Cannabinoid CB₁ allosteric ligands suppress inflammatory pain in male mice without tolerance, motor impairment, memory deficits, or respiratory depression

Abstract

Background and Purpose

Positive allosteric modulators (PAMs) engage sites distinct from the orthosteric CB₁ cannabinoid receptor binding site. GAT211, a racemic ligand, contains GAT228 (R) and GAT229 (S) enantiomers. Whether enantiomer-mediated differences in allosteric activation or positive allosteric modulation translate into differences in therapeutic efficacy or unwanted cannabimimetic effects remain unknown.

Experimental Approach

We evaluated therapeutic efficacy of the pure CB₁ allosteric agonist (GAT228 (R)) and PAM (GAT229 (S)) in inflammatory pain induced by complete Freund’s adjuvant (CFA) in male mice. We assessed whether these allosteric ligands would spare tolerance, memory deficits, motor impairment and respiratory depression associated with the orthosteric CB₁/CB₂ agonist CP55,940.

Key Results

GAT228 and GAT229 attenuated CFA-induced mechanical hypersensitivity in a CB₁-dependent manner without producing tolerance or hypothermia associated with the orthosteric CB₁/CB₂ agonist CP55,940. CP55,940 also produced memory deficits, motor impairment, and robust respiratory depression. GAT228 and GAT229 failed to do so. Unlike CP55,940, allosteric ligands did not produce motor impairment or notable respiratory depression in the presence or absence of CFA-induced inflammation. Allosteric ligands increased novel object recognition memory, whereas CP55,940 decreased time spent exploring the novel object. Intraplantar CFA increased immobility time in the tail suspension test, indicative of pain-related behavioral despair, which was normalized by allosteric ligands and exacerbated by CP55,940.

Conclusion and Implication

CB₁ allosteric ligands represent promising, safer therapeutic strategies for pain management compared to direct CB₁ orthosteric agonists. Overall, this work contributes to offering a novel pathway for pain management that minimizes the risk of adverse effects.

Graphical Abstract

1.0. Introduction

The endocannabinoid system consists of endocannabinoids, synthetic and degradative enzymes, and their receptors (Lutz, 2020). CB1 and CB2 cannabinoid receptors are preferentially activated by the endogenous ligands anandamide and 2-arachidonoylglycerol (2-AG), respectively (Hillard (2015); Guindon & Hohmann 2009). CB₁ is expressed in the central nervous system (CNS) and periphery, including in primary sensory neurons (Hohmann & Herkenham, 1999; Bridges et al., 2003). CB₂ is abundant in immune cells in the periphery, but is also expressed in the CNS, particularly in disease states (Pertwee, 2008; Guenther et al., 2025). Both CB₁ and CB₂ are implicated in modulation of pain (Finn et al., 2021; Guenther et al., 2025).

Orthosteric CB₁ agonists produce antinociception in preclinical pain models as well as unwanted on-target side effects (Deng et al., 2015a; 2015b). For example, CP55,940, a synthetic full CB₁/CB₂ agonist, and Δ⁹-tetrahydrocannabinol (THC), a partial agonist and psychoactive phytocannabinoid in cannabis, produce motor impairment, including ataxia, catalepsy, and reduced coordination in rodents (Deng et al., 2015a; 2015b) and people (Orazietti et al., 2022; Zhou et al., 2024). Recent preclinical studies demonstrate that THC induces respiratory depression at doses lower than those that produce cardinal signs of CB₁ activation (Watkins et al., 2025), suggesting a potential safety concern even for partial agonists. More severe respiratory depression is observed with full CB₁ agonists and high-efficacy synthetic cannabinoids such as CP55,940 (Watkins et al., 2025), and compounds found in illicit products (Manini et al., 2022; James et al., 2025). These drawbacks may contribute to the limited therapeutic window of CB₁-targeting drugs.

The precise resolution of protein structures and identification of allosteric binding sites has facilitated understanding of conformational changes crucial to allosteric regulation (Bueno et al., 2020) and medicinal chemistry efforts informed by structure-based allosteric design (Zhu et al., 2024). Allosteric sites present unique opportunities for selective modulation with fewer unwanted effects (Kenakin, 2013; Hryhorowicz et al., 2019).

CB₁ positive allosteric modulators (PAMs) suppress inflammatory and neuropathic pain in preclinical studies without eliciting characteristic cannabinoid-related adverse effects (Ignatowska-Jankowska et al., 2015; Cairns et al., 2017; Slivicki et al., 2018; Slivicki et al., 2020 Garai et al., 2019). We reported that the CB₁ PAM GAT211, a racemic ligand, suppressed mechanical hypersensitivity induced by both the chemotherapeutic agent paclitaxel and intraplantar injection of complete Freund’s adjuvant (CFA) (Slivicki et al., 2018). GAT211 was subsequently reported to exhibit dual activities in vitro via its respective enantiomers GAT228 (R) and GAT229 (S). Pharmacological profiling attributed the allosteric agonist profile to GAT228, whereas potent PAM profiling was attributed to GAT229 (Mitjavila et al., 2018; Laprairie et al., 2017; Thapa et al., 2020). Ago-PAMs represent a distinct class of allosteric ligands that combine properties of allosteric agonists and PAMs. These compounds enhance potency and/or efficacy of orthosteric agonists while also capable of activating the receptor in the absence of an orthosteric ligand. Thus, GAT228 acts as a partial agonist through an allosteric site (allosteric agonist), whereas GAT229 acts as a positive allosteric modulator (PAM) (Laprairie et al., 2017; Thapa et al., 2020).

Whether differences in the in vitro signaling profiles between these enantiomers translate to differences in vivo pharmacology, remains unknown. Whether CB₁ PAMs or allosteric agonists produce differential profiles of efficacy or side effects (e.g., hypothermia, memory or motor impairment, respiratory depression) has never been investigated. Here we show that GAT228 and GAT229 suppress CFA-induced inflammatory pain and pain-related behavioral despair without producing tolerance, memory deficits, respiratory depression or motor impairment observed with the orthosteric CB₁/CB₂ agonist CP55,940. Given pressing needs for safe and effective pharmacotherapies for pain (Nightingale, 2012), allosteric ligands represent a promising therapeutic strategy.

2.0. Materials and Methods

2.1. Subjects, Ethics, and Animal Welfare

One hundred and eighty-eight adult male C57BL/6J wild-type mice (Jackson Laboratories, Bar Harbor, ME) were used in these experiments. The C57BL/6J mouse strain was used for this study, due to its well-characterized behavioral and physiological responses in pain models and its common use in studies of cannabinoid pharmacology, facilitating the translational relevance. Only male mice were used to enable a broad spectrum of behavioral profiling, provide consistency with prior studies of CB₁ allosteric ligands (Slivicki et al., 2018) and reduce variability in this initial characterization. Mice were housed in a temperature and humidity-controlled facility with ad libitum access to food and water on a 12:12h light/dark cycle. All procedures were approved by the Indiana University Bloomington Animal Care and Use Committee (BIACUC) and conducted in accordance with the ARRIVE guidelines and policies on animal use of the British Journal of Pharmacology (Percie du Sert, Hurst et al. 2020). Mice were monitored daily for signs of distress or adverse effects. Intraplantar CFA injection was performed without anesthesia, as the procedure is brief and causes minimal acute distress.

2.2. Drugs and Pharmacological Specificity

GAT228 (R) and GAT229 (S) with a purity of 98.5% were synthesized by the Thakur laboratory (by SG). Complete Freund's adjuvant (CFA) was purchased from Sigma-Aldrich (St. Louis, MO). CP55,940 was purchased from Cayman Pharmaceuticals (Ann Arbor, MI). AM251 was a gift from the Makriyannis laboratory (Northeastern University, Boston, MA). Due to limited solubility, drugs were dissolved in a vehicle consisting of 10% dimethyl sulfoxide (DMSO; Sigma-Aldrich, St. Louis, MO), at a final concentration that is below levels reported to produce behavioral effects in mice, with the remaining volume consisting of kolliphur (Alkamuls EL-620; Solvay), 100% ethanol (Decon Labs; Montgomery, PA), and 0.9% saline (Baxter Pharmaceuticals; Bloomington, IN) at a 1:1:18 ratio. Injections were administered acutely (i.p.) in a volume of 10 mL/kg and chronically in a volume of 5 mL/kg. To assess mediation by CB₁ receptors, GAT229 (20 mg/kg, i.p.) and GAT228 (20 mg/kg, i.p.) were administered either alone or co-administered with the CB₁ antagonist/inverse agonist AM251 (5 mg/kg, i.p.) or vehicle (n = 5-6 per group for all groups). The AM251 dose was chosen for its ability to block anti-allodynic effects of GAT211 (Slivicki et al., 2018) and other endocannabinoid modulators (Slivicki et al., 2019) without altering nociceptive thresholds in our previous work.

2.3. General Behavioral Experimental Protocol

All behavioral experiments were conducted by a single experimenter (IS) blinded to the treatment condition. Mice were randomly assigned to treatment using computer-generated random numbers, created in Microsoft Excel, that were sorted into treatment groups to minimize allocation bias. To minimize animal discomfort in studies employing CFA injection, survival times were the shortest necessary to complete the experimental objectives. The orthosteric CB₁ agonists CP55,940 suppress CFA-induced inflammatory pain in a CB₁-dependent manner (Sain et al., 2009; Li et al., 2019) despite binding with equal affinity to CB₁ and CB₂ receptors in vitro (Felder et al., 1995). In the present study, only CP55,940 doses that exhibit CB₁-mediated pharmacological specificity in the acute inflammatory phase of the CFA model (Sain et al., 2009; Deng et al., 2015b), and in tests of locomotor activity (Orazietti et al., 2022), recognition memory (Kosiorek., 2003) and respiratory depression (Watkins et al., 2025) were used as comparator.

CFA (diluted 1:1 in saline) was administered (20 μL) via unilateral intraplantar (i.pl.) injection in the right hind paw with a 28.5-gauge needle. Mice were briefly and gently restrained for this injection and then immediately returned to their home cage. Approximately 48 hours later, vehicle or different doses of GAT228, GAT229 or CP55,940 (i.p.) were administered acutely and chronically to test for antinociceptive effects, drug tolerance and body temperature (Fig. 1A). Effects of acute drug treatments on respiration was assessed in otherwise naïve mice, followed 7 days later by assessment of effects of the same acute treatments on locomotor activity (Fig. 1B). CFA-injected mice used to assess respiratory depression (on day 3 post CFA) were additionally used to assess long term memory in the NOR test (on day 5-6 post CFA) in a state of inflammatory nociception (Fig. 1C), given that this state could potentially produce an altered endocannabinoid tone that could be modulated by allosteric ligands. This same cohort of CFA-inflamed mice was then used in assessments of locomotor activity at 14 days following intraplantar CFA injection (Fig. 1C). In a similar manner, mice receiving intraplantar CFA- or intraplantar saline- that were used in the tolerance study (i.e., to assess changes in mechanical paw withdrawal thresholds and body temperature) were also used to assess pain-related despair-like behavior using the tail suspension test (Fig. 1D). The tail-suspension test was performed 3 weeks following intraplantar injection of CFA or saline to match the post-CFA timepoint used in the literature to assess pain-related behavioral despair (Burek et al., 2022). Each group received the same drug and dosage in all studies, involving more than one (i.p.) injection. At the conclusion of all behavioral testing, all mice were euthanized humanely by CO₂ asphyxiation followed by cervical dislocation to ensure death.

Figure 1.

(A) Experimental timeline showing the assessment of acute and chronic effects of CP55,940, GAT228 and GAT229 on mechanical paw withdrawal thresholds (PWT) and body temperature (TEMP) in the CFA model. (B) Experimental timeline showing assessment of acute effects of CP55,940, GAT228 and GAT229 on respiratory depression using whole body plethysmography (WBP) in otherwise naïve mice, followed by assessments of the same ligands in the activity meter (AM) after a drug washout period of 7 days. (C) The timeline for assessing effects of CP55,940, GAT228 and GAT229 on respiratory function (WBP), novel object recognition (NOR) memory and locomotor behavior (AM). (D) The timeline for assessing effects of chronic dosing with CP55,940, GAT228 and GAT229 on paw withdrawal thresholds (PWT), body temperature (TEMP) and pain-depressive behavior in the tail suspension test in CFA-injected mice. Comparison was made with a drug naïve group that received saline (i.pl.) in lieu of CFA.

The results section is organized thematically as follows: Experiment 1 evaluates impact of acute and chronic dosing with CP55,940, GAT228 and GAT229 on pain behavior and body temperature in the CFA model. Experiment 2 evaluates the effects of CP55,940, GAT228 and GAT229 on respiratory function in the absence (i.e., naïve) or presence of CFA-induced inflammation. Experiment 3 evaluates effects of acute and chronic dosing with CP55,940, GAT228 and GAT229 on locomotor behavior in the activity meter. Experiment 4 evaluates effects of CP55,940, GAT228 and GAT229 on novel object recognition memory in CFA-inflamed mice. Experiment 5 evaluates effects of CP55,940, GAT228 and GAT229 on pain-related depressive-like behavior in the tail suspension test in CFA-inflamed mice.

2.4. Mechanical paw withdrawal thresholds

Paw withdrawal thresholds to mechanical stimulation were measured using an electronic von Frey aesthesiometer (IITC model Alemo 2390–5, Woodland Hills, CA). The force in grams (g) that produced removal of the hind paw from mechanical stimulation was measured in each paw. Mice were placed on an elevated metal mesh table and were habituated to inverted plastic cages placed on the testing platform for at least one hour to reduce exploratory behavior prior to behavioral testing; this handling protocol was consistent across all testing days. After habituation, paw withdrawal thresholds were measured in duplicate in each paw; the plantar region of each hind paw was stimulated by a rigid tip connected to the aesthesiometer, increasing force applied until a withdrawal response (paw retraction) was observed. The mechanical force was then terminated, and the force (in g) to produce paw withdrawal was recorded. Paw withdrawal thresholds were measured in both paws before (i.e., baseline) and at various time points following intraplantar CFA injection. Mechanical paw withdrawal thresholds were reported as the mean of the duplicate determinations for each hind paw, averaged across animals.

2.5. CFA-Induced Inflammatory Nociception

CFA was used to induce persistent inflammatory pain to facilitate assessments of drug efficacy on nociceptive sensitization, a key feature of human inflammatory pain, while also permitting modeling of pain-related behavioral despair (Liang et al., 2024). CFA was injected into the right hind paw without anesthesia, as the procedure is brief and causes minimal acute distress. CFA (diluted 1:1 in saline) was administered (20 μL) via unilateral intraplantar (i.pl.) injection in the right hind paw with a 28.5-gauge needle. Mice were briefly and gently restrained for this injection and then immediately returned to their home cage. Approximately 48 hours later, vehicle or different doses of GAT228 (5, 10, or 20 mg/kg, i.p.) or GAT229 (5, 10, or 20 mg/kg, i.p.) were administered. Dose selection was based upon our previous work showing that GAT211 suppressed CFA-induced mechanical hypersensitivity in mice with an ED₅₀ of 9.75 mg/kg i.p. (Slivicki et al., 2020). Paw withdrawal thresholds were measured before (i.e., pre-drug baseline) and at different timepoints (i.e., 30-, 90-, 150-, and 210-min post-injection) following pharmacological manipulation. Each dose was administered to a separate cohort of mice. (n = 5-6 per group for all groups) or vehicle (n = 5 per group). Sample sizes were determined based on previous studies conducted in our laboratory (Slivicki et al., 2018; Deng et al., 2015b). No formal a priori power analyses were performed; sample sizes were consistent with prior experiments that demonstrated sufficient statistical power to detect behavioral effects of cannabinoid ligands in similar experimental paradigms.

2.6. Efficacy following Chronic Dosing

GAT228 (20 mg/kg/day), GAT229 (20 mg/kg/day), CP55,940 (0.3 mg/kg/day) or vehicle (20% DMSO 1:1: 8 ethanol: kolliphur: saline) (n = 7-8 per group) were administered i.p. once daily over 10 consecutive days to CFA (i.pl.)-treated mice. Mechanical paw withdrawal thresholds were recorded before CFA injection, 48 h post CFA, and on days 1, 4, 7 and 10 of the 10-day chronic dosing intervals to ascertain whether tolerance developed to the anti-allodynic effects of allosteric ligands.

2.7. Assessment of Respiratory Parameters

Whole-body plethysmography (Data Sciences International (DSI); St. Paul, MN) was used to assess respiratory parameters in awake, unrestrained mice as described in our previously published work (Watkins et al., 2025). Minute ventilation represents breaths per minute, whereas respiratory frequency and tidal volume relate to rhythm and depth of breathing, respectively. Measures such as inspiratory/expiratory flow and inspiratory/expiratory time can indicate neuromuscular drive or central pattern generator disruption as described previously (Watkins et al., 2025). These dependent measures may help to infer potential sites of drug action. Major respiratory parameters evaluated were minute ventilation (volume of ventilation in mL/min), respiratory frequency (breaths per minute (min), and tidal volume (volume of individual breaths in mL). We additionally evaluated parameters that assess characteristics of individual breaths (i.e, total inspiratory time (s), total expiratory time (s), peak inspiratory flow (mL/sec), peak expiratory flow (mL/sec)) and time between respiratory phases (i.e., end-inspiratory pause (time between inspiration and expiration; ms), and end-expiratory pause (time between inspiratory cycles; ms)). Mice were habituated to recording chambers for 30 minutes within 48 hours of testing to minimize changes in respiration resulting from stress related to placement in a novel environment. On the test day, mice received air mixed with 10% CO₂ to standardize hypercapnic responses without inducing stress. Testing consisted of a 50-minute initial recording to acquire baseline respiratory parameters, after which mice were briefly removed from test chambers and given a single i.p. injection of drug or vehicle. Mice were then immediately returned to the chambers for 60 minutes to measure post-injection respiratory parameters.

In the acute study, otherwise naïve mice received i.p. injections of GAT228 (20 mg/kg), GAT229 (20 mg/kg), CP55,940 (0.1 or 0.3 mg/kg), or vehicle (i.p.) (n =8 per group). A similar procedure was followed to assess respiratory changes under inflammatory conditions using separate cohorts of mice. Respiratory effects were evaluated 3-4 days following unilateral intraplantar injection of CFA, as described above. After baseline recording, mice received i.p. injections of GAT228 (20 mg/kg), GAT229 (20 mg/kg), CP55,940 (0.3 mg/kg), or vehicle (i.p.) followed by 60-min post injection (n = 8 per group).

2.8.1. Novel Object Recognition (NOR) Test

On days 5-6 post-CFA injection, the same cohort used for whole body plethysmography (WBP) testing was tested in the NOR test to assess the impact of allosteric ligands and the low dose of the orthosteric agonist CP55,940 on recognition memory. CP55,940 (0.1 mg/kg, i.p.) was used in this study to avoid the unwanted motor effects of orthosteric agonists which could otherwise confound interpretation of performance in the NOR test. Mice were habituated to the testing arena (42×42×32 cm plexiglass chamber) for 5 min on the test day. After a 3-min rest in its home cage, the chamber was cleaned with RESCUE (Disinfectant produced by Virox Tech), and two identical objects were placed inside for a 10 min training phase. Following a 2-h interval, one familiar and one novel object (different in color and shape but similar in size) were introduced for the 10-min testing phase. GAT228 (20 mg/kg), GAT229 (20 mg/kg), CP55,940 (0.1 mg/kg) or vehicle were administered (i.p.) one hour before the testing phase (n = 8 per group). All sessions were recorded and analyzed using EthoVision XT. The discrimination index (DI) was calculated as ($(\text{Tnovel}-\text{Tfamiliar})∕(\text{Tnovel}+\text{Tfamiliar})$).

2.8.2. Locomotor Activity

After a 7-day drug wash out, the same cohort of non-inflamed mice previously used for respiratory testing was assessed for locomotor activity. Mice were brought into the room and habituated in their home cage on a table for 15 minutes before testing. Following habituation, mice received an i.p. injection of GAT228 (20 mg/kg), GAT229 (20 mg/kg), CP55,940 (0.1 or 0.3 mg/kg) or vehicle and placed in activity meters (Omnitech Superflex Nodes, Omnitech, Columbus, Ohio). Mice received the same pharmacological treatment and dose that they received in the previous assessment of respiratory function. Locomotor activity was automatically recorded by photobeams interpreted by Fusion 6.5 software starting the moment the mice enter the arena. Parameters evaluated included total distance travel, ambulatory velocity, resting time, vertical activity, horizontal activity and center time. An elevated set of sensors recorded activity of distance and time spent in exploratory positions. The chambers were illuminated with tungsten bulbs at ~80 lux, and a white noise generator provided a steady sound level of 62-63 dB in the arenas. Behavior was recorded for one hour, after which mice were promptly removed from each arena. Arenas were thoroughly cleaned with RESCUE (Disinfectant produced by Virox Tech) between animals. A similar procedure was followed in CFA-injected mice to assess locomotor behavior under inflammatory conditions. Inflamed mice that were previously subjected to respiratory testing (i.e., on day 3 post CFA) and evaluation in NOR (i.e., on day 5-6 post CFA) were also used in assessments of locomotor activity at 14 days following intraplantar CFA injection. CFA-injected mice were habituated in their home cages to the testing room as described above and subsequently received i.p. injections of GAT228 (20 mg/kg), GAT229 (20 mg/kg), CP55,940 (0.3 mg/kg), or vehicle followed by 60-min recording of locomotor activity (n = 8 per group).

2.8.3. Tail Suspension Test

The same cohorts of mice used in the tolerance study were tested for despair-like behavior associated with chronic pain (Cryan et al., 2005). Testing was initiated after a 6-day drug washout period following termination of repeated dosing, when mice were 3 weeks following intraplantar CFA injection, a timepoint associated with pain-related behavioral despair-like behavior (Cryan et al., 2005). Mechanical hypersensitivity was reassessed one hour prior to initiation of the tail suspension test to confirm the persistence of the inflammatory pain state. Testing was conducted in a separate behavioral room to avoid environmental confounds. CFA-treated mice continued to display reduced mechanical thresholds compared to saline-treated controls at this time-point, confirming the presence of sustained inflammatory pain before assessment of pain-depressive behavior. Mice were brought into the room and habituated in their home cage on a table for 15 minutes before testing. One hour prior to the test, CFA-injected mice received an i.p. injection of GAT228 (20 mg/kg), GAT229 (20 mg/kg), CP55,940 (0.3 mg/kg), or vehicle in the same injection volume used in tolerance studies. A separate group of mice that received intraplantar saline (in lieu of CFA) three-weeks prior and had an otherwise identical history of handling and in vivo testing was included in this study as a non-inflamed control group that was otherwise drug naïve but received i.p. vehicle one hour prior to testing in the tail suspension test. During testing, mice were suspended by adhesive tape applied near the tip of the tail at a safe height above the lab bench (~20 in. suspension height). All mice were suspended for a fixed period of 6 minutes, and their activity was recorded. The time that the mouse is immobile during the last 4 minutes of the test was recorded. Immobility was determined by the mouse ceasing to struggle against the unnatural position and remaining motionless. To ensure the safety of the mice, a soft towel was placed below the suspended mouse to cushion any potential falls and prevent injury. This procedure is not considered painful, as the mice do not exhibit signs of severe distress or injury (Cryan et al., 2005).

2.8.4. Rectal temperature

Rectal temperature (°C) was measured using a thermometer (Physitemp Instruments, Inc, Clifton, NJ) and rectal probe (Braintree Laboratories, Inc, Braintree, MA) in the same mice used to assess efficacy of allosteric ligands in suppressing CFA-induced inflammatory nociception following chronic dosing. Rectal temperature was measured before, and 60 minutes after measurement of paw withdrawal thresholds on each test day throughout the period of acute and chronic dosing. Rectal temperature was recorded immediately following assessment of mechanical paw withdrawal thresholds on day 1, 4, 7 and 10 of repeated i.p. administration of GAT228 (20 mg/kg), GAT229 (20 mg/kg) CP55,940 (0.3 mg/kg) or vehicle. This design permitted assessment of thermoregulatory effects of allosteric ligands compared to CB₁ orthosteric agonist over the interval spanning acute to chronic dosing.

2.9. Blinding

The investigator performing behavioral testing was blinded to treatment conditions during pharmacological manipulations and data collection to minimize bias. Respiratory parameters (DSI), measures of locomotor activity (Fusion 6.0), and cognition (Ethovision) were generated solely by commercially available equipment and software with no input from the experimenter. The threshold for paw withdrawal was revealed on a digital display and recorded by the investigator who was blinded to drug conditions. Because data collection was not subjective, the experimenter was not blinded during data analysis, and all data were analyzed using GraphPad Prism 10. 2-4 software with no experimental data excluded.

2.9.1. Data and Statistical Analysis

Data are expressed as mean ± SEM. Mechanical paw withdrawal thresholds, body temperature and respiratory data analyzed by Two-way analysis of variance (ANOVA), followed by Bonferroni’s multiple comparison post hoc test. In the case of significant main effects, Bonferroni’s post hoc test was used to compare all groups to each other and enable direct comparisons of effects of GAT228 and GAT229 to each other. In the case of significant interactions, Bonferroni’s multiple comparison post hoc test was used to compare each drug condition to vehicle to provide more statistical power by allowing a limited set of comparisons, thereby reducing type II error. Data collected with the electronic von Frey aesthesiometer was additionally log transformed at each time point for each subject given that Weber’s law suggests that mechanical sensation is perceived on a logarithmic scale; the same statistical approach as that performed with raw scores (thresholds in g) was employed. Two-way repeated measures of ANOVAs were performed separately on pre-injection and post-injection respiratory metrics. Data obtained from assessments of NOR, locomotor activity, and despair-like behavior were analyzed by One-way ANOVA, followed by either Dunnett’s or Tukey’s multiple comparison post hoc test (i.e., using the post hoc test recommended by GraphPad prism). All datasets were tested for normality using the Shapiro-Wilk test which documented homogeneity of variance. Data were analyzed using GraphPad Prism v 7.05 or GraphPad Prism 10.2-4 (GraphPad Software Boston MA). P < 0.05 was considered significant. All experimental data is provided.

2.9.2. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22 (Alexander et al., 2021).

3.0. Results

3.1. Experiment 1: CB₁ allosteric agonist GAT228 and CB₁ PAM GAT229 suppress mechanical hypersensitivity in a mouse model of CFA-induced inflammatory pain

Intraplantar CFA injection reduced mechanical paw withdrawal thresholds relative to baseline (pre-injection) thresholds prior to pharmacological manipulations in all studies (P < 0.0001). GAT228 increased mechanical paw withdrawal thresholds in the CFA-injected (ipsilateral) paw overall and in a time-dependent manner (Group: F_3,23=26.08, P < 0.0001, f = 1.676; Time: F_3,69=44.33, P < 0.0001, f = 1.371; Interaction: F_9,69=9.808, P < 0.0001, f = 1.160; n = 6 per group). The high dose (20 mg/kg) maximally increased paw withdrawal thresholds (p < 0.0001), the middle dose (10 mg/kg) trended towards efficacy (p = 0.0558), and the low dose (5 mg/kg) was inactive (p > 0.9999) relative to vehicle overall. GAT228 (20 mg/kg) also elevated paw withdrawal thresholds relative to each of the lower (10 mg/kg: p = 0.0002; 5 mg/kg: p < 0.0001; n = 6 per group) doses. The high dose increased paw withdrawal thresholds from 30-150 (p ≤ 0.0023) minutes post-injection compared to vehicle. The middle dose reliably increased ipsilateral paw withdrawal thresholds at 90 min (p < 0.0001) post-injection only compared to vehicle (Fig. 2A).

Figure 2.

CB₁ allosteric ligands GAT228 and GAT229 attenuated inflammatory pain induced by intraplantar CFA in male mice. (A) CB₁ allosteric agonist GAT228 (10 and 20 mg/kg, i.p.) increased mechanical paw withdrawal thresholds compared to vehicle in a dose- and time-dependent manner whereas GAT228 (5 mg/kg, ip) was inactive. (B) CB₁ allosteric modulator GAT229 (10 and 20 mg/kg, i.p.) increased mechanical paw withdrawal thresholds compared to vehicle in a dose- and time-dependent manner whereas GAT229 (5 mg/kg, i.p.) was inactive. (C) No differences between GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.) in suppressing CFA-induced inflammatory pain were detected. (D) The CB₁ antagonist/inverse agonist AM251 (5 mg/kg, i.p.) alone did not alter paw withdrawal thresholds in the CFA-injected paw compared to the vehicle-treated group. (E) Antinociceptive effects of the CB₁ allosteric ligands GAT228 (20 mg/kg, i.p.) (F) and GAT229 (20 mg/kg, i.p.) in the CFA-injected paw were blocked by CB₁ antagonist/inverse agonist AM251 (5 mg/kg, i.p.). Data were analyzed by Two-way ANOVA and Bonferroni’s post hoc tests. ⁺⁺⁺p < 0.001, ^^^p < 0.001, vs. vehicle for comparator identified by the symbol shown. Data are expressed as mean ± SEM. n = 5-10 C57/male mice per group.

Similarly, GAT229 (i.p.) increased paw withdrawal thresholds in the CFA-injected paw overall and in a time-dependent manner (Group: F_3,22=27.97, P < 0.0001, f = 1.737; Time: F_3,66=34.69, P < 0.0001, f = 1.333; Interaction: F_9,66=7.715, P < 0.0001, f = 1.107; n = 6 per group); both the high (20 mg/kg; p < 0.0001) and middle (10 mg/kg; p = 0.0006) doses increased paw withdrawal thresholds overall, whereas the low dose (5 mg/kg) was inactive (p > 0.9999). Effects of the high (20 mg/kg) and middle (10 mg/kg) doses of GAT229 did not differ from each other (p = 0.1022). The high dose of GAT229 increased ipsilateral paw withdrawal thresholds from 30-150 (p < 0.0001) minutes post-injection, whereas the middle dose elevated withdrawal thresholds from 90-150 minutes (p < 0.0001) post-injection, relative to vehicle (Fig. 2B).

Both GAT228 (20 mg/kg) and GAT229 (20 mg/kg) elevated ipsilateral paw withdrawal thresholds relative to vehicle overall and in a time-dependent manner (Group: F_2,19=31.50, P < 0.0001; Time: F_3,57=35.59, P < 0.0001; Interaction: F_6,57=9.557, P < 0.0001 d = 1.23); both ligands increased paw withdrawal thresholds relative to vehicle (p < 0.0001) but anti-allodynic efficacy did not differ from each other (p > 0.9999) overall (Fig. 2C).

The CB₁ antagonist/inverse agonist AM251 (5 mg/kg) did not alter paw withdrawal thresholds in the CFA-injected (ipsilateral) paw relative to vehicle (Group: F_1,14=2.646, P = 0.1261; Time: F_3,42=1.023, P = 0.3920; Interaction: F_3,42=1.169, P = 0.5895; n = 5 per group) (Fig. 2D) but reliably blocked anti-allodynic effects of GAT228 (Group: F_2,19 = 41.14, P < 0.0001, f = 1.89; Time F_3,57=14.30, P < 0.0001, f = 0.90; Interaction F_{6 57}=14.23, P < 0.0001, f = 1.28; n = 6 per group) and GAT229 (Group: F_{2 19}=30.81, P < 0.0001, f = 1.59; Time: F_3,57=13.34, P < 0.0001, f = 0.87; Interaction: F_6,57=9.308, P < 0.0001, f = 1.05; n = 6 per group). AM251 prevented GAT228- and GAT229-induced increases in paw withdrawal thresholds from 30-150 (p ≤ 0.0092) minutes postinjection. Withdrawal thresholds in the antagonist co-administration groups did not differ from vehicle (p > 0.9999) (Fig. 2E-F).

3.1.1. CB₁ allosteric agonist GAT228 and CB₁ PAM GAT229 suppress CFA-induced mechanical hypersensitivity without producing tolerance unlike the orthosteric agonist CP55,940

Repeated dosing (i.p.) with allosteric ligands GAT228 (20 mg/kg), GAT229 (20 mg/kg), and CP55,940 (0.3 mg/kg) reduced CFA-induced mechanical hypersensitivity overall but only the anti-allodynic effects of CP55,940 declined over time (Group: F_3,27=106.7, P < 0.0001, f = 1.640; Time: F_3,81=8.984, P < 0.0001, f = 1.718; Interaction: F_9,81=15.57, P < 0.0001, f = 1.640; n = 8 per group). All drug groups suppressed CFA-induced mechanical hypersensitivity relative to vehicle overall (p < 0.0001) whereas effects of GAT228 and GAT229, which did not differ from each other, each showed greater anti-allodynic efficacy than CP55,940 (p < 0.0001). CP55,940 displayed initial efficacy in reducing CFA-induced mechanical hypersensitivity (P < 0.001; day 1 vs. vehicle, d = −3.54). However, complete tolerance developed to the anti-allodynic efficacy of CP55,940 (p > 0.999, day 10 vs. vehicle; d = 0.01) by day 10 of repeated dosing, when effects of CP55,940 did not differ from vehicle. By contrast, both GAT228 (p < 0.0001, d = −2.64) and GAT229 (p < 0.0001, d = −2.80) exhibited initial anti-allodynic efficacy upon acute administration compared to vehicle and remained fully effective with no loss of antinociceptive efficacy throughout the 10-day chronic dosing period (mechanical: P < 0.0001; d = −3.05; −2.59 vs. vehicle for all days) (Fig. 3A).

Figure 3.

CB₁ allosteric ligands GAT228 and GAT229 attenuated CFA-induced inflammatory pain CFA in male mice with sustained efficacy. (A) Tolerance developed to anti-allodynic efficacy of orthosteric agonist CP55,940 (0.3 mg/kg/day x 10 days) but not GAT228 or GAT229 (each administered i.p at a dose of 20 mg/kg/day x 10 days). (B) No differences between CP55,940, GAT228 and GAT229 were detected in non-inflamed paw relative to vehicle. (C) CB₁/CB₂ agonist CP55,940 produced hypothermia relative to vehicle whereas GAT228 and GAT229 did not alter body temperature following acute or chronic dosing. Data were analyzed by Two-way ANOVA and Bonferroni’s post hoc tests. ⁺⁺⁺p < 0.001, ^^^p < 0.001, ****p < 0.0001, vs. vehicle for comparator identified by the symbol shown. Data are expressed as mean ± SEM. n = 8 C57/male mice per group.

Neither the allosteric ligands nor CP55,940 altered contralateral paw withdrawal thresholds at any time point relative to vehicle (Group: F_3,27=0.1000, P = 0.9593; Time: F_3,81=0.1305, P = 0.9417; Interaction: F_9,81=1.197, P = 0.30) (Fig. 3B). Similarly, none of the drug treatments altered paw withdrawal thresholds in the non-inflamed (contralateral) paw at any timepoint in any study (data not shown).

3.1.2. GAT228 and GAT229 did not reduce body temperature in CFA-inflamed mice unlike CP55,940

In the same CFA-injected mice used to assess anti-allodynic efficacy, acute but not chronic, treatment with CP55,940 reduced body temperature whereas allosteric ligands had no effect (Group: F_{3 27}=18.25, P < 0.0001; Time: F_4,108=16.06, P < 0.0001; Interaction: F_12,108=13.76, P < 0.0001 d = 0.87; n = 8 per group). CP55,940 reduced body temperature in CFA-injected mice overall compared to all other groups (p < 0.0001). CP55,940 reduced body temperature on day 1 and day 4 of repeated dosing compared to all other groups (p ≤ 0.0001). Neither GAT228 nor GAT229 altered body temperature compared to vehicle at any timepoint (p > 0.9999) (Fig 3C).

Experiment 2: The orthosteric CB₁/CB₂ agonist CP55,940 suppresses minute ventilation, respiratory frequency and tidal volume, indicative of respiratory depression, in non-inflamed and CFA-inflamed mice whereas GAT228 and GAT229 fail to do so

General results.

In all studies, respiratory parameters declined across time, as expected, irrespective of drug treatment (P < 0.0001 for all studies). Respiratory parameters did not differ between groups prior to pharmacological manipulations in any study. Consequently, results from statistical analyses detailed below were performed on the post-injection observation interval.

3.2.1. The orthosteric CB₁ /CB₂ agonist CP55,940 suppresses minute ventilation in non-inflamed and CFA-inflamed mice whereas GAT228 and GAT229 fail to do so

The orthosteric CB₁ /CB2 agonist CP55,940 suppressed minute ventilation in a dose- and time-dependent manner in otherwise naïve mice (Group: F_2,21=61.29, P < 0.0001, f = 4.40; Interaction: F_22,231=21.64, P < 0.0001, f = 1.44; n = 8 per group). Both the low (0.1 mg/kg, i.p.) and the high (0.3 mg/kg, i.p.) doses of CP55,940 suppressed minute ventilation relative to vehicle (p < 0.0001) with the high dose producing a greater suppressive effect than the low dose (p = 0.0005). The high CP55,940 dose suppressed minute ventilation as early as 5 min post-injection (p = 0.0428) and persisted from 10-60 min post-injection (p < 0.0001 for each time point). The low CP55,940 dose suppressed minute ventilation from 10-60 min post-injection (p ≤ 0.0153 at each time point) (Fig. 4A). Unlike CP55,940, neither GAT228 (20 mg/kg) (Group: F_1,14=0.4777, P = 0.5008, f = 0.27; Interaction: F_11,154=0.3214, P = 0.9803, f = 0.15; n = 8 per group) (Fig. 4B) nor GAT229 (20 mg/kg) (Group: F_1,14=3.086, P = 0.1008, f = 1.29; Interaction: F_11,154=1.082, P = 0.3794, f = 0.28; n = 8 per group) suppressed minute ventilation (Fig. 4C).

Figure 4.

CP55,940 suppresses minute volume in both naïve and CFA-inflamed mice, whereas GAT228 and GAT229 fail to do so. (A) Both high (0.3 mg/kg, i.p.) and low (0.1 mg/kg, i.p.) doses of CP55,940 suppressed minute volume in a dose- and time-dependent manner relative to vehicle in non-inflamed mice. Neither (B) GAT228 did not suppress minute volume compared with the vehicle-treated group in non-inflamed mice. (C) GAT229 (20 mg/kg, i.p.) did not suppress minute volume compared with the vehicle-treated group in non-inflamed mice. (D) CP55,940 suppressed minute volume in a time-dependent manner relative to vehicle in CFA-inflamed mice. (E) GAT228 did not suppress minute volume compared with the vehicle-treated group in CFA-inflamed mice. (F) GAT229 (20 mg/kg, i.p.) did not suppress minute volume compared with the vehicle-treated group in CFA-inflamed mice. No differences in minute volume were noted before the drug injection. Data were analyzed by Two-way ANOVA and Bonferroni’s post hoc tests. ****p < 0.0001, ***p < 0.001, **p < 0.01 vs. designated comparator. Data are expressed as mean ± SEM. n = 8 C57/male mice per group.

In CFA-inflamed mice, CP55,940 suppressed minute ventilation overall and in a time-dependent manner (Group: F_1,14=38.88, P < 0.0001, f = 5.25; Interaction: F_11,154=53.4, P < 0.0001, f = 1.96; n = 8 per group) compared with vehicle. CP55,940 suppressed minute ventilation as early as 15 min post-injection (p = 0.0016) and these suppressive effects were sustained from 20-60 min post-injection (p < 0.0001 for each time point) (Fig. 4D). Unlike CP55,940, neither GAT228 (20 mg/kg, i.p.; Group: F_1,14=2.059, P = 0.1733, f = 0.93; Interaction: F_11,154=1.469, P = 0.1484, f = 0.34; n = 8 per group) (Fig. 4E) nor GAT229 (20 mg/kg, i.p.; Group: F_1,14=4.451, P = 0.0534, f = 1.18; Interaction: F_11,154=1.624, P = 0.0969, f = 0.34; n = 8 per group) (Fig. 4F) reliably altered minute ventilation. GAT229 only trended to reduce minute ventilation relative to vehicle in CFA-injected mice, but no reliable differences in minute ventilation were observed between allosteric ligands or vehicle at any post-injection timepoint.

3.2.2. The orthosteric CB₁/CB₂ agonist CP55,940 reduced respiratory frequency in naïve and CFA-inflamed mice whereas GAT228 and GAT229 did not

In non-inflamed mice, CP55,940 dose-dependently reduced respiratory frequency across the observation interval and in a time-dependent manner (Group: F_2,21=63.23, P < 0.0001, f = 3.79; Interaction: F_22,231=21.81, P < 0.0001, f = 1.96; n = 8 per group). Both the high (0.3 mg/kg, i.p.) and the low (0.1 mg/kg, i.p.) doses of CP55,940 (p<0.0001) reduced respiratory frequency overall, with the high dose producing a greater suppressive effect than the low dose (p = 0.0001). High dose CP55,940 reduced respiratory frequency as early as 5 min post-injection (p = 0.00421), with suppression persisting across the observation interval (10-60 min: p < 0.0001 at each time point). Low dose CP55,940 similarly decreased respiratory frequency from 5-60 min post-injection (p ≤ 0.05 at each time point) (Fig. 5A). In contrast, neither GAT228 nor GAT229 suppressed respiratory frequency. GAT228 (20 mg/kg) did not alter respiratory frequency (Group: F_1,14=0.0006, P = 0.9803, f = 0.010; Interaction: F_11,154=0.2594, P = 0.9919, f = 0.14; n = 8 per group) (Fig. 5B). GAT229 (20 mg/kg) suppressed respiratory frequency in a time-dependent manner (Group: F_1,14=0.9587, P= 0.3441, f = 0.39; Interaction: F_11,154=2.465, P = 0.0072, f = 0.42; n = 8 per group) but post hoc analyses failed to reveal differences between GAT229 and vehicle-treated groups at any post-injection time point (p ≥ 0.2540) (Fig. 5C).

Figure 5.

CP55,940 suppresses respiratory frequency in both naïve and CFA-inflamed mice, whereas GAT228 and GAT229 fail to do so. (A) Both high (0.3 mg/kg, i.p.) and low (0.1 mg/kg, i.p.) doses of CP55,940 suppressed respiratory frequency in a dose- and time-dependent manner relative to vehicle in otherwise naïve mice. Neither (B) GAT228 (20 mg/kg, i.p.) nor (C) GAT229 (20 mg/kg, i.p.) reliably altered respiratory frequency compared with the vehicle-treated group in otherwise naïve mice. (D) CP55,940 suppressed respiratory frequency in a time-dependent manner relative to vehicle in CFA-inflamed mice. Neither (E) GAT228 (20 mg/kg, i.p.) nor (F) GAT229 (20 mg/kg, i.p.) altered respiratory frequency compared with the vehicle-treated group in CFA-inflamed mice. No differences in respiratory frequency were noted before the drug injection. Data were analyzed by Two-way ANOVA and Bonferroni’s post hoc tests. ****p < 0.0001, ***p < 0.001 vs. designated comparator. Data are expressed as mean ± SEM. n = 8 C57/male mice per group.

In CFA-inflamed mice, CP55,940 reduced respiratory frequency across the observation interval and in a time-dependent manner (Group: F_1,14=34.75, P < 0.0001, f = 3.99; Interaction: F_11,154=42.34, P < 0.0001, f = 1.74; n = 8 per group) compared with vehicle. CP55,940 decreased respiratory frequency in CFA-inflamed mice from 15-60 min post-injection (15 min: p = 0.0425; 20-60 min: p < 0.0001 at each time point) (Fig. 5D). In contrast, in CFA-inflamed mice, neither GAT228 (20 mg/kg, i.p.) (Group: F_1,14=0.02544, P = 0.8755, f = 0.06; Interaction: F_11,154=1.117, P = 0.3514, f = 0.28; n =8 per group) (Fig. 5E) nor GAT229 (20 mg/kg) (Group: F_1,14=0.4568, P = 0.5102, f = 0.29; Interaction: F_11,154=1.515, P = 0.1311, f = 0.33; n = 8 per group) (Fig. 5F) suppressed respiratory frequency compared with vehicle.

3.2.3. The CB₁/CB₂ agonist CP55,940 suppressed tidal volume in non-inflamed and CFA-inflamed mice whereas allosteric ligands (GAT228 and GAT229) did not

In non-inflamed mice, CP55,940 reduced tidal volume in a dose and time-dependent manner (Group: F_2,21=36.05, P < 0.0001, f = 3.65; Interaction: F_22,231=18.58, P < 0.0001, f = 1.33; n =8 per group). Both the low (0.1 mg/kg, i.p.; p = 0.0014) and the high (0.3 mg/kg, i.p.; p < 0.0001) doses of CP55,940 suppressed tidal volume relative to vehicle overall with the high dose producing a greater suppressive effect than the low dose (p = 0.0008). High dose CP55,940 reduced tidal volume at 10 min post-injection (p = 0.0063), which remained suppressed throughout the observation interval (p < 0.0001 at each time point). Low dose CP55,940 suppressed tidal volume from 20–60 min (p ≤ 0.0214) post-injection (Fig. 6A). Conversely, in non-inflamed mice neither GAT228 (Group: F_1,14=0.5663, P = 0.4642, f = 0.48; Interaction: F_11,154=0.5870, P= 0.8374, f = 0.20; n = 8 per group) (Fig. 6B) nor GAT229, (Group: F_1,14=3.055, P = 0.1024, f = 1.61; Interaction: F_11,154=0.7854, P= 0.6543, f = 0.24; n =8 per group) (Fig. 6C) altered tidal volume.

Figure 6.

CP55,940 suppresses tidal volume in both naïve and CFA-inflamed mice, whereas GAT228 and GAT229 fail to do so. (A) Both high (0.3 mg/kg, i.p.) and low (0.1 mg/kg, i.p.) doses of CP55,940 suppressed tidal volume in a dose- and time-dependent manner relative to vehicle in non-inflamed mice. Neither (B) GAT228 (20 mg/kg, i.p.) nor (C) GAT229 (20 mg/kg, i.p.) altered tidal volume compared with the vehicle-treated group in non-inflamed mice. (D) CP55,940 (0.3 mg/kg i.p.) suppressed tidal volume compared with the vehicle in CFA-inflamed mice. Neither (E) GAT228 (20 mg/kg, i.p.) nor (F) GAT229 (20 mg/kg, i.p.) altered tidal volume compared with the vehicle-treated group in CFA-inflamed mice. No differences in tidal volume were noted before the drug injection. Data were analyzed by Two-way ANOVA and Bonferroni’s post hoc tests. ****p < 0.0001, ***p < 0.001, **p < 0.01 vs. designated comparator. Data are expressed as mean ± SEM. n = 8 C57/male mice per group.

In CFA-inflamed mice, CP55,940 suppressed tidal volume overall and in a time-dependent manner (Group: F_1,14=25.36, P = 0.0002, f = 3.99; Interaction: F_11,154=37.95, P < 0.0001, f = 1.65; n = 8 per group) compared with vehicle. CP55,940 reduced tidal volume beginning at 15 min post-injection (p = 0.0497), which remained suppressed throughout the observation interval (p< 0.0001 at each time point) (Fig. 6D).

Conversely, in CFA-inflamed mice, neither GAT228 (20 mg/kg, i.p.) (Group: F_1,14=2.600, P= 0.1292, f = 1.27; Interaction: F_11,154=0.4953, P = 0.9040, f = 0.19; n = 8 per group), (Fig. 6E) nor GAT229 (20 mg/kg, i.p); (Group: F_1,14=2.438, P=0.1407, f = 1.31; Interaction: F_11,154=1.028, P = 0.4248, f = 0.27; n = 8 per group) altered tidal volume (Fig. 6F).

In non-inflamed mice, CP55,940 (0.1 and 0.3 mg/kg, i.p.) dose-dependently reduced inspiratory and expiratory flow whereas GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.) had no effect (Supplemental Fig. 3A-F). In non-inflamed mice, CP55,940 (0.1 and 0.3 mg/kg, i.p.) dose-dependently increased inspiratory and expiratory time, whereas GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.) had no effect (Supplemental Fig. 4A-F). In non-inflamed mice, CP55,940 (0.1 and 0.3 mg/kg, i.p.) dose-dependently increased inspiratory pause and expiratory pause whereas GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.) had no effect (Supplemental Figs. 5A-F).

In CFA-inflamed mice, CP55,940 (0.3 mg/kg, i.p.) reduced inspiratory flow whereas GAT228 (20 mg/kg, i.p.) failed to do so and GAT229 had only a modest effect (Supplemental Fig. 6A-B). GAT229 (20 mg/kg, i.p.) modestly but reliably suppressed post-injection inspiratory flow overall (P = 0.0285) compared to vehicle in CFA-inflamed mice (Supplemental Fig. 6C). Furthermore, CP55,940 reduced expiratory flow in CFA-inflamed mice, whereas GAT228 (20 mg/kg, i.p.) failed to do so (Supplemental Fig. 6D-E). GAT229 (20 mg/kg, i.p.) produced a modest but reliable suppression of inspiratory flow relative to vehicle overall, but the interaction was not significant (Supplemental Fig. 6F).

In CFA-inflamed mice, CP55,940 (0.3 mg/kg, i.p.) increased inspiratory time overall and in a time-dependent manner (Supplemental Fig. 7A). In CFA-inflamed mice, GAT228 (20 mg/kg, i.p.) was associated with modest time-dependent changes in inspiratory time, but post hoc comparisons failed to reveal any differences between groups at any time point (Supplemental Fig. 7B). In CFA-inflamed mice, GAT229 (20 mg/kg, i.p.) had no effect on inspiratory time (Supplemental Fig. 7C). In CFA-inflamed mice, CP55,940 also increased expiratory time, whereas GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.) failed to do so (Supplementary Figure 7A-C).

In CFA-inflamed mice, CP55,940 (0.3 mg/kg, i.p.) increased inspiratory pause and expiratory pause overall and in a time-dependent manner whereas GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.) had no effect (Supplemental Fig. 8A-F).

Experiment 3: The CB₁ agonist CP55,940 suppressed locomotor activity in non-inflamed and CFA-inflamed mice whereas GAT228 and GAT229 had no effect

In non-inflamed mice, CP55,940 (i.p.) reduced total distance traveled whereas GAT228 and GAT229 failed to do so (F_4,35=10.48, P < 0.0001, f = 1.1; n = 8 per group). The high (0.3 mg/kg) dose of CP55,940 reduced total distance traveled relative to i.p. treatment with vehicle, GAT228 (20 mg/kg) or GAT229 (20 mg/kg) whereas the low dose of CP55,940 (0.1 mg/kg) suppressed total distance travel relative to vehicle only (Fig. 7A). Similarly, high dose CP55,940 (0.3 mg/kg) reduced ambulatory activity (F_4,35=18.36, P < 0.0001, f = 1.45; n = 8 per group; Fig. 7B) relative to all other treatments, whereas low dose CP55,940 (0.1 mg/kg, i.p.) suppressed ambulatory activity relative to either vehicle or GAT229 treatment. High dose CP55,940 (0.3 mg/kg, i.p.) increased rest time (F_4,35=26.77, P < 0.0001, f = 1.75; Fig. 7C) relative to all other treatments, whereas low dose CP55,940 (0.1 mg/kg, i.p.) increased rest time compared to vehicle or GAT229 (20 mg/kg, i.p.) treatment. High dose CP55,940 (0.3 mg/kg) also reduced vertical activity (F_4,35=6.246, P < 0.001, f = 0.845Fig. 7D) relative to i.p. treatment with vehicle, GAT228 (20 mg/kg) or GAT229 (20 mg/kg) whereas low dose CP55,940 did not reliably alter vertical activity. Furthermore, high dose CP55,940 (0.3 mg/kg) also reduced horizontal activity (F_4,35=17.71, P < 0.0001, f = 1.42) relative to i.p. treatment with vehicle, GAT228 (20 mg/kg), GAT229 (20 mg/kg) or low dose (0.1 mg/kg) CP55,940 whereas low dose CP55,940 (0.1 mg/kg) reduced horizontal activity relative to either vehicle or GAT229 only (Fig. 7E). Effects of the allosteric ligands did not differ from either vehicle or each other for each metric shown in Fig. 7A-E. None of the ligands altered time spent in the center of the activity meter (F_4,35=0.135, P > 0.05, f = 0.124; Fig. 7F).

Figure 7.

The orthosteric CB₁/CB₂ agonist CP55,940 suppresses locomotor behavior in intact mice, whereas the CB₁ allosteric agonist GAT228 and CB₁ PAM GAT229 had no effect. (A) CP55,940 (0.3 mg.kg i.p.) dose-dependently reduced (A) total distance traveled and (B) ambulatory time and (C) increased resting time whereas GAT228 and GAT229 had no effect. (D) High dose CP55,940 (0.3 mg/kg, i.p.) reduced vertical activity relative to vehicle whereas GAT228 and GAT229 had no effect. (E) CP55,940 dose-dependently reduced horizontal compared to the vehicle-treated group whereas GAT228 and GAT229 had no effect. (F) No differences were observed in center time across groups. Data were analyzed by One-way ANOVA and Tukey’s post hoc tests. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.01 vs. designated comparator. Data are expressed as mean ± SEM. n = 8 C57BL6J/male mice per group.

In CFA-inflamed mice, CP55,940 (0.3 mg/kg, i.p.) reduced total distance traveled (F_3,28=16.99, P < 0.0001; d = 2.067; Fig. 8A), decreased ambulatory activity (F_3,28=23.94, P < 0.0001; d = 2.595; Fig. 8B) and increased rest time (F_3,28=68.44, P < 0.0001; d = 4.247; Fig. 8C) relative to treatment with vehicle, GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.). Allosteric ligands did not alter any of these metrics (Fig. 8A-C); effects of GAT228 and GAT229 did not differ from either vehicle or each other.

Figure 8.

The orthosteric CB₁/CB₂ agonist CP55,940 suppresses locomotor activity in CFA-inflamed mice, whereas GAT228 and GAT229 had no effect. CP55,940 reduced (A) total distance traveled and (B) ambulatory time in CFA-inflamed mice compared to all other groups whereas GAT228 and GAT229 had no effect. (A) CP55,940 increased resting time in CFA-inflamed mice compared to all other groups whereas GAT228 and GAT229 had no effect. (D) CP55,940 reduced vertical activity compared to all other groups in CFA-inflamed mice whereas vertical time was modestly lower in GAT228, and GAT229-treated groups compared to vehicle. (E) CP55,940 reduced horizontal activity compared to all other groups in CFA-inflamed mice whereas GAT228 and GAT229 had no effect. Data were analyzed by One-way ANOVA and Tukey’s post hoc tests. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.01 vs. designated comparator. Data are expressed as mean ± SEM. n = 8 C57BL6J/male mice per group.

CP55,940 (0.3 mg/kg) also reduced both vertical (F_3,28=39.44, P < 0.0001; d = 3.689; Fig. 8D) and horizontal activity (F_3,28=25.14, P < 0.0001; d = 2.660; Fig. 8E) compared to treatment with vehicle, GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg). Vertical activity was modestly lower in GAT228 (p = 0.0317; d = 1.037) and GAT229 (p = 0.0184; d = 1.118) treated groups relative to vehicle, and effects of the allosteric ligands did not differ from each other. None of the drug treatments reduced time spent in the center of the activity meter in CFA-injected mice (p > 0.05; Fig. 8F).

Experiment 4: CB₁ allosteric ligands GAT228 and GAT229 enhance memory performance in CFA-treated mice whereas the orthosteric agonist CP55,940 produces impairment

In CFA-inflamed mice, allosteric ligands increased the discrimination index in the NOR test whereas CP55,940 (0.1 mg/kg, i.p.) failed to do so (F_3,28=11.14, P < 0.0001; n = 8 per group) (Fig. 9A). Both GAT228 (20 mg/kg) (p = 0.0174; d = 1.126; n = 8 per group) and GAT229 (20 mg/kg) (p = 0.0076; d = 1.247; n = 8 per group) increased the discrimination index relative to vehicle. CP55,940 (0.1 mg/kg) did not reliably alter the discrimination index relative to vehicle in the same test (p = 0.6011; d = 0.442; n = 8 per group). In the same paradigm, only CP55,940 decreased time spent exploring the novel object (F_3,28=3.853, P = 0.0200) relative to vehicle (p< 0.05), whereas GAT228 and GAT229 had no effect (Fig. 9B). None of the ligands reliably altered total distance traveled in the NOR test (F_3,28=2.725, P = 0.0630) (Fig. 9C).

Figure 9.

The CB₁ allosteric agonist GAT228 and CB₁ PAM GAT229 enhance recognition memory in CFA-inflamed mice whereas the orthosteric CB₁/CB₂ agonist CP55,940 produced memory impairment. (A) GAT228 (20 mg/kg, i.p.) and GAT229 (20 mg/kg, i.p.) increased the discrimination index in the Novel Object Recognition (NOR) test in CFA-inflamed mice whereas CP55,940 (0.1 mg/kg, i.p.) did not. (B) CP55,940 decreased the time spent exploring the novel object in CFA-inflamed mice whereas GAT228 and GAT229 did not. (C) No difference was observed in total distance traveled between CP55,940, GAT228 and GAT229 and vehicle-treated CFA-inflamed mice in NOR test. Data were analyzed by One-way ANOVA and Tukey’s post hoc tests. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.01 vs. designated comparator. Data are expressed as mean ± SEM. n = 8 C57BL6J/male mice per group.

Experiment 5: CB₁ allosteric ligands GAT228 and GAT229 reduce pain-depressive-like behavior in CFA-treated mice, whereas the orthosteric agonist CP55,940 did not

All CFA-injected groups demonstrated robust mechanical hypersensitivity compared to saline-treated group prior to the evaluation of drug effects in the tail suspension test (i.e., on day 17 post CFA injection) (F₄₃₄=21.83, P < 0.0001; n = 7-8 per group) (Fig. 10A). Intraplantar CFA increased immobility time in the tail suspension test compared to intraplantar saline treatment and this effect was normalized by allosteric ligands but exacerbated by CP55,940 (F_4,34=36.82, P < 0.0001 f =2.07; n = 7-8 per group). Immobility time was higher in intraplantar CFA- compared to intraplantar saline-treated mice that each received i.p. vehicle (p = 0.0077). Immobility time was higher in CFA-inflamed mice receiving CP55,940 (0.3 mg/kg, i.p.) compared to either the intraplantar saline- vehicle (i.p.) group (p < 0.0001) or CFA-inflamed groups that received either GAT228 (20 mg/kg, i.p.), GAT229 (20 mg/kg, i.p.) or vehicle (p < 0.0001). Immobility time was lower in CFA-inflamed groups that received GAT228 (20 mg/kg; p = 0.0105) or GAT229 (20 mg/kg; p = 0.0053) compared to CFA-inflamed groups that received vehicle (i.p.) treatment. Notably, effects of the allosteric ligands did not differ from saline (i.pl.) groups that received vehicle (i.p.) and effects of the allosteric ligands did not differ from each other (Fig. 10B).

Figure 10.

(A) CFA-treated mice show robust mechanical hypersensitivity before drug treatment on day 17 post CFA- injection compared to saline-treated group. (B) CB₁ allosteric agonist GAT228 (20 mg/kg, i.p.) and CB₁ PAM GAT229 (20 mg/kg, i.p.) reduced CFA-induced pain-depressive-like behavior whereas CP55,940 (0.3 mg/kg, i.p.) exacerbates this effect. Data were analyzed by One-way ANOVA and Tukey’s post hoc tests. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.01 vs. designated comparator. Data are expressed as mean ± SEM. n = 7- 8 C57BL6J/male mice per group.

A summary of the impact of the CB₁/CB₂ agonist CP55,940 and allosteric ligands GAT228 and GAT229 on all dependent measures is provided in Table 1.

Table 1. Summary of In Vivo Effects of CB₁ Allosteric Ligands and CB₁/CB₂ Orthosteric Agonist CP55,940.

Dependent Measure	Orthosteric Agonist CP55,940	CB1 ago-PAM GAT228 (R)	CB1 PAM GAT229 (S)
Therapeutic Efficacy
Anti-allodynic Efficacy in CFA model (acute)	Yes	Yes	Yes
Anti-allodynic Efficacy in CFA model (chronic)	Absent or Decreased*	Yes	Yes
Unwanted Side Effects
Hypothermia (acute)	Yes	No	No
Hypothermia (chronic)	Absent or Decreased*	No	No
Locomotor impairment (non-inflamed mice)	Yes	No	No
Locomotor impairment (CFA-inflamed mice)	Yes	No	No
NOR Memory	Decreased	Increased	Increased
Immobility in TST (CFA-inflamed mice)	Increased+	Decreased#	Decreased#
Major Respiratory Parameters (Non-inflamed mice)
Minute Volume	Decreased	No Effect	No Effect
Respiratory Frequency	Decreased	No Effect	No Effect
Tidal Volume	Decreased	No Effect	No Effect
Minor Respiratory Parameters
Inspiratory Flow	Decreased	No Effect	No Effect
Expiratory Flow	Decreased	No Effect	No Effect
Inspiratory Time	Increased	No Effect	No Effect
Expiratory Time	Increased	No Effect	No Effect
Inspiratory Pause	Increased	No Effect	No Effect
Expiratory Pause	Increased	No Effect	No Effect
Major Respiratory Parameters (CFA-inflamed Mice)
Minute Volume	Decreased	No Effect	No Effect
Respiratory Frequency	Decreased	No Effect	No Effect
Tidal Volume	Decreased	No Effect	No Effect
Minor Respiratory Parameters
Inspiratory Flow	Decreased	No Effect	Decreased
Expiratory Flow	Decreased	No Effect	Decreased
Inspiratory Time	Increased	No Effect	No Effect
Expiratory Time	Increased	No Effect	No Effect
Inspiratory Pause	Increased	No Effect	No Effect
Expiratory Pause	Increased	No Effect	No Effect

Abbreviations: CFA, complete Freund’s adjuvant; NOR, Novel Object Recognition; TST, Tail suspension test

^*Indicative of tolerance

⁺Indicative of pain-related behavioral despair

^#Normalized to non-inflamed (i.e, intraplantar saline injected) control

4.0. Discussion and Conclusion

This report provides the first comprehensive in vivo characterization comparing efficacy and side effect profiles of the CB₁ allosteric agonist GAT228 and PAM GAT229 and the orthosteric CB₁/CB₂ agonist CP55,940. Both allosteric ligands exhibited strikingly similar profiles of efficacy in attenuating inflammatory pain without producing the unwanted CB₁-mediated cannabimimetic effects. The allosteric ligands suppressed CFA-induced inflammatory pain but did not produce tolerance, hypothermia, memory deficit, motor impairment or prominent respiratory side effects associated with CP55,940. The same pattern of results and conclusions were obtained from statistical analyses based upon raw scores (i.e., mechanical thresholds in g) and logarithmically transformed data. Thus, differences between the pure CB₁ allosteric agonist and the CB₁ PAM were not unmasked by logarithmic transformation of data, given that Weber’s law suggests mechanical sensation is perceived on a logarithmic scale. Both raw data and logarithmically transformed data are provided in this report to facilitate comparisons with effects of GAT211 and other reference ligands on paw withdrawal thresholds (in g) in our previously published work (Slivicki et al., 2018). Furthermore, only the allosteric ligands suppressed pain-related despair-like behavior and improved novel object recognition memory.

GAT228 and GAT229 suppressed mechanical hypersensitivity in the CFA-inflamed paw without affecting responding in the non-inflamed paw, consistent with prior observations with GAT211 (Slivicki et al., 2018). Anti-allodynic efficacy was CB₁-dependent; effects of each ligand were blocked by the CB₁ receptor antagonist/inverse agonist AM251. Off-target effects of AM251 observed in vitro (Henstridge et al., 2010; Seely et al 2012) are unlikely to explain the pattern of in vivo pharmacological specificity reported here as anti-allodynic efficacy of GAT211 in the CFA model is absent in CB₁ KO mice (Slivicki et al., 2018). CB₁-mediated antinociceptive actions of these ligands are also observed in corneal pain models (Thapa et al., 2020). Nonetheless, our experiments do not preclude the possibility that a CB₁ neutral antagonist (e.g, AM4113) could unmask subtle mechanistic differences between the two enantiomers in vitro and/or in vivo.

ZCZ011 and GAT211 are well-characterized ago-PAMs based upon in vitro activities (Mitjavila et al., 2018; Lapraire et al., 2017). As GAT228 is reported to be a pure CB₁ allosteric agonist with no PAM activity, it could theoretically indirectly activate CB₁ via an allosteric site to exert antinociceptive effects without requiring endogenous ligand engagement. GAT229, as a pure PAM, would presumably enhance the efficacy of endocannabinoid signaling only under conditions in which the endogenous ligand is already mobilized. Yet, in our studies both GAT228 and GAT229 attenuated CFA-induced mechanical allodynia with similar efficacies in vivo. It is plausible that in vitro profiles of pure PAM vs. ago-PAM activity may not translate in vivo in the behavioral assays evaluated herein. Convergence may be explained by shared engagement of CB₁ receptor activity within pain-relevant circuits where endocannabinoid tone is elevated (Jayamanne et al., 2006). CB₁ allosteric ligands can display context- and cell-type dependent signaling behaviors in vivo, which may not fully align with their in vitro pharmacological profiles (Shen et al., 2024). We acknowledge that differential interactions with orthosteric ligands in vivo may reveal differences in the allosteric ligands.

GAT229 functions as a PAM for both G-protein (cAMP inhibition) and β-arrestin₂ pathways but exhibits a model bias (2-fold) favoring cAMP, without intrinsic efficacy in either pathway (Laprairie et al., 2017). GAT228 lacks PAM activity (cAMP), shows minimal effect on arrestin recruitment, and acts as a direct allosteric agonist with low intrinsic efficacy in both pathways (Laprairie et al., 2017). Ligand-specific signaling bias, including G-protein vs. β-arrestin pathway preferences, may not always translate directly to distinct behavioral outcomes (Shen et al., 2024). These divergent profiles complicate attribution of reduced adverse effects of allosteric ligands to biased signaling alone. GAT228 could potentially act as a context-sensitive partial agonist or ago-PAM under conditions of enhanced endocannabinoid signaling (Thapa et al., 2020). Paucity of adverse side effects observed with allosteric ligands may be explained by their partial activation of CB₁, leading to more localized or pathway limited signaling. This hypothesis is consistent with emerging discussions in GPCR pharmacology where partial agonism, rather than signaling bias alone, accounts for improved therapeutic windows such as in the case of G-protein biased MOR agonists (Manglik et al., 2020; Gillis et al., 2020). These findings underscore the importance of studying CB₁ allosteric ligands in vivo, where their pharmacodynamic behavior may be influenced by receptor localization, endogenous tone, and signal compartmentalization. Intraplantar CFA, by altering endocannabinoid tone, could potentially equalize in vivo efficacy of allosteric ligands. These mechanisms may explain overlapping in vivo profiles of GAT228 and GAT229 despite different in vitro signaling profiles.

In CFA-inflamed mice, anti-allodynic efficacy of GAT228 and GAT229 was maintained following repeated dosing without the development of tolerance or emergence of hypothermia. The lack of effect of the allosteric ligands on the non-inflamed (contralateral) paw further suggests that acute pain sensation is spared by allosteric modulation/agonism, whilst sustained/acute inflammatory pain behavior is alleviated, and seemingly without the development of significant tolerance. This is the target of any therapeutic pharmaceutical intervention for the treatment of pain. By contrast, tolerance developed to both anti-allodynic and hypothermic effects of CP55,940 in the same paradigm. Tolerance also developed to CP55,940 and THC, but not GAT211, in a paclitaxel-induced neuropathy model (Slivicki et al., 2018; Slivicki et al., 2020; Deng et al., 2015a; 2015b). These observations support possible long-term suitability of CB₁ allosteric ligands as analgesics compared to CB₁ agonists (Deng et al., 2015b). Absence of tolerance may reflect a lack of CB₁ desensitization or downregulation with chronic dosing and/or enhanced endocannabinoid tone and signaling following CFA injection. Previous findings suggest that allosteric ligands can preserve CB₁-dependent responsiveness without inducing functional tolerance (Slivicki et al., 2018; Ignatowska et al., 2015).

Narcotic analgesics and CB₁ agonists produce respiratory depression (Watkins et al., 2025; 2026). No treatment for cannabinoid overdose-induced respiratory depression exists (Alon and Saint-Fleur 2017). Here we replicate the phenomenon of CB₁-mediated respiratory depression (Watkins et al., 2025). GAT228 and GAT229 did not alter the key metrics used to measure respiratory depression (i.e., minute ventilation, respiratory frequency, or tidal volume). CP55,940 suppressed all these respiratory metrics. Unlike CP55,940, GAT228 failed to alter inspiratory flow, expiratory flow, inspiratory time, expiratory time, in either the absence or presence of CFA. GAT229 did not alter any respiratory parameter in non-inflamed mice, although GAT229, but not GAT228, produce a minor suppression of respiratory effort (i.e., inspiratory and expiratory flow) in CFA-inflamed mice. This observation suggests possible compensatory changes in the characteristics of individual breaths rather than time between individual breaths. One caveat should be noted. It is necessary to remove mice from whole body plethysmography chambers to perform i.p. injections, which necessarily re-exposes mice to room air, and may impact post-injection inspiratory and expiratory parameters. More work is necessary to determine whether the observed effect of GAT229 on inspiratory and expiratory time is physiologically relevant or supports superiority of GAT228 over GAT229 as an analgesic intervention. Our CB₁ allosteric ligands provide a more circumscribed and beneficial profile compared to either CB₁ orthosteric or MOR agonists (Slivicki et al., 2020). CB₁ (ago)PAMs could potentially represent safe and effective pharmacotherapies or adjuvant analgesics in future clinical trials.

GAT211 did not produce catalepsy or impair motor coordination in prior studies employing otherwise normal animals (Slivicki et al., 2018). However, differences in side effect profiles could potentially emerge in states associated with elevated endocannabinoid tone. It is, consequently, noteworthy that GAT228 and GAT229 did not produce motor impairment, unlike CP55,940, in either non-inflamed or CFA-inflamed mice. Decreased vertical activity (e.g., rearing) may reflect anxiolytic effects of allosteric ligands given absence of other signs of motor impairment. The separation of anti-allodynic efficacy from nonspecific behavioral suppression with CB₁ allosteric ligands enables dissection of analgesic efficacy from motor suppressive effects that could otherwise confound interpretation of behavioral studies measuring reflexes or motor-based outputs.

One limitation of our study is that we did not test efficacy or side effect profiles of our allosteric ligands in female mice. Use of male mice facilitated broad behavioral profiling and maintained continuity with prior studies of CB₁ PAMs (Slivicki et al., 2018). CB₁/CB₂ orthosteric agonists CP55,940 and THC produce robust respiratory suppression in both male and female mice (Watkins et al., 2025) without apparent sex differences. Future studies will specifically assess female cohorts. Even in saline-treated or non-injected mice, a minor decrease in respiratory parameters is observed relative to baseline, after animals are removed from respiratory chambers containing CO₂ and re-exposed to room air during i.p. injections. Locomotor activity also declines over time within the confines of the respiratory chambers (Watkins et al., 2025). Modest rundown of respiratory parameters over time and/or vehicle effects are unlikely to mask detection of drug effects as analysis of post-injection data failed to reveal any significant effects of allosteric ligands (in either a main effect or interaction) on major respiratory parameters (i.e., minute ventilation, respiratory frequency, tidal volume).

Preclinical and clinical studies link memory deficits with chronic pain, including deficits in attention, recognition memory and executive planning (Gasper et al., 2021; Zhang et al., 2021). Here, CB₁ allosteric ligands enhanced memory function in NOR test whereas time spent exploring the novel object was reduced by a dose of CP55,940 (0.1 mg/kg, i.p.), that did not itself reduce locomotor activity. Without a non-inflamed control group, it is not possible to conclude that allosteric ligands reduced CFA-induced memory deficits. Given the established role of the endocannabinoid system in stress regulation and the biphasic effects of cannabinoid signaling (Morena et al., 2016), it is possible that direct CB₁ agonism by CP55,940 produced anxiogenic-like effects under these conditions, whereas allosteric modulation enhanced endogenous signaling in a manner consistent with anxiolysis; such factors could influence interpretation of object familiarity in the NOR test. Allosteric ligands may exhibit a distinct mechanistic profile from CB₁ orthosteric agonists, enhancing CB₁ activity in the context of inflammatory pain in a manner distinct from that activated by orthosteric CB₁ agonists (Leepakshi et al., 2017).

Chronic inflammatory pain induced by CFA is associated with affective disturbance including despair-like behavior, which reflects the comorbid depression-phenotype often seen in chronic pain (Wei et al., 2025). A systematic review and meta-analysis of preclinical studies employing the CFA model revealed that CFA heightens immobility in the tail suspension test (Burek et al., 2022). Our findings showed that CB₁ allosteric ligands GAT228 and GAT229 reduced immobility in the tail suspension test, suggesting an antidepressant-like effect in mice three weeks following intraplantar CFA, whereas the CB₁/CB₂ full agonist CP55,940 either exacerbated this effect or blunted motivational behavior. The differential effects observed between the orthosteric cannabinoid agonist and CB₁ allosteric ligands likely stem from their distinct pharmacological actions. CB₁ allosteric ligands enhances endocannabinoid signaling in a context dependent manner, promoting balanced CB₁ activation when endocannabinoids are present or elevated (Toczek et al., 2018; Ignatowska et al., 2015; Lapraire et al., 2017) while the orthosteric cannabinoid agonist produces a global activation leading to CB₁ desensitization and may reduce the dopaminergic transmission that manifests as depressive-like behavior. Given that CFA alters endocannabinoid tone and alters CB₁ receptor signaling, CB₁ allosteric ligands GAT228 and GAT229 may restore endocannabinoid balance and improve affective outcomes without the adverse motivational effects seen with CB₁ agonists. CP55,940 is notably a full agonist at both CB₁ and CB₂ receptors. Although the acute antinociceptive effects of CP55,940 are absent in CB₁ knockout mice and preserved in CB₂ knockout mice in the CFA model (Sain et al., 2009; Slivicki et al., 2018), CB₂ receptor expression is known to increase under inflammatory conditions. Therefore, contribution of a CB₂-mediated component to the behavioral effects of CP55940 observed at later time points (e.g., in the tail-suspension test), cannot be excluded.

Our studies collectively support the potential of CB₁ allosteric ligands, specifically GAT228 and GAT229 as analgesic agents with absent or minimal adverse effects. In addition to suppressing CFA-induced mechanical sensitivity, both allosteric ligands improved recognition memory in NOR and reduced pain-related depression-like behavior, favorable therapeutic effects not produced by the orthosteric CB₁/CB₂ agonist CP55,940. The conserved nature of CB₁ expression and signaling across mammalian species suggests that the pharmacodynamic properties of these ligands may generalize to higher-order species, including humans (Stincic et al., 2008; Parnell et al., 2023). However, extrapolation to human biology must be approached cautiously given known interspecies differences in cannabinoid pharmacokinetics and endocannabinoid system tone. Moreover, possible sex-specific aspects of allosteric ligand efficacy and side effect profiles remain to be explored. These limitations notwithstanding, our findings support the continued development of CB₁ allosteric modulators as a promising analgesic strategy, with the potential to overcome limitations of the current CB₁ full or partial agonists and opioid analgesics.

Bullet Point Summary.

What is already known?

• Activation of CB₁ cannabinoid receptors suppresses pathological pain but also produces undesirable side effects

• CB₁ positive allosteric modulators (PAMs) enhance signaling without adverse cannabimimetic effects of orthosteric CB₁ agonists

What this study adds?

• CB₁ allosteric agonist (GAT228) and PAM (GAT229) alleviate inflammation-induced mechanical hypersensitivity without producing tolerance

• Allosteric ligands spared motor and memory impairment and respiratory depression associated with CB₁/CB₂ orthosteric agonist

Clinical significance

• Preclinically, CB₁ allosteric ligands offer a safer therapeutic approach than CB₁ agonists

ABBREVIATION

BL

Baseline

f

Cohen f

d

Cohen d

CB

Cannabinoid

SEM

Standard error mean

CB₁

Cannabinoid receptor 1

CB₂

Cannabinoid receptor 2

PAM

Positive allosteric modulator

WBP

Whole body plethysmography

AM

Activity Meter

Temp

Temperature

VF

Vonfrey

TST

Tail suspension test

PWT

Paw withdrawal threshold

CFA

Complete Freund adjuvant

Data availability statement

The data that support the findings of this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.20162608