Introduction

Since the 90’s the mind-altering psychedelics have become again the subject of an intriguing stream of research in psychiatric disorders^1,2. Particularly, in times where the development and the approval of new medications is decreasing and the number of patients suffering from psychiatric disorders is rising there is a huge need for new therapeutical interventions^3,4. Compared to classical drugs the broad therapeutical effect of psychedelics is rapid, robust and can be long-lasting even after single administration⁵. In particular, the 5-hydroxytryptamine receptor 2A (5-HT2A-R) targeting psilocybin, the compound of the so-called “magic mushrooms”, is discussed as fast-acting and long-lasting antidepressant in treatment-resistant depression (TRD), anxiety, obsessive-compulsive disorder and addiction^6–11. Not surprisingly psilocybin is stated as “breakthrough therapy” by the United States Food and Drug Administration (FDA) in the treatment of depression since 2019. While we know that 5-HT2A-R activation plays a principal role in serotonergic psychedelic-mediated behavioral and cellular response^12–22 the molecular and cellular changes induced in the brain - on a single cell and network level are barley understood. The recent demonstration of intracellular 5-HT2A-Rs additionally increased the complexity of psychedelic actions²³. The administration of psychedelics may enable brain network resetting¹ by generating a plastic cellular state in which synaptic remodeling and augmentation of neuroplasticity-associated proteins and genes are likely²⁴. Indeed, biological evidence for the “resetting” hypothesis comes from a pioneering study by Olson and colleagues in 2018 which showed that the treatment with serotonergic psychedelics increases the synthesis of synaptic proteins, strengthens synaptic responses, and fosters neurito- and synaptogenesis in rat cortical neurons²⁵. The group therefore introduced the term “psychoplastogen” (greek: psych- (mind), - plast (molded), - gen (producing) for underlining their plasticity-promoting properties. Moreover, serotonergic psychedelics already have been shown to promote the 5-HT2A-R-mediated growth of dendritic spines and modulate neurotransmission^25–27. As molecular and cellular psychedelic research is recently based nearly exclusively on animal studies, the question emerged whether those insights can be translated to the human brain. Psychiatric disorders and psychedelic effects are complex, multisymptomatic and therefore often difficult to study in non-human model organisms. Most importantly, drugs that are efficiently tested in psychiatric animal models might not be necessarily transferable to the human system²⁸. In that context, induced human pluripotent stem cells (IPSC) have emerged as a powerful tool to generate neurons and neuronal circuitries, model brain disorders and identity the molecular mechanisms of drug interventions²⁹.

Here, we explored the molecular, transcriptional, morphometric and functional consequences of the psychoactive 5-HT2A receptor agonist psilocin, the active metabolite of psilocybin in human iPSC-derived cortical neurons. We demonstrate that psilocin leads to a set of molecular, morphological and functional changes that start shortly after administration and manifest in time. These include an increase in brain derived neurotrophic factor (BDNF) expression and an activation of gene expression programs associated with neuromodulation and plasticity resulting in increased neuronal complexity, synaptogenesis and changes in neuronal network function. Thus, our study provides first evidence that the 5TH2A receptor agonist psilocybin activates widespread neuroplastic programs in human neurons.

Results

Psilocin causes 5-HT2A receptor internalization and redistribution in human cortical neurons

As an experimental model to study the effects of psilocin in human neurons and neuronal networks we differentiated human iPSCs (expressing the pluripotency-associated transcription factors 0CT4 and S0X2) into neural progenitor cells which we further differentiated into neurons of mainly dorsal forebrain identity (Fig. 1A and Fig. S1A-B)^30–33. Neural progenitor cells express typical neural stem I progenitor cell markers such as NESTIN or S0X2 as well as the dorsal forebrain-associated transcription factors PAX6 and are negative for F0XA2, as floorplate/ midbrain marker (Fig. S1B). Neurons differentiated from these progenitors for >5 weeks express the neuronal antigen NeuN, the dendritic marker MAP2, and the axonal marker TAU. Most neurons express TBR1 and I or CTIP2, transcription factors typically found in Layer 5/6 neurons of the human cortex (Fig. 1A). Neuronal activity in mature neurons is indicated by expression of the activity-dependent immediate early gene cFOS (Fig. 1A). Astrocytes expressing glial fibrillary acid protein (GFAP) were found occasionally (Fig. 1A). These represent the only contamination in the otherwise purely neuronal culture, which contains no residual progenitors (Fig. S1C). 5-HT2A-R expression was confirmed by PCR (Fig. S1D) and by immunocytochemistry targeting an extracellular epitope which showed presentation of the receptor in the somatodendritic and axonal compartment (Fig. 1B, upper row). 5-HT2A-R agonists such as psilocin are hypothesized to cause immediate (within minutes) internalization of the 5-HT2A-R in vivo and in vitro and are described to cause desensitization by reduced receptor abundance^34,35. We thus analyzed the effects of psilocin on the 5-HT2A-R in human neurons 10 minutes (min) and 24 hours (hrs) following exposure to psilocin. By 10 min, 5-HT2A-R expression at the neuronal surface significantly decreased in the axonal compartment (Tau-positive; from MCtrl = 0.4 ± 0.3; M10min = 0.3 ± 0.3), but not in the somatodendritic compartment (MAP2-positive; measured as 5-HT2A-R particles per μm; Fig. 1B-C, F).

Fig. 1

The decrease in cell surface-located receptor signals was largely prevented by inhibiting clathrin-mediated endocytosis (CME) using dynamin inhibitor dynasore, indicating that this acute down regulation is at least in part mediated by receptor internalization. Dynasore-mediated internalization of 5-HT2A-R resulted in accumulation of receptor-dependent fluorescence signals in roundish aggregates along the axonal compartment (Fig. S1E-F). At 24 hrs, both, receptor densities on axons (from MCtrl = 0.4 ± 0.3; M24hrs = 0.1 ± 0.2) and dendrites (from MCtrl = 0.6 ± 0.4; M24hrs = 0.2 ± 0.2) were significantly decreased (Fig. 1D-E, F), suggesting a subacute downregulation and/or internalization of the receptor also in the somatodendritic compartment. Thus, psilocin treatment induces a two-phasic decrease of cell-surface located 5-HTR2A-R abundance, first in axons and then in the somatodendritic compartment (Fig. 1G).

Psilocin induces BDNF expression and downstream signaling in human neurons

Psychedelics have been shown to efficiently increase levels of neurotrophic factors, such as BDNF^12,25,36,37. We thus addressed BDNF expression in human neurons following exposure to psilocin. We analyzed the effect of different psilocin concentrations (ranging from 10 nM to 10μM) exposed for 10 min and also at 24 hrs following exposure. We observed a dosage-dependent increase in the abundance of BDNF-positive particles (quantified as particles per μm neurite length) which reached significance at a concentration of 10 μM psilocin (Fig. S2A-B). A longer treatment (6 hrs instead of 10 min) with a lower concentration (100 nM instead of 10 μM) of psilocin also induced BDNF abundance significantly, but less strikingly (Fig. S2C-D). We thus continued to use 10 μM psilocin in the following experiments. The clear induction of BDNF particle density was reproduced in four biological independent control cell lines and several differentiation batches for each line (from MCtrl = 1.5 ± 1.2; MPsi = 2.6 ± 1.9) (Fig. 2A-C). When comparing BDNF densities independently in dendrites (MAP2-positive) and axons (Tau-positive), the significant increase in BDNF could be assigned to both neuronal compartments to a comparable extent (Fig. S2E-F). Also, at 48 hrs following the exposure to psilocin, a significant increase of BDNF particles was observed compared to the untreated condition (Fig. S2G). The increase in BDNF abundance was 5-HT2A-R-mediated as treatment of the neurons with the specific 5-HTR2A-R antagonist ketanserin prevented BDNF upregulation by psilocin (Fig. 2D-E) (from MCtrl = 2.5 ± 1.4; MPsi = 4 ± 2.3; MPsi + Ket = 2.1 ±1.4; MKet = 1.4 ± 1.2). Upregulation of BDNF could also be prevented by blocking CME with dynasore or protein kinase C (PKC) with the PKC inhibitor chelerythrine, indicating that CME and PKC activation are critically involved in this process (Fig. 2F-G) (from MCtrl= 3.1 ± 1.5 to MPsi = 4.6 ± 2.4; MPsi + Chel = 3 ± 1.7; MPsi + D = 3.4 ± 1). Upregulation of BDNF was also validated by Western immunoblot where we observed an increase in the levels of pro-BDNF and were able to detect mature BDNF at approximately 14 kDa, which, under baseline conditions, was below the detection level (Fig. S2H). Increased BDNF signaling in psilocin-treated neurons is indicated by an increase in the phosphorylation of the BDNF receptor TrkB (Fig. S2H). Furthermore, as a downstream target of BDNF signaling we observed an increase of the phosphorylation of AKT at Ser 473 1 day (Fig. S2I) and 3 days after psilocin exposure (Fig. 2H-I) which was reversed by ketanserin treatment (from MCtrl = 0.8 ± 0.2; MPsi = 1.3 ± 0.2; MKet = 0.9 ± 0.2).

Fig. 2

Psilocin induces gene expression changes associated with axonal and synaptic plasticity

To analyze the influence of psilocin on global gene expression in cortical neurons, we performed whole transcriptome sequencing one day and three days following a single 10 min administration with 10 μM psilocin in neurons from two independent genetic backgrounds. GO enrichment and KEGG pathway analysis comparing cells one day and three days after psilocin administration with the respective controls revealed an enrichment of significantly affected genes in many ontologies and pathways associated with axonal growth and synaptic remodeling, plasticity and learning, memory and cognition (Fig. 3A, S3A). Most significant changes occurred within the first 24 hrs which is reflected when plotting the significantly changed genes only for selected GO term enriched at both time points (Fig. 3B, S3B). When looking at a more global gene expression level of all genes included in the respective GOs, gene expression of GOs associated with axonal outgrowth show a generally higher expression at day 1 which is further increased at day 3. In contrast genes associated to synaptic organization, plasticity and learning, memory and cognition show a trend towards decreased expression at day 1 and a generally stronger increase at day 3 (Fig. 3C, S3C). The overlapping assignment of significant differentially expressed genes to the aforementioned GOs underlines their importance in long-term plasticity (e.g., CDK5, CAMK), synapse formation (e.g., SYN1) and axonal and neurite growth (e.g., GAP43) and neuronal structure marker (e.g., MAPT) (Fig. 3D). The latter constent with increasing staining intensity for TAU (Fig. 2). We further found a strong significant upregulation of immediate early genes (lEGs) (e.g., FOSL2, JUND, EGR1, ARC, FOSB) and of the excitatory glutamatergic AMPA/NMDA receptor genes (GRIA/GRIN) after psilocin administration (Fig. 3E). One day but not three days of psilocin treatment more stongly upregulated the AMPA genes GRIA1,2,3 than the NMDA genes GRIN1,2B,2C (Fig. 3E), supporting modulation of glutamatergic excitatory signalling genes upon psilocin treatment. These effects could be reversed upon co-treatment with antagonist ketanserin, indicating that most changes are 5-HT2A receptor mediated (Fig. 3F).

Fig. 3

Psilocin increases neunte complexity

The gene expression analysis indicated an induction of morphometric changes by psilocin in particular on neurites. Therefore, we assessed neuronal complexity performing Sholl analysis in neurons differentiated for >40 days at 24 hrs and 48 hrs following psilocin exposure. To identify the dendritic arbor of single mature neurons in these cultures, the cultures were transduced with adeno-associated vectors (AAVs) coding for mCherry under control of the CaMKIla promoter (Fig. 4A-D).

Fig. 4

We observed an increase of primary neurites at 25 μm (significant at 48 hrs and a trend at 24 hrs from MCtrl = 4.1 ± 2.1; M24hrs = 5.4 ± 3.7; M48hrs = 5.5 ± 2.4) and of intersections at 50 μm (significantly increased at 24 and 48 hrs post treatment from MCtrl = 2.7 ± 1.4; M24hrs = 4 ± 2.3 to M48hrs = 3.9 ± 1.8; Fig. 4B). As a result, the calculated total neurite length was significantly increased at 24 and 48 hrs post treatment from MCtrl = 272.5 ± 142.7; M24hrs = 383.2 ± 209.9; M48hrs = 352.8 ± 162.5 (Fig. 4C). And we observed a significant increase in the total number of intersections in psilocin-exposed neurons quantified in steps of 25 μm up to a distance of 125 μm (MCtrl = 10.9 ± 5.7; M24hrs = 15.3 ± 8.4, M48hrs = 14.1 ± 6.5; Fig. 4D). These data indicate, that gene expression changes elicited by psilocin resulted in neurite outgrowth and an increase of global dendritic complexity as early as 24h following a single 10 min exposure.

Psilocin increases synaptic strength and synaptogenesis

To address changes on neuronal function, we exposed neuronal cultures 10 min or 24 h to 10 μM psilocin and performed whole-cell patch-clamp experiments one week later. Following the 10 min exposure, we observed a increasing trend in the number of evoked action potentials (APs), having amplitudes of around 100 mV (Fig. 5A). Extended exposure of the cultures to psilocin (24h) increased this effect reaching significance (total number of APs from MCtrl = 20.1 ± 12; MPsi 10min; day 7 = 28.9 ± 16.9; MPsi 24hrs; day 7 = 30 ± 14.5; AP amplitude from MCtrl = 101.1 ± 13.3; MPsi 10min; day 7 = 101.1 ± 13.2; MPsi 24hrs; day 7 = 101.3 ± 16.7). To find out whether the increase in AP firing enhanced synaptic network activity, we recorded spontaneous AP-dependent excitatory postsynaptic currents (sEPSCs). Indeed, the frequency of sEPSCs increased in both psilocin-treated conditions to some extent (MCtrl = 0.6 ± 0.7; MPsi 10min; day 7 = 1.0 ± 1.3; MPsi 24hrs; day 7 = 1.1 ±1.2; Fig. 5B). And, we observed a slight increase in sEPSC amplitude (MCtrl = 14.2 ± 4.7; MPsi 10min; day 7 = 16.7 ± 5.1; MPsi 24hrs; day 7 = 15.5 ± 5.3; Fig. 5B), consistent with the observed GRIA upregulation (Fig. 3E). The latter amplitude increase was next examined on AP-independent miniature EPSCs (mEPSCs) which are generally caused by the spontaneous release of single vesicles (Figs. 5C, S4A). The amplitudes of mEPSCs increased 7 days after the start of 24 hrs psilocin administration consistently for two cell lines (MCtrl = 11.1 ± 2.2; MPsi 24hrs; day 7 = 12.8 ± 4.2, Fig. 5C; MCtrl = 9.3 ± 2.8; MPsi 24hrs; day 7 = 11.2 ± 3.7, Fig. S4A), whereas the frequencies of the mEPSCs were rather stable (MCtrl = 0.8 ± 0.5; MPsi 24hrs; day 7 = 0.9 ± 0.6, Fig. 5C; MCtrl = 0.9 ± 0.7; MPsi 24hrs; day 7 = 0.7 ± 0.5, Fig. S4A). When examining mEPSCs from separately differentiated neurons exposed to psilocin for 24 hrs and analyzed at day 11, the increase in mEPSC amplitude was somewhat attenuated compared with 7 days (MCtrl = 11.7 ± 4.3; MPsi 24hrs; day 11 = 12.4 ± 3.6; frequency from MCtrl = 0.8 ± 0.8; MPsi 24hrs; day 11 =0.5 ± 0.3, Fig. 5C). Repeated LSD administration altered gene and protein expression related to neuroplasticity signaling in the mouse prefrontal cortex and increased dendritic spine density^38,39. We therefore hypothesized that extended psilocin administration fosters synaptic network activity and synaptogenesis and thus examined effects of a 96 hrs exposure. In line with the above-mentioned hypothesis, the amplitudes of the sEPSCs increased significantly in two lines (MCtrl = 22.5 ± 6.5; MPsi 96hrs; day 10 = 30.7 ± 4.8, Fig. 5D; MCtrl = 15.9 ± 4.5 ; MPsi 96hrs; day 10 = 27.6 ± 10, Fig. S4B), and the frequencies showed a trend to increase similar to Fig. 5B, reflecting enhanced network activity (MCtrl = 1.0 ± 1.0; MPsi 96hrs; day 10 = 1.5 ± 1.0, Fig. 5E; MCtrl = 1.5 ± 1.7; MPsi 96hrs; day 10 = 1.0 ± 0.7, Fig. S4B).

Fig. 5

This enhanced synaptic network activity should affect the abundance of either the presynaptic marker synapsin and/or the postsynaptic marker PSD-95 and may even affect their co-localization at day 4 and at day 10 following 4 day permanent 10 μM psilocin administration. In line with above electrophysiological results for the 10 day psilocin exposure (Fig. 5D) and the gene expression data, which indicated changes associated with synaptogenesis and synaptic plasticity (Fig. 3) we observed for the pre-synaptic marker synapsin, a trend towards an increase of the density (particles per neurite length) (from MCtrl = 0.25 ± 0.2; M4days = 0.28 ± 0.2; MIOdays = 0.29 ± 0.2) and intensity (mean fluorescence intensity of the particles) (from MCtrl = 21 ± 13.3; M4days = 24.1± 15; MIOdays = 19.8 ±10.9) at both time points (Fig. 5E-H). In line with the gene expression data (Fig. 3A, postsynaptic density), we observed a significant increase in the density of the postsynaptic marker PSD-95 (from MCtrl = 1.1 ± 0.5; M4days = 1.4 ± 0.6; MIOdays = 1.3 ± 0.5) and in its intensity (from MCtrl = 31.8 ± 12,9; M4days = 36.8 ± 15.6; MIOdays = 38 ± 16.4) for both time points (Fig. 5I-J). These effects were accompanied by a significant increase in the co-localization of both markers (from MCtrl = 0.2 ± 0.2; M4days = 0.3 ± 0.2; MIOdays = 0.3 ± 0.2; Fig. 5K). Together, these experiments indicate that psilocin augments synaptic strength primarily via an increase in the postsynaptic receptor density, which lasts at least for 6 days from the start of psilocin withdrawal and that extended exposure of neurons to psilocin pronounces the effects on synaptic strength.

Discussion

The neuroplasticity-promoting psychoplastogen²⁵ psilocybin is currently developed as a new medication in the treatment of psychiatric disorders^1,2,40. “Brain network resetting”¹ by psilocybin , i.e. the restoration of neuronal and synaptic dysfunction associated with the pathophysiology of mental disorders^41–45, could be achieved by stimulating signaling pathways associated with neuroplasticity and BDNF signaling^24,41. So far, experimental evidence for this hypothesis is based on data from non-human model organisms. In our experimental setting, human neurons exposed to psilocin present with a biphasic downregulation of cell surface-located 5-HT2A-Rs, which starts in the axonal compartment and expands to the dendritic compartment. Stimulation of 5-HT2A-Rs can lead to different adaptive processes including internalization, downregulation and recycling⁴². Interestingly, pychedelics have been already shown to reduce 5-HT2A-R density in animal models^35,43,44 We see a decrease in receptor surface density, prevented by inhibiting the GTPase dynamin arguing that clathrin-mediated endocytosis is involved in this process, which suggests internalization of cell surface-located 5-HT2A-R. Given the potential of psychedelics to diffuse across cell membranes and to activate intracellular pathways²³, enhanced internalization of 5-HT2A-R may result in stronger intracellular signaling processes caused by the combination of retrograde trafficking and the activation of internally residing receptors adding both to serotonin-independent and psychedelic-dependent signaling strength⁴⁵.

We also observed a pronounced psilocin-induced BDNF upregulation in human cortical neurons and show that PKC activation and endocytosis, two mechanisms contributing to 5-HT2A-R internalization^46,47, are involved as chelerythrine and dynasore inhibited BDNF augmentation. BDNF plays an important role in neurogenesis, synaptogenesis, and the formation of synaptic interactions. Particularly the mature BDNF/phosphorylated-TrkB receptor complex is involved in multiple pathways linked to prosurvival and synaptic plasticity⁴⁸. Recently it was shown that psychedelics also directly bind to the BDNF TrKB receptor, thereby affecting TrkB dimerization and facilitating the effect of endogenous BDNF, which underlines the importance of the BDNF system for the action of psychedelics⁷². In keeping with this, we show that psilocin leads to a fast and enduring upregulation of proteins and genes linked to neuronal complexity, synaptogenesis and synaptic transmission, as also demonstrated in psychedelic-mediated plasticity in animal studies^25,26,49. Of note, BDNF upregulation and changes in gene expression where reversed by ketanserin confirming that these effects are medated by 5-HT2A-Rs.

In our model, augmentation of neuronal complexity is an outcome which can be detected as early as 24 hrs after psilocin administration. In rodents, serotonergic psychedelics also foster synaptogenesis and spinogenesis^25,26,49. In line with this we show that psilocin also promotes synaptogenesis, measured by the increase in PSD-95, synapsin and their colocalization. As a result, we observe changes in intrinsic neuronal properties and network function. More specifically, we see an increased excitability and an increase in postsynaptic current frequency but in particular amplitude. The higher number of action potentials generated by current injections could be due to increased dendritic excitability as reviewed by Kwan and colleagues⁴¹. The increase in synaptic strength lasted at least 6 days and is in agreement with frequency and amplitude increases of miniature and spontaneous synaptic currents observed in acute brain slices of mice after administering psilocybin or DMT in v/vo^25,26. These drugs may increase extracellular glutamate levels, similar to ketamine, LSD and DOI administration^24,50–53. To close the circle, 5-HT2A-R stimulation, subsequent glutamate release and AMPA receptor activation⁴⁰ activates the BDNF-associated TrkB and mTORCI pathway which promotes BDNF expression itself^54,55. Of note, PSD-95, the postsynaptic marker that we found to be upregulated, controls activity-dependent AMPA receptor incorporation in the postsynapse^56,57 which can modify the strength of excitatory synaptic transmission⁵⁸.

Finally, an enrichment of differentially expressed genes associated with GO terms of learning, memory and cognition suggest behavioral long-term effects of the drug which are based on the above mentioned structural and functional modifications. The effects were dependent on the duration of drug exposure suggesting that repetitive administration of psilocybin might elicit beneficial effect with respect to brain plasticity.

Together, our work confirms the postulation by the Olson group that psychedelics may act through an evolutionary conserved mechanism as we can replicate the results from animal studies in our human system²⁵. Moreover, our study highlights the importance of a human model system for understanding the mode of action of psychedelics and potentially for disease modelling.

Supplementary information

Materials and Methods

Human material and generation of fibroblast-derived human iPSCs

The study was approved by the local ethics committee. All experiments with human material were in accordance with the Declaration of Helsinki. All healthy participants gave written informed consent. Reprogramming was done using CytoTune®-iPS 2.1 Sendai Reprogramming Kit (Thermo Fisher Scientific). Cells were cultured at 37°C, ambient O2 and 5% CO2 concentration under sterile conditions in an incubator (Binder). Applications with the cells were carried out in a sterile flow hood (Scanlaf Mars) by using sterile and autoclaved instruments and medium. Chromosomal alterations were excluded by genome-wide single nucleotide polymorphism (SNP) analysis and mycoplasma testing was performed on a regular basis. iPSCs were screened for their stem cell properties by the capability of in vitro differentiation into the three embryonic germ layers, ectoderm, mesoderm and endoderm and for pluripotent stem cells markers in vitro.

iPS cell culture

Fibroblast-derived ¡PS cells were kept as colonies under feeder free-conditions in stem cell state on 5% (v/v) Geltrex™-coated plates (Thermo Fisher Scientific) containing 1% (v/v) Pen/Strep (Thermo Fisher Scientific) in Essential 8 medium (DMEM/F12 with L-glutamine and HEPES (Thermo Fisher Scientific) supplemented with 1% (v/v) Pen/Strep (v/v), 64 μg/ml LAAP (Sigma-Aldrich), 14 ng/ml sodium selenite (Sigma-Aldrich), 200 ng/ml FGF-2 (154) (Cell Guidance Systems), 2 ng/ml TGF-β1 (Cell Guidance Systems), 20 μg/ml insulin (Sigma-Aldrich), 11 μg/ml transferrin (Sigma-Aldrich) that was changed daily. For passaging confluent iPS colonies were washed twice with phosphate buffered saline (DPBS) and then incubated with 0.5 mM EDTA (Thermo Fisher Scientific) for 5 – 10 min at RT until the colonies started to detach. After aspirating the EDTA solution the fractured colonies were gently resuspended in E8 medium containing 5 μM Rho-associated protein kinase (ROCK) inhibitor (Cell Guidance Systems).

Differentiation of iPS cells into neuronal progenitors

For in vitro differentiation into neural progenitors E8 medium was replaced by neural progenitor induction medium phase 1 (advanced DMEM/F-12 medium with glutamine (Thermo Fisher Scientific), with 1% (v/v) Pen/Strep (v/v), 2 mM GlutaMAX™ (Thermo Fisher Scientific), 1% (v/v) B-27 supplement with RA (Thermo Fisher Scientific), 10 μM SB-431542 (Cell Guidance Systems), 1 μM LDN-193189 (Stemcell Technologies), 2 μM XAV939 (Cell Guidance Systems) and 5 μM cyclopamine (Cayman Chemical) when ¡PS cells reached 70-80% confluency. At day 4, cells were dissociated 1:2 to a monolayer, continue using aforementioned medium. Upon day 8 cells were dissociated 1:2 and medium was changed to induction medium phase 2 (advanced DMEM/F-12 medium with glutamine with 1% (v/v) Pen/Strep, 2 mM GlutaMAX™, 1% (v/v) B-27 supplement with RA, 200 nM LDN-193189) for 8 more days. Finally, medium was changed to induction medium phase 3 (advanced DMEM/F-12 medium with glutamine with 1% (v/v) Pen/Strep, 2 mM GlutaMAX™, 1% (v/v) B-27 supplement with RA, 20 ng/ml FGF-2 (147) (Cell Guidance Systems). Dissociation into single cells was done using TrypLE™ (Thermo Fisher Scientific) for 5 -15 min at 37 °C until the cells start to detach. Cells were then transferred with wash medium to a 15 ml tube and centrifuged for 4 min at 1000 g. The pellet was gently resuspended in corresponding neural progenitor induction medium supplemented with 5 μM ROCK inhibitor.

In vitro differentiation of neural progenitors into human cortical neurons

In vitro differentiation into human cortical neurons was launched by the change to neural differentiation medium phase 1, defined as day 0 of differentiation. Neural differentiation phase 1 medium consisted of neural base medium 1 (DMEM/F-12 medium with glutamine, 0.1% (v/v) GlutaMAX™, 1.8 mM CaCb (Sigma-Aldrich), 1% (v/v) Pen/Strep, 1% (v/v) B-27, 0.5% (v/v) N2 supplement (prepared in the laboratory as follows: DMEM/F-12 with glutamine, 500 μg/ml insulin, 10 mg/ml transferrin, 520 ng/ml sodium selenite, 1.611 mg/ml putrescine (Sigma-Aldrich) and 630 ng/ml progesterone (Sigma-Aldrich)), 1% (v/v) NEAA (Thermo Fisher Scientific), 1.6 mg/ml glucose (Carl Roth)) supplemented with 200 μM ascorbic acid (Sigma-Aldrich), 1 μM LM22A (Sigma-Aldrich), 1 μM LM22B (Tocris Bioscience), 2 μM PD-0332991 (Selleckchem), 5 μM DAPT (Cell Guidance Systems) and was applied until day 3 of differentiation. The cells were split using TrypLE™ into phase 2 medium on polyethylenimine (Sigma Aldrichyiaminin (Sigma Aldrich) coated plates (1:2000 of 1 % (PEI) in 25 mM boric acid (pH 8.4) (Thermo Fisher Scientific); 3.75 μg/ml laminin (Sigma Aldrich) in DPBS (Thermo Fisher Scientific). Neural differentiation medium phase 2-4 media consisted of neural base medium 2 with Neurobasal™ (Thermo Fisher Scientific) (plus 1.6 mg/ml glucose, 1% (v/v) Pen/Strep, 0.1% (v/v) GlutaMAX™ and 1% (v/v) B-27) and was supplemented additionally to supplements of phase 1 medium with 3 μM CHIR99021 (Cell Guidance Systems), 10 μM forskolin (Cell Guidance Systems) and 300 μM GABA (Sigma-Aldrich). At day 10 of differentiation medium was changed to medium phase 3 containing, additionally to supplements of phase 1 medium, 3 μM CHIR99021. From day 17 on, CHIR99021 was removed in phase 4 and the medium contained 200 μM ascorbic acid, 1 μM LM22A, 1 μM LM22B and 2 μM PD-0332991. At day 24 medium was changed to neural differentiation phase 5 medium in neural base medium 3 (advanced MEM (Thermo Fisher Scientific), 1.6 mg/ml glucose, 1% (v/v) Pen/Strep, 0.1% (v/v) GlutaMAX, 1% (v/v) B-27) and 0.27 nM Bryostatin 1 (Merck Millipore) for the next 14 days. Medium was changed every 3-4 days half.

Co-culture of human cortical neurons with mouse astrocytes

For electrophysiological experiments mouse astrocytes were plated in a density of around 80,000 cells/well on PEI/laminin coated coverslips on 24-well plates. Mouse astrocytes were cultured in astrocyte standard medium, consisting of astrocyte base medium (1% (v/v) Pen/Strep, 0.5% (v/v) N2, 1.6 mg/ml glucose) supplemented with 0.5% (v/v) N2, 1.6 mg/ml glucose, 0.1% B-27, 100 ng/pl epidermal growth factor (Cell Guidance Systems), 10 ng/ml FGF-2 (147) until they reached 80 – 90% density. Medium was changed every other day. 3 days before phase 1 neurons were split in a density of 200,000 cells onto the astrocytes the medium was changed to astrocyte differentiation medium, consisting of astrocyte base medium and 0.5% (v/v) N2,1.6 mg/ml glucose and 10 ng/pl BMP-4 (Miltenyi Biotec). Medium was changed daily. Since the moment the cortical neurons were plated on the astrocytes the medium was changed to neural differentiation medium and differentiated was continued as described.

Neuroplasticity and 5-HT2A-R experiments in cortical neurons

For immunostainings around 200,000 cells/well (50,000 for Sholl analysis) were plated onto coverslips in 24-well plates. For WB, DNA- and RNA-based experiments around 2-6 million cells/well were plated on 6 well plate dishes or 10 cm dishes. Treatment of cortical neurons started at day 42 of differentiation. 3 days prior to experiments, for drug dissolving and during course of the experiment culture medium was replaced with neural base medium 3 containing 1% (v/v) Pen/Strep, 1.6 mg/ml glucose, 0,1% (v/v) GlutMAX™, 0,1% (v/v) B-27 and 200 μM ascorbic acid to avoid distorting effects of small molecules on cellular signal transduction pathways. For 7 days long-term treatment cells medium changes twice a week were performed with neural differentiation phase 5 medium.

Treatment of cortical neurons with psilocin

10 mM DMSO stock solution of light sensitive psilocin was diluted 1:1000 (final DMSO concentration = 0,001%) to a final concentration of 10 μM psilocin (THC-Pharm). For the short trigger mature cortical neurons were treated for 10 min with psilocin, washed once with advanced MEM containing 1% (v/v) Pen/Strep and were further cultured and then fixed or harvested after indicated time points post treatment. For the permanent treatment condition cells were treated for 24 (96) hrs permanently with 10 μM psilocin. For the 96 hrs plus 6 days washout condition cells were treated for 96 hrs permanently with 10 μM psilocin (fresh psilocin was replaced once), washed once and were further cultured in neural differentiation phase 5 medium. For analyzing psilocin induced effects on 5-HT2A-R presentation cortical neurons were treated for 10 min with 10 μM psilocin, fixed immediately (acute stimulation condition) or respectively after 24 hrs.

Treatment of cortical neurons with ketanserin

To determine whether psilocin induced neuroplasticity effects were 5-HT2A-R dependent 75 mM DMSO stock solution of ketanserin (ApexBio) was diluted 1:1000 to a final concentration of 75 μM ketanserin (final DMSO concentration = 0,001%). Cells were pretreated with 75 μM ketanserin for 10 min to block the 5-HT2A-R, then treated for another 10 min with psilocin and ketanserin (final psilocin concentration = 10 μM, final ketanserin concentration = 75 μM, final DMSO concentration = 0,002%), washed once with advanced MEM containing 1% (v/v) Pen/Strep and further cultured for 24 hrs post treatment.

Treatment of cortical neurons with dynasore

To determine whether psilocin induced neuroplasticity effects were endocytosis-mediated 100 mM DMSO stock solution of dynasore (Cayman Chemical) was diluted 1:2000 to a final concentration of 50 μM dynasore (final DMSO concentration = 0,0005%). Cells were pretreated with dynasore for 50 min to block endocytosis, then treated for another 10 min with psilocin (final psilocin concentration = 10 μM) and dynasore (final dynasore concentration = 50 μM; final DMSO concentration = 0,0015%), washed once with advanced MEM containing 1% (v/v) Pen/Strep and further cultured and then fixed after 24 h post treatment.

Immunocytochemistry

After a first washing with PBS, cells were fixed in 4% PFA (Sigma-Aldrich), washed three times with PBS and blocked and/or permeabilized in blocking solution containing 10% FBS (Thermo Fisher Scientific) in PBS. Blocking was performed for one h. Dilution of primary and secondary antibodies were performed in blocking solution. Please refer to Table S1 for blocking/ permeabilization solution and dilution of primary antibodies. Primary antibodies were incubated over night at 4°C. Samples were washed three times with corresponding blocking solution. Secondary antibodies, conjugated to Alexa Fluor 488, 568 or 647 (Thermo Fisher Scientific) were diluted 1:1000 and were applied for 1 h at room temperature (Table S2). Samples were then washed once in PBS to remove unbound antibodies. Counterstaining of cell nuclei was carried out by using 300 nM DAPI (Biolegend, Table S3), incubated for 5 min at room temperature and washed three times in PBS and once with ddhteO. Slides were mounted with mounting solution (100 mM Tris-HCI (pH 8.5), 25% glycerol, 10% mowiol (Carl Roth), 0.6% DABCO® (Carl Roth) on glass coverslip and air-dried overnight. Cells were observed with the Inverted Leica DMIL LED Microscope. Stem cell and neural progenitor properties were imaged with the Leica DM6 B microscope with Thunder imaging software. For neuroplasticity experiments Leica confocal TCS SP5 II microscope was used. Brightfield images were made with Fluorescence Microscope Celldiscoverer 7 microscope from Zeiss. For analysis of immunofluorescence a polygon selection of ROIs was chosen. The protein density and co-localization of particles was measured with the ComDet v.0.5.3 plug-in for Imaged (National Institutes of Health (NIH), open source). A particle size (depending on experimenters’ evaluation), constant for overall experiments (5-HTR2A: 1.0; BDNF: 1.0; PSD-95: 3.0; Synapsin: 3.0) and an intensity threshold (in SD), adjusted within each experiment, was defined. For co-localization analysis the channels of proteins that have to be co-localized were merged in advance. As maximal distance between co-localized spots 4.00 pixels were stated. The number of particles was divided by the total length of the neurite in μm measured with the straight-line tool spanning the polygonal ROL Multiple ROIs were set per image. N = 1 corresponds to one neurite segment detected by one ROL

Analysis of neuronal complexity

For Sholl analysis around 50,000 cells/well were cultured on 24 well PEI/laminin coated coverslips. Cells were transduced 2 weeks before psilocin treatment with adeno-associated virus AAV CamKIla p-

hCHR2(134a)-mCherry (friendly provided by AG Grinevich, Central Institute of Mental Health, Mannheim, Germany), when mCherry expression was visible by microscope. Cells were fixed after indicated treatment timepoints. Images were analyzed using the Sholl analysis plug-in of Imaged (circle radii). Start radius: 25 μm, step size: 25 μm, end radius: 125 μm, set center from active ROI, preview). Crossings were calculated manually. Based on that, total length and sum of intersections were calculated. N = 1 corresponds to one neuron.

Table S1

Table S2

Table S3

Protein isolation and measurement of protein concentration

Cells (1 well of 6 well plate) were washed with ice cold DPBS, harvested by using a cell lifter and were transferred into a 1.5 ml tube on ice. In the following the cells were centrifuged for 5 min at 5000 g at 4 °C. The supernatant was removed and the sample was stored at − 20°C. For protein isolation, cells were resuspended in 150 μl of protein lysis buffer (50 mM Tris-HCL (pH 7.4), 150 mM NaCI, 25 mM EDTA, 1% (v/v) SDS with one protease and one phosphatase inhibitor mini tablet (Thermo Fisher Scientific) per 10 ml of protein solution) and incubated for 10 min at room temperature and then incubated 1 h on ice. To shear gDNA and reduce viscosity samples were sonicated (20% duty cycles, 50% output, five pulses) from a Branson Ultrasonics™ sonifier 250 (Thermo Fisher Scientific). Samples were centrifuged for 5 min for 5000 g at 4°C. Supernatant was transferred to a new tube. Protein concentration was measured with the BCA protein assay (Thermo Fisher Scientific) following the manufacturer’s instructions. Samples were diluted 1:5 in ddHzO and the absorption at 562 nm was measured in a PowerWave™ XS microplate reader (BioTek). The protein concentration of samples was calculated based on the included BSA standard dilution series.

SDS-polyacrylamide gel electrophoresis and western immunoblotting

15 – 30 μg of protein was mixed with 6x denaturing protein sample buffer (93.75 mM Tris-HCL (pH 6.8), 6% SDS, 6% Glycerol, 9% 2-Mercaptoethanol, 0.25% Bromophenol blue) and boiled for 5 min at 95 °C. For the separation of proteins an SDS-PAGE (Bio-Rad Laboratories) was performed in a Mini-PROTEAN 2-D electrophoresis cell chamber and with the PowerPac™ Basic Power Supply. First, the polyacrylamide gel (Gel buffer for polyacrylamide gels, SDS-PAGE: 3M Tris-HCI (pH 8.5), 0.3% (w/v) SDS; SDS-Polyacrylamide, stacking gel: 24.8% (v/v) Gel buffer, SDS-PAGE, 3.84% (v/v) Bis/ acrylamide, 0.00672% (w/v) APS, 0.224% TEMED; SDS-Polyacrylamide, separating gel: 33.3% (v/v) Gel buffer, SDS-PAGE, 10% (v/v) Bis/ acrylamide, 10% (v/v) glycerol, 0.028% (w/v) APS, 0.09% (v/v) TEMED) runs for about 30 min with 30 V for stacking of proteins and then for about 1.5 – 2 hrs at 110 V. A semi-dry blotting was performed using the Trans-Blot Turbo™ transfer system (Bio-Rad Laboratories) performed with 20 V and 1 A for 30 – 60 min. A reference protein marker (PS10 plus, GeneON Bioscience) was included to determine the size of the protein bands. For blotting 0.45 μm pore size nitrocellulose blotting membrane (Sigma-Aldrich) was used. The membrane was surrounded from both sides by 6 layers of WB filter tissue (VWR). Filter tissue and membrane were wetted in transfer buffer (10% (v/v) Tris-glycine buffer, 20% (v/v) methanol, 0.08% (v/v) SDS). Membranes were blocked for 60 min in 5% BSA (Sigma-Aldrich) in TBS-T (10% (v/v) of TBS with 248 mM Tris-HCI (pH 7.4), 1.37 M NaCI, 26.8 mM KCI plus 0.1% (v/v) Tween®20 (Sigma-Aldrich)) in a 50 ml tube. Primary antibodies (Table S4) were diluted in 5% BSA in TBS-T and were incubated over night at 4°C or for 1 h at room temperature on a rolling mixer. The membrane was washed three times in TBS-T for 10 min at room temperature on the next day. 1:15,000 infrared DyLight™ IR-conjugated secondary antibodies (Cell Signaling Technologies, Table S5) in TBS-T were applied for 1 h at room temperature. Washing in TBS-T was carried out three times and once in TBS for 10 min at room temperature. Visualisation of tagged proteins was carried out with Odyssey IR WB imaging system (Li-COR). Signals were normalized by 1: 20,000 P-actin levels. Antibodies for Western blotting are listed in Supplementary Table S2. For quantification region of interests (ROIs) were set around lanes and intensity was measured with the densitometry analysis in Imaged.

Table S4

Table S5

Ribonucleic acid isolation

For Ribonucleic acid (RNA) isolation cells were (2 wells of Swell plate) resuspended in 500 μl peqGold Trifast™ (VWR) and incubated for 10 min at room temperature. 100 μl of chloroform (Sigma-Aldrich) was added to the lysate and incubated for 10 min at room temperature. Tubes were then centrifuged for 5 min at 12,000 g at 4°C. The upper clear, aqueous nucleic acid phase was transferred into a new tube. 250 μl of isopropanol (Th. Geyer) was added to each sample. For RNA precipitation tubes were kept overnight at − 20°C. Then, tubes were centrifuged for 15 min at 4°C of 12,000 g and supernatant was discarded. The pellet was washed twice with 75% ethanol (Th. Geyer) in DEPC water (Carl Roth) and centrifuged for 10 min at 4 °C at 12,000 g. After the last washing step all ethanol had to evaporate by first discarding the ethanol and afterwards letting air-dry the pellet for about 30 min. The pellet was resuspended in 20 μl DEPC-treated H2O and shook at 400 rpm at 37 °C. To prevent DNA contamination the DNAase I Amplification grade kit (Sigma-Aldrich) was used.

Synthesis of complementary DNA

Complementary DNA (cDNA) was synthesized via the ¡Script™ cDNA synthesis kit (Bio-Rad Laboratories) according to manufacturer’s instructions. 500 ng – 1 μg cDNA was used to reversely transcribe RNA into cDNA. Standard cycling program for cDNA synthesis: 25 °C for 5 min, 46 °C for 20 min, 95°C for 1 min. The resulting cDNA was diluted 1:5 in nuclease-free H2O. (1x) Taq reaction buffer (Biozym), each 200 μM 10 mM dNTP mix (Steinbrenner Laborsysteme), each 400 nM 10 μM primer mix, 0.625 U Taq DNA polymerase (5 U/μl) (Biozym), 0.4 – 10 ng Template DNA). Standard RT-PCR cycling program: Samples were denatured by heating at 95 °C for 1 min followed by 35 cycles of amplification and quantification (95°C for 15 s and 60 °C for 15 s and 72°C for 10 s) and by a final extension cycle (72 °C, 5 min). Primers for real-time (RT)- polymerase chain reaction (PCR) are listed in Supplementary Table S6.

Table S6

For expression control total human adult (BioCat) and fetal brain (Tebu Bio) was included.

Agarose gel electrophoresis

Samples were mixed with 10 x DNA sample buffer (50 mM Tris-HCI, pH 7.6, 0.25% (w/v) Bromphenol blue, 60% Glycerol). DNA fragments were separated in a 1 – 2% (w/v) agarose gel in 1x TAE buffer (40 mM Tris, 20 mM acetic acid, 1 mM EDTA), containing 1: 15,000 peqGreen (VWR) for staining of DNA. A reference DNA marker (100 bp or 1 kbp, New England Biolabs) was included to determine the size of the amplicons. Gels run for 50 min at 100 V and were imaged on a GeneFlash imaging system (Syngene).

RNA Bulk Sequencing and Analysis

For each sample, 30 μl of RNA solution (60 ng/μl) in ddhW was sent to the High Throughput Sequencing (seq) Unit of the Genomics & Proteomics Core Facility at the DKFZ (Heidelberg, Germany) to be processed for RNA Bulk Sequencing. Samples were run through in-house quality control and only samples with an RNA integrity number (RIN) ≥ 5.0 were used for cDNA library preparation according to the TruSeq Stranded protocol (Illumina). Libraries were sequenced to 50 bp paired reads on the Illumina NovaSeq 6K platform to an average of 40 M reads per sample. The DKFZ Omics IT and Data Management Core Facility performed the RNAseq processing workflow. Total counts per feature were imported to R⁵⁹ (v. 4.0.3-4.2.2) and analysed using DESeq2⁶⁰ (v. 1.30.1-1.38.3). All features without any counts were removed, for differential testing with DESeq2 the formula “~Batch+Cell_line+Condition_simple” was used. Differentially expressed genes (padj<=0.05) were used to perform gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis with the clusterprofiler v 4.8.3 package⁶¹. The organism database used for this analysis was org.Hs.eg.db⁶² (v. 3.17.0). GO Chord Plots were created using GOplot (v. 1.0.2)⁶³.

Zscores in GO plots showing up/downregulated genes were calculated with the following formula: zscore=(up-down)/sqrt(count). Heatmaps for RNAseq expression data show z-scaled DESeq2 normalized counts. Biological batch replicates of each samples ensured intra-cell lines stability of results. The R code used to analyze RNAseq data is available at https://qithub.com/ahoffrichter/Schmidt_et_al_2024.

Electrophysiology

Whole-cell patch-clamp recordings from iPSC-derived cortical neurons on PEI/laminin-coated coverslips cultured on mouse astrocyte cells were made using an EPC9 amplifier and PatchMaster (HEKA Elektronik GmbH). The neurons were identified with a Zeiss Axioskop using infrared differential interference contrast video microscopy in the recording chamber. Neurons were perfused at 2 ml/min (peristaltic pump; Ismatec GmbH) with carbogen (95% O2; 5% CO2)-saturated artificial cerebrospinal fluid (ACSF) containing 125 mM NaCI, 1 mM MgCb, 2 mM CaCb, 2.5 mM KOI, 10 mM D-glucose, 25 mM NaHCOs, 1.25 mM NaH2PO4 (pH 7.3, osmolarity 300 mOsm). Pipettes were pulled from borosilicate glass capillaries (GB150F-8P, 1.5 mm o.d., 0.86 i.d.; Science Products GmbH) with a P-97 micropipette puller (Sutter Instrument). Pipettes had resistances between 4-6 MD when filled with intracellular solution containing 115 mM K-gluconate (or cesium-gluconate), 20 mM KOI, 10 mM Na-phosphocreatine, 4 mM Mg-ATP, 0.3 mM GTP, 0.2 mM ethylene glycol tetraacetic acid (EGTA) and 10 mM HEPES (pH 7.3, osmolarity 300 mOsm). In voltage-clamp, input resistance was determined at -70 mV based on currents evoked by small voltage steps (-3 mV; 300 ms). In current-clamp, the resting membrane potential (RMP) was determined and APs were evoked with increasing current injections (10 μA; 300 ms). The amplitude of the first evoked AP was analyzed, and the number of all APs evoked by 10 depolarizing steps was summed up. sEPSCs were recorded at -70 mV with K (Fig, 7B)- or Cs (Fig. 7D and Fig. S7B) -based intracellular solution. Miniature EPSCs (mEPSCs) were recorded at -70 mV in 1 μM TTX and with Cs-based intracellular solution. Recordings were made at room temperature and were sampled at 20 kHz. Offline analysis was done with FitMaster (HEKA Elektronik GmbH). Spontaneous synaptic activity was analyzed using MiniAnalysis (Synaptosoft).

Statistical analysis visualization

Unless indicated otherwise, data for quantitative analysis was based on at least two genetically independent cell lines with two independent biological replicates. For statistical analysis the RStudio for macOS (Version 1.4.1106© 2009) was used. Graphs were generated with RStudio and show mean including single data points ± standard error of mean (SEM) or violin blots including single data points (scatterplot) with mean ± standard deviation (SD) and boxplot with median. The Kolmogorov-Smirnov test for equality of a probability distribution and the Levene test for homogeneity of variance were calculated prior statistical analysis. In case the data does not meet the assumption for parametric testing Kruskal-Wallis test for more than two group comparisons (post hoc Wilcoxon rank sum test, p-value adjustment Bonferroni correction), or for a two group comparison a Mann-Whitney-U-test for independent samples was calculated. Significance levels against the respective controls were not significant (n.s) if p > 0.05, or significant *p < 0.05. Individual figure legends contain detailed information regarding the number of replicates, sample size and the applied statistical tests.

Suppl. Fig. 1

Suppl. Fig. 2

Suppl. Fig. 3

Suppl. Fig. 4

Acknowledgements

We thank Isabell Moskal, Gina Tillmann and Helene Schamber for the excellent technical support. This work was supported by the Bundesministerium fur Bildung und Forschung (BMBF, SysmedSUD, Grant 01EK2101B), and the Hector Stiftung II. This work was previously published as an abstract and poster at the following conferences: ISSCR 2021, eMed 2021 + 2022, GSCN 2022, ICPR 2024.

Additional information

Author contributions

MS designed the study, performed the experiments, analyzed the data and wrote the manuscript. AH conducted processing, analysis and visualization of bulk RNA sequencing data. MD and GK performed electrophysiological characterization of neuronal cultures supplied by MS. SH generated the iPS cell lines Ctrl 1, Ctrl 2, Ctrl 3, Ctrl 4. JL, GK and TL edited the manuscript. TL was scientific consultant. MM and RS provided psilocin and edited the manuscript. PK conceptualized the study, supervised the work, provided resources and edited the manuscript.