Prolonged epigenomic and synaptic plasticity alterations following single exposure to a psychedelic in mice

Depressive, anxiety and stress-related disorders are highly prevalent and current pharmacotherapies, including monoamine reuptake inhibitors, are limited by delayed onset of action and a substantial proportion of treatment resistance. Psychedelic compounds (phenethylamines such as DOI and mescaline, and tryptamines such as psilocybin and DMT) produce marked acute alterations in perception and cognition, and recent clinical pilot studies have reported rapid and sometimes long-lasting improvements in mood and anxiety measures following single administrations. Despite promising clinical signals, the molecular mechanisms that might underlie durable antidepressant-like effects of psychedelics—and whether these rely on the same receptor targets implicated in acute hallucinogenic effects—remain incompletely understood. This study set out to characterise post-acute molecular and synaptic plasticity changes in mouse frontal cortex following a single systemic dose of the phenethylamine psychedelic DOI, and to test the extent to which those changes depend on the serotonin 5-HT 2A receptor (5-HT 2A R). The investigators combined behavioural assays relevant to depression, anxiety and fear learning with dendritic spine imaging, electrophysiological measures of synaptic plasticity, cell-type-specific epigenomic profiling (H3K27ac ChIP-seq) and RNA-seq from neuronal nuclei, and computational analyses including enhancer clustering, motif and gene-network analysis. The goal was to link persistent chromatin and synaptic alterations to lasting behavioural effects observed after the drug has been cleared from brain tissue.

Adult male mice (10–20 weeks) were used across experiments. Where relevant, the study employed wild-type strains (129S6/SvEv and C57BL/6) and 5-HT 2A R knockout (5-HT 2A R −/−) mice derived from heterozygote crossings; animals were randomly allocated to treatment groups and housed under standard conditions. DOI (2,5-dimethoxy-4-iodoamphetamine hydrochloride, 2 mg/kg, i.p.) or saline vehicle was administered, with behavioural testing typically conducted 24 h after injection; for molecular time courses frontal cortex tissue was collected at 24 h, 48 h and 7 days post-injection. The investigators state that no formal sample-size calculation was performed; experimenters performing imaging were blind to treatment. Behavioural assays included locomotor activity (open-field monitoring), light–dark box preference, novel-object recognition, prepulse inhibition of startle (PPI), forced-swim test (immobility scored in last 4 min of 6 min session), and contextual fear acquisition, generalisation and extinction using two distinct contexts. For fear paradigms DOI was given either 10 min after the expression stage or 24 h before acquisition depending on the cohort. Dendritic spine structure was assessed by injecting AAV8-CaMKIIa-eYFP into frontal cortex pyramidal neurons, waiting ≥3 weeks for expression, then imaging apical dendritic segments in fixed slices and classifying spines automatically into thin, stubby or mushroom using NeuronStudio and a validated rayburst algorithm. Functional synaptic plasticity was measured by whole-cell patch-clamp recordings of L2/3 pyramidal neurons with an L4→L2/3 pairing protocol to induce long-term potentiation (LTP). Recordings were performed 24 h after DOI or saline, with EPSC amplitudes normalised to baseline. For epigenomic and transcriptomic profiling, neuronal nuclei were isolated from frontal cortex, NeuN+ nuclei were sorted by FACS, and H3K27ac ChIP-seq (MOWChIP-seq) and Smart-seq2 RNA-seq were performed using low-input protocols; six animals per condition were reported for genomic assays. Bioinformatic analyses included peak calling and enhancer prediction from H3K27ac, K-means clustering of differential enhancers across time points (FDR < 0.05), motif and gene-ontology enrichment, differential gene-expression analysis with DESeq2 (FDR < 0.05), weighted gene co-expression network analysis (WGCNA), and overlap testing between H3K27ac peaks and human GWAS loci using liftOver to hg19 and Monte Carlo randomisation. Statistical tests reported in the manuscript included two-way ANOVA, repeated-measures ANOVA and Student's t tests as appropriate.

Behavioural findings: DOI (2 mg/kg, i.p.) produced a reduction in immobility time in the forced-swim test 24 h after administration (t18 = 3.80, p < 0.01; 10 mice per group), with a trend for reduced immobility detectable at 7 days in a smaller cohort (t8 = 2.96, p < 0.05 reported for a 5-per-group comparison). DOI did not alter locomotor activity in a novel environment, preference in the light–dark test, novel-object recognition, or prepulse inhibition measured 24 h after injection. In contextual fear paradigms, DOI accelerated context fear extinction in 5-HT 2A R +/+ mice: DOI-treated wild-type animals showed a faster progressive reduction in freezing during extinction compared with vehicle-treated controls (extinction F[1,96] = 5.51, p < 0.05). This facilitation of extinction was absent in 5-HT 2A R −/− mice, supporting receptor dependence. When DOI was administered 24 h prior to acquisition in a separate cohort of 5-HT 2A R +/+ mice, acquisition and expression were not affected but DOI reduced freezing during fear generalisation and extinction, indicating lingering effects on associative learning beyond the acute presence of the drug. Dendritic spine and electrophysiology results: Structural analysis of frontal cortex pyramidal neurons revealed that a single DOI administration increased the density of immature/transitional spines (thin and stubby categories) 24 h post-dose, while mature mushroom spine density was unchanged. Vehicle-injected 5-HT 2A R −/− mice exhibited lower thin and stubby spine density relative to vehicle wild-type controls, and the DOI-induced spine increases were not observed in knockout mice, indicating that spine remodelling required 5-HT 2A R expression. Functionally, DOI enhanced cortical synaptic potentiation: LTP induced by pairing L4 stimulation with brief postsynaptic depolarisation produced greater magnitude potentiation in DOI-treated mice than in vehicle-treated mice (two-way ANOVA F[1,10] = 9.84, p < 0.05; averaged post-induction magnitude t9 = 3.05, p < 0.05). Reported LTP recordings came from n = 5 DOI-treated neurons (from 5 mice) and n = 6 vehicle neurons (from 5 mice). Epigenomic and transcriptomic results: Cell-type-specific H3K27ac profiling in NeuN+ frontal cortex nuclei revealed widespread, time-dependent changes at enhancer regions after DOI. K-means clustering divided differential enhancers into six clusters: clusters 1–3 showed net increases after DOI, clusters 4–6 decreases. Clusters 1 and 2 peaked at 48 h then declined by 7 days; cluster 3 showed a slower increase peaking at 7 days. Notably, 32.7% of differential enhancers (clusters 3 and 6) remained altered at 7 days, indicating persistent epigenomic remodelling that outlasted the drug. Gene-ontology enrichment linked many clusters to synapse organisation, axonogenesis and cell junction assembly; motif analysis highlighted transcription factors implicated in neuronal growth and plasticity (including Mef2B family members), and NFIL3 was a top motif in the cluster with circadian-process enrichment. By contrast, transcriptomic changes were largely transient. Only 3.7% of differentially expressed genes (DEGs, FDR < 0.05) showed sustained variation at 7 days, and 574 genes (4.2% of total DEGs) displayed expression patterns that correlated with enhancer dynamics (Spearman r > 0.4). Weighted gene co-expression network analysis (WGCNA) identified ten modules; the blue module correlated positively with 48 h post-DOI and negatively with control, containing 1,933 genes with GO terms including axonogenesis and learning/memory and hub genes such as Inpp4a, Nfasc and Cntnap2. The yellow module, positively correlated with control, was enriched for synaptic vesicle recycling and HIF-1 pathway genes. GWAS overlap: The authors tested overlap between H3K27ac peaks and human GWAS loci. Of 12 GWAS datasets examined, three psychiatric sets—schizophrenia, depression and ADHD—showed statistically significant overlap (p < 0.05) with both differential and static peak groups. Two additional sets (white-matter structure and Crohn disease) showed significant overlap only with differential peaks, suggesting DOI-induced epigenomic changes intersect human loci previously associated with some psychiatric and non-psychiatric traits.

De La and colleagues interpret their results as providing evidence that a single systemic dose of the phenethylamine psychedelic DOI produces post-acute increases in immature dendritic spines, enhances cortical LTP and accelerates contextual fear extinction in mice, and that these synaptic and behavioural outcomes require expression of the 5-HT 2A R. They further argue that DOI induces persistent reshaping of the frontal cortex epigenomic landscape—particularly at enhancer regions marked by H3K27ac—that endures for at least 7 days and is larger in magnitude and longer-lasting than transcriptomic changes measured in bulk neuronal nuclei. The authors position these data as complementary to earlier findings of rapid structural plasticity following psychedelic exposure, and suggest a mechanistic schema in which transient transcriptional responses to acute 5-HT 2A R activation may engage downstream chromatin and enhancer changes that sustain synaptic remodelling and lasting behavioural effects. Enrichment of enhancer clusters in synapse- and axonogenesis-related GO terms and the identification of related TF motifs (for example, Mef2B and NFIL3) are cited as molecular correlates tying chromatin remodelling to plasticity and potentially to circadian effects. Limitations acknowledged by the investigators include analysis of the entire neuronal population rather than specific neuronal subtypes, which could obscure cell-type-specific signatures, and uncertainty about extrapolation to other brain regions implicated in addiction or psychosis. They also note unresolved questions about whether subjective hallucinogenic experiences are necessary for clinical benefits in humans and whether non-hallucinogenic 5-HT 2A R agonists (or effects of tryptamine-class psychedelics with different signalling profiles) produce similar plasticity and epigenomic outcomes. Finally, the authors highlight the overlap between DOI-linked epigenomic marks and GWAS loci for schizophrenia as a reason for caution when considering psychedelic use in individuals at risk of psychosis. Overall, they conclude that 5-HT 2A R plays a central role in the persistent synaptic and chromatin changes observed after DOI and that these findings may help to understand the long-lasting effects reported in clinical studies, while emphasising the need for further work to delineate cell-type specificity, broader brain-region effects and safety considerations.