This study examines the specific roles and activities of noribogaine at the opioid receptors in relation to physiological outputs in order to characterize noribogaine to the mu (OPRM) and the kappa (OPRK) opioid receptors. The study observed that the biased agonist/antagonist pharmacology is distinctive to noribogaine in comparison to other ligands including ibogaine, nalmefene, 18-MC, and 6′-GNTI. It predicted that noribogaine promoted some analgesic effects and anti-addictive response.
Noribogaine is the long-lived human metabolite of the anti-addictive substance ibogaine. Noribogaine efficaciously reaches the brain with concentrations up to 20 μM after acute therapeutic dose of 40 mg/kg ibogaine in animals. Noribogaine displays atypical opioid-like components in vivo, anti-addictive effects and potent modulatory properties of the tolerance to opiates for which the mode of action remained uncharacterized thus far. Our binding experiments and computational simulations indicate that noribogaine may bind to the orthosteric morphinan binding site of the opioid receptors. Functional activities of noribogaine at G-protein and non G-protein pathways of the mu and kappa opioid receptors were characterized. Noribogaine was a weak mu antagonist with a functional inhibition constants (Ke) of 20 μM at the G-protein and β-arrestin signaling pathways. Conversely, noribogaine was a G-protein biased kappa agonist 75% as efficacious as dynorphin A at stimulating GDP-GTP exchange (EC50 = 9 μM) but only 12% as efficacious at recruiting β-arrestin, which could contribute to the lack of dysphoric effects of noribogaine. In turn, noribogaine functionally inhibited dynorphin-induced kappa β-arrestin recruitment and was more potent than its G-protein agonistic activity with an IC50 of 1 μM. This biased agonist/antagonist pharmacology is unique to noribogaine in comparison to various other ligands including ibogaine, 18-MC, nalmefene, and 6′-GNTI. We predict noribogaine to promote certain analgesic effects as well as anti-addictive effects at effective concentrations >1 μM in the brain. Because elevated levels of dynorphins are commonly observed and correlated with anxiety, dysphoric effects, and decreased dopaminergic tone, a therapeutically relevant functional inhibition bias to endogenously released dynorphins by noribogaine might be worthy of consideration for treating anxiety and substance related disorders.
Papers cited by this study that are also in Blossom
Alper, K. R., Lotsof, H. S., Frenken, G. M. N. et al. · The American Journal on Addictions (2010)
Papers in Blossom that reference this study
Ona, G., Reverte, I., Rossi, G. N. et al. · Journal of Psychopharmacology (2023)
Nardou, R., Sawyer, E., Song, Y. J. et al. · Nature (2023)
Havel, V., Kruegel, A. C., Bechand, B. et al. · Biorxiv (2023)
Noribogaine is the principal human metabolite of ibogaine, an alkaloid historically used in west-central African traditional medicine and investigated for anti-addictive properties. Previous preclinical and some human data indicate that noribogaine reaches high concentrations in brain tissue after ibogaine administration, produces ibogaine-like reductions in self-administration of multiple drugs in animals, potentiates morphine analgesia and can prevent development of morphine tolerance. Despite these pharmacological effects, the molecular mechanisms at opioid receptors have remained unclear: earlier work reported neither conventional mu nor kappa agonism/antagonism in vivo, and prior in vitro findings suggested complex, non-classical interactions with opioid signalling pathways. To address this gap, Maillet and colleagues set out to characterise noribogaine's interactions with the mu (OPRM) and kappa (OPRK) opioid receptors using a combination of in vitro binding and functional assays and in silico docking and molecular dynamics. The study quantified binding affinities, measured G-protein activation using [35S]GTPgS stimulation and non-G-protein signalling using b-arrestin-2 recruitment assays, and compared noribogaine's pharmacology with ibogaine, 18-MC and several well characterised opioid ligands (for example nalmefene and 6'-GNTI). The aim was to determine whether noribogaine displays pathway-selective (biased) activity at these opioid receptors that could explain its atypical in vivo profile.
The experimental programme combined radioligand competition binding, functional G-protein activation assays, b-arrestin-2 recruitment assays and computational docking with molecular dynamics. Human OPRM- and OPRK-expressing CHO-K1 cell membranes and rat midbrain membranes were used for biochemical assays. Radioligand competition used [3H]U69,593 for OPRK and [3H]DAMGO for OPRM, with increasing concentrations of test compounds to derive IC50 and apparent Ki values (Cheng–Prusoff transformation). Non-specific binding was defined with 1 mM naloxone. Data fitting was performed in GraphPad Prism. G-protein pathway activity was assessed by [35S]GTPgS binding to Gα proteins in membrane preparations. Membranes were pre-incubated with GDP and exposed to test ligands; bound tracer was separated by filtration and counted. EC50 and EMax values were obtained by non-linear regression. Antagonist functional inhibition constants (Ke) were calculated from agonist EC50 shifts or concentration-inhibition curves using standard equations. For the non-G-protein pathway, b-arrestin-2 recruitment was measured using the PathHunter enzyme complementation assay in CHO-K1 cells co-expressing human opioid receptors and an enzyme fragment–tagged b-arrestin-2; agonist-treated cells were typically incubated 180 min prior to readout, and antagonist incubations used a 30 min pre-treatment. Quantitative descriptors were derived to compare binding and functional potencies: pKi-pEC50 (activation coupling efficiency or e-coupling) links apparent affinity and potency, e-signal (efficacy efficiency) compares maximal responses against reference full agonists, and bias-coupling quantified pathway preference by comparing G-protein versus b-arrestin readouts after correcting for intrinsic assay efficiencies. Computational work used the mouse mu opioid receptor co-crystal (PDB 4dkl) as a template (94% identity to human OPRM in the binding site). Ligands were prepared and docked into the orthosteric site using Schrodinger tools (Glide, Prime) and representative complexes were subjected to 12 ns molecular dynamics in an explicit membrane and solvent environment in Desmond. Trajectories were analysed for key interactions and representative frames extracted for visualisation.
Binding affinities showed that noribogaine bound more tightly to human OPRK than to OPRM: Ki for OPRK was 720 ± 128 nM, while for OPRM it was 1.52 ± 0.3 mM. By comparison, ibogaine and 18-MC had weaker apparent affinities at OPRK (ibogaine Ki ~3.68 ± 0.22 mM; 18-MC Ki ~1.84 ± 0.12 mM) and at OPRM. These values were broadly consistent with assays on calf receptors and with historical literature values presented by the authors. In the G-protein activation assay ([35S]GTPgS), noribogaine acted as a partial agonist at OPRK: it stimulated GTPγS binding with an EMax of 72 ± 3.8% of the full agonist U69,593 and an EC50 of 8.75 ± 1.09 mM. By contrast, ibogaine showed much lower efficacy (E Max 18 ± 1.4%) and 18-MC failed to stimulate G-protein activation. Full agonists dynorphin A and U69,593 produced EMax values of ~94% and ~91% respectively. Noribogaine’s agonist effect was also observed in rat midbrain membranes and was sensitive to the kappa antagonist nor-BNI, supporting activity at native OPRK. When tested as a competitor in G-protein assays, classical kappa antagonists (naloxone, nor-BNI) and partial agonist nalmefene produced rightward shifts in agonist concentration–response curves with functional Ke values close to their Kis, consistent with competitive behaviour. Noribogaine behaved atypically: at concentrations manyfold higher than its Ki it poorly shifted the EC50 of dynorphin A and morphine, yielding estimated Ke values (~40 ± 16 mM and ~15 ± 4 mM) roughly 40-fold higher than its Ki. Noribogaine reduced the maximal activation of more efficacious agonists to its own EMax (~70%) but only at high concentrations (IC50 in the 100–300 mM range), a profile the authors describe as atypical and possibly protean. In b-arrestin-2 recruitment assays at OPRK, noribogaine was markedly weak: EMax was 13 ± 3% with an estimated EC50 of 110 nM, whereas dynorphin A had an EC50 of 11 ± 2 nM. Quantitative coupling analyses indicated a strong activation bias in favour of G-protein signalling: noribogaine showed an activation coupling preference of approximately 1:630 (G-protein vs b-arrestin) under the experimental conditions, and its activation efficacy bias was greater than that of the known biased ligand 6'-GNTI. Noribogaine also acted as an inhibitor of agonist-induced b-arrestin recruitment: it reduced dynorphin A–induced b-arrestin signalling by up to ~60% with an IC50 of 1 ± 0.16 mM. In contrast, in parallel assays assessing inhibition of dynorphin-induced G-protein signalling, noribogaine produced only minimal inhibition (~5%) and had much weaker apparent potency (IC50 ~150 mM). Thus noribogaine functioned as a more potent inhibitor of the b-arrestin pathway than of G-protein activation at OPRK. At OPRM, noribogaine was a marginal G-protein agonist (EMax ~10% of DAMGO/[met]-enkephalin) with an approximate EC50 of 16 mM. Functional inhibition experiments showed noribogaine behaved as a weak antagonist at mu receptors, shifting EC50s of mu agonists and yielding Ke values around ~20 mM. Noribogaine inhibited mu b-arrestin recruitment with an IC50 of ~100 ± 25 mM and, overall, was characterised as an unbiased mu antagonist in terms of pathway preference although it was an outlier in inhibition coupling efficiency compared with classical mu antagonists. Computational docking and 12 ns molecular dynamics of noribogaine and ibogaine in an inactive OPRM crystal structure placed both ligands in the morphinan orthosteric pocket, forming a hydrogen bond via the tertiary amine to Asp147 and multiple hydrophobic and pi interactions with surrounding residues (for example Tyr148, His297). Noribogaine uniquely formed a transient water bridge with Tyr148 during part of the simulation. The in silico interaction patterns aligned with the experimental affinity ranking.
Maillet and colleagues interpret their findings as demonstrating that noribogaine is a G-protein biased kappa opioid receptor agonist and a weaker mu receptor antagonist. The authors note that the potencies and biased profile they observed are compatible with brain concentrations reported after therapeutic ibogaine dosing, supporting physiological relevance. Because noribogaine stimulated G-protein signalling at OPRK while poorly recruiting b-arrestin, the researchers suggest this pharmacology could underlie noribogaine’s capacity to produce certain analgesic effects without the dysphoria typically associated with kappa agonists; the paper also links central OPRK activation to previously reported prolactin release after noribogaine. The study group highlights that noribogaine’s ability to inhibit dynorphin-induced b-arrestin recruitment while preserving G-protein signalling could modulate overactive dynorphin/kappa pathways observed in stress, anxiety and substance dependence. They propose this functional inhibition bias as a mechanistic basis for potential anxiolytic or antidepressant-like effects and for noribogaine’s modulation of opioid tolerance and morphine analgesia, citing parallels with genetic deletion of b-arrestin-2 and with reports that low-dose kappa agonists can reduce morphine tolerance. Comparisons with ibogaine and 18-MC emphasise distinct pharmacologies: ibogaine is a stronger mu antagonist but weaker kappa agonist than noribogaine, whereas 18-MC behaved as a non-specific micromolar competitive antagonist in these assays. The authors caution that 18-MC and noribogaine should not be considered pharmacologically equivalent. Mechanistically, noribogaine does not appear to act as an allosteric ligand because it competes for orthosteric radioligand binding and its behaviour in functional assays is consistent with orthosteric interaction. Instead, the authors propose that noribogaine stabilises a unique set of receptor conformations—accessible states that favour G-protein coupling but disfavour b-arrestin recruitment—consistent with multistep activation models of GPCRs. Finally, they acknowledge that further work is needed: direct in vivo studies in wild-type and genetically modified animals, deeper molecular investigations into receptor conformational dynamics and discrete drug–receptor interaction studies would help clarify the mechanistic and therapeutic implications of noribogaine’s biased pharmacology.
The study concludes that noribogaine acts as a dual ligand at mu and kappa opioid receptors, exhibiting pronounced G-protein biased agonism at kappa receptors and weak antagonism at mu receptors. By modulating dynorphin signalling at OPRK in a pathway-selective manner, noribogaine’s cellular pharmacology provides a plausible mechanistic basis for some of its reported in vivo effects and suggests avenues for further therapeutic exploration.
Noribogaine (Fig.) is the primary human metabolite of ibogaine, an alkaloid derived from the African shrub, iboga (Tabernanthe iboga). The therapeutic and oneirophrenic properties of iboga roots are known for centuries in Equatorial Africa where iboga continues to be used as natural medicine and for ceremonial purposes. Naranjo, in collaboration with Bocher, issued a patent in 1969 based on 54 clinical cases featuring the usefulness of ibogaine for psychotherapy and anti-drug purposes. A few decades later, the benefits of iboga (ibogaine) in the treatment of addiction for multiple drugs of abuse were highlighted by different groups. Preclinical studies show that ibogaine is a polypharmacological drug that can reduce self-administration to many drugs of abuse in rodents, including cocaine, morphine, heroin, alcohol, and nicotine; and further experimentation in humans supported its usefulness to treat addiction. Noribogaine displayed a slow pharmacokinetic clearance rate in humans, being detected for several days in blood after ibogaine ingestion, and was proposed to be responsible for many of the human in vivo effects seen after ibogaine therapy. Noribogaine produced ibogaine-like anti-addictive effects in animals and the systemic administration of noribogaine induced long-lasting decrease of morphine and cocaine self-administration. Noribogaine also decreased ethanol self-administrationand nicotine self-administration in rats. The brain levels of noribogaine in female/male rats were approximately 20/ 13, 10/7 and 0.8/0.1 mM at 1, 5, and 19 h after intra-peritoneal injection of 40 mg/kg ibogaine whereas plasma levels were 10e20 fold less. This indicated that noribogaine had excellent drug permeability across the bloodebrain barrier and reached high levels in the brain. Unlike ibogaine, noribogaine did not produce tremors and ataxia in rodents, suggesting that it is better tolerated than its parent compound and a better drug candidate for clinical development. Recently, a study performed in healthy volunteers indicated that single oral doses of noribogaine from 3 to 60 mg were safe and well tolerated. Of particular interest with regards to their acknowledged roles in modulating opiate dependence are ibogaine and noribogaine's effects on the opioid system. In addition to effects on the pursuit and administration of drug of abuse, these drugs were shown to modulate the analgesic power and the development of physical tolerance to morphine. Pre-administration of noribogaine (40 mg/ kg, 19 h) had a moderate but detectable effect of potentiation on both basal nociception and morphine-induced analgesia. When noribogaine (40 mg/kg) was co-administered with 4 mg/kg morphine, it amplified the duration of morphineinduced analgesia. Noribogaine (40 mg/kg) enhanced anti-nociception when added to morphine but did not enhance anti-nociception when combined with U50,488 (kappa opioid agonist) or DPDPE (delta opioid agonist). At lower doses of 10e20 mg/kg, noribogaine has also been shown to greatly potentiate (or restore) morphine anti-nociceptive activity in morphine-tolerant mice but remained inactive in naïve mice. Finally, noribogaine was shown to prevent the development of tolerance to the analgesic effects of morphine. The mode of action for these effects, as well as the anti-addictive effects, remained largely uncharacterized thus far. Noribogaine has principal known binding affinities to the opioid receptors, the nicotinic acetylcholine receptors, and the SERT and DAT transporters, but marginal affinities to NMDA, sigma 2 and 5-HT2 receptors in comparison to the parent compound ibogaine2001a;. Ibogaine and noribogaine were found to have potencies in the micro-to high nano-molar range for the opioid receptorsbut they do not appear to be conventional opioid agonists or antagonists. Noribogaine, and ibogaine to a lesser extent, were shown to induce naloxone-sensitive GTPgS binding in rat brain preparations. These drugs however did not induce conditioned place preference like mu agonists or conditioned place aversion like kappa agonists or mu antagonistsin animals (and unpublished results). In addition, neither morphine nor the kappa agonist U50,488 substituted for the discriminative stimulus of noribogaine treatment in rats. In a sense, noribogaine was neither a mu or a kappa opioid agonist or antagonist in vivo and it remained largely unclear what the specific roles and activities of noribogaine at the opioid receptors would be in relation to physiological outputs. To address this gap of knowledge, we carried out a pharmacological study to characterize noribogaine to the mu (OPRM) and the kappa (OPRK) opioid receptors. We performed comprehensive in vitro and in silico experimental designs encompassing both binding and functional assays for G-protein signaling using GTPgS binding stimulation, b-arrestin signaling using protein recruitment assay, in addition to performing docking simulations. We then compared the pharmacology of noribogaine to its parent drug ibogaine, the synthetic indole alkaloid 18-methoxycoronaridine (18-MC), and a series a specific opioid ligands including the kappa partial agonist nalmefene, and the biased kappa agonist 6 0 -GNTI(Fig.). Our study reveals an atypical biased agonism of noribogaine at the kappa receptors and provides important pharmacological basis for its mechanism of action as a drug therapeutic.
[Phenyl-3, 4-3 H]-U-69,593 (43.6 Ci/mmol), [Tyrosyl-3, 5-3 H(N)]-DAMGO ([D-Ala 2 , N-MePhe 4 , Gly 5 -ol]-enkephalin) (50 Ci/mmol) and [35S]GTPgS (Guanosine 5 00 -(gamma-thio)triphosphate) (1250 Ci/mmol) were purchased from PerkinElmer Life Sciences (Boston, MA). U69,593, naloxone, nor-binaltorphimine (nor-BNI), morphine, nalmefene, dynorphin A, DAMGO, GTPgS, GDP and all buffer constituents were purchased from SigmaeAldrich Corp (St. Louis, MO). CHOeK1 cell lines expressing human opioid receptors were provided by Dr. Toll at Torrey Pines Institute (Port St. Lucie, FL). Ibogaine was provided by Dr. Mash at the University of Miami (Miami, FL). 18-methoxycoronaridine (18-MC) was purchased from Obiter Research LLC (Champaign, IL). Noribogaine hydrochloride was purchased from Sigma Aldrich Chemie GmbH (Buchs, Switzerland).
Membrane preparations of rat midbrain tissues were purchased from Chantest (Cleveland, OH). Membrane preparation of human OPRK CHOeK1 cells were purchased from PerkinElmer Life Sciences (Boston, MA) and membranes of human OPRM CHOeK1 cells were prepared as described below. Adherent cells were harvested on ice, with cold PBS and a cell scraper, pelleted and frozen at À80 C overnight. Cell lysis was performed at 4 C in 50 mM Tris (pH 7), 2.5 mM EDTA and cOmplete protease inhibitor cocktail (cOmplete, F. Hoffmann-La Roche Ltd). Cells were homogenized with a polytron, centrifuged at 2500 rpm for 10 min at 4 C and the supernatant was recovered. The process was repeated once. Supernatant was centrifuged at 21,000 rpm for 90 min at 4 C and pellets were re-suspended in 50 mM Tris (pH 7) and 0.32 M sucrose. Total protein concentration was determined using a Ther-moScientific NanoDrop spectrophotometer or by Bradford assay. Membrane samples were stored at À80 C at 1e5 mg/mL protein concentration. Membranes from brain tissues were stored in 50 mM Tris (pH 7), 1 mM EDTA and 0.32 M sucrose with protease inhibitor cocktail.
Competitive binding experiments were performed using conditions recommended by the supplier (Perkin Elmer). Membranes were thawed on ice and diluted in binding buffer 50 mM Tris-HCl pH 7.4, 5 mM MgCl 2 at 5 mg of membrane per reaction. Competition binding assays were performed in 500 mL total volume containing [ 3 H]U69,593 (0.88 nM) for OPRK membranes or [ 3 H] DAMGO (0.75 nM) for OPRM membranes in the presence of increasing concentrations of each unlabeled drug (noribogaine, ibogaine, 18-MC, U69,593, morphine, DAMGO, naloxone) for 60 min at 25 C. Nonspecific binding was defined in the presence of 1 mM naloxone. Bound and free radiolabeled ligands were separated by filtration using a MicroBeta FilterMate-96 Harvester and wash 6 Â 1 mL with ice cold wash buffer (50 mM Tris-HCl pH 7.4) over GF/B filter (presoaked in 0.5% BSA) (Perkin Elmer, Waltham, MA). Radioactivity counts were determined using Perkin Elmer MicroBeta microplate counter with scintillation cocktail Micro-Scint-20™ according to manufacturer recommendations. Data were collected and the half maximal inhibitory concentration (IC 50 ) and apparent binding affinity (K i ) for all data sets were calculated with GraphPad Prism 5.04.
[ 35 S]GTPgS binding to Ga proteins was determined using a procedure modified fromand carried out in a 96well format. Cell membranes (10 mg per reaction) were incubated in a binding buffer (20 mM HEPES, pH 7.4, 100 mM NaCl, 10 mM MgCl 2 x6H 2 O, 0.2% bovine serum albumin, and GDP 10 mM, pH 7.4) containing 80 pM [ 35 S]GTPgS and varying concentrations of opioid agonists (U69,593, DAMGO, morphine, dynorphin A, nalmefene, or noribogaine) in a total volume of 100 mL for 60 min at 25 C. Membranes were pre-incubated with the GDP for 15 min on ice prior to the addition of ligands. Antagonists were added to the membrane solution 20 min prior the addition of the agonist, and [ 35 S]GTPgS was added 5 min after the agonist. Non-specific and basal levels of GTPgS binding were evaluated by using 10 mM cold GTPgS or binding buffer, respectively. Bound and free [ 35 S]GTPgS were separated by filtration using a MicroBeta FilterMate-96 Harvester and wash 4 Â 1 mL with ice cold wash buffer (20 mM Tris, pH 7.4, and 2.5 mM MgCl 2 x6H 2 O, pH 7.4) over presoaked GF/B filter (Perkin Elmer, Waltham, MA). Radioactivity counts were determined using Perkin Elmer MicroBeta microplate counter with scintillation cocktail MicroScint-20™ according to manufacturer recommendations. Data were collected and the half maximal effective concentration (EC 50 ) and maximal responses (E max ) values were calculated using non-linear regression with GraphPad Prism 5.04.
The PathHunter enzyme complementation Arrestin-2 Recruitment assays were performed at DiscoveRx Corporation (Fremont, CA) and at the DemeRx Laboratory (Miami, FL). This assay utilized CHOeK1 cells stably transfected to overexpress b-arrestin-2 fused to a b-galactosidase fragment together with human OPRK gene (NM_000912.3, encoding human KOR) or human OPRM gene (NM_000914.3, encoding human MOR). Briefly, when b-arrestin-2 travels to active receptor, the complementary b-galactosidase fragments fused to the receptor and b-arrestin interact to form a functional enzyme with activity that is detected by chemiluminescence. For all in vitro assays, data were normalized as a percentage of control agonist responses, typically defined by dynorphin A stimulated activity in the OPRK assays, and [met]enkephalin stimulated activity in the OPRM assays. For agonist concentration-response experiments, cells were treated with test compound for 180 min prior to assessment of enzyme complementation. For antagonist concentration-inhibition experiments, the cells were incubated with the test compound for 30 min prior to agonist addition.
The IC 50 and K i values for ligands in the radioactive binding assays were determined by fitting competition binding data of individual experiments normalized to buffer (total binding) and 1 mM naloxone (nonspecific binding) to a single site competition model in GraphPad Prism 5.04 using the transformation of Cheng and PrusoffGTPgS binding and b-arrestin-2 translocation were determined by fitting data from individual experiments to sigmoidal concentration-response curves with variable slope in GraphPad Prism 5.04. I Max was calculated as the % of inhibition from the E Max of the full agonist U69,593 (OPRK) or DAMGO (OPRM) to the basal signal. Functional inhibitory potency (K e ) values for agonist concentration-response displacement experiments were calculated using the Gaddum/Schild EC 50 shift calculation in GraphPad Prism or with the following equation: K e ¼ [A]/(DR-1), where [A] is the concentration of antagonist and DR is the ratio of the EC 50 for an agonist in the presence and absence of the antagonist. K e values from concentration-inhibition experiments were calculated with a modified CFeq: K e ¼ IC 50 /(1þ[S]/EC 50 ) where [S] is the concentration of agonist, EC 50 is the functional potency of the agonist, and IC 50 is the half-maximal inhibition concentration of antagonist. Final mean and S.E. were calculated using individual values from each experiment. Activation coupling efficiency (e-coupling) values describe the relationship between the apparent binding affinity K i versus the apparent functional potency EC 50 of a given agonist ligand and used the equation pKi-pEC 50 where "p" represents a elog10 transformation. For the functional inhibitory components of antagonists and partial agonists, e-coupling represents the relationship between the K i versus the K e of a given inhibitor against a given agonist ([ϞDyn-A] and used the equation pKi-pK e . Efficacy efficiency (e-signal) values describe the ratio of the E Max to a tested ligand versus the E Max to reference full agonist dynorphin A (or U69,593) for OPRK and DAMGO or [met]-enkephalin for OPRM and used the equation E Max (test compound)/E Max (control agonist). For inhibitory ligands, e-signal was calculated using maximal level of inhibition (I max ) normalized from 0 (basal, buffer) to 1 (agonist without inhibitor). Bias-coupling (quantification of pathway bias for the coupling efficiency) was calculated by dividing the EC 50 (activation bias) or the K e (inhibition bias) issued from the G-protein pathway assays by those issued from b-arrestin pathway assays for a given ligand, rectified by the intrinsic functional assay efficiency (defined as the average of the cluster values of functional coupling efficiency of typical full agonists and antagonists). The intrinsic functional assay efficiency is dependent of the assay and the experimental conditions (for instance the GDP content in the GTPgS assay, the temperature and time of incubation) and independent of the ligands binding affinity (the potency ranking of high to low affinity ligands remain constant regardless of the assay conditions). Bias-efficacy in favor of the G-protein pathway was evaluated by dividing the functional activation and the functional inhibition maximum responses (e-signal) from the G-protein pathway by the beta-arrestin pathway assays for a given ligand.
We used the mouse m-opioid receptor OPRM co-crystal structure available in the Protein Data Bank (PDB), PDB accession 4dkl, Uniprot accession P42866. The mouse OPRM has 94% (global) sequence identity to the corresponding human receptor (Uniprot accession P35372) and all residues in the binding site are identical. The receptor was crystallized as a fusion protein (OPRMeT4L) with an irreversible morphine antagonist ligand (bound to Lys233, PDB numbering). All simulations were performed using the Schrodinger 2014.2 and Desmond 2014.2 software suite. For initial docking studies we imported the PDB file into Maestro 9.5 (Schrodinger) and ran the standard protein preparation workflow to assign bond orders and clean up the structure including hydrogen bond optimization and constrained minimization. In the preparation process missing side chains were added using Prime. The fusion protein was manually cut and removed between residues Val262 and Glu270 to leave just the GPCR transmembrane domain; the cut residues were capped as primary amide (C-terminal) and acetate (N-terminal). A (non-covalent) ligand entry (separate from the chain) was manually created in Maestro. The resulting protein complex was again processed via the protein preparation workflow. A docking grid was created around the co-crystal ligand using Glide (standard settings). Several small molecules including the morphinan co-crystal ligand (unbound), ibogaine and noribogaine were imported as 2D SDF into Maestro and 3D structure representations were generated using LigPrep (default settings); two representations (inverted at the tertiary bridgehead nitrogen) were generated for each ligand. These were docked using Glide SP (standard settings except keeping 5 poses per compound out of 30 for post-minimization). The docked morphinan ligand reproduced the co-crystal almost perfectly. This docked complex was then optimized using Prime Refine ProteineLigand complex (default settings). This complex was then used to generate another docking grid using Glide (default settings around the ligand) followed by Glide SP docking of the prepared ligands. In these results, the top poses of noribogaine and ibogaine aligned well the morphinan antagonist (hydrophobic ibogaine and noribogaine bicyclic system and ethyl substituent with morphinan cyclopropyl residues and the positively charged tertiary amines, which all form a hydrogen bond to the site chain of Asp147). The m-OR noribogaine docking complex was then used in a 12 ns molecular dynamics (MD) simulation. The MD system generation and simulations were performed in Desmond using an all atom system with a membrane model and explicit water model (ASP). The Desmond software automatically sets up the systems (adjust charges, adds water molecules) and performs several rounds of minimization and short simulations before the 12 ns production run. MD was run on the Pegasus 2 cluster at the Center for Computational Science at the University of Miami () using 48 processors and completed in less than 19 h. Simulation analysis was performed using the Desmond trajectory analysis software. A representative frame with these most prevalent interactions throughout the simulation was extracted from the trajectory, processed via protein preparation (including constrained minimization) to remove overlapping atoms, and visualized using PyMol.
Competitive inhibition of [ 3 H]-U69,593 to human OPRK and of [ 3 H]-DAMGO to human OPRM by noribogaine was conducted and compared to ibogaine, 18-MC, and various control ligands (Fig., Table). Noribogaine exhibited the highest apparent affinity for OPRK with a K i value of 720 ± 128 nM. At OPRK, ibogaine displayed a K i of 3.68 ± 0.22 mM, while 18-MC had a K i a 1.84 ± 0.12 mM. At the OPRM, noribogaine displayed a K i of 1.52 ± 0.3 mM, while ibogaine and 18-MC K i values were 6.92 ± 0.83 mM and 2.26 ± 0.35 mM respectively. Values of both noribogaine and ibogaine for the human OPRM/K receptors were comparable to that of the calf OPRM and OPRK receptors (1.52 and 0.96 mM, Table) where noribogaine was also previously shown to have ~30-fold less affinity at OPRD than at OPRK. In our assays, 18-MC had no OPRM preferential binding with a DpK i of 0.1, contrary to the previously reported 5-fold selectivity (DpK i ¼ À0.7) for OPRM over OPRK. Experimental values, historical values from the literature, and control ligands, are displayed in Tablefor agonists, partial agonists, and antagonists used in this study.
[ 35 S]GTPgS binding to membrane preparations of CHO cells stably transfected with OPRK was examined in response to noribogaine, ibogaine, morphine, and nalmefene drug treatment and the activation of the G-protein pathway by kappa agonists was measured (Fig.). The prototypical full agonist, U69,593, and the endogenous ultra-potent agonist, dynorphin A, were used as controls for OPRK function. Calculated EC 50 and E Max values are enumerated in Table. Noribogaine was a partial agonist at stimulating [ 35 S]GTPgS binding to OPRK with an E Max of 72 ± 3.8% of U69,593, and an EC 50 of 8.75 ± 1.09 mM (Fig., Table). Ibogaine displayed a notably lower agonist efficacy than noribogaine at OPRK with an E Max of 18 ± 1.4%, while 18-MC failed to stimulate [ 35 S]GTPgS binding to OPRK. In our assays, morphine and dynorphin A displayed E Max values of 91 ± 7% and 94 ± 7% respectively, and nalmefene, a partial agonist of OPRK, maximally stimulated at 35 ± 4.7%, which is similar to formally reported values. Noribogaine stimulation of [ 35 S]GTPgS binding was also observed in membrane preparations from rat midbrain tissues and this signal was nor-BNI-sensitive, confirming the agonist activity of noribogaine in native brain tissues (data not shown). The apparent coupling efficiencies of agonists DAMGO, U69,593, morphine, dynorphin A, nalmefene, 6 0 -GNTI, noribogaine and ibogaine at the G-protein pathway were calculated (pK i e pEC 50 ) and found to be congruent with EC 50 values shifted by ~1 log in comparison to apparent affinity K i for all agonists (Tablesand). The activation coupling efficiency of dynorphin A (0.6, this work) and 6 0 -GNTI (0.2,) were superior to other agonists in this assay, indicating that a lower fraction of bound receptors was sufficient to activate the pathway toward its maximal efficiency. Conversely, the hillslope for dynorphin A concentration response curves was below 1 (~0.7), indicating slower kinetics and probably a longer residence time in the receptor active conformations, producing the observed heightened coupling efficiency. TableBinding affinity of noribogaine and other drugs at the human mu (OPRM) and kappa (OPRK) opioid receptors. K i values of noribogaine, ibogaine, and 18-MC (n ! 3). Values for control ligands morphine, naloxone DAMGO, U69,593, dynorphin A, [met]-enkephalin, nalmefene, and buprenorphine were determined and/or gathered from the literature. Specificity for the OPRK receptor was evaluated using the equation DpK i ¼ pK i (OPRK)-pK i (OPRM). Agonists (red), partial agonists (purple), and antagonists (green)..
Inhibitory effects of kappa antagonists naloxone, nor-BNI, 18-MC, and the partial agonist nalmefene on the agonist-induced [ 35 S]GTPgS binding by full or partial agonists dynorphin A, U69,593, morphine, noribogaine and nalmefene were investigated. Concentration-response curve shifts of agonists were gathered in the absence and presence of fixed antagonist concentration (30 nM naloxone, 5 nM nor-BNI, 100 mM 18-MC, and 3 nM nalmefene) or in some instances with iterative concentrations of the antagonist (Fig., Table). All antagonists tested right-shifted the concentration-response curves of noribogaine, dynorphin A, morphine or U69,593 in a concentration-dependent manner. This finding is consistent with these ligands being surmountable competitors of the noribogaine and other tested agonists' binding site. Functional inhibition constants (K e ) of the antagonists are shown in Tablewith the assumption of ideal conditions of competitiveness and equilibrium. In all instances, the functional inhibition constants for these inhibitors were close to their K i , regardless of the agonist they were competing with, indicating that noribogaine was no different than other agonists tested.
Nalmefene is a partial agonist and can compete as a rival agonist against more efficacious agonists and inhibits their response up to the reduced efficacy of nalmefene itself, a pharmacologically TableA/B: Noribogaine modulation of [ 35 S]GTPgS binding (A) and b-Arrestin 2 recruitment (B) in CHOeK1 stably expressing human OPRK. The maximal stimulation (E Max ) and the functional activation constant (EC 50 ) of tested ligands dynorphin A (Dyn-A), U69,593, morphine, noribogaine, nalmefene, ibogaine, 18-MC, nor-BNI, and naloxone were measured using non-linear regression analysis. The functional inhibition constant (K e ) for GTPgS assay was calculated using EC 50 shifts with the equation: K e ¼ [Inhibitor]/(DR-1), where DR is the ratio of the agonist EC 50 in the presence and the absence of inhibitor or from concentration-inhibition curves in the presence of agonist at EC 80 using a modified Cheng-Prusoff equation: K e ¼ [IC 50 ]/(1 þ[agonist]/EC 50 ). Coupling efficiency (e-coupling, e-cpl) indicated the relationship between the apparent binding affinity (K i ) versus the EC 50 or the functional inhibition potency (K e ) of a given ligand and used the equations p(Ki/EC 50 ) and p(Ki/K e ), where p represents a elog10 transformation. [Ϟagonist] represents the agonist used to test inhibitory compounds. Outliers in comparison to reference compounds dynorphin A and U69,593 (activation), and nor-BNI and naloxone (inhibition) are underlined. Data are shown as the mean ± SE of (n) experiments. coherent phenomenon also known as mixed agonism/antagonism property. Noribogaine was a partial agonist at OPRK in the [ 35 S] GTPgS binding stimulation assays (Fig.). Therefore, we set out to determine if noribogaine was also a rival agonist that can functionally compete with and reduce the activity of more efficacious agonists than itself. Dynorphin A and morphine concentration-response curves were performed in the presence and absence of rival agonists nalmefene or noribogaine at concentrations of 36-fold and 79-fold their respective K i (nalmefene 3 nM, noribogaine 50 mM) (Fig.). Nalmefene readily shifted the EC 50 of dynorphin A and morphine to the right, with a functional inhibition constant K e of 0.077 ± 0.016 nM and 0.11 ± 0.005 nM. Calculated K e were within a close range of the Ki (0.08 nM), similar to a competitive antagonist such as nor-BNI or naloxone (Table, Fig.). Noribogaine, on the other hand, poorly shifted the EC 50 of these agonists and the K e estimates in these conditions were 40 ± 16 mM and 15 ± 4 mM respectively, about 40fold its K i (Table, Fig.). Noribogaine against dynorphin A, U69,593, morphine, and nalmefene, was an outlier rival agonist in all instances compared to the typical partial agonist nalmefene and antagonists like naloxone, nor-BNI, and 18-MC (Table, underlined values, Fig.). These data showed that noribogaine was an atypical partial kappa agonist because it was unable to functionally surmount receptor activation of other agonists of the OPRK G-protein pathway.
In another set of experiments (Fig.), noribogaine and nalmefene concentration-inhibition curves were produced in the presence of more efficacious agonists such as U69,593, dynorphin A, or morphine. In the case of nalmefene, noribogaine was used as the agonist. Nalmefene readily reduced the signal of moderate to high concentrations of rival full (U69,593) or partial (noribogaine) agonists to its own reduced levels (30%) with an apparent IC 50
Bias quantification at the kappa and mu opioid receptors. E-coupling for activation and inhibition is a measure of the coupling efficiency, using pKi-pEC 50 and pKi-pK e , where p is elog10 transformation. Bias-coupling represents the relative preference of coupling between the G-protein or b-arrestin pathways and used the difference in e- proportional to the rival agonist concentration (including noribogaine) (Fig.) and a K e consistent with its K i (Table). Noribogaine reduced the signal of more efficacious agonists to its own signal (70%), but the IC 50 values were high (100e300 mM range) and the calculated K e were consistent with poor functional inhibition. The apparent functional activation constant (apparent EC 50 ), of dynorphin A, U69,593 and morphine were then estimated in the presence of noribogaine or nalmefene (Table). Apparent EC 50 for all agonists tested in the presence of nalmefene were close to their experimental EC 50 . However, in the presence of noribogaine, apparent EC 50 of common kappa agonists were significantly lower than their actual experimental EC 50 (Table, underlined values). This atypical phenomenon may reflect the stabilization of a differential set of active receptor conformations by noribogaine compared to other kappa agonists tested. Altogether, these data suggest a potential protean agonism of noribogaine, a predictive functional biasthat we set to assess further.
PathHunter b-Arrestin GPCR assays detecting the interaction of b-arrestin 2 with the activated receptor were used to measure non-G protein OPRK activity in CHOeK1 live cells. Concentration-response curves to noribogaine were compared to full endogenous agonist dynorphin A drug treatment (Fig.). Calculated EC 50 values, maximal responses and coupling efficiencies are shown in Table. Dynorphin A displayed an EC 50 of 11 ± 2 nM at OPRK. Noribogaine exhibited a profound functional bias at OPRK and was marginally efficacious at inducing b-arrestin recruitment with an E Max of 13 ± 3% and an estimated EC 50 of 110 nM. For comparative purposes, nalmefene was tested once and displayed an E Max of 30% as expected for this unbiased partial agonist. EC 50 and E Max values fromwere collected for 6 0 -GNTI, and showed that noribogaine was activating at the same level as 6 0 -GNTI. Ibogaine and 18-MC were tested once; they did not display any bias and their pharmacological behavior was comparable to typical kappa ligands. The functional activation coupling efficiencies in the b-arrestin assay were calculated in a similar manner as the [35S]GTPgS binding assay and showed that dynorphin A and U69,593 displayed a similar coupling efficiency of 2, reflecting the intrinsic assay efficiency (Table, Fig.). Noribogaine had a negative activation coupling efficiency of À0.9 for the b-arrestin pathway, indicating a profound coupling bias of 1:630 in favor of this pathway in comparison to the G-protein pathway (Table). In comparison, the biased ligand 6 0 -GNTI did not display activation coupling bias between the G-protein and the b-arrestin pathways (Table). The activation efficacy bias between the G-protein pathway and the b-arrestin pathway was evaluated by comparing the maximal efficacy (E Max ) for each pathway in comparison to reference ligand dynorphin A or U69,593 (1e1 maximal efficacies ratio) (Table). Although nalmefene is a partial agonist, it did not display an activation efficacy bias, reflecting that partial agonism is not an indication of bias per se. Both noribogaine and 6 0 -GNTI displayed an activation efficacy bias in favor of the G-protein pathway and noribogaine was almost twice more biased than 6 0 -GNTI (5.5e1 versus 3 to 1).
Because noribogaine was unable to induce the recruitment of barrestin following activation of the OPRK receptors, it was then tested for its ability to inhibit b-arrestin recruitment induced by efficacious agonists of this pathway (Fig.). In these assays, barrestin recruitment was induced by the endogenous agonist dynorphin A sets at its EC 80 concentration and challenged with increasing concentrations of noribogaine. Noribogaine inhibited dynorphin A-induced b-arrestin recruitment up to ~60%, with an IC 50 of 1 ± 0.16 mM (Fig., Table). In contrast, when a similar TableNoribogaine modulation of [ 35 S]GTPgS binding (A) and b-Arrestin 2 recruitment (B) in CHOeK1 stably expressing human OPRM. The maximal stimulation E Max and the functional activation constant EC 50 of tested ligands [met]-enkephalin (Met-EnK), DAMGO, morphine, noribogaine, ibogaine, 18-MC, and naloxone were calculated using nonlinear regression analysis. The functional inhibition constant K e for GTPgS assays was calculated using EC 50 shifts with the equation: K e ¼ [Inhibitor]/(DR-1) where DR is the ratio of the agonist EC 50 in the presence and the absence of inhibitor. For b-Arrestin assays, K e was calculated from concentration-inhibition curves in the presence of agonist at EC 80 using a modified Cheng-Prusoff equation: K e ¼ [IC 50 ]/(1 þ[agonist]/EC 50 ). Coupling efficiency (e-coupling, e-cpl) indicated the relationship between the apparent binding affinity K i versus the apparent functional activation potency (EC 50 ) or the functional inhibition potency (K e ) of a given ligand and used the equations pKi-pEC 50 and pKi-pK e where p represents a elog10 transformation.[Ϟagonist] represents the agonist that was used to produce concentration-response curves in the presence and the absence of tested inhibitory compounds noribogaine, ibogaine, naloxone, and 18-MC. Data are shown as the mean ± SE of (n) experiments. noribogaine concentration-inhibition experiment was performed on the G-protein pathway with dynorphin A at its EC 80 , noribogaine only inhibited agonist-induced activation of the receptor by ~5%
with an IC 50 of ~150 mM (Fig.). Noribogaine was thus 150-fold more potent at inhibiting dynorphin A-induced b-arrestin 2 recruitment than at inhibiting dynorphin A-induced G-protein activation (Table). For comparison, typical partial agonist nalmefene inhibited dynorphin A-induced b-arrestin recruitment up to 70% with an approximate K e of 0.32 nM, whereas it inhibited dynorphin A-induced GTP-binding with a K e of 0.08 nM and was apparently equipotent with only a 4-fold difference. Notably, the apparent inhibition coupling efficiencies for antagonists of the barrestin pathway were all clustered in a 0.5 range (Fig.) and indicated that the current experimental conditions were producing an assay slightly less sensitive to antagonists than the GTP-binding assay (coupling efficiency close to 0) with a robust assay's intrinsic coupling efficiency reproducibility. Finally, the inhibition efficacy bias for noribogaine was 1:2 in favor of the b-arrestin pathway, which is similar to 6 0 -GNTI (Table), whereas other antagonists were unbiased.
[ 35 S]GTPgS binding to membrane preparations of CHO cells stably expressing human OPRM were examined in response to noribogaine, ibogaine, and morphine drug treatment (Fig., Table). The prototypical full agonist DAMGO and the endogenous agonist [met]-enkephalin ([met]-Enk) were used as reference agonists. Morphine was a partial agonist with an E Max of 80 ± 4.5% and an EC 50 of 32 ± 1.2 nM. The partial agonist buprenorphine stimulated OPRM with an E Max of 26 ± 2.2% in our assays and in previously reported assays. Noribogaine marginally stimulated [ 35 S]GTPgS binding to OPRM, with an E Max of 10% of the full agonist DAMGO or [met]-enkephalin (Fig., Table) and comparable to the level of activation previously reported. Ibogaine and 18-MC did not stimulate the OPRM G-protein pathway. Concentration-response curves of barrestin recruitment of noribogaine at the OPRM receptors were compared to full agonist [met]-enkephalin (Fig.). [Met]enkephalin displayed an EC 50 of 159 ± 38 nM at OPRM. As expected given its marginal efficacy at the G-protein pathway, noribogaine was not a mu agonist of the b-arrestin pathway. Calculated EC 50 values, maximal responses and coupling efficiencies are shown in Table.
Noribogaine marginally stimulated [ 35 S]GTPgS binding via OPRM with an approximate EC 50 of 16 mM (Fig.). Therefore, we investigated if noribogaine was an antagonist of OPRM. [Met]enkephalin, DAMGO and morphine dose responses were carried out in the presence and absence of 15 and 150 mM of noribogaine (Fig.and). Noribogaine was an inhibitor of all agonists tested and shifted their EC 50 to the right by a magnitude of ~1Àlog. The calculated functional inhibition constant (K e ) values were ~20 mM (Table, and Tablefooter). In a similar design, naloxone displayed a K e of 1.2 ± 0.3 nM, a value close to its K i at OPRM, like its inhibition constant at OPRK (Tablesand). On the contrary, noribogaine was an outlier antagonist with an inhibitory coupling efficiency of 1.1, while ibogaine, naloxone, and 18-MC were still behaving like typical antagonists with clustered e-coupling values close to 0 (Fig.and Table). Noribogaine was then tested for its ability to inhibit [met]-enkephalin-induced b-arrestin-2 recruitment at OPRM (Fig.and Table). Noribogaine inhibited agonist responses by 80e100% with an IC 50 of 100 ± 25 mM. Regardless of the pathway tested (G-protein and b-arrestin pathways), noribogaine displayed similar functional inhibition constants and was an unbiased ligand of the mu opioid receptor (Table), albeit an inhibitory coupling efficiency outlier (Table/ B) in comparison to other mu antagonists. In fact, noribogaine also decreased the E Max of [met]-enkephalin, DAMGO and morphine (Fig.and), indicating a degree of unsurmountable antagonism by noribogaine in functional assays. Apparent unsurmountable antagonism can encompass several distinct molecular mechanisms such as (a) irreversible competitive antagonism, (b) noncompetitive antagonism, or (c) functional antagonism; for review. 3.9. Binding model of noribogaine and ibogaine to the inactive conformation of the mu opioid receptor Noribogaine showed a profile of unbiased antagonist at the OPRM and stabilized the inactive conformation of the receptor. Thus, it was possible to use an available crystal structure of the inactive form of the OPRM to perform in silico binding experiments with noribogaine. We developed an in silico binding model based on the mouse OPRM co-crystal structure [PMID 22437502]as described in methods. The mouse and human OPRM share 94% (global) sequence identity and all binding site residues are identical. After initial optimization of the model, the top docking poses of noribogaine and ibogaine were pharmacophorically aligned with the co-crystal morphinan antagonist as one would expect: the hydrophobic ibogaine and noribogaine bicyclic system and ethyl substituent with morphinan cyclopropyl residues were spatially aligned and the positively charged tertiary amines were superimposed with each forming a hydrogen bond to the site chain of Asp147. Then, the noribogaine and ibogaine OPRM complexes were each used in a 12 ns all atom explicit water molecular dynamics simulation (see methods). Trajectory analysis revealed the most prevalent interactions of noribogaine (Fig.) and ibogaine. Both ligands formed a stable hydrogen bond with Asp147 via their tertiary amine. Noribogaine and ibogaine formed pi-cation interaction with Tyr148 (64% and 56%, respectively), and hydrophobic interactions with His297 (64 and 93%, respectively). Further hydrophobic interactions were observed between Val236 (~40 and ~60%, respectively), Tyr326 (~20 and ~40% respectively), Met151 (~20% and ~30%, respectively) and also Trp293, Ile296, Val300. Characteristically, noribogaine, but not ibogaine, formed a water bridge with Tyr148 for 34% of the simulation time. Both ligands showed a hydrogen bond with His297 for about 20% of the simulation. Movies of the simulations were generated and are available as supporting material. A representative illustration frame of noribogaine in the OPRM was extracted from the simulation and is shown in Fig.. Supplementary data related to this article can be found online at.
Historically, in vivo studies excluded the possibility of prototypical morphine-like agonistic mechanism of ibogaine and its metabolite noribogaine at the mu and/or kappa opioid receptors (see introduction). In the other hand, naloxone-like antagonistic mechanisms could not be demonstrated either. These ambiguous results lead to a gap of knowledge between the correlation of receptor-ligand mechanistics and the intriguing and beneficial effects of ibogaine and its metabolite noribogaine with opiate drugs and to the opioid system. Our current study now demonstrates that noribogaine is in fact a G-protein biased kappa agonist, and a weaker mu antagonist. The described potencies in our study closely match physiologically relevant brain concentrations of therapeutically effective doses of noribogaine. Thus, noribogaine's atypical modulatory properties of the kappa opioid system, and to a certain extend the mu opioid system, in the central nervous system (CNS) must be taken into account for its mechanism of action in vivo. Our study shows that noribogaine is a kappa agonist. Intravenous noribogaine potently triggered the release of prolactin in rats and it was reported that this effect was centrally mediated. We propose that noribogaine-induced prolactin release is mediated by central OPRK, similar to what was described for the kappa agonist nalmefene. Our data also show that noribogaine was a moderately potent mu antagonist. Thus, noribogaine also belongs to the class of dual agonist/antagonist kappa-mu opioid ligands. In comparison, ibogaine is a more potent mu antagonist and a much weaker kappa agonist than noribogaine. Ibogaine is metabolized to noribogaine in mammals and, as a consequence, the acute and delayed biological effects of ibogaine treatment can be attributed to the ibogainenoribogaine mixture rather than to ibogaine only. Intermediate levels of drug-related interoceptive stimuli generalization were observed with the mixed action opiates SKF10,047 (79%), pentazocine (74%), nalorphine (70%), and diprenorphine (75%). These results provide support to the notion that the ibogaine-noribogaine mixture could be recognized as a mixed opioid agonist-antagonist in vivo. 18-MC was a micro-molar non-specific competitive antagonist of the kappa and mu receptors; however it should remain clear that noribogaine belongs to a different class of opioid ligand than ibogaine or 18-MC. In consequence, the in vivo effects of 18-MC, experimentally demonstrated or hoped for, cannot rely on the pharmacological knowledge and demonstrated effects of noribogaine on the opioid system. A series of studies (see introduction) described the biological effects of noribogaine treatment in reducing the stimulating and reinforcing properties of morphine, possessing a modest analgesic power on its own, and having the capability to potentiate morphine analgesia, especially in morphine tolerant animals. Our study shows that noribogaine inhibited b-arrestin-2 recruitment of both mu and kappa agonists. This could to a certain extent mimic the effects of a functional deletion of the b-arrestin-2 gene which resulted in remarkable potentiation and prolongation of the analgesic effect of morphine in mice. Additionally, it was shown that dynorphin A attenuated morphine tolerance at doses devoid of observable effects on nociception. U50,488 also attenuated morphine tolerance and potentiated morphine analgesia in morphine tolerant animals at doses devoid of analgesic activity when co-administrated with morphine. Taken together, these studies show that low doses of OPRK agonists can reduce morphine tolerance and desensitization. Future studies assessing the comparative activity of noribogaine in wild type and genetically modified opioid receptor animal models should provide further understanding of the contribution of the kappa receptors in the effects of noribogaine. Ultra-low doses of opioid antagonists (naloxone and naltrexone), which selectively inhibit the excitatory effects of opioid agonists, have been reported to augment systemic morphine analgesia and inhibit the development of tolerance/physical dependence. In vitro assays demonstrated that these phenomena are reproducible at the cellular level where ultralow concentrations of opioid ligand effectively synergized the binding and functional response of agonists via OPRM-OPRD heteromers. In this regard, the moderate OPRM inhibitory activity of noribogaine at moderate to low therapeutic doses could be of biological relevance and warrants further investigation at the molecular level on OPRM-OPRD heteromers. In line with their localization in the hippocampus, amygdala, hypothalamus, striatum and spinal cord, the function of the kappa opioid receptors are related to learning and memory, emotional control, stress response and pain. Kappa agonists hold therapeutic potential for mood and addiction disorders, psychiatric comorbidities, and pain management, however they also induce undesirable on-target side effects such as place aversion, dysphoria and anhedonia; and hallucinations at high doses. On the other hand, kappa antagonists hold therapeutic potential as antidepressants and anxiolytics, but may induce hyperalgesic states. Thus, until recently, kappa ligands were not considered for therapeutic development due to their undesirable side effects. This view changed with the introduction of a more elaborate understanding of ligand-receptor pharmacology and the characterization of receptor drugs exhibiting functional selectivity, for review. As reviewed in, certain opioid ligands were reported to be G protein-biased agonists, notably at the mu receptor, and exhibited interesting physiological properties. Recent elegant studies in rodents have mechanistically linked the activation of p38 MAPK to stress-mediated OPRK stimulation via the b-arrestin mediated transduction pathway. In this frame, G-protein biased kappa agonists were described as hypothetical analgesic drugs without aversive and dysphoric components. In our study, noribogaine exhibited a profound functional bias and was not an agonist of the OPRK b-arrestin pathway. Therefore, noribogaine appears to carry the prerequisite pharmacological characteristics of an analgesic kappa opioid drug devoid of aversive and dysphoric effects which may become important during opiate detoxification and painful states. Our study also shows that noribogaine is a b-arrestin biased kappa antagonist, more potent and efficacious at inhibiting agonistinduced signaling to this pathway than to the G-protein pathway. At a concentration corresponding to levels in the brain of rats (e.g. 0.5e5 mM) several hours after noribogaine dosing in animals, or days after ibogaine treatment in humans, noribogaine tested in vitro preserved the signaling of dynorphin A to the G-protein pathway while markedly inhibiting b-arrestin recruitment. Thus noribogaine in the presence of dynorphin was able to modulate the functional selectivity of dynorphin. This peculiar pharmacological property could contribute to antagonist-like anti-depressive and anxiolytic activities against stress-induced and over-active dynorphin/kappa system, as seen during drug dependence, drug withdrawal and cravings. This functionally selective inhibitory kappa activity may also mediate positive effects against stress and anxiety, and atypical depression that will be considered in future studies. Ligand-induced functional selectivity of otherwise unbiased agonists was previously demonstrated for some receptors of the GPCR family interactions with conformation-specific allosteric modulators of the receptors (i.e. the allosteric ligand LPI805 for the NK2 receptor; for review). However, in the present study noribogaine does not appear to be an allosteric ligand: 1) it directly competed with the binding of orthosteric radiolabeled agonists DAMGO, U69,593; 2) it displayed functional competitive behavior with opioid orthosteric antagonists in GTPgS assays; 3) it was docked to the morphinan orthosteric binding site of the OPRM inactive state in silico with good stringency: in silico binding experiments provide basis for noribogaine's interacting moieties with the inactive conformational state crystal structure of the mu receptor and in silico interaction scores of noribogaine and ibogaine are congruent to their affinity ranking in in vitro radioligand assays. In fact, our data suggest that noribogaine would induce functional selectivity to dynorphin A via the interplay of a set of active and inactive conformational states. Certain conformations would be easily accessible to other agonists (e.g. the inactive conformations and active G-protein conformations) and other conformations would be energetically challenging to populate in place of noribogaine (e.g. the non-recruiting barrestin conformation). Multiple studies provide evidence for the existence of intermediate conformational states linking the inactive receptor to the fully active receptor. Agonist binding and activation of GPCRs has been proposed to occur through a multistep process; for review. The intermediate conformational states generated during multistep agonist binding may have unique functional properties as it is known that GPCR can couple to different Gproteins and also activate non-G protein dependent pathways depending on their conformational state(s); for review. Interestingly, recent investigations in drug design described an allotropic binding mode for certain OPRK agonists, which encompassed sequential drugereceptor interaction mechanisms. In this regard, noribogaine merits further investigation at a deeper molecular level using discrete drugereceptor interactions and conformational dynamic designs as well as cellular designs to assess further these potential allotropic binding modalities and their functional consequences.
This study shows that noribogaine is a dual ligand of both mu and kappa opioid receptors. Noribogaine exhibits a profound Gprotein biased agonism at the opioid receptors and can modulate dynorphin signaling via the kappa receptor. This study clarifies the mechanisms of noribogaine at modulating the function of opioid receptors at the cellular level, providing ground for explanatory mechanisms at the opioid system in vivo as well as new avenues of therapeutic development and applicability.
Create a free account to open full-text PDFs.
Forstmann, M., Kettner, H. S., Sagioglou, C. et al. · Journal of Psychopharmacology (2023)
Köck, P., Frölich, K., Walter, M. et al. · Journal of Substance Abuse Treatment (2022)
Lason, W., Carnicella, S., Mash, D. C. et al. · Frontiers in Pharmacology (2019)
Wasko, M. J., Witt-Enderby, P. A., Surratt, C. K. · ACS Chemical Neuroscience (2018)
Wilkins, C., Dos Santos, R. G., Solá, J. et al. · Journal of Psychedelic Studies (2017)
Mash, D. C., Ameer, B., Prou, D. et al. · Journal of Psychopharmacology (2016)