The Mitragyna speciosa (kratom) alkaloid mitragynine: Analysis of adrenergic α₂ receptor activity in vitro and in vivo

Abstract

Mitragynine, an alkaloid present in the leaves of Mitragyna speciosa (kratom), has a complex pharmacology that includes low efficacy agonism at μ-opioid receptors (MORs). This study examined the activity of mitragynine at adrenergic α₂ receptors (Aα₂Rs) in vitro and in vivo. Mitragynine displaced a radiolabeled Aα₂R antagonist ([³H] RX821002) from human Aα_2ARs in vitro with lower affinity (K_i = 1260 nM) than the agonists (−)-epinephrine (K_i = 263 nM) or lofexidine (K_i = 7.42 nM). Mitragynine did not significantly stimulate [³⁵S]GTPγS binding at Aα_2ARs in vitro, but in rats trained to discriminate 32 mg/kg mitragynine from vehicle (intraperitoneally administered; i.p.), mitragynine exerted an Aα₂R agonist-like effect. Both α₂R antagonists (atipamezole and yohimbine) and MOR antagonists (naloxone and naltrexone) produced rightward shifts in mitragynine discrimination dose-effect function and Aα₂R agonists lofexidine and clonidine produced leftward shifts. In the mitragynine trained rats, Aα₂R agonists also produced leftward shifts in discrimination dose-effect functions for morphine and fentanyl. In a separate rat cohort trained to discriminate 3.2 mg/kg i.p. morphine from vehicle, naltrexone produced a rightward shift, but neither an Aα₂R agonist or antagonist affected morphine discrimination. In a hypothermia assay, both lofexidine and clonidine produced marked effects antagonized by yohimbine. Mitragynine did not produce hypothermia. Together, these data demonstrate that mitragynine acts in vivo like an Aα₂R agonist, although its failure to induce hypothermia or stimulate [³⁵S]GTPγS binding in vitro, suggests that mitragynine maybe a low efficacy Aα₂R agonist.

1. Introduction

Opioid overdose is a leading cause of death for individuals under age 50 in the United States (US) and has decreased US life expectancy (Butelman et al., 2023; Crimmins and Zhang, 2019). Current medications approved by the US Food and Drug Administration (FDA) for opioid use disorder (OUD) fall into two pharmacological categories: μ-opioid receptor (MOR) and adrenergic-α₂ receptor (Aα₂R) ligands. Methadone and buprenorphine (MOR agonists) are both approved to treat opioid withdrawal symptoms and used to prevent relapse, although these MOR agonists have abuse liabilities (Hiranita et al., 2014; Timko et al., 2016). The Aα₂R agonist lofexidine, is indicated for short-term use to alleviate opioid withdrawal symptoms during the initial phase of detoxification. In general, Aα₂R agonists including off label use of clonidine are effective treatments of opioid withdrawal symptoms (Doughty et al., 2019; Gold et al., 1979).

Mitragyna speciosa (kratom), a plant native to Southeast Asia, is used as a short- or longer-term substitute for opioids (Garcia-Romeu et al., 2020; Singh et al., 2022; Vicknasingam et al., 2010). Mitragynine is the most abundant kratom alkaloid, i.e., ~40% of the total alkaloid content (Chear et al., 2021; Hassan et al., 2013), is a low efficacy MOR agonist (Váradi et al., 2016), and produces antinociception in mice (Macko et al., 1972) that is blocked by the MOR antagonist naloxone (Matsumoto et al., 1996b) and is absent in transgenic mice lacking MORs (Kruegel et al., 2019). In rats discriminating morphine, naltrexone blocks the capacity of mitragynine to exert morphine-like discriminative stimulus effects (Harun et al., 2015; Obeng et al., 2021) but not mitragynine-induced decreases in food-maintained responding; suggesting that some of mitragynine’s effects are not MOR mediated (Hiranita et al., 2019; Obeng et al., 2021).

In vitro, mitragynine displaces tritiated Aα₂R antagonists with a 10-fold lower affinity than its affinity for MORs (Ellis et al., 2020; Obeng et al., 2020). In vivo, the antinociceptive and anti-allodynic effects of mitragynine are blocked by Aα₂R antagonists (Farkas et al., 2022; Foss et al., 2020; Matsumoto et al., 1996a) suggesting that mitragynine might be an Aα₂R agonist. These preclinical data raise the possibility that mitragynine exerts agonist activity not only at MORs but also at Aα₂Rs, thereby representing a novel and valuable pharmacological profile for treatment of OUD. The current study determined the in vitro pharmacology of mitragynine at Aα₂R with affinity and intrinsic efficacy assessments using competition binding and [³⁵S]GTPγS binding assays. Sprague-Dawley rats were trained to discriminate mitragynine and we determined whether rats could distinguish between mitragynine and the Aα₂R agonist lofexidine (alone and in combination with mitragynine) and examined the effects of the Aα₂R selective antagonist atipamezole on mitragynine discrimination. For comparison we examined whether morphine or the MOR antagonist naltrexone substituted for mitragynine. A control cohort of rats were also trained to discriminate a dose of morphine; the effects of Aα₂R ligands on morphine discrimination were examined. Finally, the hypothermic and antinociceptive effects of Aα₂R and MOR agonism were assessed (Minor et al., 1989; Obeng et al., 2022).

2. Materials and methods

2.1. Compounds

Salt and enantiomeric forms of the following drugs used were: [³H] RX821002 (PerkinElmer, Boston, MA), [³⁵S]GTPγS (PerkinElmer), atipamezole hydrochloride (HCl: Antisedan^®, Zoetis U.S., Durham, NC), clonidine HCl (Duraclon^®, Mylan N.V., Canonsburg, PA), (−)-epinephrine (Sigma-Aldrich Co., St. Louis, MO), fentanyl HCl (National Institute on Drug Abuse, Drug Supply Program, Rockville, MD), lofexidine HCl (Sigma-Aldrich Co.), (−)-mitragynine HCl [extracted as published (Maxwell et al., 2020)], (−)-morphine sulfate (National Institute on Drug Abuse), (−)-naltrexone HCl (Sigma-Aldrich Co.), (−)-naloxone HCl dihydrate (Sigma-Aldrich Co.), and yohimbine HCl (Sigma-Aldrich Co.). (−)-Mitragynine HCl was available in our alkaloid library and was isolated and structural elucidated through ¹H NMR, ¹³C NMR, and HRMS using Bruker model AMX 500 and Avance NEO 600 NMR spectrometers operating at 500 and 600 MHz in ¹H and 126 and 151 MHz in ¹³C, respectively. HRMS and purity (≥95%) were determined using an Agilent 1290 Infinity series ultraperformance liquid chromatography (UPLC) system equipped with photodiode array detector and quadrupole-time-of-flight (QTOF) Agilent 6540 mass spectrometer. We have previously reported all spectra data for (−)-mitragynine HCl (Obeng et al., 2020). Doses/concentrations are expressed as the weight of the salt form listed above, or as a base if no salt form is noted. For the in vitro studies, compounds were dissolved in dimethyl sulfoxide (Sigma-Aldrich Co.) to form 10 mM stock concentrations. For the in vivo studies, the vehicle consisted of sterile water containing 5% Tween 80 (polyoxyethylenesorbitanmonooleate, Sigma-Aldrich Co.) and 5% propylene glycol (Sigma-Aldrich Co.); each solution was filtered with a 0.2-μm pore size syringe filter (Millex-LG, 0.20 μm, SLLG025SS), and compound doses and vehicle were administered intraperitoneally (i.p.) in a volume of 1.0 mL/kg of body weight except mitragynine, which was prepared in volumes of 1.0–10 mL/kg due to limited solubility. Doses and pretreatment times were chosen from previous studies (Jalal et al., 2018; Obeng et al., 2020, 2021).

2.2. In vitro procedures

Aα_2AR was selected for affinity and efficacy assays because it is the major Aα₂R subtype in the brain (Bücheler et al., 2002).

2.2.1. In vitro receptor binding procedure

Monoclonal Chinese hamster ovary (CHO)–K1 cells expressed human Aα_2AR (PerkinElmer). The Bradford protein assay was used to determine and adjust the concentration of protein required for the assay (Tal et al., 1985). The K_d (mean ± SEM) and B_max (mean ± SEM) values for [³H] RX821002 were determined using a saturation assay as 0.433 ± 0.0140 nM and 6.75 ± 0.371 pmol/mg, respectively, which were comparable to previously reported values (O’Rourke et al., 1994). Ten μg of Aα_2AR membrane protein was incubated with 1.0 nM [³H]RX821002 in the presence of different concentrations of test compounds in TME [50 mM Tris (Sigma-Aldrich Co.), 3.0 mM MgCl₂ (Sigma-Aldrich Co.), and 0.2 mM ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA, Sigma-Aldrich Co.), pH 7.7] buffer for 60 min at room temperature. The bound radioligand was separated by filtration using the Connectorate filtermat harvester for 96-well microplates (Inotech, Dietikon, Switzerland) and counted for radioactivity using a Microbeta2 microplate counter (PerkinElmer). Specific Aα_2AR binding was determined as the difference in binding obtained in the absence and presence of 10 μM lofexidine. The assay was conducted in triplicate and repeated at least three times.

2.2.2. [³⁵S]GTPγS functional assay

A [³⁵S]GTPγS functional assay (Harrison and Traynor, 2003) was employed to compare the intrinsic activity of mitragynine to Aα_2AR ligands. Twenty μg of the human Aα_2A-CHO-K1 membrane protein was incubated with 10 μM guanosine diphosphate (GDP), 0.1 nM [³⁵S] GTPγS, and varying concentrations of test compounds for 90 min at room temperature. To determine antagonism at the Aα_2AR, test compounds were incubated together with 3.0 μM (−)-epinephrine. Nonspecific binding was determined using 40 μM unlabeled GTPγS. TME buffer (50 mM Tris-HCl, 9.0 mM MgCl₂, 0.2 mM EGTA, pH 7.4) with 150 mM NaCl and 0.14% bovine serum albumin (BSA) was used to increase agonist-stimulated binding; the final volume in each well was 300 μL. Ten μM of (−)-epinephrine was included in the assay as the maximum effective concentration at the Aα_2AR. After incubation, the bound radioligand was separated from the free radioligand by filtration through a GF/B glass fiber filter paper and rinsed three times with ice-cold wash buffer (50 mM Tris-HCl, pH 7.2) using the Unifilter-96 cell harvester (PerkinElmer). Radioactivity was measured using the Microbeta2 microplate scintillation counter. All assays were determined in triplicate and repeated at least three times.

2.3. In vivo procedures

2.3.1. Animals

Adult female and adult male Sprague Dawley rats (Taconic Biosciences, Germantown, NY, N = 8; 4 females and 4 males), weighing approximately 250 and 300 g upon arrival, respectively, were housed individually in a temperature- (mean ± SEM: 21.9 ± 1.9 °C) and humidity-controlled (mean ± SEM: 53 ± 14 %) vivarium with a 12-h light/dark cycle (lights on at Eastern Daylight Time 07:00 h). Reverse osmosis-purified water was available in the vivarium at all times. Free-feeding body weights were determined whilst each subject had continuous access to 2918 Teklad global 18% protein rodent diets (Envigo, Frenchtown, NJ). Weights were maintained at ≥85% of free-feeding weight by adjusting the amount of food (Dustless Precision Pellets Grain-Based Rodent Diet, Bio-Serv, Frenchtown, NJ) provided in the home cages 30 min following daily experimental sessions, in addition to the 45-mg sucrose pellets (maximum 50) that could be obtained during operant conditioning sessions. The Body Conditioning Score (Ullman-Culleré and Foltz, 1999) was maintained at ≥2.5 for each subject. Studies were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Florida (Protocol number: 201, 810,396) and were in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals, which was fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC).

2.3.2. Apparatus for in vivo studies

2.3.2.1. Hotplate.

A 1440 Analgesia Hot Plate with RS-232 Port and Software (Columbus Instruments, Columbus, OH) was surrounded by a clear acrylic cubicle with a lid. Temperature on the plate surface was maintained at 52 ± 0.1 °C.

2.3.2.2. Rectal thermometer.

An uninsulated microprobe (50313 Rat Rectal Probe, Stoelting, Wood Dale, IL) and a digital thermometer (50315 Body Temperature Thermometer, Stoelting) were used to measure rectal temperature. Veterinary ophthalmic ointment (Puralube^®, Dechra Veterinary Products, Overland Park, KS) was applied to the tip of the microprobe prior to each use.

2.3.2.3. Drug discrimination.

Each operant-conditioning chamber (Model ENV-022MD; Med Associates Inc., Fairfax, VT) measured 45 cm × 63 cm with a height of 41 cm and was enclosed in a sound-attenuating cubicle equipped with a fan. On the front wall of each chamber were two retractable, 5-cm-long levers positioned 5 cm from the midline and 9 cm above the grid floor. One amber light-emitting diode (LED) was positioned above each lever. A response was registered by a downward displacement of the lever with 0.20 N force. A dispenser (Model ENV-203–20; Med Associates Inc.) delivered 45-mg sucrose pellets to a receptacle positioned between the levers and 2 cm above the floor. A house light was mounted on the upper, opposite wall. Each chamber was connected to a Dell desktop computer (Intel^® Core^™ i7-7700 3.60 GHz processor, 16.0 GB of RAM, Microsoft^® Windows 10) by an interface (MED-SYST-8, Med Associates Inc.). Med-PC software version V (Med Associates Inc.) controlled experimental events and provided a record of responses.

2.3.3. In vivo experimental procedures

2.3.3.1. Hotplate and rectal temperature.

Experiments were conducted in the light cycle and body weights were measured daily prior to each experiment. Each subject was placed on the hotplate and response latency was measured manually using a stopwatch (Martin Stopwatch, Martin Sports, Carlstadt, NJ) by trained and experimentally blinded raters until the subject jumped, licked or shook a back paw or up to 60 s, whichever occurred first. Immediately thereafter, the 2.0-cm long microprobe was inserted for 10 s and rectal temperature was measured. Baseline values were measured sequentially for hotplate response latency and rectal temperature in this order prior to the drug discrimination test sessions. Following determination of each baseline, a dose or vehicle was administered i.p. Immediately following the drug discrimination session, hotplate response latency and rectal temperature were measured again. In naïve rats, drugs were administered cumulatively (every 15 min). Doses were increased by a quarter log unit per injection.

2.3.3.2. Drug discrimination procedures

2.3.3.2.1. Lever-response shaping.

Sessions started with the presentation of one of the retractable levers and the illumination of the LED above the presented lever. Initially, under a fixed-ratio (FR) 1 schedule of reinforcement, each downward deflection of the lever turned off the LEDs and activated the pellet dispenser for 0.1 s followed by a 0.1-s timeout period during which LEDs were turned off, the house light was illuminated, and responding had no scheduled consequences. The number of lever presses required for food delivery was increased systematically across daily sessions to a final value of FR10. For these sessions only the right or left lever was presented and the available lever was alternated across sessions. Sessions for drug discrimination training commenced after 50 reinforcers per session were delivered within 20 min for two consecutive sessions.

2.3.3.2.2. Drug discrimination training.

Subjects were divided equally into one of two training groups: morphine (3.2 mg/kg, i.p., administered 15 min prior to sessions, 4 females and 4 males) or mitragynine (32 mg/kg, i.p., administered 30 min prior to sessions, 4 females and 4 males). Following the pretreatment interval, subjects were placed into their assigned chamber. Initially, one lever was presented with illumination of its associated LED; completion of 10 responses turned off the LED, activated the pellet dispenser, and produced a 0.1-s timeout, during which the house light was illuminated and responding had no scheduled consequence. Training sessions ended after 20 min or delivery of 50 pellets, whichever occurred first. For each subject, one lever was assigned to the training dose and the other lever to vehicle (e. g., left lever paired with drug; right lever paired with vehicle); assignments were counterbalanced among subjects and remained the same for that subject. The order of drug and vehicle training followed a double-alternation sequence (i.e., …, right, left, left, right, …). After 12 training sessions under these parameters, both levers were presented, and 10 responses were required to complete the ratio. Test sessions for a subject commenced when the following criteria were met for at least four consecutive sessions: 1) a minimum of 80% of the total responses were correct; and 2) the total of incorrect responses made prior to delivery of the first reinforcer was less than ten. After the first test, sub-sequent test sessions were conducted after the criteria were satisfied for at least one drug and one vehicle training session.

2.3.3.2.3. Testing.

Test sessions were identical to the training sessions, except ten responses on either lever resulted in delivery of food and doses of test compounds administered. Dose-effect assessments of each training drug were obtained once at the beginning of the study and a second time after tests with all other drugs were completed. Doses of test compounds were administered from doses that produced less than group averages of 20% drug-appropriate responding up to doses that produced greater than or equal to group averages of 80% drug-appropriate responding, decreased response rate to less than 20% of the vehicle control per subject, were deemed potentially toxic, or could not be increased further due to solubility limitations. For the substitution and combination tests in both training groups, each test compound was injected 15 min prior to sessions except mitragynine and naltrexone, which were administered 30 min prior to sessions. The following ranges of drug doses (mg/kg) increased in one quarter log unit were assessed: atipamezole (3.2–17.8), clonidine (0.01–0.56), fentanyl (0.0056–0.32), lofexidine (0.0178–1.0), mitragynine (3.2–56), morphine (0.32–56), naloxone (17.8–100), naltrexone (17.8–100), and yohimbine (3.2–17.8). Naloxone (0.1 mg/kg) and naltrexone (0.032 mg/kg) were combined with morphine (3.2–56 mg/kg) in morphine-trained rats, and mitragynine (17.8–56 mg/kg) in mitragynine-trained rats. Atipamezole (3.2 mg/kg) and yohimbine (3.2 mg/kg) were combined with morphine (0.56–56 mg/kg) in morphine-trained rats, and mitragynine (17.8–100 mg/kg) in mitragynine-trained rats. Clonidine (0.032 mg/kg) and lofexidine (0.1 mg/kg) were combined with fentanyl (0.0056–0.32 mg/kg) and morphine (0.32–56 mg/kg) in morphine-trained rats and combined with fentanyl (0.001–0.32 mg/kg), morphine (0.056–56 mg/kg), and mitragynine (0.56–56 mg/kg) in mitragynine-trained rats.

2.4. Data analyses

2.4.1. In vitro data analyses

To calculate binding affinity, the IC₅₀ values were determined using average values from at least three experiments conducted in triplicate and calculated using a nonlinear, least-squares regression analysis (Prism 9; GraphPad Software, Inc., San Diego, CA). IC₅₀ values were converted to Ki values using the Cheng-Prusoff equation (Cheng and Prusoff, 1973). Percent (−)-epinephrine-stimulated [³⁵S]GTPγS binding (E_max) was defined as [(net-stimulated binding by a test compound)/(net-stimulated binding by 10 μM (−)-epinephrine] × 100%. All data were normalized to obtain % stimulation values relative to 10 μM (−)-epinephrine. IC₅₀ values for antagonists obtained from the functional assays were then converted to K_b values using a functional version of the Cheng-Prusoff equation (Leff and Dougall, 1993). EC₅₀ and IC₅₀ values were obtained using a nonlinear, least-squares regression analysis (GraphPad Prism version 9). Comparisons were considered significant at P < 0.05. A one-way analysis of variance (ANOVA) and post hoc Bonferroni t-test were used to analyze the effects of the compound concentration using SigmaPlot version 14.0 (Systat Software Inc., Chicago, IL).

2.4.2. In vivo data analyses

The percentage of drug-appropriate responding was calculated by dividing the total number of responses on the drug-appropriate lever by the total number of responses on the drug- and vehicle-appropriate levers. The rates of responding were normalized to baseline, defined as the mean rate of responding following an injection of vehicle during all the inter-test sessions. Percentage of drug-appropriate responding for a particular test was not plotted or analyzed when: 1) rates of responding were less than 20% of the control and 2) less than half of the sample size (e.g., five out of eight rats) did not produce more than or equal to 20% of the control rate of responding. The hotplate latencies were normalized to the percentage of the maximum possible effect (%MPE), using the following formula and each animal as its own control: %MPE = 100 × (post-injection latency - pre-injection baseline latency)/(maximum latency 60 s - pre-injection baseline latency). Changes in rectal temperature were calculated per animal as the test value subtracted from the averaged baseline value in °C. Baseline values for antinociceptive latency and rectal temperature were calculated as the average for all test sessions. A within-subjects design was used with a sample size of 8 per group (N = 4 per sex) per treatment. Data are expressed as mean values (±SEM) as a function of dose. Two- and three-way repeated-measures ANOVAs followed by post-hoc Bonferroni t-tests (P < 0.05) were conducted using GraphPad Prism version 9 for Windows and SigmaPlot version 14.0; factors included dose, sex, training drug, and multiple determinations of a dose-effect function. Standard linear regression (Snedecor and Cochran, 1967) determined using GraphPad Prism version 9 was used to calculate ED₅₀ and ED_−2°C values and corresponding 95% confidence intervals (95%CI) from dose-effect functions when the mean percentage of drug-appropriate responding or maximum possible effect was increased to greater than 50%, or when response rates were decreased to less than 50%. The linear portion of the dose-effect functions were used for the analysis. All data for response rates, maximum possible effects, and changes in rectal temperature were plotted and analyzed. Potency ratios, slope ratios, and corresponding 95% CIs were calculated as the ratio of ED₅₀ or slope values calculated from the dose-effect function (Tallarida, 2002). If the 95% CIs of the ED₅₀ and slope values did not overlap or if the 95% CIs of the potency ratio did not include 1, the potencies were considered to be significantly different. The dose-effect functions for discriminative-stimulus and rate-decreasing effects for both training drugs (mitragynine and morphine) determined at the beginning and end of the study were compared by evaluating the potency ratios; because the ED₅₀ values were not significantly different from each other (Supplemental Figs. S1 and S2), data from the first dose-effect assessment were used for further analyses.

3. Results

3.1. Aα₂R binding affinity and intrinsic activity

The binding affinity of mitragynine at the Aα_2AR was compared to those of reference Aα₂R ligands (Fig. 1, leftmost panel; Table 1). The reference Aα₂R ligands atipamezole, clonidine, (−)-epinephrine, lofexidine, and yohimbine concentration-dependently displaced [³H] RX821002 bound at the Aα_2AR. The rank order of binding affinities at the Aα_2AR was yohimbine > lofexidine ≈ atipamezole > clonidine > (−)-epinephrine > mitragynine. The intrinsic activity (E_max) and potency (EC₅₀ or K_b) of mitragynine and reference Aα₂R ligands were compared to that of (−)-epinephrine (Fig. 1, middle panel; Table 1). (−)-Epinephrine was a full agonist (E_max ± SEM = 102 ± 1.95 %) at the Aα_2AR whereas clonidine was a partial agonist (E_max ± SEM = 33.2 ± 4.03 %). Lofexidine had a lower intrinsic activity than previously reported (Raffa et al., 2019), but the 10.8 ± 0.88 % maximum stimulation was significantly different from vehicle (F10,52 = 3.67; P < 0.001). Mitragynine up to 100 μM and the reference Aα_2AR antagonist, atipamezole and yohimbine up to 10 μM did not significantly stimulate [³⁵S] GTPγS binding at the Aα_2AR. (Fig. 1, middle panel; p < 0.05). Yohimbine, atipamezole, lofexidine, and mitragynine all substantially decreased (−)-epinephrine-stimulated [³⁵S]GTPγS binding (Fig. 1, rightmost panel; Table 1), although mitragynine at 100 μM (highest concentration tested) did not decrease the efficacy of (−)-epinephrine by more than 50%.

Fig. 1.

Displacement of radioligand (leftmost panel), [³⁵S]GTPγS stimulation (middle panel), and antagonism of (−)-epinephrine stimulation (rightmost panel) at Aα_2AR. Abscissae: concentrations of compounds in nM (log scale). Ordinates: leftmost panel, percentage of [³H]RX821002 bound to membrane preparations; middle and rightmost panels, percentage of maximum stimulation of [³⁵S]GTPγS binding normalized to maximum stimulation by (−)-epinephrine as 100%. In rightmost panel, varying concentrations of test compounds were co-incubated with an EC₉₀ concentration (3.0 μM) of (−)-epinephrine. Each data point represents the mean results of three repeated experiments; vertical bars represent SEM (N = 3) from at least three independent triplicate replications per sample. Details for statistical analyses are shown in Table 1.

Table 1.

Binding affinity (K_i), potency (EC₅₀ and K_b), and intrinsic activity (E_max) at Aα_2ARs. The data are plotted in Fig. 1.

Compound	Ki ± SEM (nM)	${\text{EC}}_{50}^{\#}\pm \text{SEM}$ or Kb* ± SEM (nM)	Emax ± SEM [% of (−)-epinephrine stimulation]
Atipamezole	7.85 ± 0.384	6.82 ± 1.21*	0.623 ± 3.04
Clonidine	45.7 ± 4.16	37.9 ± 4.68#	33.2 ± 4.03
(−)-Epinephrine	263 ± 27.5	443 ± 104#	102 ± 1.95
Lofexidine	7.42 ± 0.512	37.5 ± 0.30*	10.8 ± 0.88
Mitragynine	1260 ± 90.2	Not Applicable	−4.88 ± 1.29
Yohimbine	1.13 ± 0.089	4.16 ± 0.314*	1.85 ± 2.20

Not Applicable: The EC₅₀ value of mitragynine could not be calculated because of its low intrinsic activity, and the K_b value for mitragynine could not be calculated because it produced less than 50% reduction in the intrinsic activity of 3.0 μM (−)-epinephrine. ^#EC₅₀ values and *K_b values.

3.2. Effects of training drugs mitragynine and morphine alone

The baseline hotplate response latencies (seconds), baseline rectal temperatures (°C), and mean operant response rates (responses per second) after vehicle administration in mitragynine and morphine-trained rats were not significantly different. The baseline hotplate response latencies [mean (SEM) in seconds] for morphine- and mitragynine-trained rats were 6.82 (0.095) and 8.46 (0.663), respectively. The baseline rectal temperatures [mean (SEM) in °C] for morphine- and mitragynine-trained rats were 37.4 (0.06) and 37.3 (0.14), respectively. The mean operant response rates [mean (SEM) in responses per second] after vehicle administration for morphine- and mitragynine-trained rats were 1.01 (0.064) and 0.99 (0.094), respectively. There was no sex difference in each parameter (data not shown).

The ED₅₀ and corresponding 95% CI values for discriminative stimulus, rate-decreasing, and antinociceptive effects, as well as the ED_−2°C and corresponding 95% CI values for hypothermic effects of all compounds are summarized in Tables 2 and 3 and Supplemental Tables S1–S3. In morphine-trained rats, morphine dose-dependently increased morphine-lever responding and hotplate latency (antinociception), and decreased response rates (Figs. 2 and 3). Morphine did not significantly alter rectal temperature (Fig. 3; Table 2). The potency of morphine [ED₅₀ (95% CIs) value] to increase morphine-lever responding [1.70 (0.793, 2.75) mg/kg] was 21-fold higher than its potency to produce antinociception [35.3 (26.9, 47.3) mg/kg], consistent with a higher degree of MOR-mediated intrinsic activity being required for antinociception as compared with discriminative stimulus effects (Table 2). The potency of morphine [ED₅₀ (95% CIs) value] to decrease rates of operant responding was 13.4 (7.03, 21.2) mg/kg.

Table 2.

Potency comparisons in mg/kg of the discriminative-stimulus, rate-decreasing, antinociceptive [ED₅₀ (95% CIs) values], and hypothermic [ED_{−2C° (95% CIs) values]} effects of compounds in rats trained to discriminate 3.2 mg/kg morphine. The data are plotted in Figs. 2–5.

Test drug	ED50			ED−2°C	Potency Ratio
	Discrimination	Response Rate	Antinociception	Hypothermia	Rate-Decreasing/Discrimination	Antinociception/Discrimination	Hypothermia/Discrimination
Atipamezole	Inactive	12.2 (11.0, 13.7)	14.3 (10.1, 31.9)	Not Determined	Not Applicable	Not Applicable	Not Applicable
Lofexidine	Inactive	0.382 (0.246, 0.560)	Not Determined	0.373 (0.292, 0.466)	Not Applicable	Not Applicable	Not Applicable
Mitragynine	20.9 (9.17, 118)	35.7 (32.9, 38.8)	Inactive	Not determined	1.71 (0.280, 4.23)	Not Applicable	Not Applicable
Morphine	1.70 (0.793, 2.75) Females: 1.66 (1.11, 2.40) Males: 1.77 (0.962, 2.79)	13.4 (7.03, 21.2) Females: 11.9 (1.50, 22.7) Males: 14.8 (9.47, 21.9)	35.3 (26.9, 47.3) Females: 36.5 (30.7, 44.1) Males: 34.1 (22.6, 52.0)	Inactive	7.85 (2.56, 26.7) Females: 7.00 (2.54, 28.7) Males: 8.65 (3.45, 27.6)	20.7 (9.80, 59.7) Females: 21.4 (11.2, 55.6) Males: 20.0 (8.24, 65.6)	Not Applicable
Morphine + 3.2 mg/kg Atipamezole	1.96 (1.76, 2.17)	18.7 (1.70, 36.7)	34.5 (27.3, 45.4)	Inactive	9.53 (0.781, 20.8)	17.6 (12.6, 25.7)	Not Applicable
Morphine + 0.1 mg/kg Lofexidine	1.61 (1.35, 4.33)	7.17 (1.36, 14.0)	35.2 (29.6, 42.9)	Inactive	4.46 (0.314, 10.4)	21.9 (6.84, 31.7)	Not Applicable
Morphine + 0.032 mg/kg Naltrexone	19.9 (15.2, 37.4)	34.1 (27.0, 45.5) Females: 52.1 (48.3, 57.1) Males: 22.5 (9.93, 37.0)	Not determined	Inactive	1.71 (0.721, 2.99) Females versus Males: 2.32 (1.31, 5.74)	Not Applicable	Not Applicable
Naltrexone	Inactive	62.3 (53.9, 72.4)	Inactive	Inactive	Not Applicable	Not Applicable	Not Applicable

Each value is a combination of females and males unless otherwise noted. Potency ratios (95% CIs) are calculated by dividing the ED₅₀ values for the rate-decreasing or antinociceptive effects or the ED_−2C° values for the hypothermic effects by the ED50 values for the discriminative stimulus effects. Significant differences are in bold.

Table 3.

Potency comparisons in mg/kg of the discriminative-stimulus, rate-decreasing, antinociceptive [ED₅₀ (95% CIs) values], and hypothermic [ED_−2C° (95% CIs) values]

Test drug	ED50			ED−2°C	Potency Ratio
	Discrimination	Response Rate	Antinociception	Hypothermia	Rate-Decreasing/Discrimination	Antinociception/Discrimination	Hypothermia/Discrimination
Atipamezole	Inactive	12.8 (10.9, 15.6)	15.1 (11.4, 24.1)	Not Determined	Not Applicable	Not Applicable	Not Applicable
Lofexidine	0.182 (0.140, 0.242)	0.223 (0.145, 0.320)	0.573 (0.415, 0.813)	0.305 (0.272, 0.345)	1.22 (0.599, 2.28)	3.13 (1.72, 5.79)	1.67 (1.12, 2.46)
Mitragynine	14.3 (8.20, 20.9)	50.9 (44.8, 59.4)	Inactive	Inactive	3.55 (2.15, 7.24)	Not Applicable	Not Applicable
Mitragynine + 3.2 mg/kg Atipamezole	Not Determined	60.7 (47.8, 77.5)	Inactive	Inactive	Not Applicable	Not Applicable	Not Applicable
Mitragynine + 0.1 mg/kg Lofexidine	2.45 (1.26, 3.87)	16.2 (10.8, 22.1)	Not determined	45.8 (32.4, 94.0)	6.63 (2.78, 17.5)	Not Applicable	18.7 (8.38, 74.6)
Mitragynine + 0.032 mg/kg Naltrexone	Not determined	Not determined	Inactive	Inactive	Not Applicable	Not Applicable	Not Applicable
Morphine	12.7 (10.5, 15.3)	21.3 (3.25, 40.6)	32.9 (28.4, 38.6)	Inactive	1.68 (0.213, 3.85)	2.60 (1.86, 3.67)	Not Applicable
Morphine + 0.1 mg/kg Lofexidine	0.546 (0.194, 0.992)	16.0 (4.67, 30.6)	27.8 (25.1, 30.9)	Inactive	29.2 (7.16, 158)	50.9 (38.5, 159)	Not Applicable
Naltrexone	Inactive	97.5 (81.3, 126)	Not Determined	Inactive	Not Applicable	Not Applicable	Not Applicable

Each value is a combination of females and males unless otherwise noted. Potency ratios (95% CIs) are calculated by dividing the ED₅₀ values for the rate-decreasing or antinociceptive effects or the ED_{−2° C} values for the hypothermic effects by the ED₅₀ values for the discriminative-stimulus effects. Significant differences are in bold.

Fig. 2.

Substitution of Aα₂R and MOR ligands for 3.2 mg/kg morphine (left panels) or 32 mg/kg mitragynine (right panels). Abscissae: Vehicle and compound dose in mg/kg (i.p., log scale). Ordinates: Upper panels, percentage of responses on the training drug-appropriate lever; lower panels, mean rates of responding expressed as a percentage of vehicle control. Atipamezole (squares with crosshatch), lofexidine (diamonds), and morphine (circles) were administered 15 min before sessions while mitragynine (squares) and naltrexone (circles with crosshatch) were administered 30 min before sessions. Each gray symbol indicates a significant difference from vehicle in the corresponding training group. Each point represents the mean ± SEM (N = 4 rats/sex/dose). Details for statistical analyses are shown in Tables 2 and 3.

Fig. 3.

Antinociceptive and hypothermic effects of Aα₂R and MOR ligands. Abscissae: Vehicle and compound dose in mg/kg (i.p., log scale). Ordinates: Upper panels, percent maximum possible effects; lower panels, changes in rectal temperature from baseline. The left panels are from morphine-trained rats while the right panels are from mitragynine-trained rats. Atipamezole (squares with crosshatch), lofexidine (diamonds), and morphine (circles) were administered 15 min before sessions while mitragynine (squares) and naltrexone (circles with crosshatch) were administered 30 min before sessions. Gray symbols indicate significant differences from vehicle in the corresponding training group. Each point represents the mean ± SEM (N = 4 rats/sex/dose). Details for statistical analyses are shown in Tables 2 and 3.

In morphine-trained rats, mitragynine did not produce significant antinociception up to the largest dose that could be safely studied (56 mg/kg) (F1.51,9.06 = 3.76; P = 0.0731, Fig. 3, upper left panel). Mitragynine produced a maximum of 73% drug lever responses in the morphine-trained rats while morphine produced a maximum of 70% drug lever responses in the mitragynine-trained rats. The potency of mitragynine [ED₅₀ (95% CIs) value] to increase the percentage of mitragynine-lever responding was 14.3 (8.20, 20.9) mg/kg (Fig. 2, upper right panel) and its potency to decrease rates of operant responding was 50.9 (44.8, 59.4) mg/kg (Table 3).

At the end of the study, the dose-effect functions for each training compound were re-assessed (Supplemental Figs. S1 and S2). In both morphine- (F_1.81,10.9 = 1.16; P = 0.345, F_2.49,14.9 = 0.996; P = 0.409, F_2.27,13.6 = 0.746; P = 0.509, and F_2.60,15.6 = 0.904; P = 0.449, respectively) and mitragynine-trained (F_1.49,8.96 = 0.309; P = 0.680, F_3.60,21.6 = 1.41; P = 0.266, F_3.15,18.9 = 1.12; P = 0.369, and F_2.99,17.9 = 2.77; P = 0.0717, respectively) groups, there were no significant differences between the first and second determination for the discriminative stimulus, rate-decreasing, antinociception, or hypothermia dose-effect functions. Moreover, the antinociceptive and hypothermic effects of mitragynine and morphine in drug-naïve rats were not significantly different from those in the morphine- and mitragynine-trained rats (Supplemental Fig. S3 and Supplemental Table S1). Thus, morphine produced similar antinociceptive and hypothermic effects in drug naïve rats and in trained rats that received repeated administrations of morphine. Mitragynine produced similar antinociceptive and hypothermic effects in drug naïve rats and mitragynine-trained rats, indicating that the lack of antinociceptive effects of mitragynine in the mitragynine-trained rats is not due to tolerance.

3.3. Effects of Aα₂R and MOR ligands alone

Neither naltrexone (up to 100 mg/kg) nor naloxone (up to 100 mg/kg) significantly increased morphine- or mitragynine-lever responding (Fig. 2 and Supplemental Fig. S4). Both naltrexone and naloxone reduced response rates (Fig. 2 and Supplemental Fig. S4) but neither naltrexone nor naloxone significantly altered hotplate response latency or rectal temperature (Fig. 3 and Supplemental Fig. S5). In mitragynine-trained rats, the MOR agonist fentanyl produced greater than 80% mitragynine-paired lever responses, greater than 80% antinociception, and decreased response rates (rightmost panels in Supplemental Figs. S6 and S7).

In morphine-trained rats, lofexidine did not significantly increase drug-lever responding or produce antinociception up to a dose (1.0 mg/kg) that significantly decreased operant response rates (Figs. 2 and 3, left panels). Lofexidine produced marked hypothermia, with a maximum decrease of 5.0 °C (±0.42 °C SEM) (Fig. 3, lower left panel). In mitragynine-trained rats, lofexidine increased drug-lever responding up to a maximum of 77% (±16% SEM), decreased response rates, and induced hypothermia (Figs. 2 and 3, right panels). In mitragynine-trained rats, lofexidine produced significant antinociception with a maximum possible effect of 62% (±5.1% SEM) (Fig. 3, upper right panel, F_2.25,13.5 = 97.0; P < 0.001). There were no significant differences in the antinociceptive effects of lofexidine in mitragynine-trained and drug naïve rats (Fig. S3 and Table S1). Clonidine increased drug-lever responding to a maximum of 9.7% in morphine-trained rats and 55% in mitragynine-trained rats; in both groups, clonidine significantly decreased operant response rates and produced significant antinociceptive and hypothermic effects (Supplemental Figs. S4 and S5, Supplemental Tables S2 and S3). Atipamezole (up to 17.8 mg/kg) and yohimbine (up to 17.8 mg/kg) did not significantly increase drug-lever responding in either morphine- or mitragynine-trained rats (Fig. 2 and Supplemental Fig. S4; Tables 2 and 3 and Supplemental Tables S2 and S3). Atipamezole and yohimbine produced significant antinociception in both groups of rats (Fig. 3 and Supplemental Fig. S5; Tables 2 and 3 and Supplemental Tables S2 and S3). Moreover, atipamezole and yohimbine produced significant hypothermia in the mitragynine-trained rats but not in morphine-trained rats, although the decreases in rectal temperature produced by atipamezole (−1.4 °C) and yohimbine (−2.6 °C) were lower than that produced by clonidine (−4.5 °C) or lofexidine (−3.8 °C). The ED₅₀ and ED_−2°C values of atipamezole and lofexidine are shown in Tables 2 and 3, while those for clonidine and yohimbine are shown in Supplemental Tables S2 and S3.

3.4. Effects of MOR or Aα₂R antagonists in combination with morphine and mitragynine

Naltrexone (0.032 mg/kg) antagonized the discriminative stimulus effects of morphine, as evidenced by a significant rightward shift in the morphine dose-effect function (Fig. 4, upper leftmost panel; Table 2). Naltrexone also significantly antagonized the discriminative stimulus effects of mitragynine, as evidenced by a rightward and downward shift of the mitragynine dose-effect function (Fig. 4, upper middle panel; Table 3). Naltrexone did not significantly modify the dose-effect function of mitragynine on rates of responding but produced a 2.5-fold rightward shift in the dose-effect function of morphine on rates of responding (Fig. 4, lower middle and leftmost panels, respectively; Tables 2 and 3) and antagonized the antinociceptive effects of morphine (56 mg/kg) in morphine-trained rats (Fig. 5, upper leftmost panel; Table 2). Naloxone (0.1 mg/kg) also antagonized the discriminative stimulus effects of morphine and mitragynine (Supplemental Fig. S6; Supplemental Tables 2 and 3), and the antinociceptive effects of morphine (Supplemental Fig. S7; Supplemental Tables 2 and 3).

Fig. 4.

Effects of atipamezole, lofexidine, and naltrexone to modify the discriminative-stimulus and rate-decreasing effects of morphine and mitragynine. Abscissae: Vehicle and compound dose in mg/kg (i.p., log scale). Ordinates: Upper panels, percentage of responses on the training drug-appropriate lever; lower panels, mean rates of responding expressed as a percentage of vehicle control. The leftmost panels show data obtained in morphine-trained rats while the middle and rightmost panels show data obtained in mitragynine-trained rats. Open circles and open squares represent data for morphine and mitragynine administered alone, respectively. Squares with crosshatch, open diamonds, and circles with crosshatch represent data for morphine or mitragynine in the presence of 3.2 mg/kg atipamezole, 0.1 mg/kg lofexidine, and 0.032 mg/kg naltrexone, respectively. All compounds were administered 15 min before sessions except mitragynine and naltrexone (30 min before sessions). Gray symbols indicate significant differences from vehicle in the corresponding training compound. Each point represents the mean ± SEM (N = 4 rats/sex/dose). Details for statistical analyses are shown in Tables 2 and 3.

Fig. 5.

Effects of atipamezole, lofexidine, and naltrexone in combination with morphine or mitragynine in hotplate antinociception and rectal temperature. Abscissae: Vehicle and compound dose in mg/kg (i.p., log scale). Ordinates: Upper panels, percent antinociceptive maximum possible effects; lower panels, changes in rectal temperature from baseline. All compounds were administered 15 min before sessions except mitragynine and naltrexone (30 min before sessions). The leftmost panels show data obtained in morphine-trained rats while the middle and rightmost panels show data obtained in mitragynine-trained rats. Open circles and open squares represent data for morphine and mitragynine, respectively, administered alone. Squares with crosshatch, open diamonds, and circles with crosshatch represent data of morphine or mitragynine in the presence of 3.2 mg/kg atipamezole, 0.1 mg/kg lofexidine, and 0.032 mg/kg naltrexone. The antinociceptive and thermoregulatory effects of mitragynine in combination with the Aα_2AR antagonist atipamezole (3.2 mg/kg) and MOR antagonist naltrexone (0.032 mg/kg) are not shown because mitragynine alone did not significantly alter percent antinociceptive maximum possible effects or rectal temperature. Gray symbols indicate a significant difference from vehicle in the corresponding training compound. Each point represents the mean ± SEM (N = 4 rats/sex/dose). Details for statistical analyses are shown in Tables 2 and 3.

In morphine-trained rats, neither atipamezole (3.2 mg/kg) nor yohimbine (3.2 mg/kg) significantly altered the discriminative stimulus or rate-decreasing effects of morphine (Fig. 4 and Supplemental Fig. S6, upper leftmost panels; Table 2 and Supplemental Table S2). However, both antagonists produced rightward and downward shifts in the dose-effect functions of mitragynine to produce discriminative stimulus effects (Fig. 4, upper middle panel; Supplemental Fig. S6, upper second panel from the left; Table 3 and Supplemental Table S3). However, neither atipamezole (3.2 mg/kg) nor yohimbine (3.2 mg/kg) modified the capacity of mitragynine to decrease operant response rates (Fig. 4, lower middle panel; Supplemental Fig. S6, lower second panel from the left; Table 3 and Supplemental Table S3).

3.5. Effects of Aα₂R agonists on discriminative stimulus, rate-decreasing, and antinociceptive effects of morphine

In morphine-trained rats, neither lofexidine (0.1 mg/kg) nor clonidine (0.032 mg/kg) significantly altered the dose-effect function of morphine to produce discriminative stimulus effects (Fig. 4 and Supplemental Fig. S6, upper leftmost panels; Table 2 and Supplemental Table S2). However, both lofexidine and clonidine shifted leftward the dose-effect function of morphine to decrease operant response rates (Fig. 4 and Supplemental Fig. S6, lower leftmost panels; Table 2 and Supplemental Table S2). Neither lofexidine nor clonidine significantly altered morphine antinociception (Fig. 5 and Supplemental Fig. S7, upper leftmost panels; Table 2 and Supplemental Table S2).

3.6. Effects of Aα_2AR agonists on discriminative stimulus and rate-decreasing effects of mitragynine

In mitragynine-trained rats, lofexidine (0.1 mg/kg) produced 5.8- and 3.1-fold leftward shifts in the dose-effect functions of mitragynine to produce discriminative stimulus and rate-decreasing effects, respectively (Fig. 4, middle panels; Table 3), while clonidine (0.032 mg/kg) produced 1.5- and 5.0-fold leftward shifts in the dose-effect functions of mitragynine on the discriminative stimulus and rate-decreasing effects, respectively (Supplemental Fig. S6, second panels from the left; Supplemental Table S3).

In mitragynine-trained rats, lofexidine (0.1 mg/kg; Fig. 4, upper rightmost panel; Table 3) and clonidine (0.032 mg/kg; Supplemental Fig. S6, upper third panel from the left; Supplemental S3) produced 23- and 15-fold leftward shifts, respectively, in the dose-effect function of morphine to increase mitragynine-lever responding. In addition, lofexidine (Fig. 4, lower rightmost panel; Table 3) and clonidine (Supplemental Fig. S6, lower third panel from the left; Supplemental S3) significantly shifted the morphine dose-effect functions for the rate-decreasing effects to the left. However, neither lofexidine (Fig. 5, upper rightmost panel; Table 3) nor clonidine (Supplemental Fig. S7, upper third panel from the left; Supplemental S3) significantly changed the morphine dose-effect functions for the antinociceptive effects. To examine the generality of the findings with Aα₂R agonists in combination with morphine in mitragynine-trained rats and not in morphine-trained rats, fentanyl was tested in combination with Aα₂R agonists in both groups. In mitragynine-trained rats (Supplemental Fig. S6, rightmost panels; Supplemental S3), co-administration of fentanyl with lofexidine (0.1 mg/kg) or clonidine (0.032 mg/kg) produced 7.1- and 7.7- fold leftward shifts, respectively, in the dose-effect functions of fentanyl on the discriminative stimulus effects, and 5.0- and 3.6- fold leftward shifts, respectively, in the rate-decreasing dose-effect function of fentanyl. On the contrary, neither lofexidine (0.1 mg/kg) nor clonidine (0.032 mg/kg) produced leftward shifts in the dose-effect functions of fentanyl to produce discriminative stimulus and rate-decreasing effects in morphine-trained rats (data not shown).

3.7. Comparison of effects in females and males

There were no sex differences in any effects tested. As a result, all data points plotted are a combination of females and males unless otherwise noted.

4. Discussion

In the current study, mitragynine at the human Aα_2ARs exhibited low micromolar binding affinity (see also Obeng et al., 2020), and did not stimulate [³⁵S]GTPγS binding. The drug discrimination results reported here are consistent with mitragynine exhibiting low efficacy agonism at Aα_2ARs and MORs. This was evidenced not only by antagonism of mitragynine by Aα_2ARs and MOR antagonists, but also by enhancement of the effects of mitragynine, morphine, and fentanyl by Aα_2AR agonists in rats discriminating mitragynine. The apparent low efficacy agonism exhibited by mitragynine at both MORs and Aα₂Rs may provide advantages over currently approved FDA medications for OUD inasmuch as multiple medications (e.g., buprenorphine and lofexidine) are currently required to achieve this polypharmacology.

The affinity and intrinsic activity of the reference Aα₂R agonists and antagonists obtained in this study were compared to previously reported results. Aα_2AR agonists clonidine and lofexidine inhibited isoproterenol-stimulated cyclic adenosine monophosphate (cAMP) production in bovine ciliary process membranes (Jin et al., 1989). The intrinsic activity of lofexidine and clonidine at human Aα_2ARs were reported to have a higher maximum intrinsic activity [70 and 90%, respectively (Raffa et al., 2019)] than observed in the current study (11 and 33%). The low in vitro intrinsic activity of lofexidine observed here could be due to a lower level of Aα_2AR expression/reserve in our heterologous cell model [B_max (mean ± SEM) = 6.75 ± 0.371 pmol/mg] compared to human Aα_2ARs employed by Raffa et al. (2019) and unfortunately there was no information on B_max values in the study by Raffa et al. (2019). When a low efficacy agonist binds, the probability of resultant receptor activation is lower than for a full agonist. However, when receptor expression is high relative to the effector population, i.e. “high reserve”, such as can be found in some heterologous models, even low efficacy agonists can activate a high enough number of receptors to saturate the downstream effector. Such assay conditions obscure the differences between high and low efficacy agonists (Kelly, 2013; Niedernberg et al., 2003). Moreover, a [³⁵S]GTPγS assay was used in the current study versus the adenylate cyclase assay previously used and adenylate cyclase has a faster turnover rate than G-proteins resulting in signal amplification (Costa et al., 1988; Newton et al., 2016). In our laboratory, mitragynine did not stimulate Aα_2AR-mediated [³⁵S]GTPγS binding, although we previously observed insignificant human MOR-mediated [³⁵S]GTPγS binding (Obeng et al. (2021), while mitragynine’s pharmacological profile in rats discriminating morphine was consistent with low efficacy MOR agonism. In addition, we did observe mitragynine antagonism of (the full agonist) epinephrine-induced [³⁵S]GTPγS binding in the current study consistent with mitragynine acting as a low efficacy agonist (or antagonist) in vitro.

The current results extend previous findings demonstrating that at least some of the behavioral effects of mitragynine are mediated by MORs. As reported previously in separate groups of rats trained to discriminate either morphine or mitragynine (Harun et al., 2015; Obeng et al., 2021), tests of cross-substitution demonstrated that discriminative stimulus effects partially overlapped. That is, when tested in rats discriminating mitragynine, morphine produced a maximum of 70% drug-appropriate responding. Conversely, mitragynine produced a maximum of 73% drug-appropriate responding in rats discriminating morphine. The MOR agonist fentanyl exhibited a similar profile, producing a maximum of 84% drug-appropriate responding in the mitragynine discrimination assay. These results, coupled with antagonism of the discriminative stimulus effects of mitragynine by naltrexone (Obeng et al., 2021) in the previous and present studies, support the hypothesis that mitragynine is a MOR agonist in vivo.

Lofexidine and clonidine produced significantly higher levels of drug-appropriate responding in the mitragynine discrimination assay as compared to the morphine discrimination assay. While these results may suggest greater involvement of Aα_2AR in the effects of mitragynine as compared with morphine, more parsimonious explanations must be considered and ruled out. For example, increases in drug-lever responding can sometimes indicate that a discrimination assay lacks pharmacological selectivity, i.e., there is an effect of the test compound to produce responding on both drug- and vehicle-appropriate levers that is not due to shared pharmacological mechanisms between the training and test compounds. However, this explanation might be ruled out by the results obtained with other test compounds, e.g., the Aα_2AR antagonists atipamezole and yohimbine produced negligible mitragynine-appropriate responding up to doses that markedly decreased rates of responding. Partial substitution of clonidine and lofexidine was accompanied by leftward shifts in the dose-response function for not only mitragynine as a training drug, but also for morphine and fentanyl in their capacity to substitute for the mitragynine discriminative stimulus. However, this enhancement of the mitragynine-like discriminative stimulus of morphine and fentanyl occurred only in the mitragynine discrimination assay, i.e., clonidine and lofexidine did not enhance the discriminative stimulus effects of morphine in the morphine discrimination assay. Selective enhancement of the mitragynine-like discriminative stimulus of morphine by clonidine and lofexidine appears to rule out pharmacokinetic interactions as the underlying cause insofar as the same enhancement would have been expected in both the morphine and mitragynine discrimination assays. Collectively, these results implicate a pharmacodynamic interpretation whereby the discriminative stimulus effects of mitragynine are mediated at least in part by Aα_2AR agonism.

Pharmacological mechanisms of mitragynine at Aα_2ARs were further assessed with the antagonists atipamezole and yohimbine. Both antagonists attenuated the discriminative stimulus effects of mitragynine, but not morphine. The specific antagonism by atipamezole and yohimbine occurred at doses that appear to be selective for Aα₂Rs based on previous results (Obeng et al., 2020, 2022). In addition, the opioid antagonists naltrexone and naloxone attenuated the discriminative stimulus effects of mitragynine and morphine. These findings suggest that the discriminative stimulus effects of mitragynine are mediated by both Aα₂Rs and MOR, whereas there is little or no evidence for the involvement of Aα₂Rs in the discriminative stimulus effects of morphine. Previous studies spanning decades have consistently demonstrated that the discriminative stimulus effects of morphine are selectively mediated by MORs (Colpaert, 1978; Obeng et al., 2021; Suzuki et al., 1995; Young et al., 1992). Thus, the discriminative stimulus effects of mitragynine are due to the dual activation of the Aα₂Rs and MORs which explains why both Aα₂R (mitragynine-lever responding for clonidine and lofexidine were 55% and 77%, respectively) and MORs (mitragynine-lever responding for morphine and fentanyl were 70% and 84%, respectively) agonists substituted or partially substituted for mitragynine and also why the Aα₂R and MOR antagonists decreased mitragynine-appropriate lever responding. There is a reasonable scientific premise that shows that MOR and α₂AR agonists interact, including both anatomical and physiological relationships between the neurotransmitters associated with these receptor types (Dang et al., 2012; Jordan et al., 2003; Tan et al., 2009; Vilardaga et al., 2008). MOR and α₂AR are co-localized in brain and spinal pathways that control pain processing, and under some conditions α₂AR and MOR agonists act synergistically to increase the latency of reflexive behavioral responses induced by application of a noxious heat stimulus (Drasner and Fields, 1988; Ossipov et al., 1990; Stone et al., 2014). The pharmacodynamic interactions between the Aα₂R and MOR could explain why lofexidine and clonidine produced leftward shifts in the dose-effect function of fentanyl and morphine to produce mitragynine-like discriminative stimulus effects but not morphine-like discriminative stimulus effects.

We have previously shown that both lofexidine and clonidine decreased rectal temperature, whereas morphine did not significantly alter rectal temperature (Obeng et al., 2022). The hypothermic effects of both lofexidine and clonidine were enhanced by 17.8 mg/kg mitragynine but antagonized by 1.0 mg/kg yohimbine (Supplemental Fig. S8, Obeng et al., 2022). However, both here and as previously reported, mitragynine alone did not significantly alter rectal temperature, suggesting that mitragynine in rats may be a low efficacy Aα_2AR agonist as compared with clonidine and lofexidine, just as mitragynine is a low efficacy agonist at MOR. The differences in the in vitro and in vivo efficacy of mitragynine could be due to species differences (rats used in vivo versus human Aα_2ARs used in vitro). While affinity and intrinsic activity of mitragynine have been reported for human and mouse MORs, no data are published from rat MORs, perhaps indicating a low level of interaction or response at the rat MORs. While the Aα_2AR antagonists atipamezole and yohimbine at high doses (17.8 mg/kg) also produced hypothermia on their own, the main effect of these Aα_2AR-selective ligands was to antagonize the effects of lofexidine and clonidine. Mitragynine, on the other hand, did not antagonize the effects of Aα_2AR agonists but rather increased their potency, while exerting minimal effect on its own. This suggests that mitragynine binds to Aα_2AR and is a low efficacy agonist in rats.

In contrast to the mitragynine-like discriminative stimulus effects, neither atipamezole nor naltrexone altered the rate-decreasing effects of mitragynine, indicating that the rate decreasing effects of mitragynine are not mediated by MOR or Aα_2AR. Moreover, the doses of yohimbine and atipamezole that produced hypothermia and antinociception also produced significant decreases in response rates indicating that there may be a lack of pharmacological specificity (i.e., non-Aα_2AR mediated hypothermia) at the higher doses (17.8 mg/kg) of atipamezole and yohimbine tested.

5. Conclusion

In summary, the present results suggest that mitragynine is not only a MOR agonist, but also an Aα_2AR ligand with low efficacy agonist activity in vivo. The low efficacy MOR agonist (buprenorphine) is approved by the FDA to treat OUD, as is the Aα₂R agonist lofexidine. Concurrent activation of either or both of these receptor types by mitragynine may contribute to the purported use of kratom and related products as a self-treatment for opioid withdrawal (Boyer et al., 2008). The potential dual activation of both MOR and Aα_2AR by mitragynine may improve upon currently approved drugs for OUD by including both mechanisms in a single novel compound.

Data availability

The data that support the findings of this study are available within the paper and its Supplemental Data. Any additional information is available on request from the corresponding authors.