mobility) behaviors and can be used alongside traditional pain-stimulated assessments to reduce spurious results and improve translation to human work. In addition to quantifying thermal preference, the TGR simultaneously monitors locomotor activity, allowing for the dissociation of true analgesic effects and motor impairment. Previous studies have observed that CIA results in thermal hyperalgesia (Inglis et al., 2007; Kinsey et al., 2011) and mechanical sensitivity (Feng et al., 2022). However, pain-stimulated behaviors are often inconsistently observed in this model, underscoring the lack of sensitive measures for CIA-related pain. Although it is worth noting that the observed shift in temperature preference may be pain-independent, the present data are consistent with reports of pain-depressed behaviors. For example, CIA reduces voluntary wheel-running and increases preference for warmer temperatures in a two-choice thermal place preference test (Oto et al., 2019). Similarly, the individual variability observed in the inverted grid test is consistent with differences in other measures related to paw function, including the wire-climb test. The present study is the first to report that climbing behavior and thermal preference in the TGR are adversely affected by CIA, providing novel outcome measures that will be useful in the behavioral assessment of arthritic mice.
To assess a potential physiologic mechanism through which Δ8-THC prevented CIA-induced paw inflammation, cytokine levels were quantified in whole hind paw homogenates. As observed in previous work (Marinova-Mutafchieva et al., 1997; Lu et al., 2000), CIA altered levels of cytokines and immunomodulatory proteins in the hind paws. In rheumatoid arthritis patients, serum and synovial fluid levels of IL-1β, IL-6, and VEGF-A are upregulated and correlated with disease activity (Eastgate et al., 1988; Horneff et al., 1993; Kasama et al., 2000). Furthermore, IL-6 can promote synovitis and joint destruction (Srirangan and Choy, 2010), while VEGF-A is a marker of angiogenesis, or the formation of new blood vessels, which contributes to inflammation and pannus formation (Taylor, 2002). Cytokine inhibitors, like IL-1 and tumor necrosis factor-α blockers, are likely to slow disease progression in inflammatory arthritis (Gabay, 2002). However, the immunosuppressant effects of most biologic and antirheumatic medications put patients at a higher risk of infection (Gupta et al., 2022). In contrast to the clinically relevant steroid dexamethasone, Δ8-THC did not fully block the CIA-induced influx of cytokines. Although not directly measured here, these data may indicate that Δ8-THC reduces inflammatory arthritis symptoms while still allowing for healthy immune functioning, thus protecting the host against threats from viruses and bacteria commonly observed with glucocorticoid or DMARD treatments.
One limitation of this work is that females were not included in the study design, due to established resistance of female DBA mice developing CIA (Holmdahl et al., 1986; Jansson and Holmdahl, 2001). Examining sex as a biologic variable is an important consideration in preclinical research that has historically been overlooked. Further, arthritis is more prevalent in women (20.9%) than men (16.3%) (Fallon et al., 2023), and thus, evaluation of novel therapeutics should prioritize including females in the experimental design. Other models exist, such as the collagen antibody-induced arthritis model (Khachigian, 2006), in which anti-collagen antibodies are passively transferred to experimental animals, leading to rapid onset of joint inflammation. Given the autoimmune nature of most forms of inflammatory arthritis, including rheumatoid arthritis, the CIA model more closely resembles the human disease state. However, as methods for experimentally-induced arthritic models continue to develop and be optimized, we are hopeful that females may soon also be studied in preclinical autoimmune inflammatory arthritis models. In addition, the potential effects of repeated administration of Δ8-THC or dexamethasone on thermal preference were not assessed in non-arthritic control mice in the present study. Thus, we cannot exclude the possibility that repeated dosing alters locomotor activity or a shift in thermal preference.
Developing Δ8-THC as an alternative or adjunctive therapy for inflammatory arthritis would help fulfill the need for medications with significantly reduced side effect profiles. Indeed, the US Food and Drug Administration has placed black box warnings on anti-tumor necrosis factor agents due to increased risk of fatal infection and certain cancers such as non-Hodgkins lymphoma (Wolfe and Michaud, 2007; Askling et al., 2009; Mariette et al., 2011; https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-drug-labels-tumor-necrosis-factor-alpha-tnfa-blockers-now-include). Steroids and nonsteroidal anti-inflammatory drugs are common alternatives to biologic drugs such as tumor necrosis factor inhibitors; however, their prolonged use can trigger compounding health issues, including bone weakening and organ damage. Cannabinoids, on the other hand, are generally well-tolerated in humans. Take, for example, the 1:1 THC:CBD combination drug inhaler nabiximols (i.e., Sativex), which reliably produces analgesic effects in placebo-controlled trials with few reports of adverse events (Barnes, 2006; Blake et al., 2006). Additionally, there were no reports of serious adverse events in two studies that investigated the effects of Δ8-THC in adult males (Hollister and Gillespie, 1973) and children (Abrahamov et al., 1995). By contrast, the US Food and Drug Administration received 104 reports of adverse events involving products that purportedly contained Δ8-THC between 2020 and 2022 and has since raised concerns regarding variability in product formulations and labeling (https://www.fda.gov/consumers/consumer-updates/5-things-know-about-delta-8-tetrahydrocannabinol-delta-8-thc). Major side effects of cannabis-based medicines are rare, although dizziness, somnolence, and dry mouth are commonly reported (Grotenhermen and Müller-Vahl, 2016; Vermersch and Trojano, 2016). Nonetheless, these off-target effects are slight in comparison with those induced by DMARDs and biologic drugs. Furthermore, tolerance quickly develops to cannabinoid receptor type 1agonism and, therefore, side effects are typically resolved after a short time. Yet, future studies may choose to implement local versus systemic administration of Δ8-THC to circumvent any undesired central nervous system effects.
Acknowledgments
The authors thank UConn Animal Care Services for excellent animal care, and Carl Rodriguez, Shiv Patel, Zoe Kassapidis, and Diondra Owusu, for excellent technical assistance.
Data Availability
The authors declare that all the data supporting the findings of this study are contained within the paper.
Authorship Contributions
Participated in research design: Vanegas, Dealy, Kinsey.
Conducted experiments: Vanegas, Zaki, Dealy, Kinsey.
Performed data analysis: Vanegas, Dealy, Kinsey.
Wrote or contributed to the writing of the manuscript: Vanegas, Zaki, Dealy, Kinsey.
Footnotes
- Received February 16, 2023.
- Accepted April 11, 2024.
This work was supported financially by the National Institutes of Health (NIH) National Center for Complementary and Integrative Health (NCCIH) [Grant AT010773] and National Institute on Drug Abuse (NIDA) [Grant DA052690].
No author has an actual or perceived conflict of interest with the contents of this article.
↵
This article has supplemental material available at jpet.aspetjournals.org.
Abbreviations
- CBD
- cannabidiol
- CIA
- collagen-induced arthritis
- DEX
- dexamethasone
- DMARD
- disease-modifying antirheumatic drug
- IFA
- incomplete Freund’s adjuvant
- IL
- interleukin
- SMASH
- standardized microscopic arthritis scoring of histological sections
- TGR
- thermal gradient ring
- Δ8-THC
- delta-8-tetrahydrocannabinol
- Δ9-THC
- delta-9-tetrahydrocannabinol
- VEGF-A
- vascular endothelial growth factor-A
- Copyright © 2024 by The Author(s)
This is an open access article distributed under the CC BY-NC Attribution 4.0 International license.
