Psychedelics promote neuroplasticity through the activation of intracellular 5-HT2A receptors

Dysregulation of the cortex, including reduced dendritic arbor complexity and lower dendritic spine density, is implicated in several neuropsychiatric disorders and is a target for therapeutics that promote structural plasticity. Psychoplastogens such as ketamine and serotonergic psychedelics produce rapid and sustained increases in cortical structural and functional plasticity; for psychedelics this effect requires activation of 5-hydroxytryptamine 2A receptors (5-HT2ARs). However, a puzzling observation is that some 5-HT2AR agonists promote cortical neuroplasticity while others, including endogenous serotonin, do not, despite serotonin being a potent, balanced 5-HT2AR agonist in canonical signalling assays. The physicochemical differences between ligands led the investigators to consider location bias — the idea that receptor signalling differs depending on whether ligands access plasma-membrane versus intracellular receptor pools — as a possible explanation. Vargas and colleagues set out to test whether activation of an intracellular population of 5-HT2ARs is necessary for psychedelics to induce cortical structural plasticity and related antidepressant-like behavioural responses. To do so they combined in vivo pharmacology and genetics (including 5-HT2AR knockout mice and viral expression of transporters), in vitro neuronal culture assays and structure–activity studies of tryptamine analogues, fluorescence- and biophysical-reporter assays of receptor conformation and signalling, subcellular localisation experiments, and chemical modification of ligands to alter membrane permeability.

The study used a mixture of in vivo, ex vivo and in vitro approaches. In vivo, wild-type and 5-HT2AR knockout mice received systemic administration of 5-methoxy-N,N-dimethyltryptamine (5-MeO-DMT or 5-MeO) and changes in layer 5 pyramidal neurons of the prefrontal cortex (PFC) were assessed 24 hours later by Golgi–Cox staining and ex vivo electrophysiology to measure spontaneous excitatory postsynaptic currents (sEPSCs). Behavioural assays included the head-twitch response (HTR), novelty-induced locomotion, and the forced swim test (FST). Viral and AAV constructs were used to express serotonin transporter (SERT) selectively in cortical pyramidal neurons (CaMKII promoter) in vitro (electroporated cultures) and in vivo (mPFC injections) to permit intracellular import of serotonin. In vitro experiments employed primary rat embryonic cortical neurons to assess structural plasticity using Sholl analysis of dendritic arbor complexity and spine density measures. Structure–activity relationship (SAR) studies treated neurons with various tryptamine derivatives (serotonin, tryptamine, 5-MeO-tryptamine and their N-methyl congeners) and with ketamine as a positive control. Membrane-permeable ligands (e.g. DMT, psilocin) were chemically converted into permanently charged, membrane-impermeable congeners (TMT, psilocybin, methylated ketanserin) to test requirements for intracellular access; electroporation was used to transiently permit intracellular access for charged compounds. Receptor signalling and localisation were interrogated using several reporter assays and imaging tools. PsychLight2 and psychLight1 fluorescent biosensors were used to detect 5-HT2AR conformational changes; bioluminescence resonance energy transfer (BRET) assays measured Gq activation and β-arrestin-2 recruitment; [3H]-inositol phosphate ([3H]-IP) accumulation and IP1 assays assessed canonical signalling. Subcellular localisation employed Myc–5-HT2AR–ECFP and β2AR–ECFP constructs expressed in HEK293T cells and cortical neurons, colocalisation with a membrane dye, and immunocytochemistry for organelle markers (Rab5, Rab7, Golgi). Antibody specificity was validated with overexpression controls and 5-HT2AR knockout tissue. Radioligand competition binding was performed in intact cells and membrane preparations to compare ligand access to extracellular versus intracellular receptor pools. Where reported, analyses included correlation assessments between ligand lipophilicity (calculated LogP, cLogP) and psychoplastogenic efficacy, comparison of efficacy across assays (psychLight, [3H]-IP, BRET), and statistical tests for spine density, dendritogenesis and electrophysiological endpoints. The extracted text does not clearly report exact sample sizes, full statistical models, or all p-values for each experiment.

Activation of 5-HT2ARs was necessary for psychedelic-induced structural and functional plasticity in vivo: systemic 5-MeO increased dendritic spine density in PFC layer 5 pyramidal neurons in both sexes, an effect absent in 5-HT2AR knockout mice. Ex vivo electrophysiology showed that 5-HT2AR activation was required for sustained increases in both frequency and amplitude of sEPSCs after 5-MeO treatment. SAR studies in cultured rat cortical neurons showed that increasing N-methylation of tryptamine derivatives enhanced the ability to promote dendritic arbor complexity, with N,N-dimethyl compounds (e.g. DMT) producing the greatest effect. However, measures of canonical 5-HT2AR efficacy did not predict psychoplastogenicity: [3H]-IP accumulation assays and psychLight2 biosensor efficacy tended to anticorrelate with dendritogenesis (psychLight2 anticorrelation P = 0.06). BRET assays of Gq activation and β-arrestin-2 recruitment also showed a strong negative correlation with psychoplastogenicity; both Gq activation and β-arrestin recruitment correlated well with psychLight efficacy (coefficient of determination R2 = 0.9; P < 0.0001 and P = 0.0006, respectively). By contrast, calculated lipophilicity (cLogP) correlated positively with psychoplastogenic effects — more lipophilic agonists promoted greater structural plasticity than polar compounds. Subcellular imaging revealed that, unlike in HEK293T cells, 5-HT2ARs in cortical neurons were predominantly intracellular and showed high colocalisation with Rab5 and Rab7 and markedly greater colocalisation with Golgi markers. Control experiments with β2AR constructs and untagged fluorescent proteins argued against tag-driven mislocalisation, and a validated 5-HT2AR antibody showed a native intracellular pattern that matched tagged receptor localisation. Chemical modification experiments established a requirement for membrane permeability to engage the plasticity-promoting receptor population. Permanently charged, membrane-impermeable congeners retained 5-HT2AR binding affinity and psychLight efficacy but could only promote dendritogenesis when electroporation was used to permit intracellular access. Conversely, a membrane-permeable 5-HT2AR antagonist blocked DMT-induced plasticity regardless of electroporation, whereas a membrane-impermeable antagonist only blocked plasticity when electroporated. In more mature neurons, membrane-permeable agonists increased dendritic spine density while membrane-impermeable analogues did not; the permeable antagonist blocked these effects whereas the impermeable antagonist did not. Additional functional separations between ligand populations were observed in psychLight1 and binding experiments: pretreatment with a membrane-impermeable antagonist nearly fully antagonised serotonin-induced psychLight1 responses but only partially blocked responses to the more lipophilic 5-MeO-DMT. Radioligand competition in intact cells versus membrane preparations showed that 5-MeO-DMT was substantially more potent than serotonin when intact cells with intracellular receptors were used. IP1 assays gave concordant results: in HEK293T cells with largely plasma-membrane receptors serotonin and 5-MeO-DMT had comparable potencies, whereas in cortical neurons serotonin failed to elicit an agonist response while 5-MeO-DMT remained a potent agonist. Directly permitting serotonin access to the intracellular compartment produced psychoplastogenic effects. Serotonin promoted dendritogenesis only when applied with electroporation in vitro. Electroporation or viral expression of SERT (under a CaMKII promoter to target excitatory pyramidal neurons) allowed selective import of serotonin into SERT-positive neurons; under these conditions serotonin increased dendritic complexity and spine density only in SERT-positive neurons. The selective SERT inhibitor citalopram blocked serotonin’s effects but did not affect DMT-induced plasticity, while the 5-HT2AR antagonist KTSN blocked both serotonin and DMT effects, indicating reliance on intracellular 5-HT2AR activation. In vivo, expression of SERT in the mPFC followed by administration of the serotonin-releasing agent para-chloroamphetamine (PCA) led to increased dendritic spine density on mPFC pyramidal neurons compared with controls expressing mCherry. Behaviourally, mice with mPFC SERT expression showed no baseline differences in locomotion or FST immobility, but after PCA they exhibited a head-twitch response and reduced immobility in the FST 24 hours post-administration, consistent with an antidepressant-like effect linked to intracellular serotonin access and 5-HT2AR activation. The extracted text does not consistently report sample sizes or full statistical details for each individual experiment.

Vargas and colleagues interpret their results as evidence that a substantial intracellular population of 5-HT2ARs in cortical neurons — notably localised to the Golgi and endosomal compartments — mediates the neuroplasticity-promoting effects of psychedelic 5-HT2AR agonists. They extend the concept of location bias to 5-HT2AR ligands: ligands that are membrane permeable can access intracellular receptor pools and thereby elicit sustained signalling that drives dendritic growth and increases in synaptic strength. The authors suggest that the slightly acidic environment of intracellular organelles such as the Golgi could cause protonation and retention of weak-base psychedelics, supporting prolonged intracellular signalling; persistent growth after transient extracellular exposure is a characteristic of serotonergic psychoplastogens. The investigators acknowledge that the precise signalling cascade linking intracellular 5-HT2AR activation to structural plasticity remains incompletely defined, although prior evidence implicates AMPA receptor, TrkB and mTOR pathways and these are likely involved here. They also note that intracellular 5-HT2ARs might contribute to subjective or hallucinogenic effects: increasing cortical serotonin availability did not elicit an HTR unless SERT was expressed in the mPFC, a region required for the HTR, suggesting that intracellular receptor activation can influence behaviour. Several limitations and open questions are acknowledged. The experiments demonstrate that membrane permeability is essential for a ligand to engage intracellular 5-HT2ARs in cortical neurons, but they do not distinguish whether psychedelics act via bona fide intracellular signalling or by serving as pharmacological chaperones that facilitate receptor export to the plasma membrane where canonical signalling would occur. The reason for the marked difference in 5-HT2AR subcellular localisation between neurons and heterologous cells (HEK293T) is unclear; the authors propose that cell-specific interacting proteins or heterodimeric complexes could direct trafficking and thereby affect ligand responses in a circuit-specific manner. Finally, Vargas and colleagues raise the intriguing possibility that serotonin may not be the endogenous ligand for intracellular 5-HT2ARs in cortex; methylated tryptamines such as DMT, which are more membrane permeable and have been detected in brain tissue, could be candidate endogenous activators and warrant further investigation. The authors recommend future work to delineate the detailed intracellular signalling mechanisms, to distinguish intracellular signalling from chaperone-like effects, to explore the role of intracellular 5-HT2ARs in hallucinogenic versus therapeutic responses, and to assess whether other psychotropic weak bases act at intracellular targets to promote plasticity. The extracted text does not present a separate concluding section.