This mice study investigated the antidepressant and neuroplastic effects of psychedelics, specifically lysergic acid diethylamide (LSD) and psilocin. The trial centres around their high-affinity binding to TrkB, a BDNF receptor implicated in antidepressant mechanisms. These substances bind to TrkB with affinities 1,000-fold higher than other antidepressants (SSRIs), and their effects on neurotrophic signalling, plasticity, and antidepressant-like behaviour in mice depend on this TrkB binding and promotion of endogenous BDNF signalling. Interestingly, these effects are independent of serotonin (5-HT) 2A receptor activation, suggesting that TrkB is a primary target for antidepressants.
Psychedelics produce fast and persistent antidepressant effects and induce neuroplasticity resembling the effects of clinically approved antidepressants. We recently reported that pharmacologically diverse antidepressants, including fluoxetine and ketamine, act by binding to TrkB, the receptor for BDNF. Here we show that lysergic acid diethylamide (LSD) and psilocin directly bind to TrkB with affinities 1,000-fold higher than those for other antidepressants, and that psychedelics and antidepressants bind to distinct but partially overlapping sites within the transmembrane domain of TrkB dimers. The effects of psychedelics on neurotrophic signaling, plasticity and antidepressant-like behavior in mice depend on TrkB binding and promotion of endogenous BDNF signaling but are independent of serotonin 2A receptor (5-HT2A) activation, whereas LSD-induced head twitching is dependent on 5-HT2A and independent of TrkB binding. Our data confirm TrkB as a common primary target for antidepressants and suggest that high-affinity TrkB positive allosteric modulators lacking 5-HT2A activity may retain the antidepressant potential of psychedelics without hallucinogenic effects.
Depression incidence has risen substantially, and there is a pressing need for faster, more effective treatments. Preliminary clinical trials suggest that psychedelics such as lysergic acid diethylamide (LSD) and psilocybin (via its metabolite psilocin, PSI) can produce rapid and sustained antidepressant effects; however, their acute hallucinogenic effects mediated by serotonin 2A receptors (5-HT2A) limit clinical use. Earlier work showed that diverse antidepressants, including fluoxetine and ketamine, can bind to TrkB, the receptor for brain-derived neurotrophic factor (BDNF), and thereby potentiate neurotrophic signalling that underpins plasticity and therapeutic actions of antidepressants. Moliner and colleagues set out to determine whether psychedelics directly bind to TrkB and whether such binding mediates their plasticity-promoting and antidepressant-like effects. The study aims to localise the binding site, characterise binding affinities and structural consequences for TrkB, and test whether psychedelic-induced neurotrophic signalling, synaptic plasticity and behavioural effects depend on TrkB binding and BDNF signalling or on 5-HT2A activation. This work combines biochemical binding assays, structural NMR and atomistic molecular dynamics (MD) simulations, cellular imaging and electrophysiological assays, and in vivo behavioural and plasticity paradigms using wild-type and TrkB TMD mutant mice.
The investigators used an array of in vitro, in silico and in vivo approaches to interrogate psychedelic–TrkB interactions and downstream consequences. Binding of radiolabelled LSD (3H-LSD) to TrkB was measured in lysates of HEK293T cells transiently expressing TrkB and in embryonic brain lysates, with saturation, competition and dissociation radioligand assays performed to obtain Kd, Bmax, Ki and koff values. Site-specificity was tested using TrkB mutants (Y433F, V437A, S440A) and a chimera replacing the TrkB transmembrane domain (TMD) with that of TrkA (TrkA.TM). Complementary microscale thermophoresis (MST) assays used GFP-tagged TrkB in crude lysates to detect binding in near-native conditions, and nuclear magnetic resonance (NMR) spectroscopy of isotope-labelled TrkB TMD in lipid bicelles was used to map chemical shift perturbations (CSPs) upon LSD titration. Atomistic MD simulations and free energy calculations were carried out on TrkB TMD dimers embedded in POPC:cholesterol membranes to identify spontaneous binding sites and to estimate binding free energies for LSD, PSI and related ergoline compounds; the total MD sampling reported was approximately 730 µs. Protein–protein interaction assays in live cells used a split luciferase complementation (PCA) system to monitor TrkB dimerization and interactions with a raft reporter (Fyn) in N2a cells. Biochemical readouts included ELISAs and co-immunoprecipitation to quantify TrkB phosphorylation (pY816), PLCγ1 interaction, ERK and mTOR phosphorylation and cell-surface retention of TrkB. Single-molecule localisation microscopy (dSTORM/TIRF) assessed TrkB:PLCγ1 colocalisation in dendritic spines. Primary cortical and hippocampal neuronal cultures (DIV21) were used for FRAP experiments to track TrkB trafficking, for spinogenesis and dendritogenesis assays, and for BDNF quantification by ELISA and RT‑qPCR. In vivo work employed heterozygous Y433F TrkB knock-in mice (Y433F +/-) and wild-type littermates for behavioural and plasticity assays: head-twitch response, repeated forced swim test (rFST) as an antidepressant-like readout, contextual fear conditioning and extinction, BrdU labelling for long-term dentate gyrus neuron survival, and ocular dominance plasticity (monocular deprivation with intrinsic signal optical imaging). Pharmacological manipulations included 5-HT2A antagonists (ketanserin, M100907) and TrkB.FC to sequester extracellular BDNF. Statistical analyses used t-tests or one-/two-way ANOVA with Bonferroni post hoc tests; detailed statistics are reported in the Supplementary Table (not provided in the extracted text).
Psychedelics bind TrkB with high affinity and specificity. Radioligand saturation binding showed 3H-LSD binds human, rat and mouse TrkB with Kd values around 0.93 nM, 0.66 nM and 0.43 nM respectively and measurable Bmax values; these affinities are comparable to LSD's canonical affinity for 5-HT2A and are up to 1,000-fold higher than affinities reported for conventional antidepressants such as fluoxetine and ketamine to TrkB. Competition experiments found PSI displaces 3H-LSD with a Ki of 6.73 ± 3.05 nM, while fluoxetine and ketamine display much lower TrkB affinities (Ki = 4.77 µM and 20.82 µM respectively). R,R-HNK displaced LSD at higher nanomolar concentrations (Ki = 166.00 ± 21.15 nM). Residence-time measurements indicated a slow LSD koff (0.0085 ± 0.0011 min−1) that increased with the Y433F mutation (koff = 0.0240 ± 0.0037 min−1). Lisuride bound TrkB with reduced affinity relative to LSD, whereas cabergoline and dihydroergotamine showed poor binding. Binding localises to the TrkB transmembrane domain and involves key residues. Replacement of the TrkB TMD with TrkA TMD abolished LSD binding, and point mutations Y433F and V437A markedly reduced binding, whereas S440A (a fluoxetine-interfering mutation) did not affect LSD binding. MST assays confirmed direct binding of LSD, PSI and lisuride to native TrkB but not to Y433F or TrkA.TM mutants, though apparent potencies were lower in MST owing to lysate effects. NMR spectroscopy of isotope-labelled TrkB TMD in bicelles revealed CSPs in residues including Y433 and V437 upon LSD addition, changes that were specific to the dimeric state. Atomistic MD simulations identified an extracellular-facing crevice formed by crisscrossed TrkB TMD dimers as the binding pocket for LSD and PSI, requiring both TMDs for stable association. Free energy estimates indicated higher affinity of LSD and PSI to TrkB dimers than fluoxetine. Structurally, LSD and PSI stabilise a tighter cross-shaped TMD dimer conformation with Cα distances around 17 Å (favoured for BDNF activation), whereas fluoxetine binds deeper and stabilises a more open conformation (Cα ~20 Å). LSD forms specific interactions including a hydrogen bond between its diethylamide oxygen and Y433 on one monomer and aromatic pi-stacking with Y433 on the other monomer. These distinct but partially overlapping pockets explain differential effects of TrkB mutations on drug binding. Psychedelics potentiate BDNF–TrkB signalling and synaptic localisation. Using a split-luciferase PCA, LSD and PSI rapidly and persistently increased TrkB dimerization at nanomolar concentrations; this effect was abolished in heterodimers containing Y433F. LSD increased TrkB phosphorylation at Y816 in cortical and hippocampal cultures and in mouse brain after a single administration. Psychedelics increased TrkB interaction with PLCγ1, promoted ERK and mTOR phosphorylation, increased Bdnf mRNA at 1 h and BDNF protein at 24 h, and enhanced neuronal surface retention of TrkB and its colocalisation with raft-associated Fyn—effects lost in Y433F +/- samples. TrkB.FC, which sequesters extracellular BDNF, blocked psychedelic-induced TrkB dimerization, indicating that psychedelics act allosterically to facilitate endogenous BDNF signalling rather than directly agonising TrkB. Pretreatment with 5-HT2A antagonists ketanserin or M100907 did not prevent psychedelic-induced TrkB dimerization or downstream signalling in systems lacking 5-HT2A expression. Psychedelics promote structural plasticity in a TrkB- and BDNF-dependent manner. FRAP experiments showed that LSD and PSI rapidly increased recovery of GFP‑TrkB fluorescence in dendritic spines (+42.4% for LSD, +22.7% for PSI relative to vehicle), indicating increased trafficking into spines; these effects were absent for Y433F TrkB. In mature neuronal cultures, LSD and PSI increased spine density and dendritic arbor complexity in wild-type but not in Y433F +/- cultures. TrkB.FC prevented LSD-induced spinogenesis and dendritogenesis, whereas the 5-HT2A antagonist M100907 did not, demonstrating dependence on extracellular BDNF/TrkB but not on 5-HT2A activation. In vivo plasticity and behavioural effects require TrkB binding but hallucinogenic-like responses do not. A single LSD administration doubled the number of surviving dentate granule cells at 4 weeks in wild-type but not in Y433F +/- mice. LSD promoted ocular dominance plasticity after monocular deprivation. Behaviourally, LSD produced sustained antidepressant-like effects in a repeated forced swim paradigm and facilitated contextual fear extinction when paired with extinction training; both effects were present in wild-type but absent in Y433F +/- mice. The head-twitch response—used as a rodent proxy of psychedelic hallucinogenic action—remained intact in Y433F +/- mice but was blocked by the 5-HT2A antagonist M100907, indicating that hallucinogenic-like actions depend on 5-HT2A whereas plasticity and antidepressant-like effects depend on TrkB binding. Across multiple assays, 5-HT2A antagonism failed to block the neurotrophic, structural or antidepressant-like effects of LSD, supporting a dissociation between 5-HT2A-mediated hallucinogenic actions and TrkB-mediated plasticity.
Moliner and colleagues interpret their results as evidence that psychedelics are high-affinity positive allosteric modulators of TrkB that facilitate the actions of endogenously released BDNF at active synapses. They emphasise that psychedelics are not direct TrkB agonists—TrkB dimerization and downstream signalling elicited by LSD and PSI required extracellular BDNF and were blocked by TrkB.FC—but by binding to a dimeric TMD pocket they stabilise conformations that favour BDNF-mediated activation, thereby promoting activity-dependent Hebbian plasticity. This mechanism helps explain why psychedelics can induce rapid and persistent plasticity and behavioural effects after single doses, given their high brain penetration and much higher TrkB affinity than conventional antidepressants. The authors position these findings within ongoing debates about the role of 5-HT2A receptors in psychedelic therapeutics, arguing that the TrkB-dependent plasticity and antidepressant-like effects can be detached from 5-HT2A-mediated hallucinogenic responses. They note that lisuride, a nonpsychedelic 5-HT2A agonist, also binds TrkB but with lower affinity, consistent with prior data showing it can promote dendritogenesis in vitro. At the structural level, MD and NMR data indicate that psychedelics and conventional antidepressants engage distinct but partially overlapping pockets in the TrkB TMD and stabilise different dimer conformations, which may underlie differences in pharmacology and time course. Key limitations and uncertainties acknowledged by the authors include the polypharmacology of psychedelics—TrkB and 5-HT2A are not the only targets that contribute to overall drug effects—and the need for further research to fully reconcile discrepant findings (for example, previous reports that ketanserin blocks some LSD effects in vitro). They also stress that environmental input is required for many long-term behavioural outcomes: LSD facilitated but did not produce fear extinction in the absence of extinction training, mirroring the clinical importance of ‘‘set and setting.’' Finally, the authors propose translational implications: TrkB emerges as a common and central target for antidepressant-induced plasticity, and structure-based design of high-affinity TrkB-selective positive allosteric modulators that lack 5-HT2A activity could, in principle, retain plasticity-promoting and antidepressant efficacy while avoiding hallucinogenic effects. They call for further work to pursue such therapeutic avenues and to clarify remaining mechanistic questions.
We used atomistic molecular dynamics (MD) simulations of the TrkB TMD dimers embedded in cholesterol (CHOL)-enriched membranesto identify the binding site of psychedelics. LSD and PSI spontaneously associate with TrkB TMD dimers. Consistent with our NMR structural data and mutagenesis binding studies, simulations revealed binding sites for LSD and PSI located in the extracellular-facing crevice of the crisscrossed TMD dimers such that both TMDs of a dimer are required for binding (Fig.and Extended Data Fig.). Subsequent binding free energy estimations showed that LSD and PSI bind to the TMD of TrkB dimers with higher affinity than fluoxetine (Fig.and Extended Data Fig.). In agreement with our experimental data, Y433, A434 and V437 stabilize the binding of LSD and PSI (Fig.-h and Extended Data Figs.and). Y433F and V437A mutations markedly decrease the binding affinity of psychedelics (Fig.and Extended Data Fig.). At the high CHOL concentration typical for synaptic membranes, LSD stabilizes the TrkB TMD dimer in a conformation where the distance between its Cα residues (L451-L453) is ~17 Å (Fig.and Extended Data Fig.). Thus, LSD stabilizes TrkB dimers in a conformation that is more favorable to activation by BDNF, akin to the mechanism previously identified for fluoxetine. Notably, we found that the binding pockets for fluoxetine and LSD are distinct, even though their binding involves shared residues. Fluoxetine binds deeper within the TMD and requires membrane lipid stabilization, locking the TMD dimer in a more open cross-shaped conformation (Cα distance ~20 Å). In contrast, LSD binds closer to the membrane surface and establishes more stable interactions with the dimer: a hydrogen bond between the oxygen of LSD's diethylamide group and the Y433 residue of one monomer, and pi-stacking of the ergoline aromatic backbone with the Y433 residue of the second monomer, locking TMD dimers in a tighter cross-shaped conformation (Cα distance ~17 Å) compared with fluoxetine (Fig.). This is consistent with the deeper-located mutation S440A interfering with fluoxetinebut not with LSD binding (Fig.). PSI displays a binding pocket, interacting residues and stable drug:TrkB complex conformation (Cα distance ~17 Å) similar to that of LSD (Fig.and Extended Data Fig.). Atomistic simulations and free energy calculations show that lisuride binds to TrkB with reduced affinity when compared with LSD (Extended Data Fig.), but interacting residues and TMD dimer conformation (Cα distance ~17 Å) of lisuride are similar to those of LSD and PSI (Extended Data Fig.). In contrast, control compounds cabergoline and dihydroergotamine display very poor TrkB binding (Extended Data Fig.). Thus, psychedelics and antidepressants have distinct binding pockets and stabilize different conformations of the TMD dimer, which may underlie the disparities observed among TrkB ligand affinities. retain the antidepressant effects of psychedelics, but are devoid of the hallucinogenic effects. It is, however, unclear how these separate effects are mediated. Essentially all antidepressant drugs, including psychedelics, promote neuroplasticity, which is considered a critical component of their therapeutic effect. Brain-derived neurotrophic factor (BDNF) and its receptor TrkB (neurotrophic receptor tyrosine kinase, Ntrk2) are central mediators of plasticity and the therapeutic action of antidepressants. Recent findings show that antidepressants, including conventional antidepressants such as fluoxetine and imipramine, as well as the rapid-acting ketamine, directly bind to TrkB and allosterically potentiate BDNF signaling. BDNF and TrkB have also been implicated in the action of psychedelics as downstream effectors of 5-HT 2A activation. In this Article, because psychedelics produce robust spinogenesis and dendritogenesisand these effects are known to require intact TrkB signaling, we set out to explore whether direct binding to TrkB might mediate the neuroplastic effects behind their therapeutic potential.
Create a free account to open full-text PDFs.
Papers cited by this study that are also in Blossom
Vollenweider, F. X., Preller, K. H. · Nature Reviews Neuroscience (2020)
Carhart-Harris, R. L., Giribaldi, B., Watts, R. et al. · New England Journal of Medicine (2021)
Johnson, M. W., Richards, W. A., Griffiths, R. R. · Journal of Psychopharmacology (2008)
Van Der Walt, A., Ford, H., Fraser, C. L. et al. · Frontiers in Neurology (2022)
Cameron, L. P., Tombari, R. J., Lu, J. et al. · Nature (2020)
Hesselgrave, N., Troppoli, T. A., Wulff, A. B. et al. · PNAS (2021)
Shao, L-X,, Liao, C., Gregg, I. et al. · Neuron (2021)
Ly, C., Greb, A. C., Cameron, L. P. et al. · Cell Reports (2018)
Greb, A. C., Vargas, M. V., Duim, W. C. et al. · ACS Pharmacology and Translational Science (2020)
Zanos, P., Moaddel, P. J., Morris, P. J. et al. · Nature (2016)
Halberstadt, A. L., Chatha, M., Klein, A. K. et al. · Neuropharmacology (2020)
Yaden, D. B., Earp, D., Graziosi, M. et al. · Frontiers in Psychology (2022)
Yaden, D. B., Griffiths, R. R. · ACS Pharmacology and Translational Science (2020)
Andersen, K. A. A., Carhart-Harris, R. L., Nutt, D. J. et al. · Acta Psychiatrica Scandinavica (2020)
Papers in Blossom that reference this study
Urban, M. M., Zillich, L., Rieser, N. M. et al. · Translational Psychiatry (2026)
Avram, M., Menegaux, A., Müller, F. et al. · Cell Reports (2026)
Lyons, T., Spriggs, M. J., Kerkelä, L. et al. · Nature Communications (2026)
Baker, J. J., Kogan, E., Lu, J. et al. · Biorxiv (2026)
Mäki-Marttunen, V. · Communications Biology (2026)
We first tested whether psychedelics bind to TrkB in lysates of HEK293T cells transiently expressing TrkB and immunoprecipitated with a TrkB antibody. We observed that radiolabeled LSD ( 3 H-LSD) binds directly to human (K d = 0.930 ± 0.414 nM, B max = 4.028 ± 0.326 pmol mg -1 protein), rat (K d = 0.656 ± 0.093 nM, B max = 3.053 ± 0.099 pmol mg -1 protein) and mouse TrkB (K d = 0.425 ± 0.296 nM, B max = 7.388 ± 0.771 pmol mg -1 protein) (Fig.and Extended Data Fig.) with high affinity, which is similar to that for its canonical target 5-HT 2A (K i = 3.5 nM) (ref. 14). Remarkably, the affinity of LSD to TrkB was up to 1,000-fold higher than that of other antidepressants such as fluoxetine and ketamine to TrkB. To verify the binding and to localize it, we tested the binding ofH-LSD to a set of TrkB mutants (Fig.). Replacement of the transmembrane domain (TMD) of TrkB with the corresponding sequence of TrkA (TrkA.TM) completely abolishes LSD binding, which restricts the LSD binding site to the TMD of TrkB. A tyrosine-to-phenylalanine point mutation (Y433F) in the TMD of TrkB that disrupts binding and plasticity-related effects of antidepressants also impairs LSD binding. V437A has similar effects, but S440A, a mutation that interferes with fluoxetine binding to TrkB, did not influence LSD binding (Fig.). These findings indicate that the LSD binding site in the TrkB TMD partially overlaps with that of other antidepressants. LSD displays a high residence time at TrkB (k off = 0.0085 ± 0.0011 min -1 ), consistent with high affinity, but residence time is markedly reduced by Y433F (k off = 0.0240 ± 0.0037 min -1 ) (Extended Data Fig.). We also found that PSI displaces 3 H-LSD bound to TrkB with high affinity (K i = 6.73 ± 3.05 nM, Fig.), while fluoxetine (K i = 4.77 ± 1.50 µM, Fig.) and ketamine (K i = 20.82 ± 8.86 µM, Extended Data Fig.) display substantially lower TrkB affinities. 2R,6R-hydroxynorketamine (R,R-HNK), an active metabolite of ketamine that binds to TrkB but shows low affinity to NMDA receptors, displaces LSD with high nanomolar concentrations (K i = 166.00 ± 21.15 nM, Extended Data Fig.). In turn, biotinylated R,R-HNK is displaced from TrkB by psychedelics at low nanomolar concentrations (Extended Data Fig.). To further define the TrkB binding specificity, we tested displacement of 3 H-LSD by other ergoline compounds. We found that lisuride, a nonpsychedelic ergoline 5-HT 2A agonist, binds to TrkB at higher nanomolar concentrations in a linear manner, but other LSD-related compounds cabergoline and dihydroergotamine fail to displace 3 H-LSD, as do the negative controls chlorpromazine and diazepam (Extended Data Fig.). 5-HT 2A antagonists ketanserin and M100907 also fail to displace LSD binding to TrkB (Extended Data Fig.). Importantly, no other known targets for psychedelics or antidepressants were identified by mass spectrometry Article
We used a split-luciferase complementation assay to investigate TrkB dimerization and interactions in response to psychedelics. LSD and PSI induce a fast and long-lasting increase in dimerization of wild-type (WT) TrkB at nanomolar concentrations, but this effect is abolished in WT:Y433F heterodimers (Y433F +/-) (Fig.). Consistently, LSD increases phosphorylation of TrkB tyrosine 816 (pY816) residues in cortical and hippocampal cultures and in mouse brain samples after a single administration (Extended Data Fig.,j,l,q), indicating facilitated dimerization. Lisuride similarly increases TrkB dimerization at higher concentrations (Extended Data Fig.). Pretreatment with ketanserin or M100907 fails to prevent the effects of LSD and BDNF on TrkB dimerization in N2a cells that do not express 5-HT 2A messenger RNA (Fig., Extended Data Fig.and Extended Data Table), indicating that psychedelics induce TrkB dimerization in the absence of 5-HT 2A and ketanserin and M100907 do not interfere with this effect. TrkB.FC, which sequesters extracellular BDNF, prevents the effects of LSD and BDNF on TrkB dimerization (Fig.). Furthermore, LSD potentiates the effects of very low concentrations of BDNF (0.1 ng ml -1 ) on TrkB dimerization (Fig.), presumably by stabilizing TrkB dimerization. No differences in total TrkB expression were observed between WT and Y433F +/-or following treatment with LSD or BDNF (Extended Data Fig.). These data demonstrate that psychedelics do not directly activate TrkB, but their effects on TrkB dimerization are dependent on release of endogenous BDNF, consistent with an allosteric effect. TrkB is mainly localized in intracellular vesicles and only transiently translocated to the cell surface where it is exposed to BDNF. Psychedelics rapidly increase the neuronal surface retention of TrkB, independently from 5-HT 2A activation, and promote TrkB interaction with raft-restricted Src family kinase Fyn, indicating increased localization to raft-like synaptic membranes. This effect is lost in Y433F +/- heterodimers (Extended Data Fig.). Psychedelics also promote BDNF downstream signaling. In neuronal cultures, LSD increases TrkB interaction with phospholipase C gamma 1 (PLCγ1) (Extended Data Fig.), which docks on pY816 and regulates intracellular Ca 2+ signaling and antidepressant action. TrkB interaction with PLCγ1 is also increased in the prefrontal cortex (PFC) and hippocampus of WT but not Y433F +/-mice, without affecting total TrkB expression (Extended Data Fig.). Tracking single-molecule localizations of TrkB and PLCγ1 in neurons by direct stochastic optical reconstruction microscopy coupled with total internal reflection fluorescence (dSTORM/TIRF) confirmed that LSD and PSI increase TrkB:PLCγ1 colocalization in dendritic spines (Fig.and Extended Data Fig.). In addition, LSD rapidly increases phosphorylation of ERK (pERK), independently of 5-HT 2A activation (Fig.), and promotes phosphorylation of mTOR for at least 1 h (Extended Data Fig.). Bdnf mRNA is upregulated at 1 h after LSD treatment and BDNF protein levels are elevated at 24 h (Fig.) without any changes in Ntrk2 mRNA levels (Extended Data Fig.). In summary, the effects of psychedelics on neurotrophic signaling are dependent on BDNF and TrkB, and do not require 5-HT 2A activation.
We used fluorescence recovery after photobleaching (FRAP) in neuronal cultures expressing GFP-tagged TrkB to investigate the effects of psychedelics on TrkB trafficking. FRAP revealed that psychedelics induce rapid TrkB trafficking into dendritic spines. LSD robustly increases fluorescence recovery (+42.42% increment, t 1/2 = 13.53 s), followed closely by PSI (+22.72% increment, t 1/2 = 13.99 s), when compared with vehicle treatment (E max = 62.88%, t 1/2 = 16.93 s). However, LSD (+4.80% increment, t 1/2 = 16.00 s) and PSI (+8.63% increment, t 1/2 = 15.18 s) failed to increase fluorescence recovery of Y433F compared with vehicle treatment (E max = 70.82%, t 1/2 = 11.08 s) (Fig.,b and Extended Data Fig.). Previous findings have reported that psychedelics increase neurite outgrowth and spine formation in cultured neurons. Consistent with these findings, LSD and PSI increase spine density of mature neuronal cultures from WT mice at 24 h after treatment, but not of neuronal cultures from Y433F +/-mice. Notably, TrkB.FC but not M100907 prevented the spinogenic effect of LSD (Fig.), indicating that this effect is dependent on BDNF release but not on 5-HT 2A activation. Psychedelics also increase dendritic arbor complexity in neuronal cultures derived from WT but not Y433F +/-mice. Again, TrkB.FC but not M100907 prevented LSD effects on dendritogenesis (Fig.). Ketanserin has previously been reported to prevent the effects of LSD on dendritic arbor complexity in cultured neurons; more research is needed to resolve this apparent discrepancy with present results, but the various off-target effects of ketanserin might at least partially explain this contradiction. Importantly, Y433F +/-neuronal cultures treated with BDNF showed increased dendritic arbor complexity and were indistinguishable from WT cultures (Extended Data Fig.), indicating that the heterozygous Y433F +/-mutation does not interfere with the TrkB response to BDNF. Furthermore, Y433F +/-mice do not show any baseline differences from WT TrkB mice, and Y433F +/-neuronal cultures and mice respond normally to BDNF in plasticity-related and behavioral assays. These results indicate that Y433F +/-does not compromise TrkB function or BDNF responses; it only abrogates the binding of psychedelics and antidepressants to TrkB and their effects on plasticity. Together these findings are consistent with previous reports of TrkB involvement in psychedelic drug-induced plasticity, and support the notion that psychedelics are not direct TrkB agonists, but they facilitate the effects of endogenous BDNF, independently of 5-HT 2A activation.
Chronic treatment with conventional antidepressants or a single administration of ketamine enhance neurogenesis and long-term neuronal survival of dentate granule cells (DGCs) of the hippocampus, and these effects are dependent on BDNF signaling through TrkB. We found that LSD doubles the number of surviving DGCs of WT mice, but not of Y433F +/-mice, at 4 weeks after a single administration (Fig.and Extended Data Fig.), indicating that the effects of LSD on DGC survival are mediated by their binding to TrkB. We and others have previously shown that antidepressantsand ketaminereactivate critical period-like plasticity in the adult visual cortex. We found that LSD also promotes visual plasticity and facilitates a shift in ocular dominance (OD) in favor of the open eye in the primary visual cortex of mice following a week of monocular deprivation (Extended Data Fig.)..
The head-twitch response has been used in rodents as a reporter of the hallucinogenic effects of psychedelics in humans and this response is known to be mediated by 5-HT 2A receptors. We found that the head-twitch response occurs normally in Y433F +/-mice treated with LSD, but is blocked by M100907 (Fig.and Extended Data Fig.). These results suggest that, as far as the head-twitch response reliably reports hallucinogenic actions of LSD in humans, these effects are dependent on 5-HT 2A activation but detached from TrkB activity, which indicates a dissociation between the hallucinogenic and plasticity-promoting effects of psychedelics. In a rodent model of chronic stress consisting of repeated sessions of the forced swimming test (rFST), LSD produced a sustained antidepressant-like effect 7 days after a single administration in WT but not Y433F +/-mice. Pretreatment with M100907 did not prevent the antidepressant-like effect of LSD (Fig.). LSD also facilitates contextual extinction of conditioned fear response 72 h after a single administration, which persists for at least 4 weeks. However, these short-and long-term effects are lost in Y433F +/-mice. Importantly, LSD alone does not bring about fear extinction, as extinction training is required to produce a sustained decrease in behavioral freezing after LSD treatment (Fig.-k and Extended Data Fig.). This is reminiscent of the effects of other antidepressants that depend on environmental input to exert their therapeutic-like effects, and are consistent with the importance of 'set and setting' in modulating psychedelic effects in humans. Taken together, these results strongly suggest that TrkB mediates the plasticity-related and antidepressant-like effects of LSD at the network and behavioral levels but is not involved in its hallucinogenic-like action.
Recent clinical studies suggesting that psychedelics produce rapid and long-lasting antidepressant effects 2 have received tremendous attention. Mystical experiences induced by psychedelics have been associated with their clinical efficacy. However, the hallucinogenic effects of psychedelics limit their widespread clinical application, as their administration is restricted to clinical settings that often require intensive monitoring. A number of recent observations suggest that the antidepressant and plasticity-promoting effects of psychedelics may be dissociable from their hallucinogenic effects. First, plasticity-like effects of psilocybin are not blocked by the 5-HT 2A antagonist ketanserin in mice, while it blocks the head-twitch response, a 5-HT 2A -mediated behavior in rodents used as a proxy for hallucinogenic action of psychedelics in humans. Second, derivatives of psychedelic compounds have recently been introduced that promote plasticity and antidepressant-like behavior but do not seem to produce head-twitch responses. However, detailed information concerning the sites that mediate these differential actions have remained unclear. Here we show that psychedelics promote neuroplasticity and plasticity-related behavioral effects through their high-affinity binding to TrkB (Extended Data Fig.). Psychedelics are not direct TrkB agonists, as extracellular BDNF is necessary for their effects on TrkB dimerization and plasticity, but they act allosterically by facilitating the effects of endogenous BDNF released in active synapses, similar to what we recently found for other antidepressants. Activity-dependent release of BDNF in stimulated synapses helps to selectively stabilize active synapses at the expense of inactive ones, which is critical for Hebbian-type plasticity. Direct TrkB agonists are expected to indiscriminately activate TrkB in active and inactive synapses, which leads to gradual reduction of signal-to-noise ratio within neuronal networks. Therefore, through a positive allosteric modulation of BDNF signaling, psychedelics selectively promote, maintain and strengthen activity-dependent plasticity in active synapses. A point mutation in the critical Y433 residue that impairs psychedelics binding to TrkB also abolishes induction of neuroplasticity and long-term plastic responses, but does not affect the head-twitch response. Moreover, 5-HT 2A antagonists fail to prevent psychedelic-induced TrkB dimerization and neurotrophic signaling, spinogenesis, dendritogenesis and antidepressant-like behavioral effects. These findings are perhaps not completely surprising, as there is still some debate in the field about whether 5-HT 2A receptors mediate Detailed statistics reported in Supplementary Table.
the therapeutic effects of psychedelics. Overall, our results suggest that the TrkB-dependent effects of psychedelics on plasticity can be detached from their hallucinogenic-like effects mediated by 5-HT 2A . We further show that lisuride also binds to TrkB, albeit with a lower affinity than LSD and PSI, which is consistent with recent evidence suggesting that lisuride increases dendritogenesis and spinogenesis in culture neurons. We have recently found that a dimer of TrkB TMDs contains a binding site for antidepressants, and that their plasticity-promoting and antidepressant-like effects are dependent on this interaction. However, the affinity of conventional antidepressants to TrkB, albeit pharmacologically relevant, is low. Brain concentrations required for selective serotonin reuptake inhibitors binding to TrkB are achieved only after long-term treatment, which may partially explain the delay in the onset of their antidepressant action. Psychedelics readily penetrate into the brain after a single dose and, as shown here, bind to TrkB with much higher affinities than antidepressants in current clinical use. This may contribute to the fast and potent induction of neuroplasticity and more persistent behavioral effects produced by psychedelics when compared with other antidepressants. We have here characterized at an atomistic level of detail the binding sites of LSD, PSI and lisuride at TrkB TMD dimers, and used several control compounds to define binding specificity. We have also described the TrkB TMD conformational changes induced by psychedelics. Notably, the TMD binding pockets and TrkB conformational changes induced by antidepressants and psychedelics are distinct. However, it is clear that these compounds have different polypharmacological profiles beyond TrkB and 5-HT 2A that probably contribute to the overall effects and timeframes of each particular drug. In conclusion, our findings support TrkB as the key target for psychedelic drug-induced plasticity. These data confirm TrkB as a common binding target for antidepressants and open an avenue for structure-based design of high-affinity TrkB-selective ligands with fast and long-lasting antidepressant action, but potentially devoid of hallucinogenic-like activity.
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at.
Female and male C57BL/6NTac-NTrk2em6006 (Y433F +/-) mice were generated and genotyped as previously described, and their WT littermates were used as controls. Female and male C57BL/6JRccHsd WT mice (Envigo) were also used where indicated.
HEK293T cells transiently expressing GFP-tagged TrkB (WT, Y433F, V437A and TrkA.TM) were used for binding assays and MST, and N2a cells transiently expressing phGLuc-TrkB (WT and Y433F) and phGLuc-LR/Fyn were used for protein-fragment complementation assay (PCA). Cell lines were cultured and maintained as previously described. Cortical and hippocampal primary neuronal cultures derived from rat or mouse embryos (embryonic day 18, E18) were dissected, plated and maintained as previously described. Experiments were performed with mature neuronal cultures (21 days in vitro, DIV21), unless otherwise specified. Lipofectamine 2000 (Invitrogen) was used for transfection of cell lines and primary neuronal cultures.
Ketamine hydrochloride (ketamine, Tocris, 3131), R,R-HNK (Tocris, 6094), (+)-LSD (Sigma-Aldrich, L7007), PSI (Biosynth Carbosynth, P-7800), lisuride maleate (Tocris, 4052), cabergoline (Sigma-Aldrich, C0246), ketanserin (+)-tartrate (ketanserin, Sigma-Aldrich, S006) and M100907 (Sigma-Aldrich, M3324) were solubilized in dimethyl sulfoxide (DMSO) and diluted 1:1,000 (final concentration 0.1% DMSO) for in vitro experiments. Drugs were diluted in sterile 0.9% saline and injected intraperitoneally (10 ml kg -1 ) for in vivo experiments. All controlled psychotropic substances were obtained, stored, handled and disposed of as regulated under the permit issued by the Finnish Medicines Agency to the Faculty of Medicine, Medicum (FIMEA/2022/000945).
Binding assays were performed as previously described with minor modificationsThe use and validation of the anti-TrkB antibody in these binding assays has been previously described in the literature. Samples from mouse embryo (E18) brains were also lysed, vortexed, sonicated and centrifuged for binding assays. Plates were then washed with PBS. For radioligand saturation binding assays, we incubated 3 H-LSD (PerkinElmer, NET638250UC) in PBS at 1 pM to 100 nM final concentrations for 1 h at room temperature to determine the K d and B max . Cold LSD at 10 µM final concentration was used to determine nonspecific binding in wells with matching concentrations of the radioligand. For radioligand competition binding assays, 10 nM final concentration of 3 H-LSD was incubated with 1 nM to 10 µM of drugs in PBS for 2 h at room temperature to calculate the K i values. Nonspecific binding was determined with wells containing the same concentration of radioligand with 10 µM of cold LSD. For radioligand dissociation binding assays, 10 nM final concentration of 3 H-LSD in PBS (2% BSA) was pre-incubated for at least 2 h at room temperature before the addition of excess cold LSD (500 nM final concentration) at the designated timepoints (2 min, 15 min, 30 min, 1 h, 2 h, 4 h and 8 h) and finally washed with PBS. Radioactive emission of bound 3 H-LSD for saturation, competition and dissociation binding assays was determined by scintillation (PerkinElmer, OptiPhase SuperMix, 1200-437). For the competition binding assays with biotinylated R,R-HNK, the amino-biotinylation of R,R-HNK was performed using a commercial kit (Thermo Scientific, EZ-Link NHS-PEG4 Biotinylation Kit, 21455) and the reaction monitored by MS. A mixture of biotinylated R,R-HNK (10 µM) and unlabeled compounds (1 pM to 10 µM) in PBS was added to the 96-well plates containing the immunoprecipitated TrkB and incubated for 1 h at room temperature. For biotin-labeled BDNF (bBDNF, Alomone Labs, #B-250-B) assays, the immunoprecipitated TrkB was incubated with bBDNF (0.01-10 ng ml -1 ) for 1 h. Luminescence was determined via streptavidin conjugated with horseradish peroxidase (HRP; Thermo Fisher Scientific, 21126; 1:5,000, 1 h at room temperature) in reaction with the Pierce ECL substrate (Thermo Fisher Scientific, 32106). Luminescence from blank wells that contained all reagents but the sample lysates was used to determine nonspecific binding. To obtain the binding curves, results from binding assays were fitted to the equation of specific binding with Hill slope from GraphPad Prism 9.2. To calculate k off values, results were fitted to the equation of dissociation of one phase exponential decay from GraphPad Prism 9.2.
MST experiments were performed using the Monolith NT.115 (Blue/ Red) instrument (NanoTemper Technologies). Lysates of HEK293T cells expressing GFP-tagged TrkB (WT, Y433F, TrkA.TM) were lysed, vortexed, sonicated and centrifuged 24 h after transient transfection and used as a source of fluorescently labeled TrkB, similarly to previously described protocols. The advantage of this approach is the possibility of studying the proteins in close-to-native conditions. However, the presence of other proteins, lipids, nucleic acids and other components of crude cell lysates can alter the diffusion of biomolecules and therefore binding affinities. High viscosity and background signal typically results in the affinity values being orders of magnitude lower than those of more traditional binding assays. To evaluate binding of LSD, PSI, lisuride and BDNF (Peprotech, 450-02) to TrkB, cell lysates were diluted 1:2 with MST buffer (10 mM Na-phosphate buffer, pH 7.4, 1 mM MgCl 2 , 3 mM KCl, 150 mM NaCl and 0.05% Tween-20) to provide optimal level of initial fluorescence. The lysates of nontransfected HEK293T cells were used to evaluate background fluorescence, which was undetectable in the MST setup used in the present study. Titration series of LSD, PSI, lisuride or BDNF were mixed 1:1 and incubated with diluted cell lysates. The measurements were performed in premium coated capillaries (NanoTemper Technologies, MO-K025) using a LED source with 470 nm and 50% infrared-laser power at 25 °C. Analysis was performed using MO Affinity Analysis 2.3. Results were fitted to specific binding with Hill slope equations from GraphPad Prism 9.2.
1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) and 1,2-dihexanoyl-sn-glycero-3-phosphocholine (DHPC) were purchased from Avanti Polar Lipids. Trifluoroethanol (TFE) was provided by Alfa Aesar. Deuterium oxide and deuterated TFE were products of the equipped with a cryogenically cooled 1 H,C,N triple-resonance probe head, using 450 µl of TrkB TMD in DMPC/DHPC bicelles sample. First, a set of reference spectra, including one-dimensional (1D) 1 H as well as 2D 1 H,N-TROSY-HSQC for theN TrkB TMD protein without the ligand were measured. Then the sample was taken from the NMR tube and mixed with the dry aliquot of LSD pre-equilibrated at room temperature for 30 min. The solution was gently mixed by pipetting and centrifuged at 17,000g for 10 s and placed back in the NMR tube. The sample preparation was the same for each LSD concentration point, and the same set of spectra was collected as described above. The increasing amount of LSD in the sample was additionally controlled by following a well-separated NH resonance from the fused indole ring at δ H 10.15 p.p.m. using 1D NMR. All the samples contained 0.25 mM of TSP-d4 as an internal chemical shift reference standard. The obtained data were processed using the CSP approach. Chemical shifts of the individual signals were tracked in 1 H,N-TROSY-HSQC spectra, and the perturbations of peak positions in both dimensions of NMR spectra were combined to obtain the generalized chemical shift change: where Δδ 2 H and Δδ 2 N are the chemical shift changes of protons and nitrogens, respectively, while 100 is introduced to take into account the difference in gyromagnetic ratios of these two nuclei. The resultant CSPs for all TrkB TMD residues were then processed using the previously reported approach. In brief, CSPs were approximated by a half-normal distribution, and residues with CSPs exceeding two standard deviations were considered as outliers (Extended Data Fig.). To take into account the nonspecific chemical shift changes, most likely occurring due to the drug-induced variation in the properties of the lipid environment, we used the CSP difference approach. Since LSD and citalopram provide different magnitude of CSPs, we took the titration points with the closest average value of CSPs and measured the distance between the peak positions in the NMR spectra of TrkB TMD samples with LSD and citalopram using equation (1), where N are the chemical shift changes of protons and nitrogens between these two samples (Extended Data Fig.). Such an approach allows the contrasting of chemical shift changes that are specific for LSD. As a result, we observed that substantial differences in the effects of the two drugs are seen only for S432-Y434 and V437-I439, and the CSPs of the rest of residues seem to be nonspecific due to alteration in lipid packing and/or effective lipid-to-detergent ratio caused by the penetration of the drug into bicelles.
Proteins associated with TrkB in HEK293T transfected cells were determined by MS following immunoprecipitation of TrkB. For immunoprecipitation, the raw lysate of HEK293T cells transiently expressing TrkB (WT, Y433F and V437A) were incubated overnight (4 °C, under light rotation) with anti-TrkB (R&D systems, AF1494) at 1 µg of antibody per 500 µg of total protein. Following incubation with 30 µl of Protein G-Sepharose (Life Technologies, 101242), the samples were centrifuged and the supernatant discarded. The precipitate was washed six times with Tris-buffered saline (TBS) and stored at -80 °C. Samples were then washed several times with a detergent-free buffer, and bound proteins were eluted with urea buffer (8 M urea in 50 mM ammonium bicarbonate). Proteins were reduced with TCEP (Sigma-Aldrich, C4706), alkylated with iodoacetamide (Sigma-Aldrich) and digested with Sequencing Grade Modified Trypsin (Promega, V5111), after which the samples were desalted using C18 microspin columns (Nest Group). Liquid chromatography coupled to tandem MS analysis was done with a Q Exactive ESI-quadrupole-orbitrap mass spectrometer coupled Cambridge Isotope Laboratories. SDS, Tris, tris(2-carboxyethyl)phosphine (TCEP) and ethylenediaminetetraacetic acid were purchased from Sigma-Aldrich. The gene encoding transmembrane residues of human TrkB (TrkB TMD, T423-G466, UniProt ID Q16620) was amplified by PCR based on the TrkB gene (Cloning Facility). The PCR product was treated with endonucleases NdeI and HindIII and was cloned into the pGEMEX1 vector.N-and 15 N-13 C-labeled TrkB TMD was produced using the cell-free continuous exchange expression system based on the S30 Escherichia coli extract as described earlierTrkB TMD was purified using the size-exclusion chromatography. Precipitate from 3 ml of cell-free reaction mixture was washed two times with aqueous buffer (20 mM Tris, pH 8.0, and 150 mM NaCl) and solubilized with 500 µl of buffer (20 mM Tris, pH 8.0, 150 mM NaCl, 1.5% lauryl sarcosine, 10 mM 2-mercaptoethanol and 1 mM ethylenediaminetetraacetic acid). After centrifugation (60 min at 25,000g at room temperature) the clarified protein solution was applied onto a 10/600 Tricorn column packed with Superdex 200 prep grade (GE Healthcare) and pre-equilibrated with SEC buffer (20 mM Tris, pH 8.0, 50 mM NaCl, 0.5% lauryl sarcosine and 5 mM 2-mercaptoethanol). Protein-containing fractions were combined and precipitated by a trichloroacetic acid/acetone procedureN-and 15 N-13 C-labeled TrkB TMD as a dry precipitate was dissolved in 450 µl of TFE/water 2:1 mixture. Then DMPC, 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate, TCEP and water were added up to TFE/water vol/vol ratio of 1:1 with subsequent freeze-drying and resuspension in 450 µl of an aqueous buffer (20 mM NaPi, 1 mM TCEP and 0.01% vol/vol NaN 3 , pH 6.0) forN-labeled samples and 330 µl for the 15 N-13 C-labeled one. Samples for the chemical shift assignment contained 1.1 mM 15 N-13 C-labeled TrkB TMD, 9.6 mM DMPC and 42.7 mM DHPC. The effective DMPC/DHPC ratio was equal to 0.3, taking into account the concentration of monomeric DHPC in solution, determined using our previously published approach based on the NMR diffusion measurement. Lipid-to-protein ratio was equal to 40. Samples for the LSD titration contained 0.2 mM 15 N-labeled TrkB TMD, 3.6 mM DMPC and 21.8 mM DHPC, resulting in the effective DMPC/ DHPC ratio of 0.3 as well. Lipid-to-protein ratio was equal to 80 to ensure the presence of both the dimeric and monomeric state of the protein in NMR spectra. NMR spectra for the chemical shift assignment of TrkB TMD were recorded at 40 °C using the 600 MHz Bruker Avance III NMR spectrometer equipped with a triple-resonance cryogenic probe. Chemical shift assignment of the TrkB TMD backbone resonances, highlighted in the two-dimensional (2D) 1 H,N-TROSY-HSQC spectrum (Extended Data Fig.), was accomplished with three-dimensional (3D) HNCO, HN(CO)CA, HNCA and 1 H,N-NOESY-HSQC spectra. Protein side chains were assigned using 3D 1 H,C-NOESY-HSQC and HCCH-TOCSY experiments. The obtained chemical shifts were processed using the TALOS-N software to determine the secondary structure of the peptide. Initial analysis of NMR spectra revealed that the protein adopts the alpha-helical conformation in the region spanning from H430 to H459 (Extended Data Fig.) and may reside in two distinct states, which is manifested by the presence of two sets of signals in HSQC NMR spectra. Populations of states were depending on the lipid-to-protein ratio (Extended Data Fig.), which corresponds to the monomer-dimer transitions, and the state, which is abundant at high excess of lipids, was assigned to the monomeric form of TrkB TMD. To analyze the interaction between the TrkB TMD and LSD, 0.2 mM samples ofN-labeled TrkB TMD in DMPC/DHPC bicelles were titrated with the drug as follows. Lyophilized protein was resuspended using 95% H 2 O:5% D 2 O. LSD was first solubilized in acetonitrile at a concentration of 1 mg ml-1. Then LSD solution was aliquoted to have in the final sample LSD:TrkB TMD ratios of 0.25, 0.5, 1, 2, 4, 8 and 17.6. The aliquots were vacuum-dried using Eppendorf Concentrator Plus in 1.5 ml tubes and stored in the dark at -20 °C until usage. The titration was carried out using 5 mm tubes in the 850 MHz Bruker Avance III HD spectrometer (CGenFF). The systems were solvated and neutralized with 150 mM KCl solution. The equilibration protocol with subsequent simulations in the canonical (NVT) and isothermal-isobaric (NPT) ensembles as suggested by CHARMM-GUI were performed before the production runs. All production runs were performed using GROMACS 2020.6 (ref. 63) at 310 K and 1 bar pressure with the same parameters given in previous studies. FEP/HREX simulationswere also performed as previously described for in silico free energy estimations. The binding free energies and their errors were estimated using the multistate Bennet acceptance ratio methodas implemented in gmx bar and discarding 4-20 ns from the beginning of each window. The distributions of the distance between the center of mass of L451-L453 Cα atoms at the C-terminus of the TM dimer shown in Fig.were computed as an average over the separate kernel density estimates performed using the last 200 ns and a sampling interval of 10 ns (200 samples) for each of ten independent simulations. The contact probability maps shown in Fig.,h were calculated using a cutoff of 5 Å between the heavy atoms from the fully coupled state of the FEP/ HREX simulations after discarding the first 20 ns. The total timescale of atomistic MD simulations was about 730 µs. Marvin 66 was used for drawing 2D chemical structures, LigPlot+ (ref. 67) for creating 2D diagrams of protein-ligand interaction, and VMDfor creating 3D representations. The putative structural model of full-length TrkB dimer (Fig.) was built on the basis of the structures of PDB IDs 1hcf (ref.
PCA was used as a protein-protein interaction assay in live N2a cells to evaluate TrkB (WT and Y433F) dimerization and interaction with the lipid raft reporter protein cleaved Fyn (ref. 22) in response to treatment with LSD, PSI or lisuride (1 nM to 1 µM). PCA was also used to assess the effects of pretreatment with ketanserin (100 nM), M100907 (100 nM) or TRKB.FC (200 ng ml -1 , R&D Systems, 688-TK-100) (ref. 74) on TrkB dimerization induced by LSD or BDNF (0.1 ng ml -1 or 10 ng ml -1 ). Ketanserin, M100907 and TRKB.FC were added to the media 10 min before treatment. The PCA method was performed as previously described. Split humanized Gaussia princeps luciferase (hGLuc) was cloned on the C-terminus of the reporter proteins TrkB (WT and Y433F) and lipid raft Fyn. If the reporter proteins that carry complementary halves of the luciferase interact, the enzyme can refold in its functional conformation and, upon addition of its substrate, produce luminescence proportionally to the amount of reporter proteins interacting at a given time. N2a cells were seeded in 96-well plates with opaque walls and transparent bottom (PerkinElmer, ViewPlate 96). Cells were transfected with Lipofectamine 2000 according to the manufacturer's instructions 24 h post-plating with the PCA pair plasmids: phGLuc(1C)-TrkB(WT) or the mutant phGLuc(1C)-TrkB(Y433F) and phGLuc(2 C)-TrkB(WT) for the dimerization experiments; phGLuc(1C)-lipid raft Fyn and phGLuc(2C)-TrkB(WT) or phGLuc(2C)-TrkB(Y433F) for the lipid raft localization experiments. Forty-eight hours post transfection, cells were treated (40 min, 37 °C) in phenol red-free Dulbecco's modified Eagle medium without serum. For the time course effect of LSD on TrkB dimerization, the medium was changed to phenol red-free Dulbecco's modified Eagle medium 8 h before the measurement for all conditions, and treatments were added separately at designated timepoints (8 h, 4 h, 2 h and 1 h). The PCA signal was detected with the substrate for the luciferase, native coelenterazine, which was injected into each well directly in the plate reader at a final concentration of 25 mM. One data point was excluded from Fig.by applying Grubb's method for outlier detection test, which detected it as a significant outlier. This was a pre-established exclusion criteria and a statistical formality, as not excluding the value did not alter the statistical results of the analysis in any significant manner. to an EASY-nLC 1000 nanoflow LC (Thermo Fisher Scientific), using the Xcalibur version 3.1.66.10 (Thermo Scientific). The peptides were separated by a 60 min linear gradient using the C18-packed analytical column (Acclaim PepMap 100, 75 µm × 15 cm, 2 µm, 100 Å). The MS analysis was performed as data dependent acquisition in positive-ion mode. MS spectra were acquired from m/z 200 to m/z 2000 with a resolution of 70,000. The ten most abundant ions with charge states from 2+ to 7+ were selected for subsequent fragmentation (higher-energy collisional dissociation), and tandem MS spectra were acquired with a resolution of 17,500. Dynamic exclusion duration was 30 s. For each sample three biological replicates were analyzed. The raw MS data were processed with Andromeda search enginevia Maxquant version 1.6.0.16 (ref. 54). Searches were done against the reviewed human Uniprot database (release 03/2020) supplemented with porcine trypsin and green fluorescence protein sequences (total of 20,303 entries). Carbamidomethylation (+57.021 Da) of cysteine residues was used as static modification and oxidation (+15.994 Da) of methionine was used as dynamic modification. Precursor mass tolerance and fragment mass tolerance were set to less than 20 p.p.m. and 0.1 Da, respectively. A maximum of two missed cleavages was allowed, and the results were filtered to a maximum false discovery rate of 1%. Contaminant Repository for Affinity Purification database (CRAPome, www.crapome.org/) were used to identify high-confidence interactors on the basis of the tandem MS counts.
The TM dimer of TrkB (residues 427-459) was modeled as previously described. An extensive set of atomistic MD simulation systems (Supplementary Table) was prepared on the basis of this model to (1) explore the LSD binding site and mode, (2) optimize the binding modes of each drug (LSD, LSD + , PSI, PSI + , lisuride + , lisuride, cabergoline and dihydroergotamine) in the WT and mutant (Y433F and V437A) dimers, (3) estimate the drug binding affinities and (4) explore the effects of CHOL and the drugs on the conformation of the protein. The binding mode and site for neutral LSD was characterized using 100 independent simulations, each 1 µs long (system 1). In each of these simulation repeats, starting with the TrkB dimer embedded in a membrane (80:20 mol:mol POPC:CHOL), LSD was initially placed within the vicinity of the extracellular crevice in a different random orientation. The simulation data were then analyzed for stable association of the drug. The identified binding mode and site for neutral LSD was later used as an initial estimation for LSD + , PSI, PSI + , lisuride + , lisuride, cabergoline and dihydroergotamine in systems 2-9, where the binding site for each drug was optimized in separate 2-µs-long simulations. For both psychedelics, the predominant titration state in an aqueous solution at neutral pH is the protonated form (pK a > 7.8) (ref. 55). LSD and PSI display higher affinity for TrkB in protonated form compared with their neutral form. The drug binding modes were further optimized for selected homozygous mutants (Y433F and V437A) in systems 10-15. The binding affinities for each drug for the respective optimized binding mode were estimated using the free energy perturbation (FEP) method with 24 windows (60 ns each) coupled via the Hamiltonian replica exchange scheme (FEP/HREX, Systems 16-29). To explore the effect of CHOL and the drugs on the conformation of the protein, additional systems were prepared by re-embedding the drug-bound proteins in 60:40 (mol:mol) POPC:CHOL membrane and/or also removing the bound drug (systems 30-34). Ten repeats were performed, each initiated with different velocities. For neutral PSI, only optimization of the binding mode and binding affinity estimation for the WT protein were performed, which showed low binding affinity. All membrane-protein systems were constructed using CHARMM-GUIusing the Charmm36(m) force field for the proteinand lipids, the TIP3P model for water, and a compatible parameter set for ions. All drugs were parametrized using the CHARMM-GUI ligand reader and modelerwith the CHARMM general force field
N2a cells were transfected to express phGLuc-TrkB using the same protocol as the one described for PCA. Forty-eight hours after transfection, mRNA was extracted using QIAzol Lysis Reagent (Qiagen, 79306). Potential double-strand DNA contamination was removed by dsDNase treatment, and complementary DNA was synthesized using Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, K1672) following the manufacturer's protocol. RT-qPCR was carried out in the CFX96 Real-Time PCR System (Bio-Rad) with 0.25 µM of each primer (Extended Data Table), using SYBR Green PCR Master Mix (Thermo Fisher Scientific, 4309155). All reactions were run in triplicate for 44 cycles using standard cycling conditions-initial 10 min at 95 °C, followed by 44 cycles of 15 s denaturation at 95 °C, 30 s annealing at 54 °C (for 5-HT 5A and 5-HT 1F primers) or 65 °C (for all other primers), 30 s extension at 72 °C-and finished with a melting curve analysis (5 s at 65 °C). The sequence of primersused can be found in Extended Data Table. Cortical neurons (DIV21) were treated with DMSO (0.1 %) or LSD (100 nM) for 1 h, 24 h or 72 h. The same protocol was followed as with N2a cells for RNA extraction and cDNA synthesis. For the RT-qPCR, annealing temperature for Bdnf was 62 °C, and 63 °C for Ntrk2 and Actb. Bdnf and Ntrk2 values were normalized to control beta actin (Actb). The ΔΔCt method was applied and the fold changes of treatment groups (LSD) over the control groups (vehicle) were analyzed within each timepoint. No template and no reverse transcriptase enzyme controls were included for each reaction. The following primers were used: Bdnf total (forward: 5′-GAAGGCTGCAGGGGCATAGACAAA-3′, reverse: 5′-TACACAGGAAGTGTCTATCCTTATG-3′) (ref. 34); Ntrk2 (forward: 5′-GAAGGGAAGTCTGTGACCATTT-3′, reverse: 5′-GTGTGTGG CTTGTTTCATTCAT-3′); Actb (forward: 5′-TGTCACCAACTGGGAC GATA-3′, reverse: 5′-GGGGTGTTGAAGGTCTCAAA-3′) (ref. 78).
Rat hippocampal neuronal cultures were grown in 1.5H glass coverslip bottom dishes (Ibidi, 81158) and at DIV21 treated with 0.1% DMSO, LSD or PSI (100 nM, 1 h, 37 °C). Then, dishes were fixed with 4% paraformaldehyde (PFA) containing 0.1 µm TetraSpeck beads (Invitrogen, T7279; 1:100). After washing with PBS, dishes were incubated with the blocking buffer (0.1% Triton X-100, 0.05% Tween-20, 10% normal donkey serum, 1% BSA and 0.1% gelatin in PBS) for 1 h at room temperature. Next, dishes were incubated with goat anti-TrkB (R&D Systems, AF1494; 1:100) and rabbit anti-PLCγ1 (Cell Signaling Technologies, 5690S; 1:100) primary antibodies diluted in detergent-free blocking buffer overnight at 4 °C. After washing with PBS, dishes were incubated with Alexa Fluor 647-conjugated donkey anti-goat (Invitrogen, A32849; 1:500) and CF568-conjugated donkey anti-rabbit (Biotium, 20098; 1:500) secondary antibodies in PBS. Dishes were washed and stored in PBS until imaging. For dSTORM/TIRF, we used fresh blinking buffer (0.5 mg ml -1 glucose oxidase, 40 µg ml -1 catalase, 10% glucose, 10 mM NaCl, 10 mM cysteamine and 50 mM Tris-HCl, pH 8.0). We used a Nikon Eclipse Ti-E N-STORM/TIRF microscope equipped with an ApoTIRF 100×/1.49 numerical aperture (NA) objective, a Perfect Focus System for Z focus stabilization, an Andor iXon + 897 back-illuminated EMCCD camera (512 × 512), laser lines 405 nm (100 mW), 561 nm (150 mW) and 647 nm (300 mW) and a 405/488/561/647 nm laser quad band filter cube. Image acquisition was performed with the central 256 × 256 pixels and a total of 20,000 frames (exposure time of 15 ms) for both channels using the NIS-Elements Advanced Research software (Nikon). The 405 nm laser was progressively raised to maintain a constant number of localizations per frame. After chromatic aberration correction by correlation with the fluorescent bead fiducials, fitting of the blinking events to build the single-molecule localization coordinate map and drift correction were performed using the Huygens Localizer 21.04 software (Scientific Volume Imaging). Before identifying particles, the first 512 frames of each channel were discarded, a temporal filter length (factor 30) was applied to determine background signal, and data were prefiltered through the standard deviation of a Gaussian filter (σ = 1.0). Next, a threshold calculated from the standard deviation of unfiltered data was multiplied by a factor of 0.75 for the red and 1.5 for the far-red channels and applied to distinguish particles from the background. Pixels from filtered data with values larger than the threshold were considered as candidates for single-molecule localizations. Pixels in filtered data with higher values than surrounding pixels were detected as local maxima. The least-squares fitting was applied for subpixel localization of detected particles, with a fixed full width at half maximum of the Gaussian point spread function of 256 nm. Finally, the resulting single-molecule localization coordinate maps were analyzed with Coloc-Tesseler 1.0 (ref. 79) by an experimenter blinded to treatment to compute the Manders colocalization coefficients of TrkB (M TrkB ) and PLCγ1 (M PLCγ1 ) in dendritic spines on the basis of their overlapping Voronoi diagrams.
The phosphorylation of TrkB (pY816) and interaction with PLCγ1 were assessed by enzyme-linked immunosorbent assay (ELISA) as previously describedwith minor modifications. Briefly, white 96-well plates (PerkinElmer, OptiPlate 96F-HB) were precoated with goat anti-TrkB (R&D systems, AF1494; 1:500, overnight, 4 °C) antibody in carbonate buffer (pH 9.8) and blocked with 3% BSA in PBS with 0.1% Tween-20 (PBS-T) for 2 h at room temperature. Next, lysates from rat cortical or hippocampal neuronal cultures (DIV21) that had been treated with DMSO or LSD (1 nM to 100 nM, 1 h, 37 °C) were loaded into the plates. Alternatively, we used lysates containing equal amounts of total protein from PFC or hippocampus dissected from adult mice (WT and Y433F +/-) that were killed 1 h after a single intraperitoneal administration of LSD (0.1 mg kg -1 ). Samples were incubated overnight at 4 °C, and 24 h later plates were washed with PBS-T and incubated with either rabbit anti-phospho-TrkB(Tyr816) (Cell Signaling Technology, 4168S; 1:2,000), rabbit anti-PLCγ1 (Cell Signaling Technologies, 5690S; 1:2,000) or rabbit anti-Trk (Cell Signaling Technologies, 92991; 1:2,000) antibodies overnight at 4 °C. After washing with PBS-T, plates were incubated with HRP-conjugated goat anti-rabbit (Bio-Rad, 170-5046; 1:5,000, 1 h, room temperature) antibody, followed by more washing with PBS-T. Finally, the chemiluminescent signal was read after brief incubation with Pierce ECL substrate (Thermo Fisher Scientific, 32106). Luminescence from blank wells that contained all reagents but the sample lysates was used to determine background signal. The cell surface retention of TrkB was measured as previously described. Rat cortical neurons (DIV21) cultivated in clear bottom 96-well plates (PerkinElmer, ViewPlate 96) were treated with DMSO (0.1%) or LSD (100 nM) at different timepoints (5 min, 15 min, 30 min and 1 h) and fixed with 4% PFA. For antagonist experiments, cultures were pretreated with M100907 (100 nM) for 10 min and remained present for the rest of the experiment. After washing with PBS and incubating for 1 h with blocking buffer (5% nonfat dry milk and 5% BSA) at room temperature, plates were incubated with with goat anti-TrkB antibody (R&D Systems, AF1494; 1:1,000, overnight, 4 °C). After further PBS washes, plates were incubated with HRP-conjugated goat anti-rabbit antibody (Bio-Rad, 170-5046; 1:5,000, 1 h, room temperature). Finally, wells were washed with PBS and the chemiluminescent signal was read after brief incubation with Pierce ECL substrate (Thermo Fisher Scientific, 32106). To determine BDNF protein concentrations, ELISA was run as previously describedwith minor modifications. Rat cortical neurons (DIV21) were treated with DMSO (0.1%) or LSD (100 nM) for 5 min, 1 h, 24 h or 72 h, lysed and centrifuged. Samples were diluted in Hanks buffer (125 mM NaCl, 5 mM KCl, 1.2 mM NaH 2 PO 4 , 1 mM CaCl 2 , 1.2 mM MgCl 2 , 1 µM ZnCl 2 , 10 mM glucose, 25 mM HEPES and 0.25% BSA, pH 7.4). Samples and BDNF standards (10-1,000 pg ml -1 , Peprotech) wereintensity baseline was established. Next, spines were bleached with high laser power until the fluorescence signal decreased to 50% of baseline intensity. Fluorescence recovery in spines was monitored for 1 min after bleaching at a rate of 1 frame per second (f.p.s.). Zen blue 3.2 was used for image analysis. The recovery curve was obtained after subtracting the background and normalizing the bleached spine signal with that of a neighboring area that was not bleached and had equivalent fluorescence intensity at baseline. Finally, the normalized fluorescence intensity values of bleached spines were plotted as percentage of baseline signal and fitted to the exponential one-phase association equation in GraphPad Prism 9.2 to build the recovery curve.
Spinogenesis experiments were performed similarly to previous studieswith modifications. Neuronal cultures of WT and Y433F +/-E18 mouse embryos grown in 1.5H glass coverslips were treated with DMSO, LSD or PSI (100 nM) at DIV20. In a separate set of experiments to study the effects of extracellular BDNF and 5-HT 2A activation, cultures were also pretreated for 10 min with TrkB.FC (200 ng ml -1 ) or M100907 (100 nM), respectively. TrkB.FC and M100907 remained in the medium for the rest of the experiment. Twenty-four hours later, coverslips were fixed with 4% PFA, washed with PBS, incubated with blocking buffer (0.1% Triton X-100, 0.05% Tween-20, 10% normal donkey serum, 1% BSA and 0.1% gelatin in PBS) for 1 h at room temperature and then overnight with rabbit anti-MAP2 (Invitrogen, PA5-17646; 1:100) primary antibody at 4 °C. After washing with PBS, coverslips were incubated with CF568-conjugated donkey anti-rabbit (Biotium, 20098; 1:500) for 1 h at room temperature. Another round of PBS washes was followed by incubation with Alexa Fluor 647-conjugated phalloidin (Invitrogen, A22287; 1:40) overnight at 4 °C. Next, coverslips were briefly washed in phosphate buffer (PB) and mounted on slides with ProLong Glass Antifade Mountant (Invitrogen, P36982). We used a Zeiss LSM 880 confocal microscope with an AiryScan GaAsP detector equipped with a PlanApochromat 63×/1.4 NA objective. We used the BP 495-550 + LP 570 AiryScan filter and Zen Black in super-resolution mode for image acquisition. Pinhole size was adapted accordingly, and frame size was set to optimal. Sensors were set to continuous adjustment across the imaging sessions to ensure optimal sensor alignment and signal detection. Raw data were processed into AiryScan images with the autofilter mode on using Zen Blue. Finally, image analysis was performed in ImageJ by experimenters blinded for treatment and genotype. Spines were quantified in 15-µm-long dendrite sections.
Neurons from WT or Y433F +/-mouse embryo (E18) were plated and treated with 100 nM of psychedelics (LSD and PSI) or vehicle (0.1% DMSO) once a day for three consecutive days starting at DIV5 (cumulative doses). In a separate set of experiments to study the effects of extracellular BDNF and 5-HT 2A activation, cultures were also pretreated for 10 min with TrkB.FC (200 ng ml -1 ) or M100907 (100 nM), respectively. Similar to treatment with the psychedelics, TrkB.FC and M100907 pretreatments were administered in cumulative doses and remained in the medium for the rest of the experiment. At DIV7, neurons were pre-incubated for 1 h with 10 µM MgCl 2 to prevent excitotoxicity, then neurons were transfected for the heterologous expression of mCherry. To minimize any possible interference of the transfection protocol upon the drug effect, transfection was done in the morning while treatment with psychedelics was always added in the evening. Twenty-four hours after transfection, neurons were fixed with 4% PFA for 20 min at room temperature, washed in PBS-T, blocked for 30 min with 3% BSA in PBS-T and incubated with chicken anti-MAP2 (Abcam, ab5392; 1:5,000, 1 h, room temperature). Then, coverslips were washed in PBS-T and incubated with Alexa Fluor 647-conjugated antibody goat anti-chicken (Invitrogen, A-21449; 1:1,000, 1 h, room temperature). Next, coverslips were washed in PBS-T and MilliQ water, and mounted acidified at room temperature to pH 3 by adding 1 M HCl and neutralized to pH 7 after 15 min with 1 M NaOH. A Maxisorb Nunc ELISA plate (Thermo Scientific) previously coated with a primary BDNF antibody (1:1,000 in carbonate buffer, overnight, 4 °C) and blocked (2% BSA and 0.1% Triton X-100 in Hanks buffer, 2 h, room temperature). Then, 170 µl of samples and standards were added to the plate with 30 µl of the secondary POD-conjugated BDNF antibody (1:1,000 in 6.66% BSA and 0.66% Triton X-100 in Hanks buffer). The BDNF antibodies were received from Roche and have been previously described. After incubating overnight at 4 °C the plate was washed with PBS-T. BM Blue POD substrate (Roche, 54827-17-1) was added to each well, and the colorimetric reaction was stopped after 20 min by adding 50 µl of 1 M H 2 SO 4 . Absorbance was measured at 490 nm with a plate reader (Varioskan Flash, Thermo Scientific).
To quantify TrkB protein expression in N2a cells, plates were transfected to express phGLuc-TrkB using the same protocol as described for PCA. Forty-eight hours after transfection, cells were treated with DMSO (0.1 %), LSD (100 nM) or BDNF (10 ng ml -1 ) for 40 min at 37 °C, lysed and centrifuged. To examine the effects of LSD treatment on the phosphorylation of mTOR and ERK, rat cortical neurons (DIV21) were treated with DMSO (0.1%) or LSD (100 nM) for 5 min, 1 h or 24 h. In a separate set of experiments to examine the effect of 5-HT 2A activation, M100907 (100 nM) was administered as a 10 min pretreatment and remained in the medium for the rest of the experiment. Next, cells were lysed and centrifuged. In an additional experiment to study the effect of LSD on TrkB interaction with PLCγ1, we used lysates from PFC dissected from adult WT mice that were killed 1 h after a single intraperitoneal administration of LSD (0.1 mg kg -1 ). The co-immunoprecipitation was carried out as previously describedwith minor modifications. Briefly, Magnetic Dynabeads Protein G (Invitrogen, 10004D) were coated with anti-TrkB antibody (R&D Systems, AF1494) at 200 ng of antibody per 1 µl of beads in rotation (2 h, 4 °C). Finally, lysates were mixed with the coated beads (100 µg of total protein per 15 µl of beads) and incubated in rotation (overnight, 4 °C). Proteins were separated according to their molecular weight by SDS-PAGE and transferred to a polyvinylidene fluoride membrane. Membranes were incubated with the following primary antibodies (all 1:1,000 in 3% BSA in TBS-T, overnight, 4 °C): anti-TrkB (R&D Systems, AF1494), anti-β-actin (Sigma-Aldrich, A1978), anti-phospho-mTOR (Cell Signaling Technology, 5536S), anti-phospho-p44/42 MAPK (Erk1/2) (Cell Signaling Technology, 9106S), anti-mTOR (Cell Signaling Technology, 2972S) and anti-p44/42 MAPK (Erk1/2) (Cell Signaling Technology, 9102S). Membranes were washed with TBS-T and incubated with corresponding secondary HRP-conjugated antibodies (all 1:10,000 in 3% BSA in TBS-T, 1 h, room temperature): HRP-conjugated RaG (Invitrogen, 61-1620), HRP-conjugated GaR (Bio-Rad, 1705046) and HRP-conjugated GaM (Bio-Rad, 170-6515). Protein bands were visualized using enhanced chemiluminescent substrate (Thermo Scientific, 32132, ECL Plus) and imaged with a charge-coupled device camera (G:BOX, Syngene).
FRAP was performed as previously describedwith minor modifications. Briefly, hippocampal neurons (DIV16) derived from rat embryos (E18) were grown in glass coverslips and transfected to express GFP-tagged TrkB (WT, Y433F). For FRAP we used a Zeiss LSM 710 confocal microscope equipped with a W PlanApochromat 63×/0.9 NA objective and adapted with an environmental control chamber for live cell imaging set at 37 °C and 5% CO 2 . Twenty-four hours after transfection, coverslips were transferred into pre-acclimatized 35 mm plates containing phenol red-free Hank's balanced salt solution. Plates were then administered 0.1% DMSO or psychedelics (LSD and PSI) at a final concentration of 100 nM. FRAP was performed during the first hour after treatment. Dendritic spines were localized, and a prebleachingin Dako fluorescence mounting medium (Agilent, S302380-2). Ten optical sections with a Z-step of 1.3 µm were acquired using a Zeiss LSM 710 confocal microscope with a LCI Plan/Neofluar 25×/0.8 lmm Korr DIC M27 objective. In the mCherry channel, the center of the soma was manually assigned and the number of dendritic intersections at fixed distances (every 5 µm) from the soma was automatically counted using the Sholl analysis plugin of ImageJ.
WT and Y433F +/-male mice were administered four doses of 5-bromo-2′-deoxyuridine (BrdU, 75 mg kg -1 per injection, every 2 h) intraperitoneally to tag newborn neurons in the hippocampus and returned to their home cage. Twenty-four hours later, animals were injected a single dose of saline or LSD (0.1 mg kg -1 ) intraperitoneally. Four weeks later, animals were intracardially perfused with PBS followed by 4% PFA. Brains were dissected and postfixed with 4% PFA (overnight, 4 °C). Next, brains were transferred to 30% sucrose (0.1% PBS) solution and stored at 4 °C for 48-72 h. Brains were moved into disposable vinyl specimen molds (Sakura, Tissue-Tek Cryomold standard) and covered in Tissue-Tek O.C.T. Compound (Sakura). Trays were placed on dry ice and brains covered once more in TissueTec to ensure full coverage. Once frozen, samples were stored at -80 °C. Brains were sliced coronally at 40 µm thickness, collecting every sixth slice. Slices were stored in cryoprotectant fluid (30% sucrose, 30% ethylene glycol, 1% PVP-40, and 0.05 M PB) in -20 °C until staining. Every third hippocampal slice containing dentate gyrus (DG) was collected for washing with 0.1 M PBS and staining. BrdU immunofluorescence was conducted similarly as previously reported in the literaturewith modifications. Hippocampal slices were incubated in 2.0 M HCl (30 min, 37 °C) and washed with PBS. Slices were then incubated in 0.1 M borate buffer (pH 8.0) at room temperature for 10 min, followed by further washes in PBS and 1 h incubation in 0.1% PBS blocking buffer (0.5% Triton X-100, 5% normal donkey serum, 5% normal goat serum, 1% BSA and 0.1% gelatin). Sections were incubated with primary antibodies rat anti-BrdU (Abcam, ab6326; 1:200) and mouse anti-NeuN (Millipore, MAB377; 1:1,000) for 48 h at 4 °C with gentle shaking. Slices were then washed with PBS, followed by incubation with secondary antibodies Alexa Fluor 488 donkey anti-rat (Invitrogen, A21208; 1:300) and Alexa Fluor 568 goat-anti-mouse (Invitrogen, A11004; 1:500) for 2 h at room temperature and protected from light. Slices were washed in PBS and PB before mounting with Dako fluorescence mounting medium (Agilent, S302380-2) and slides stored at 4 °C until imaging. A total of eight slices containing both hemispheres and covering the whole DG (every third slice) were imaged with a Zeiss LSM 710 confocal microscope. Fifteen optical sections with a Z-step of 2 µm were acquired per slice with a PlanApochromat 10×/0.45 M27. BrdU + and BrdU + /NeuN + cells in the DG were counted manually by experimenters blinded to treatment and genotype using ZEN Blue 3.2. Similarly to Ma et al., we estimated the number of mature neurons (BrdU + /NeuN + ) in the DG of both hemispheres by multiplying by 6 the total count derived from every third hippocampal slice. The percentage of mature neurons over total BrdU + cells in the DG was also calculated.
Female and male mice were 17 weeks old at the time of monocular deprivation, since the physiological critical period in the binocular region of the visual cortex is closed by then. Transparent skull surgeries were executed as described before. Briefly, mice were injected with an anesthetic mixture containing 0.05 mg kg -1 fentanyl (Hameln), 5 mg kg -1 midazolam (Hameln) and 0.5 mg kg -1 medetomidine (Orion Pharma) diluted in saline. Eyes were protected from drying with eye-protective gel (Alcon), and the head was shaved and fixed in a stereotaxic frame with a heating pad maintained at 37 °C. A mixture of lidocaine and adrenaline (Orion Pharma) was applied locally on the head, and the scalp above the visual cortex was cut. The periosteum was removed gently, and the skull was rapidly cleaned from fat with acetone. The surface of the skull was covered with a thin layer of cyanoacrylate glue (Henkel, Loctite 401), followed by two layers of acryl, obtained by mixing colorless acrylic powder (EUBECOS) with liquid methacrylate (Densply). Mice were awakened with 1 mg kg -1 atipamezole (VetMedic) and postoperative analgesia provided with 0.1 mg kg -1 buprenorphine (Indivior) and 5 mg kg -1 carprofen (VetScan). After 24 h, mice were anesthetized with isoflurane and the acryl polished. A metal bar holder was glued and fixed with a mixture of cyanoacrylate glue and dental cement (Densply) to the skull, and transparent nail polish (Electron Microscopy Sciences, 72180) was applied above the visual cortex, which was visible at the center of the holder. Monocular deprivation lasted for 1 week. Mice were anesthetized with isoflurane, antibiotic gel (Dechra, Isathal Vet 1%) was applied to the eye contralateral to the hemisphere being imaged, eyelashes were cut and eye was sutured shut with three mattress sutures (Ethiconn). Antibiotic ointment (Anten, Oftan Dexa-Chlora) was applied on the sutured eye and 5 mg kg -1 carprofen was injected subcutaneously for postoperative analgesia. All animals were checked daily, and the sutured eye was opened right before the second imaging. Mice showing corneal injury were excluded from experiments. LSD (0.1 mg kg -1 ) was administered intraperitoneally three times on alternate days during the week-long monocular deprivation period following previously described protocols. Optical imaging of the intrinsic signal was performed to assess visually evoked cortical activity by measuring the changes in the oxygen saturation of hemoglobin in the primary visual cortex. Briefly, two imaging sessions were performed: one before LSD administration and monocular deprivation (IOS I) and a second session at the end of monocular deprivation (IOS II). Anesthesia was induced with 1.8% isoflurane and maintained with 1.2% isoflurane in a 1:2 mixture of O 2 :air. Animals were kept on a heating pad, 25 cm in front of the stimulus monitor, where a 2° horizontal bar moving upwards was shown with temporal and spatial frequency of 0.125 Hz and 1/80 degree, respectively. The visual stimulus was limited to the central part of the monitor (width: -15° to 5° azimuth, relative to the animal visual field) to stimulate only the binocular part of the visual field. The continuous-periodic stimulation was synchronized with a continuous frame acquisition, collected independently for each eye. To acquire a vascular map, the skull was illuminated with a green light (540 ± 20 nm). The camera was then focused 600 µm below the cortical surface, and a red light (625 ± 10 nm) was used to record the intrinsic signal. The signal was recorded from one eye at a time, with the other eye covered by a patch. The continuous-periodic stimulation was synchronized with a continuous frame acquisition, and frames were collected at a rate of 30 f.p.s. for 5 min and stored as a 512 × 512 pixel image. Following the acquisition for both contralateral (C) and ipsilateral (I) eyes, the cortical maps were computed after spatial binning and thresholding by using a custom-made analysis software package that performs Fourier decomposition on the cortical maps. The OD score as (C - I)/(C + I) was calculated, and the ocular dominance index (ODI) was determined as the mean of the OD score for all responsive pixels, as described by Cang et al.. The ODI values range between -1 and +1, where positive values correspond to a contralateral bias and negative ones to ipsilateral bias. ODI values of 0 indicate that inputs from ipsilateral and contralateral eyes are equally strong.
WT and Y433F +/-mice were administered a single intraperitoneal dose of saline or LSD (0.1 mg kg -1 ). Another independent cohort of WT mice received a single intraperitoneal injection of M100907 (0.5 mg kg -1 ) or vehicle followed 30 min later by another injection of LSD (0.1 mg kg -1 ) or vehicle. Immediately after the last injection, the animals were placed inside a transparent chamber and their activity video-recorded for 30 min. Similarly to previous studies, head twitches were scored as counts every 2 min by observers blinded to treatments and genotypes. rFST rFST was conducted in WT and Y433F +/-mice as described in literaturewith minor modifications. Briefly, female and male adult mice were exposed to five consecutive pretest sessions (5 min of duration, 24 h apart). In the first experiment, WT or Y433F +/-mice received a single intraperitoneal injection of LSD (0.1 mg kg -1 ) or vehicle 24 h after the last pretest session. In a second independent experiment, WT mice received an intraperitoneal injection of M100907 (0.5 mg kg -1 ) or vehicle 24 h after the last pretest session, followed by an intraperitoneal injection of LSD (0.1 mg kg -1 ) or saline 30 min later. The animals were submitted to a 5 min test session 7 days after the drug administration. The apparatus consisted of 5 l glass beaker cylinders (19 cm diameter, 20 cm water column at 23 ± 2 °C). The water was changed between each trial, for each animal. After each session, the animals were dried and kept in a warmed cage before returning to their home cages. The pretest and test sessions were recorded and analyzed by a trained observer blind to treatment and genotype.
Male WT mice (C57BL/6JRccHsd, Envigo) were used for contextual fear conditioning experiments with and without extinction training as previously describedwith modifications. Briefly, animals were conditioned in context A (metallic bar flooring, transparent Plexiglass walls, 23 × 23 × 35 cm arena) with a series of five scrambled foot shocks (0.6 mA, 2 s) along a 480 s session. Twenty-four hours later, animals were administered a single dose of saline or LSD (0.1 mg kg -1 ) intraperitoneally and returned to their home cage. During the following 3 days, mice were exposed to extinction training in context A for 480 s every 24 h or placed for the same duration and number of sessions in context B (black planar floor, black nontransparent walls, 23 × 23 × 35 cm arena), which we categorized as extinction and no extinction trainings, respectively. Afterwards, animals were placed back into their home cage and left undisturbed for 3 weeks. Finally, all mice were reconditioned with a new set of five scrambled foot shocks (0.6 mA, 2 s) along a 480 s session in context A, and their fear reinstatement response was measured 24 h later for the same duration also in context A. All sessions were conducted under constant illumination (100 lux). For the contextual fear conditioning and fear reinstatement experiment with male WT and Y433F +/-, we used the same protocol as for animals undergoing extinction training in context A above. Baseline freezing was established during the first conditioning session. Movement was measured by infrared beams placed on the sides of the conditioning chamber, and freezing was automatically determined by the fear conditioning apparatus and software (TSE Systems) and expressed as the percentage of time spent freezing.
Data were analyzed with GraphPad Prism 9.2. Detailed statistical information for all presented data is available in Supplementary Table. No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those reported in previous publications. Data met the assumptions for normality and equal variances for the statistical tests used, and when this was not the case, data were assumed to be normal and individual data points are displayed. Sample and animal distribution among groups was randomized. Single comparisons among groups were made using Student's t-test (two-tailed, paired or unpaired), and one-way or two-way analysis of variance followed by Bonferroni's post hoc test was used for multiple comparisons unless indicated otherwise. Data are presented as mean ± standard error of the mean (s.e.m.) and *P < 0.05, unless otherwise indicated.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. widely expressed across many brain areas and cell-types, the enhanced state of plasticity induced by psychedelics and other antidepressants that bind to TrkB may allow for a relatively widespread rewiring of neuronal networks (for example PFC, visual cortex, hippocampus). However, the enduring changes at the network and behavioral levels following psychedelic-induced plasticity depend on environmental input (for example extinction training, monocular deprivation). Thus, the therapeutic potential of psychedelics for treating psychiatric disorders may lie in their combination with environmental support, and not in their induction of plasticity alone. Network representations in this panel were modified from Castrén, E..
Cao, D., Yu, J., Wang, H. et al. · Science (2022)
Kaplan, A. L., Confair, D. N., Kim, K. et al. · Nature (2022)
Rodriguiz, R. M., Nadkarni, V., Means, C. R. et al. · Scientific Reports (2021)
Thuery, G., Crossen, F., Mc Loone, D. et al. · Therapeutic Advances in Psychopharmacology (2026)
Fabiano, N., Stubbs, B., Lawrence, D. W. et al. · Discover Mental Health (2026)
Jacobs, E., Zahid, Z., Hinkle, J. et al. · BMJ (2026)
Aarrestad, I. K., Cameron, L. P., Fenton, E. M. et al. · Nature Neuroscience (2025)
Jain, M. K., Gumpper, R. H., Slocum, S. T. et al. · Neuron (2025)
Peck, S. K., Knatz Peck, S., Brewerton, T. D. et al. · Journal of Eating Disorders (2025)
Zuiker, R. G. J. A., Otto, M. E., Bryan, C. S. et al. · Clinical and Translational Science (2025)
Lu, O. D., White, K., Raymond, K. et al. · Biorxiv (2025)
Calder, A. E., Hase, A., Hasler, G. · Molecular Psychiatry (2024)
Agnorelli, C., Spriggs, M. J., Godfrey, K. et al. · Preprints (2024)
Sabé, M., Sulstarova, A., Glangetas, A. et al. · Molecular Psychiatry (2024)
Enriquez-Geppert, S,, Lietz, M. P., O'Higgins, F. · Philosophical Transactions of the Royal Society B (2024)
Fleury, S., Nautiyal, K. M. · Biorxiv (2024)
Avram, M., Fortea, L., Wollner, L. et al. · Molecular Psychiatry (2024)
Anderson, T. L., Keady, J. V., Songrady, J. et al. · Progress in Neurobiology (2024)
Salvetti, G., Saccenti, D., Moro, A. S. et al. · Brain Sciences (2024)
Juliani, A., Chelu, V., Graesser, L. et al. · Biorxiv (2024)
Tap, S. C. · Addiction Biology (2024)
Wong, S., Kwan, A. T. H., Teopiz, K. M. et al. · Journal of Affective Disorders (2024)
De Filippo, R., Schmitz, D. · Neuroscience and Biobehavioral Reviews (2024)
Murray, C., Frohlich, J, Haggarty, C. J., Tare, I. et al. · Neuropsychopharmacology (2024)
Allen, J., Dames, S., Foldi, C. J. et al. · Molecular Psychiatry (2024)
Molla, H. M., Lee, R., Tare, I. et al. · Neuropsychopharmacology (2023)
Cameron, L. P., Benetatos, J., Lewis, V. et al. · Journal of Neuroscience (2023)
Enriquez-Geppert, S,, Krc, J., O'Higgins, F., Lietz, M. P. · OSF Preprints (2023)
Jaster, A. M., González-Maeso, J. · Molecular Psychiatry (2023)
Wall, M. B., Harding, R., Zafar, R. et al. · Molecular Psychiatry (2023)
Tyagi, R., Saraf, T. S., Canal, C. E. · ACS Pharmacology and Translational Science (2023)
Kovacevich, A., Weleff, J., Claytor, B. et al. · Journal of Psychoactive Drugs (2023)