Serotonergic Psychedelics Rapidly Modulate Evoked Glutamate Release in Cultured Cortical Neurons

Abstract

The serotonergic psychedelics psilocybin, LSD and DMT hold great promise for the development of new treatments for psychiatric conditions such as major depressive disorder, addiction and end-of-life anxiety. Previous studies in both animals and humans have confirmed the effects of these drugs on neuronal activity and plasticity. However, the understanding of the mechanisms of action of these substances is limited. Here we show rapid effects of psychedelics on presynaptic properties, using live cell imaging at the level of single synapses in primary rat cortical neurons. Using the genetically encoded reporter of synaptic vesicle fusion synaptopHluorin, we detected a reduced fraction of synaptic vesicles that fused in response to mild or strong electrical stimulation 3-30 min after application of serotonergic psychedelics. These effects were transient and no longer present 24 h after treatment. While DMT only reduced the total recycling pool, LSD and psilocin also reduced the size of the readily releasable vesicle pool. Imaging with the sensors for glutamate, iGluSnFR, and presynaptic calcium, synGCaMP6, showed that while psilocin and DMT increased evoked glutamate release, LSD and psilocin reduced evoked presynaptic calcium levels. Interestingly, psilocin also affected short-term plasticity leading to a depression of responses to paired stimuli. The rapid and drug-specific modulation of glutamatergic neurotransmission described in this study may contribute to distinct anxiolytic and antidepressant properties of serotonergic psychedelics.

Keywords: 5‐HT2A; fluorescent sensors; neurotransmitter release; presynaptic; short‐term plasticity; synaptic vesicles.

FIGURE 1

Serotonergic psychedelics increase nuclear localization of phosphorylated CREB after 24 h treatment. Representative images of immunofluorescent detection of pCREB Ser133 (pseudocoloured, range shown in the rectangle below) and dendritic marker MAP2 (red) overlaid with DAPI nuclear staining (blue) in mature cortical cultures treated with CTRL or drugs for 30 min (a), 2 h (c) or 24 h (e). Scale bar is 15 μm. Quantification of nuclear immunofluorescence (IF) of pCREB after treatment with CTRL or drugs for 30 min (b), 2 h (d) or 24 h (f), number of visual fields (each consisting of 10–15 analysed nuclei) from 2 to 3 independent experiments in brackets. Data did not pass Shapiro–Wilk test for normality, and were assessed by Kruskal–Wallis test (30 min: H _(3,92) = 7.020, p = 0.071; 2 h: H _(3,108) = 1.003, p = 0.801; 24 h: H _(3,128) = 40.26, p < 0.001) and Dunn's multiple comparisons test of CTRL versus each drug (24 h: p < 0.001 for all comparisons).

FIGURE 2

LSD and psilocin reduce the fusion competence of SVs after 30 min but not after 24 h treatment. Representative colour gradient (min to max range shown in the rectangle below) images from live‐cell imaging of sypHy‐expressing primary cortical cultures treated with vehicle (CTRL), LSD, DMT and PSI for 30 min (a) or 24 h (e) during baseline recording, stimulation with 40 and 900 APs at 20 Hz, and upon NH₄Cl pulse, in the presence of bafilomycin A1. Scale bars are 10 μm. (b, f) Average traces of sypHy fluorescence intensity from (a, e), respectively, normalised to minimum (baseline, F ₀) and maximum (NH₄Cl pulse, F _max). Total number of recordings (independent coverslips) per treatment from 3 to 4 independent culture preparations indicated in brackets. Quantification of the RRP (c, g) and TRP (d, h) fractions of SVs from the respective experiments. In graphs, bars show the mean, whiskers SEM, and circles all data points. Statistical significance assessed by one‐way ANOVA (30 min: RRP: F _(3,34) = 6.071, p = 0.002; TRP: F _(3,34) = 3.512, p = 0.026; 24 h: RRP: F _(3,34) = 0.189, p = 0.905; TRP: F _(3,34) = 0.694, p = 0.562) and Dunnett's multiple comparisons test CTRL versus each drug (30 min: RRP: LSD p = 0.002, DMT p = 0.205, PSI p = 0.002; TRP: LSD p = 0.010, DMT p = 0.192, PSI p = 0.040).

FIGURE 3

LSD and DMT, but not psilocin, reduce fusion competence of SVs upon acute treatment. (a, e) Representative colour gradient images (min to max range is shown in the rectangle below) from live‐cell imaging of sypHy‐expressing primary cortical cultures treated with vehicle (CTRL) and LSD (a), or CTRL, DMT and PSI (e). Drugs were added to the imaging solution immediately before (2–3 min) recording. SV fusion was evoked by 40 and 900 APs at 20 Hz and upon NH₄Cl pulse, in the presence of bafilomycin A1. Scale bars 10 μm. (b, f) Average traces of sypHy fluorescence intensity from (a, e) respectively, normalised to baseline (F ₀) and NH₄Cl response (F _max). Total number of recordings (independent coverslips) per treatment from 2 to 3 independent culture preparations indicated in brackets. Quantification of the RRP (c, g) and TRP (d, h). In graphs, bars show the mean, whiskers SEM, and circles all data points. Statistical significance assessed by unpaired t‐test for LSD (RRP: t ₍₁₂₎ = 3.757, p = 0.003; TRP: t ₍₁₂₎ = 2.956, p = 0.012) and by one‐way ANOVA (RRP: F _(2,28) = 1.397, p = 0.264; TRP: F _(2,28) = 12.35, p < 0.001) with Dunnett's multiple comparisons test for CTRL versus PSI or DMT (TRP: PSI p = 0.612, DMT p = 0.002).

FIGURE 4

5‐HT2AR antagonists ketanserin and M100907, as well as the nonselective 5‐HT2AR agonist lisuride, affect SV pools. Representative colour gradient images (range shown in the rectangle below) from live‐cell imaging of sypHy‐expressing neurons treated with vehicle (CTRL) or ketanserin (KTSR) 1 or 5 μM (a), CTRL or M100907 1 or 10 μM (e), and CTRL, serotonin (SER) 10 or 100 μM or lisuride 10 μM (i). Images show cells during baseline recording, stimulation with 40 and 900 APs at 20 Hz and upon NH₄Cl pulse, in the presence of bafilomycin A1. Scale bar is 10 μM. (b, f, j) Average traces of sypHy fluorescence intensity normalised to minimum (baseline, F ₀) and maximum (NH₄Cl pulse, F _max). Total number of recordings per treatment from 3 independent culture preparations is indicated in brackets. Quantification of the RRP (c, g, k) and TRP (d, h, l) fractions of SVs from the respective experiments. In graphs, bars show the mean, whiskers SEM and circles all data points. Statistical significance was assessed by one‐way ANOVA (KTSR: RRP: F _(2,24) = 3.294, p = 0.054, TRP: F _(2,24) = 21.59, p < 0.001; M100907: RRP: F _(2,28) = 3.476, p = 0.045, TRP: F _(2,28) = 4.757, p = 0.017; SER + LIS: RRP: F _(3,48) = 0.353, p = 0.787, TRP: F _(3,48) = 16.12, p < 0.001) with Dunnett's multiple comparisons test CTRL versus each drug (KTSR: TRP: 1 μM p = 0.676, 5 μM p < 0.001; M100907: RRP: 1 μM p = 0.212, 10 μM p = 0.029, TRP: 1 μM p = 0.242, 10 μM p = 0.010; SER + LIS: TRP: SER 10 μM p = 0.829, SER 100 μM p = 0.189, LIS p < 0.001).

FIGURE 5

Specific 5‐HT4R and 5‐HT7R antagonists reduce the size of TRP of SVs, while the 5‐HT1AR antagonist does not affect SV fusion competence. Representative colour gradient images from live‐cell imaging of sypHy‐expressing primary cortical cultures treated with vehicle (CTRL) or 5‐HT7R antagonist DR‐4485 (1 μM; a), CTRL or 5‐HT1AR antagonist NAD299 (1 or 10 μM; b), and CTRL or 5‐HT4R antagonist GR113808 (1 or 10 μM; c) for 45 min during baseline recording, stimulation with 40 and 900 APs at 20 Hz and upon NH₄Cl pulse, in the presence of bafilomycin A1. Scale bar 10 μM. (d, g, j) Average traces of sypHy fluorescence intensity from the respective experiments normalised to minimum (baseline, F ₀) and maximum (NH₄Cl pulse, F _max). Total number of recordings per treatment from 2 to 4 independent culture preparations is indicated in brackets. Quantification of the RRP (e, h, k) and TRP (f, i, l) fractions of SVs. In graphs, bars show the mean, whiskers SEM and circles all data points. Statistical significance was assessed by unpaired t‐test (DR4485: RRP: t ₍₁₄₎ = 1.098, p = 0.291, TRP: t ₍₁₄₎ = 2.160, p = 0.049), and one‐way ANOVA (GR113808: RRP: F _(2,27) = 3.065, p = 0.063, TRP: F _(2,27) = 11.41, p < 0.001; NAD299: RRP: F _(2,32) = 0.711, p = 0.499, TRP: F _(2,32) = 0.405, p = 0.670) with Dunnett's multiple comparisons test CTRL versus each drug (GR113808: TRP: 1 μM p = 0.421, 10 μM p = 0.004).

FIGURE 6

Psilocin induces paired‐pulse depression by increasing glutamate release probability. (a, e) Live‐cell imaging of mature cortical cultures expressing glutamate sensor iGluSnFR was conducted 30 min after treatment with substances. Representative colour gradient (range is shown in the rectangle) images were acquired during baseline and paired stimuli (peak 1 and 2; P1, P2) at 10 Hz in Tyrode's buffer with standard, 2 mM Ca²⁺ (a), or low, 1 mM Ca²⁺ (e). Scale bars are 10 μm. (b, f) Average traces of iGluSnFR fluorescence intensity shown as change in fluorescence (ΔF) divided by baseline fluorescence (F ₀). In brackets indicated per treatment: total number of recordings/numbers of independent coverslips from which these were taken; all from 3 independent culture preparations. (c, g) Fluorescence intensity (FI) ratio of P2/P1 (paired‐pulse ratio). Significance was assessed by one‐way ANOVA (c: F _(3,107) = 7.197, p < 0.001; g: F _(2,95) = 1.144, p = 0.323) with Dunnett's multiple comparisons test of CTRL versus each drug (c: LSD p = 0.982, DMT p = 0.252, PSI p < 0.001) in either 2 mM (c) or 1 mM (g) Ca²⁺. Amplitude of fluorescence at P1 shown as ΔF/F ₀ in 2 mM Ca²⁺ (d) or 1 mM Ca²⁺ (h). Data did not pass Shapiro–Wilk test for normality; thus, significance was assessed by Kruskal –Wallis test (d: H _(3,116) = 9.183 p = 0.027, h: H _(2,99) = 4.395, p = 0.111) with Dunn's multiple comparisons test of CTRL versus each drug (d: LSD p > 0.999, DMT p = 0.037, PSI p = 0.048). (i) Comparison of paired‐pulse ratios in standard (2 mM; c) and low (1 mM; g) Ca²⁺ concentration. Significances were assessed by two‐way ANOVA (drug effect: F _(2,177) = 4.055, p = 0.019; Ca²⁺ effect: F _(1,177) = 56.86, p < 0.001; interaction: F _(2,177) = 4.237, p = 0.016) with Šídák's multiple comparisons test for the differences between 1 mM and 2 mM Ca²⁺ per treatment (CTRL p = 0.007, DMT p = 0.015, PSI p < 0.001).

FIGURE 7

LSD and PSI reduce evoked presynaptic calcium transients. Representative colour gradient (range is shown in the rectangles below) images from live‐cell imaging of mature cortical cultures expressing calcium indicator synGCaMP6, treated with psychedelics for 30 min, during baseline, stimulus (a single stimulus, d paired‐pulse, g burst) and after the stimulation. Scale bars 10 μm. (b, f, h) Average traces of synGCaMP6 fluorescence intensity from respective experiments shown as change in fluorescence (ΔF) divided by baseline fluorescence (F ₀). Total number of recordings from 3 independent experiments indicated in brackets. Quantification of the amplitude of fluorescence intensity (as ΔF/F ₀) following the delivery of single stimulus (c), paired‐pulse (f) and burst of stimuli (30 APs; i). In graphs, bars show the mean, whiskers SEM and circles all data points. Statistical significance was assessed by one‐way ANOVA (single stimulus: F _(3,50) = 6.775, p < 0.001; paired‐pulse: F _(3,51) = 10.26, p < 0.001; burst: F _(3,50) = 4.824, p = 0.005) with Dunnett's multiple comparisons test for CTRL versus each drug (single stimulus: LSD p < 0.001, DMT p = 0.671, PSI p = 0.296; paired‐pulse: LSD p < 0.001, DMT p = 0.946, PSI p = 0.910; burst: LSD p = 0.005, DMT p = 0.577, PSI p = 0.019).