Pharmacological Mechanism of the Non-hallucinogenic 5-HT2A Agonist Ariadne and Analogs

Classic psychedelics produce transient altered states of consciousness mainly via activation of the serotonin 5-HT2A receptor, and renewed clinical interest in these compounds has prompted two enduring questions: which therapeutic effects require subjective psychedelic experiences, and whether therapeutic benefit can be achieved by compounds that lack hallucinogenic activity. Cunningham and colleagues situate Ariadne, a phenylalkylamine analogue closely related to the hallucinogen DOM but differing by one methylene at the α-position, within this conceptual frame. Historical and limited clinical reports—including Bristol-Myers trials and Shulgin's observations—suggest Ariadne lacks hallucinogenic effects while showing striking therapeutic signals across psychiatric and neurological indications, yet its molecular pharmacology has remained largely uncharacterised. This study aims to fill that gap by systematically characterising Ariadne and a series of analogs in vitro and in vivo, comparing them to hallucinogenic DOx congeners. The investigators set out to determine receptor binding and functional signalling profiles (including G protein and β-arrestin pathways), to test structure–activity relationships for α-substituent length, to model ligand–receptor interactions, and to evaluate behavioural readouts in mice, notably the head twitch response (HTR) and a pilot auxilin knockout model of Parkinson's disease (PD). The goal is to generate a molecular hypothesis that explains Ariadne's apparent absence of hallucinosis alongside its reported therapeutic effects.

The chemical series was prepared from substituted benzaldehydes via Henry condensation to nitrostyrenes followed by reduction (typically LiAlH4, with alternative alane protocols where dehalogenation risk was present). A variety of 4‑position substituents were introduced using established cross-coupling and functional-group transformations; enantiomers were separated by chiral supercritical fluid chromatography. Full synthetic and characterisation details are provided in the supporting information. Pharmacological profiling began with a 44-target radioligand displacement screen (SafetyScreen44, Eurofins-Panlabs) at 10 μM of racemic Ariadne. Functional activity across a panel of human 5-HT receptors was evaluated with a G protein activation BRET assay (the “5-HTome” screen). For detailed 5-HT2A signalling, the team used BRET-based G protein‑dissociation assays (Gα‑Rluc8 with GFP2‑γ) and β‑arrestin2 recruitment BRET assays in HEK293T cells, with transfections and measurements specified in the methods. Calcium flux was measured in tetracycline-inducible Flp-In 293 T‑Rex cell lines using Fluo‑4 and FLIPR readout. Radioligand displacement binding assays used I-DOI to estimate Ki values. Drug dilutions, incubation times, and data normalisation procedures are described; EC50 and Emax values were obtained by nonlinear regression (log[agonist] vs response) using GraphPad Prism. In vivo work used mice to assess the head twitch response (HTR) as a proxy for 5-HT2A-mediated hallucinosis potential, with DOPR (a DOx phenylalkylamine) included as a hallucinogenic reference. Pharmacokinetics (brain penetration) were measured to place behavioural time windows in context. General locomotion (open field), elevated plus maze and novelty-suppressed feeding were used to profile acute and delayed behavioural effects. A pilot reverse-translation study employed auxilin knockout (Aux KO) mice as a genetic PD model; groups were WT or Aux KO treated with vehicle or (R)-Ariadne (10 mg/kg i.p.), with pre- and post-treatment assessments including balance beam and hind-limb clasping. Group sizes for the PD pilot were provided: WT vehicle n = 5, WT (R)-Ariadne n = 5, Aux KO vehicle n = 6–7, Aux KO (R)-Ariadne n = 7. Finally, molecular docking and comparative structural modelling used active-state 5-HT2A (PDB 6WHA) and related 5-HT2B/2C structures to rationalise steric effects of α‑substituent length; docking and local minimisation were performed in ICM‑Pro.

Broad receptor profiling identified 5-HT2A and 5-HT2B as the only hits above the 50% radioligand displacement threshold at 10 μM racemic Ariadne. There was essentially no binding to monoamine transporters; SERT activity was minimal in a transporter uptake assay (IC50 > 50 μM). In the functional 5-HT panel, racemic and (R)-Ariadne displayed robust agonism at 5-HT2A/2B/2C (Emax > 70%, EC50 < 1 μM), with little or no activity at 5-HT4, 5-HT5A and 5-HT7. Activities at 5-HT1 subtypes were in the high‑hundreds of nanomolar range, except 5-HT1B where racemate EC50 was ∼4 μM. At 5-HT2A, Gq dissociation assays reported higher potency for (R)-Ariadne versus racemate (reported EC50 ≈ 149–185 nM for Gq depending on assay), and calcium flux assays showed EC50 ≈ 30 nM with Emax ≈ 96% (R)-Ariadne in one readout; the (S)-enantiomer exhibited substantially lower efficacy (e.g. calcium Emax ≈ 63%). Radioligand displacement yielded Ki values at human 5-HT2A of 120 nM (racemate), 53 nM ((R)-Ariadne) and 220 nM ((S)-Ariadne). Overall, Ariadne behaved as a partial agonist versus the reference agonist 5-HT in some assays. Comparisons with hallucinogenic analogs revealed a consistent shift: Ariadne showed approximately 4–6-fold weaker potency than DOM across Gq/G11 and β‑arrestin2 assays, together with modest but consistent reductions in signalling efficacy (example: Emax/Gq ≈ 80% for Ariadne versus 96% for DOM; β‑arrestin2 recruitment shifted from ≈103% to ≈83%). Extending the α‑substituent further (α‑propyl analogs such as 5C‑D) produced larger drops in potency (20–30-fold) and further reductions in efficacy. Docking into the active 5-HT2A structure suggested steric crowding of longer α‑chains against conserved W/F side chains in an orthosteric subpocket, providing a structural rationale for the potency/efficacy trend. In mice, DOPR elicited a robust, dose-dependent HTR peaking at 1 mg/kg s.c., with a rapid onset. Racemic Ariadne produced a markedly attenuated HTR (>3-fold lower head-twitch counts per 15 min) with a maximal effect at a one-log higher dose (10 mg/kg s.c.). Enantiomeric rank order of HTR magnitude mirrored in vitro Gq potency. Pharmacokinetics showed good brain penetration and indicated that behavioural sampling occurred during the ascending PK phase. Several Ariadne analogs (e.g. 4C‑MOM and 5C‑D) produced minimal HTR, and the HTR elicited by (R)-Ariadne and rac‑4C‑TFM was abolished by the 5-HT2A antagonist MDL100,907, confirming 5-HT2A mediation. Open-field testing showed minor sedation at the HTR peak dose and greater sedation at higher doses; elevated plus maze and centre‑time measures suggested an acute anxiogenic‑like effect, whereas novelty-suppressed feeding measured 7 days post‑drug indicated a dose-dependent trend toward reduced latency to feed at 10 mg/kg, suggestive of a lasting anxiolytic‑like effect. In the Aux KO PD pilot, a single intraperitoneal dose of (R)-Ariadne (10 mg/kg) given 15–30 min before testing produced significant and rapid improvement in balance-beam performance (increased runs/minute and shorter times to cross) and abolished hind-limb clasping in Aux KO mice, bringing measures close to WT levels. WT mice also showed some improvement after (R)-Ariadne. The investigators consider these effects comparable to previously reported L‑DOPA responsiveness in this model, but note the pilot study's mixed-age/sex cohorts and modest group sizes.

Cunningham and colleagues interpret the convergent in vitro and in vivo data as supporting a signalling‑efficacy explanation for Ariadne's non‑hallucinogenic phenotype. Specifically, Ariadne and its α‑ethyl analogs act as 5-HT2A agonists but show reduced signalling potency and a modest but consistent drop in efficacy across multiple 5-HT2A-coupled pathways (Gq/G11, calcium dynamics and β‑arrestin2 recruitment) when compared to closely related α‑methyl hallucinogens such as DOM. The attenuated mouse HTR profile aligns with these biochemical observations and with earlier preclinical and limited clinical reports that Ariadne lacks classic hallucinogenic effects despite producing other psychoactive and therapeutic signs. The authors frame a signalling‑efficacy hypothesis: reduced maximal 5-HT2A signalling in key circuits implicated in hallucinosis yields a non‑hallucinogenic clinical phenotype while preserving therapeutic activity in other circuits where receptor reserve or downstream amplification may be greater. To support this idea they cite an analogous tryptamine pair (DET versus 6F‑DET), which shows a similar pattern of minor loss of signalling potency and efficacy aligned with loss of psychedelic effects in a small human study. However, the investigators caution that the hypothesis is limited to closely related structural analogs and is not universally applicable across all 5-HT2A agonists; psilocin is noted as a counterexample with relatively modest signalling efficacy but robust psychedelic effects, likely due to differences in potency, receptor subtype pharmacology and other properties. Key limitations acknowledged include the interpretative boundaries of the HTR (efficacy versus potency confounds and scaffold dependence), incomplete disclosure and anecdotal nature of historical human data, potential pharmacokinetic differences between compound classes, and the pilot nature of the Aux KO PD study (mixed ages/sexes and small n). The authors propose several mechanistic routes for Ariadne's apparent therapeutic effects in PD, favouring an indirect enhancement of dopaminergic transmission via 5-HT2A/C‑mediated circuit effects or by promoting mitochondrial biogenesis, but they underline that these mechanisms remain speculative and require targeted follow-up. Finally, the paper situates Ariadne within an emerging class of non‑hallucinogenic 5-HT2A agonists that could offer take‑home psychiatric or neurological therapeutics, and it calls for further pharmacological, mechanistic and translational work to test the signalling‑efficacy framework and to replicate therapeutic signals in controlled clinical settings.

The authors conclude that Ariadne exemplifies a non‑hallucinogenic 5-HT2A agonist with a pharmacological profile that plausibly explains both its historical clinical signals and lack of hallucinosis. On the basis of the combined receptor profiling, functional signalling assays, structural modelling, behavioural pharmacology and a pilot PD model, they propose that modest reductions in 5-HT2A signalling potency and efficacy account for the attenuated hallucinogenic potential while allowing therapeutic effects to manifest. Ariadne is presented as a prototype for a potentially valuable new class of medicines, meriting renewed investigation.