Structural pharmacology and therapeutic potential of 5-methoxytryptamines

The structural component used cryo-EM to determine multiple drug-bound 5-HT1A–Gi signalling complex structures. Human 5-HT1A constructs were engineered (N-terminal BRIL fusion replacing residues 1–24 and an L125W stabilizing mutation) and co-expressed with a heterotrimeric G protein complex (Gβ1 and a Gγ2–Gαi1 fusion with stabilizing point mutations). Receptor–G protein complexes were stabilised with ligands (5-MeO-DMT, LSD, 4-F,5-MeO-PyrT, buspirone, vilazodone), purified in detergents (DDM then LMNG) and subjected to cryo-EM data collection on Titan Krios microscopes. Maps were built in Coot and refined in PHENIX/ServalCat; sterols were modelled as cholesterol hemisuccinate though co-purified cholesterol could not be fully excluded. Medicinal chemistry produced a panel of 5-methoxytryptamine analogues via established synthetic routes (oxalylation–amidation–reduction or Speeter–Anthony), including amine cyclisations and indole substitutions such as 4-fluorination. Structure–activity relationships were characterised in vitro using BRET-based G protein activation assays: G i1 reporters for 5-HT1A and G q reporters for 5-HT2A. Complementary assays included cAMP-based readouts, PRESTO-Tango β-arrestin recruitment screens across 5-HT receptors, and transporter uptake inhibition assays (SERT, OCT1/2, PMAT). Receptor mutagenesis probed key residues identified from structures (for example F3616.51, N386, A365). In vivo pharmacology employed male C57BL/6J mice for behavioural testing and CD-1 aggressor mice in a chronic social defeat (SD) stress paradigm. Pharmacokinetic and brain-penetration studies were performed with s.c. dosing; plasma and brain concentrations were quantified by LC–MS/MS and non-compartmental analysis. Behavioural endpoints included open-field locomotor activity (to index 5-HT1A-mediated sedation), the head-twitch response (HTR) as a 5-HT2A-mediated proxy for psychedelic-like effects, a two-phase social interaction (SI) test and a sucrose-preference test for anhedonia. WAY-100635, a selective 5-HT1A antagonist, was used to test receptor mediation of behavioural effects. Experimental blinding, group sizes and statistical approaches (two-way ANOVA with post hoc tests, Fisher’s exact test for resilience classification) are reported for the behavioural assays.

BRET signalling assays established that the tested tryptamine psychedelics are full or near-full agonists at 5-HT1A-mediated G protein signalling versus serotonin. Potency and selectivity varied across ligands: DMT, psilocin and mescaline were more potent at 5-HT2A, whereas LSD and 5-MeO-DMT showed roughly equipotent activity at 5-HT1A and 5-HT2A. In these cellular assays, 5-MeO-DMT had a G i BRET EC50 ≈ 25.6 nM. Cryo-EM structures of 5-MeO-DMT– and LSD–bound 5-HT1A–Gi complexes were solved at nominal global resolutions of 2.79 Å and 2.64 Å, respectively. Both ligands are anchored by the conserved ionic interaction with D1163.32 and bind slightly deeper (~0.8 Å) in the orthosteric pocket than serotonin, forming a hydrogen bond between the indole nitrogen and T1213.37 not observed for serotonin. These interactions produce a rotation of the ligand amine–D116 axis (≈8.5° for 5-MeO-DMT; LSD shows a larger ≈15.4° rotation compared with previously observed poses at 5-HT2 receptors) and reveal distinct binding modes for LSD at 5-HT1A versus 5-HT2A. Systematic SAR around the 5-MeO-DMT scaffold identified amine cyclisation and 4-position indole fluorination as major determinants of 5-HT1A potency and 5-HT1A versus 5-HT2A selectivity. Converting the dialkylamine of 5-MeO-DMT to a pyrrolidine (5-MeO-PyrT) increased 5-HT1A potency ≈12-fold (G i BRET EC50 = 2.1 nM) and reduced 5-HT2A potency, yielding ≈38-fold selectivity. A 3-pyrroline variant further improved potency (EC50 = 0.3 nM). Introducing a 4-fluoro substituent on the indole increased 5-HT1A potency ≈10–14-fold while decreasing 5-HT2A potency several-fold. Combining amine cyclisation and 4-fluorination produced very potent, highly selective compounds such as 4-F,5-MeO-PyrT (G i BRET EC50 = 370 pM) and 4-F,5-MeO-3-PyrrolineT (EC50 = 220 pM); 4-F,5-MeO-PyrT displayed >800-fold selectivity for 5-HT1A over 5-HT2A. A 2.85 Å cryo-EM structure of 4-F,5-MeO-PyrT–bound 5-HT1A showed the ligand occupying the same binding pose with conserved interactions to D116 and T121. Mutagenesis of residues lining the pocket implicated F3616.51, Y390 and N386 in stabilising small cyclic amines — mutation of F361 to smaller hydrophobic residues reduced potency of ligands with smaller amine substituents more than that of 4-F,5-MeO-PyrT. Substitutions at position 7.39 and the unique A365 residue of 5-HT1A influenced potency for indole-substituted tryptamines; A365→Asn reduced potency broadly, consistent with A365 accommodating diverse 5- and 4-substituents. Parallel mutation of N3436.55 in 5-HT2A suggested this residue forms a hydrogen bond with 5-hydroxyl groups in serotonin-like ligands, explaining reduced 5-HT2A potency of 4-fluorinated compounds. The authors also solved cryo-EM structures of clinically used 5-HT1A-targeting drugs buspirone and vilazodone bound to 5-HT1A (nominal resolutions 2.62 Å and 2.94 Å). Despite distinct scaffolds, their indole/aryl cores occupy the orthosteric binding pocket; vilazodone formed a similar T121 hydrogen bond and extended towards the extracellular space, whereas buspirone assumed a kinked pose that displaces F112. Functional assays showed buspirone (Emax = 93.4% of 5-HT), vilazodone (Emax = 97.4%), 4-F,5-MeO-PyrT (Emax = 102.8%) are high-efficacy 5-HT1A agonists, while aripiprazole had lower efficacy (Emax = 77.1%). Vilazodone had high potency (G i BRET EC50 ≈ 480 pM), comparable to 4-F,5-MeO-PyrT. Pharmacokinetic and in vivo pharmacology for 4-F,5-MeO-PyrT indicated brain penetration with a total brain-to-plasma ratio of 3.3 and an unbound brain-to-plasma ratio of 0.91, Cmax brain ≈ 143 ng ml−1 after 1 mg kg−1 s.c., and largely cleared within 2 h. Estimated free brain concentration at Tmax was ≈100 nM following 1 mg kg−1, predicted to strongly engage 5-HT1A (in vitro EC50 = 370 pM) while minimally engaging 5-HT2A (G q EC50 ≈ 300 nM). Behaviourally, 4-F,5-MeO-PyrT produced dose-dependent locomotor suppression (consistent with 5-HT1A activation) but did not elicit the head-twitch response up to 3 mg kg−1, indicating a lack of apparent 5-HT2A-mediated psychedelic-like activity. In contrast, 5-MeO-MET displayed HTR when 5-HT1A was antagonised, consistent with 5-HT1A-mediated suppression of HTR. In the chronic social defeat model, a single s.c. dose of 4-F,5-MeO-PyrT (1 mg kg−1) given 1 h after the final defeat reversed social interaction deficits measured 24 h later; co-administration of the selective 5-HT1A antagonist WAY-100635 blocked this rescue. Sucrose-preference testing showed that 4-F,5-MeO-PyrT reversed defeat-induced anhedonia, an effect also antagonised by WAY-100635. The compound increased the proportion of stress-resilient mice relative to vehicle-treated stressed animals. The investigators report appropriate control measures for locomotion and corner-time behaviour to exclude confounds from sedation.

Warren and colleagues interpret their findings as evidence that 5-HT1A engagement by 5-methoxytryptamines can contribute meaningfully to antidepressant- and anxiolytic-like effects while potentially dissociating therapeutic actions from classical psychedelic-like behaviours. They emphasise that 5-MeO-DMT and related tryptamines show comparable in vitro potency at 5-HT1A and 5-HT2A and that both receptors contribute to in vivo pharmacology, but their data with a highly 5-HT1A-selective analogue (4-F,5-MeO-PyrT) demonstrate that selective 5-HT1A agonism is sufficient to rescue social avoidance and anhedonia in a chronic social defeat model without producing the head-twitch response. Structurally, the authors highlight conserved anchoring to D116 and a T121 hydrogen-bond interaction for tryptamines bound to 5-HT1A, and they show how subpockets and specific residues (for example F3616.51, N386, A365) determine amine and indole substitution preferences that modulate potency and receptor selectivity. They also note that clinically used 5-HT1A ligands occupy extended binding pockets and may stabilise distinct receptor conformations, which could explain differences in signalling outputs relative to tryptamine ligands. The investigators acknowledge that prior preclinical work on the receptor contributions to therapeutic effects of psychedelics has been mixed and that human subjective data correlating 5-HT1A versus 5-HT2A engagement with experiential reports are lacking. They therefore present their selective probe as a tool to dissect 5-HT1A-mediated aspects of tryptamine pharmacology in vivo and as a structural guide for developing tryptamine-derived therapeutics. Finally, the authors caution that further experiments are required to map precise correlates between structural binding modes, distinct signalling ensembles and clinical outcomes, and they frame their results as enabling the rational design of probes and potential therapeutics that vary in 5-HT1A/5-HT2A selectivity.