Effects of repeated lysergic acid diethylamide (LSD) on the mouse brain endocannabinoidome and gut microbiome

Abstract

Background and Purpose

Psychedelics elicit prosocial, antidepressant and anxiolytic effects via neuroplasticity, neurotransmission and neuro-immunomodulatory mechanisms. Whether psychedelics affect the brain endocannabinoid system and its extended version, the endocannabinoidome (eCBome) or the gut microbiome, remains unknown.

Experimental Approach

Adult C57BL/6N male mice were administered lysergic acid diethylamide (LSD) or saline for 7 days. Sociability was assessed in the direct social interaction and three chambers tests. Prefrontal cortex and hippocampal endocannabinoids, endocannabinoid-like mediators and metabolites were quantified via high-pressure liquid chromatography with tandem mass spectrometry (HPLC-MS/MS). Neurotransmitter levels were assessed via HPLC-UV/fluorescence. Gut microbiome changes were investigated by 16S ribosomal DNA sequencing.

Key Results

LSD increased social preference and novelty and decreased hippocampal levels of the N-acylethanolamines N-linoleoylethanolamine (LEA), anandamide (N-arachidonoylethanolamine) and N-docosahexaenoylethanolamine (DHEA); the monoacylglycerol 1/2-docosahexaenoylglycerol (1/2-DHG); the prostaglandins D₂ (PGD₂) and F2α (PGF_2α); thromboxane 2 and kynurenine. Prefrontal eCBome mediator and metabolite levels were less affected by the treatment. LSD decreased Shannon alpha diversity of the gut microbiota, prevented the decrease in the Firmicutes:Bacteroidetes ratio observed in saline-treated mice and altered the relative abundance of the bacterial taxa Bifidobacterium, Ileibacterium, Dubosiella and Rikenellaceae RC9.

Conclusions and Implications

The prosocial effects elicited by repeated LSD administration are accompanied by alterations of hippocampal eCBome and kynurenine levels, and the composition of the gut microbiota. Modulation of the hippocampal eCBome and kynurenine pathway might represent a mechanism by which psychedelic compounds elicit prosocial effects and affect the gut microbiome.

Abbreviations

5-HIAA

5-hydroxyindoleacetic acid

DHEA

N-docosahexaenoylethanolamine

eCBome

endocannabinoidome

LEA

N-linoleoylethanolamine

LSD

lysergic acid diethylamide

PEA

N-palmitoylethanolamine

SEA

N-stearoylethanolamine

VEH

vehicle

What is already known

• Psychedelics elicit therapeutic effects for psychiatric disorders and affect the plasma endocannabinoidome (eCBome) in humans.

What does this study add

• In mice, the prosocial effects elicited by repeated LSD are accompanied by hippocampal eCBome and 5-HT metabolism, and gut microbiome shifts.

What is the clinical significance

• These effects could be involved in the mechanism of action of psychedelics.

1 INTRODUCTION

Serotonergic psychedelics including lysergic acid diethylamide (LSD) have shown promise as novel therapeutics for the treatment of psychiatric disorders, including those resistant to conventional therapies (Carhart-Harris et al., 2021; Davis et al., 2021; Inserra et al., 2021; Mitchell et al., 2021; Palhano-Fontes et al., 2019). Several mechanisms could explain these effects through 5-hydroxytryptamine (5-HT; serotonin) receptors, including the 5-HT_2A, 5-HT_1A as well as AMPA receptors (de Gregorio et al., 2016, 2022; Seeman et al., 2005), neural and synaptic plasticity (Li et al., 2010; Ly et al., 2018), immune function (Szabo et al., 2014, 2016) and epigenetics (de la Fuente Revenga et al., 2021; Inserra et al., 2022; Ruffell et al., 2021). Psychedelics elicit anti-inflammatory effects (Flanagan et al., 2019; Flanagan & Nichols, 2018; Szabo et al., 2014) triggered by 5-HT_2A (Flanagan et al., 2019; Flanagan & Nichols, 2018; González-Maeso et al., 2003; Nau et al., 2013, 2015) and sigma-1 (σ₁) receptor signalling (Szabo et al., 2014, 2016). Thus, they might produce beneficial effects on the altered inflammatory and oxidative milieu contributing to the development of psychiatric disorders and co-morbid gut conditions (Dinan & Cryan, 2015; Inserra et al., 2018; Kelly et al., 2015; Shoubridge et al., 2022; Wong et al., 2016).

Indeed, LSD administration in humans modulates cortisol, corticosterone, prolactin, oxytocin and adrenaline (Strajhar et al., 2016), whereas, in vitro, it elicits anti-inflammatory effects by down-regulating interleukins, IL- 2, IL-4 and IL-6, and up-regulating mitogen-activated protein kinase 1 (ERK2) (House et al., 1994). Given that inflammatory signalling influences gut microbiome composition (Elinav et al., 2011; Henao-Mejia et al., 2012; Inserra, Choo, et al., 2019; Wong et al., 2016), which is variably yet recurrently altered in psychiatric disorders (Dinan & Cryan, 2015; Foster & McVey Neufeld, 2013; Rogers et al., 2016; Strati et al., 2017), the anti-inflammatory effects of LSD could have a modulatory role over psychopathology-related gut bacteria (Inserra, Mastronardi, et al., 2019). In fact, the gut microbiome plays a fundamental role in neurophysiology, behaviour and mental health via the gut–brain axis and its therapeutic manipulation represents both a novel frontier and a research priority in psychiatry (Chinna Meyyappan et al., 2020; Inserra et al., 2018; Settanni et al., 2021; Shoubridge et al., 2022).

Another key biological system intimately connected with the immune system (Chiurchiù et al., 2015; Hemarajata & Versalovic, 2012; Pandey et al., 2009) and the gut microbiome (Cani et al., 2016; Manca et al., 2020) is the expanded endocannabinoid system or endocannabinoidome (eCBome). This is a complex lipid signalling system composed of a plethora of fatty acid-derived mediators and their receptors, including, but not limited to, cannabinoid receptors (Galve-Roperh et al., 2009; Iannotti & Di Marzo, 2021; Janero et al., 2009; Parolaro et al., 2010; van der Stelt & Di Marzo, 2003). Importantly, the eCBome mediates the association between gut microbiome composition and anhedonia/amotivation (Minichino et al., 2021). Hence, it is possible that the therapeutic effects elicited by psychedelics acting via the 5-HT system might be mediated by or result in, changes in gut the microbiota composition, brain eCBome composition or brain 5-HT metabolism (Thompson & Szabo, 2020). Notably, 5-HT, synthesized from tryptophan (L-Trp), is in competition with the kynurenine pathway, especially in the presence of inflammation (Comai et al., 2020) and this pathway is also related to the eCBome system (Gobbi et al., 2005). So far, studies exploring the brain eCBome, the gut microbiome and the 5-HT system after treatment with psychedelics have not been reported. The goal of this study was to test if the regimen of LSD (30 μg·kg⁻¹ once a day for 7 days) with prosocial (De Gregorio et al., 2021), anxiolytic (De Gregorio et al., 2022) and neurotrophic (Inserra et al., 2022) activity affects the prefrontal and hippocampal (HCC) eCBome, 5-HT metabolism and the composition of the gut microbiota.

2 METHODS

All procedures were approved by the McGill University Animal Care Ethics Committee (Protocol 5764) and are in line with the Canadian Institute of Health Research for Animal Care and Scientific Use, and the Facility Animal Care Committee of McGill University. No adverse events were encountered during the study. The primary outcome measures were (i) whether repeated LSD affects the hippocampal and prefrontal cortex eCBome, (ii) whether the treatment affects the prefrontal cortex and hippocampal 5-HT metabolism and (iii), whether the treatment affects the composition of the gut microbiota. Secondary outcome measures were (iv) whether the treatment affects sociability and the investigation of correlations between (v) behavioural phenotypes and hippocampal eCBome, (vi) behavioural phenotypes and gut microbiota composition at the genus level and (vii) gut microbiota composition and the hippocampal eCBome.

2.1 Animals

Adult male C57BL/6N mice (Charles Rivers, Saint-Constant, Quebec, Canada) (n = 53) specific pathogen-free, aged 60–70 days, weighing 25–30 g, were bred in-house and housed under standard laboratory conditions (max 5 per cage) with a 12 h light–dark cycle (lights off at 19:00) with ad libitum access to food and water, in cages enriched with nesting material. Mice were fed a standard soya-based, irradiated diet with a 6.5% fat content, 1.1% total monounsaturated fatty acids, 2.9% total polyunsaturated fatty acids, 47% carbohydrates and 19% protein (Envigo Teklad 2920X, Madison, WI, USA). The animal species, sample size and dosing regimen were chosen based on our previous studies given that in this species and with similar sample sizes, the LSD regimen employed here elicits prosocial effects while potentiating excitatory neurotransmission in the prefrontal cortex (de Gregorio et al., 2021) and elicits anxiolytic effects while increasing dendritic spine density in pyramidal neurons of the prefrontal cortex and enhancing the firing activity of 5-HT neurons in the dorsal raphé nucleus (de Gregorio et al., 2022). Besides, this dose engenders low hallucinogenic-like behaviours (de Gregorio et al., 2022) and only engages 5-HT, but not dopamine transmission (de Gregorio et al., 2016), thus increasing the translational value of the current study. No exclusion criteria were determined a priori. No humane endpoint was established for this study. Animal studies are reported in compliance with the ARRIVE guidelines (Percie du Sert et al., 2020) and with the recommendations made by the British Journal of Pharmacology (Lilley et al., 2020).

2.2 Drugs and treatment

LSD (Sigma-Aldrich, London, UK) was dissolved in a 0.9% NaCl vehicle (VEH) solution and administered intraperitoneally (i.p.) at a volume of 5 ml·kg⁻¹ body weight. Mice within the same cage received the same treatment, either VEH (n = 16) or LSD (n = 17) for 7 days (30 μg·kg⁻¹·day⁻¹), switching injection quadrant to minimize discomfort. The direct social interaction (DSI) test took place 24 h after the 7th injection. Immediately after the direct social interaction test, the mice received an additional (8th) injection and the three chambers test (TCT) was performed 24 h later. An overview of the experimental timeline is provided in Figure 1 (top).

FIGURE 1

Repeated LSD treatment induces prosocial effects in the direct social interaction (DSI) test and the three chambers test (TCT). (top) Experimental timeline. 16S, 16 Svedberg; eCBome (endocannabinoidome); hippocampus (HCC); HPLC-MS/MS (high-pressure liquid chromatography with tandem mass spectrometry); LSD (lysergic acid diethylamide); PFC (prefrontal cortex); VEH (vehicle). (bottom) Data are represented as mean ± SEM if normally distributed or represented as box (fist quartile, median, third quartile) and whisker (minimum, maximum). Individual mice represented as dots. (a) Mice treated with LSD (30 μg·kg⁻¹·day⁻¹ for 7 days) spent more time interacting with a stranger mouse in the DSI test compared to vehicle-treated mice (n = 16–17 per group). Mann–Whitney test. (b) Mice treated with LSD spent more time interacting with a stranger mouse (S1) versus the empty cage compared to mice treated with vehicle in the TCT (n = 16–17 per group). Two-way analysis of variance (ANOVA) followed by Bonferroni post hoc comparison. (c) Treatment with LSD increased the time spent interacting with the new stranger mouse (S2) versus the initial mouse (S1) compared to vehicle treatment (n = 16–17 per group). Two-way ANOVA followed by Bonferroni post hoc comparison. (d) Mice treated with LSD had a higher social preference index in the TCT compared to vehicle-treated mice. Two-tailed unpaired t test. (e) Mice treated with LSD had a higher social novelty index compared to vehicle. Mann–Whitney test. (f) The distance travelled in the TCT was not significantly different between treatments. Repeated measures ANOVA followed by Bonferroni post hoc comparison. (g) The number of rearing events in the DSI test was not different between LSD- and vehicle-treated mice. Two-tailed Student's unpaired t test. (h) Rearing time in the DSI test was not affected by the treatment. Two-tailed Student's unpaired t test. (i) Grooming bouts in the DSI test were not significantly different between mice treated with vehicle or LSD. Two-tailed Student's unpaired t test. (j) The total grooming time in the DSI test was not affected. Two-tailed Student's unpaired t test. S1, Stranger Mouse 1; S2, Stranger Mouse 2; *P < 0.05; versus vehicle. ^#P < 0.05 S1 or S2, respectively, versus empty cage or S1; ns, non significant.

Source: Created with BioRender.com

2.3 Behavioural tests

2.3.1 Direct social interaction test

This test was performed according to our standardized protocol (de Gregorio et al., 2021). After 10 min habituation, a novel weight- and sex-matched conspecific stranger mouse was introduced into the cage with the test mouse and the mice were able to freely engage in social interaction for 10 min. The interaction time, defined by the following behaviours: nose-to-anogenital sniffing, nose-to-nose sniffing and social grooming was manually scored blindly.

2.3.2 Three chambers test

This test was performed according to our standardized protocol (de Gregorio et al., 2021). Test mice were placed in the middle chamber and allowed to freely explore the empty three-chamber arena for 10 min. Immediately after, an unfamiliar mouse (Stranger 1 [S1], male C57BL/6N, weight matched) was introduced into one of the two side chambers, enclosed in a wire cage, thus allowing only the test mouse to initiate social interaction. An identical empty wire cage was placed in the other side chamber. The test mouse was allowed to explore the three-chamber arena for 10 min. Immediately afterwards, a new unfamiliar mouse (Stranger 2 [S2], male C57BL/6N, weight matched) was placed in the previously unoccupied wire cage and the test mouse was observed for an additional 10 min to assess social novelty. The location of the empty wire cage was alternated between side chambers for different test mice to prevent chamber bias. The time spent sniffing S1, S2 or the empty cage was manually scored blindly. We extrapolated a ‘sociability index’ for each mouse, calculated as 100 × (S1 − empty cage time)/(S1 + empty cage time) and a ‘social novelty index’, calculated as 100 × (S2 − S1)/(S2 + S1) (de Gregorio et al., 2021; Heifets et al., 2019). Grooming was defined by elliptical, small and bilateral strokes, as well as flank, tail and genital licks (Kalueff & Tuohimaa, 2004). Rearing was defined by the animal temporarily standing on its hind legs, supported (against a wall) or unsupported (free-standing) (Sturman et al., 2018).

2.4 Sample collection

Faecal pellets were collected by placing each mouse in a clean, standard cage without bedding on Days 15 or 16 post-introduction to the new facility and on Days 23 or 24, the latter two representing Days 8 or 9 following initiation of the repeated LSD regimen. Both collections occurred ~24 h after the last injection to avoid potential confounding acute effects of LSD on the gut microbiome. On Day 23 (Day 8 of the repeated regimen, in which the direct social interaction test was performed), the faecal pellets were collected after the direct social interaction test and prior to the last injection of LSD (as discussed in Section 2.2) with autoclaved toothpicks, placed in 1.5 ml tubes, snap-frozen on dry ice and stored at −80°C until processing. Following cervical dislocation, brains were quickly removed, dissected on a chilled metal plate on an ice bath and prefrontal cortices and hippocampi isolated, placed in 1.5 ml tubes, snap-frozen on dry ice and stored at −80°C until processing. A separate cohort of mice was used for neurotransmitters measurements.

2.5 Lipid extraction and high-pressure liquid chromatography with tandem mass spectrometry (HPLC-MS/MS) analysis of endocannabinoids and endocannabinoid-like mediators

We separately analysed the eCBome composition of the prefrontal cortex and hippocampus. Lipids were extracted from tissue samples according to the Bligh and Dyer method (Bligh & Dyer, 1959) with modifications and processed and analysed randomly and blindly. Briefly, the prefrontal cortex and hippocampus of each mouse were powdered in liquid nitrogen and about 10 mg was homogenized in 1 ml of a 1:1 Tris-HCl 50 mM pH 7:methanol solution containing 0.575% acetic acid and 5 μl of deuterated standards. One millilitre of chloroform was then added to each sample, which was then vortexed for 30 s and centrifuged at 4000 g for 5 min. The organic phase was collected and another 1 ml of chloroform was added to the inorganic one. This was repeated twice to ensure the maximum collection of the organic phase. The organic phases were pooled and evaporated under a stream of nitrogen and then suspended in 58 μl of mobile phase containing 50% of Solvent A (water + 1 mM ammonium acetate + 0.05% acetic acid) and 50% of Solvent B (acetonitrile/water 95/5 + 1 mM ammonium acetate + 0.05% acetic acid). Fifty microlitres of each sample were finally injected onto an HPLC column (Kinetex C8, 150 × 2.1 mm, 2.6 μm, Phenomenex) and eluted at a flow rate of 400 μl·min⁻¹ using a discontinuous gradient of Solvent A and Solvent B (Everard et al., 2019). Quantification of eCBome-related mediators (see Table S1 for a full list of molecules analysed) was carried out by an HPLC system interfaced with the electrospray source of a Shimadzu 8050 triple quadrupole mass spectrometer and using multiple reaction monitoring in positive ion mode for the compounds and their deuterated homologues. In the case of unsaturated monoacylglycerols, the data are presented as 1/2-monoacylglycerols (1/2-MAGs), representing the combined signals from the 2- and 1(3)-isomers because the latter are most likely generated from the former via acyl migration from the sn-2 to the sn-1 or sn-3 position.

2.6 Analysis of tryptophan metabolites via serotonin and kynurenine endocannabinoids

The quantitative analysis of 5-HT, its precursor tryptophan and its degradation metabolite 5-hydroxyindoleacetic acid (5-HIAA) and kynurenine was performed blindly in the hippocampus and prefrontal cortex of a separate cohort of mice (VEH n = 8, LSD n = 11) following our previously reported method with slight modifications. Briefly, brain areas were homogenized and sonicated for 2 min in 200 μl of a 0.2 M perchloric acid solution and then centrifuged for 6 min at 13,000 g at 4°C. The separation and then quantification of tryptophan, 5-HT and 5-HIAA were conducted with a Shimadzu LC-10AD HPLC system equipped with a Shimadzu RF-10AXL fluorometric detector set at excitation and emission wavelengths of 279 and 320 nm, respectively. The chromatographic separation was performed on an Apollo C18 (5 μm 250 mm × 4.6 mm) column (Sepachrom Mega Srl, Milan, Italy) using a mobile phase at a flow rate of 1 ml·min⁻¹ composed by milliQ water/acetonitrile (5% water, 95% acetonitrile) and milliQ water/methanol (90% water, 10% methanol) in a ratio of 5:95 v/v, respectively, that were acidified with orthophosphoric acid to a pH of 3.5. The separation and then quantification of kynurenine were conducted with a Shimadzu LC-10AD HPLC system equipped with a Shimadzu SPD-10A UV–VIS detector set at 360 nm, a Robusta C18 (5 μm 250 mm × 4.6 mm) column (Sepachrom Mega Srl) and a mobile phase as above indicated for tryptophan and 5-HT, at a flow rate of 1 ml·min⁻¹. The ratio 5-HIAA/5-HT was calculated as an index of the overall brain activities consisting of release and uptake of 5-HT, and metabolism of 5-HT into 5-HIAA (Commissiong, 1985). The ratio kynurenine/tryptophan was calculated as an indirect measure of the amount of tryptophan degraded along the kynurenine pathway (Comai et al., 2020).

2.7 16S ribosomal RNA gene amplicon sequencing

DNA was extracted from feces using the QIAmp PowerFecal DNA kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The DNA concentrations of the extracts were measured fluorometrically with the Quant-iT PicoGreen dsDNA Kit (Thermo Fisher Scientific, MA, USA) and the DNA was stored at −20°C until 16S rDNA library preparation. Briefly, 1 ng of DNA was used as template and the V3–V4 region of the 16S rRNA gene was amplified by polymerase chain reaction (PCR) using the QIAseq 16S Region Panel protocol in conjunction with the QIAseq 16S/ITS 384-Index I (Sets A, B, C, D) kit (Qiagen Canada, Montreal, QC, Canada) (Rausch et al., 2019). The 16S metagenomic libraries were eluted in 30 μl of nuclease-free water and 1 μl was qualified with a Bioanalyzer DNA 1000 Chip (Agilent, CA, USA) to verify the amplicon size (expected size ~600 bp) and quantified with a Qubit (Thermo Fisher Scientific). Libraries were then normalized and pooled to 2 nM, denatured with NaOH and diluted to a final concentration of 6 pM. Paired-end sequencing (2 × 300 bp) was performed using the MiSeq Reagent Kit V3 (600 cycles) on an Illumina MiSeq System. Sequence data were processed using the DADA2 pipeline and taxonomic assignation with reference to the Silva database (v132) (Callahan et al., 2016; Quast et al., 2013). Sequence counts from samples were normalized using the cumulative sum scaled method from the MetagenomeSeq package (McMurdie & Holmes, 2013; Paulson et al., 2013).

2.8 Statistical analysis

The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2022).

2.8.1 Behaviour, eCBome and neurotransmitters

Data are expressed as mean ± SEM for data normally distributed and as box (fist quartile, median, third quartile) and whisker (minimum, maximum) for data non-normally distributed and analysed using GraphPad Prism 9 (GraphPad Software Inc., San Diego, CA, USA). The normality of data distribution was assessed using the Shapiro–Wilk test. When comparing data between two groups, unpaired Student's t test or Mann–Whitney test was performed according to whether data are respectively normally or non-normally distributed. Repeated measures (RM) two-way analysis of variance (ANOVA) was used to analyse social interaction in the three chambers test considering the factors treatment and social interaction. Mixed effects ANOVA was used to analyse locomotion throughout the three phases of the three chambers test considering the factors treatment and three chambers test phase. When appropriate, post hoc tests with Bonferroni multiple comparison corrections were performed to examine differences among groups. Values of P < 0.05 were considered statistically significant. Post hoc tests were conducted only if F achieved P < 0.05, and there was no significant variance in homogeneity.

2.8.2 Gut microbiome

Outliers were defined as samples outside the 95% confidence interval (CI) ellipse by a principal component analysis (PCA). Asterisks (if none: ns) displayed above boxes represent Kruskal–Wallis P-values: *P < 0.05. Permutational ANOVA (PERMANOVA) was used for microbiota composition analyses. Pairwise comparisons (within group) was performed using the Wilcoxon test, with the P-values indicated above the brackets. Data are represented as bar graphs showing mean ± SEM with individual mice shown as dots if normally distributed or represented as box (fist quartile, median, third quartile) and whisker (minimum, maximum) plots with individual mice shown as dots if non-normally distributed.

The heatmap.2 package for R was used to represent bacterial family composition between treatment groups and time points. Cumulative sum scaled -normalized bacterial counts were scaled across the entire cohort using Z-scores and then treatment groups and time points were clustered using unsupervised hierarchical clustering. Differential abundance testing was assessed using two-way ANOVA (taxa ~ Group*Day) and Tukey's honestly significant difference (HSD) post hoc P-values. Spearman's correlations were computed using the cor.test R function and Spearman's rho correlation coefficients were represented into heatmaps.

2.9 Materials

LSD was obtained from Sigma-Aldrich, London, UK. Acetic acid, ammonium acetate and chloroform were obtained from Fisher Scientific (Saint-Laurent, QC, Canada). Acetonitrile (HPLC grade), methanol, perchloric acid, and orthophosphoric acid, where obtained from Sigma-Aldrich (Oakville, ON, Canada); liquid nitrogen was obtained from Messer Canada Inc. (Mississauga, ON, Canada). Details of other materials and suppliers were provided in the specific sections.

2.10 Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in the IUPHAR/BPS Guide to PHARMACOLOGY http://www.guidetopharmacology.org and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22 (Alexander et al., 2021).

3 RESULTS

3.1 Repeated LSD increases sociability in the direct social interaction test and three chambers test

Confirming our previous findings (de Gregorio et al., 2021), mice receiving repeated LSD spent significantly more time in social interaction with a stranger mouse in the direct social interaction test compared to VEH-treated mice (Figure 1a). In Phase 2 of the three chambers test, we found a significant main effect of treatment (Figure 1b), sociability (intruder mouse or empty cage), (Figure 1b) and a treatment × sociability interaction (Figure 1b). Post hoc analyses unveiled that whereas both LSD- and VEH-treated mice significantly interacted more with an unfamiliar stranger compared to an empty cage, LSD-treated mice spent significantly more time interacting with a stranger mouse compared to VEH-treated controls (Figure 1b). In Phase 3 of the three chambers test, we observed a significant main effect of treatment (Figure 1c), social novelty (Figure 1c) and an interaction between treatment and social novelty (Figure 1c). In fact, whereas both LSD- and VEH-treated mice spent significantly more time interacting with the novel conspecific mouse (S2) rather than the familiar one (S1), LSD-treated mice spent significantly more time interacting with the novel intruder than VEH-treated mice (Figure 1c). Accordingly, the social preference and social novelty indices were significantly greater in LSD-treated mice (respectively, Figure 1d and 1e). Repeated LSD did not alter locomotion (mixed design ANOVA, treatment effect, Figure 1f) in the three chambers test, nor did it alter the number of rearing events (Figure 1g), the time spent rearing (Figure 1h), the number of grooming bouts (Figure 1i) or the total grooming time (Figure 1j).

3.2 Repeated LSD administration modulates the levels of eCBome mediators in the hippocampus, but not in the prefrontal cortex

In the hippocampus, the levels of several eCBome mediators were decreased by the repeated LSD treatment as also observed by principal component analysis of the murine hippocampal eCBome of the LSD and VEH groups post-intervention (Figure S5). Specifically, we observed a statistically significant decrease in N-acylethanolamines (NAEs), that is, N-linoleoylethanolamine (LEA; Figure 2a), anandamide (N-arachidonoylethanolamine, AEA; Figure 2b), a trend that did not reach significance for N-palmitoylethanolamine (PEA; Figure 2c) and a significant decrease in N-docosahexaenoylethanolamine (DHEA; Figure 2d). On the contrary, N-stearoylethanolamine (SEA), N-oleoylethanolamine (OEA) and N-eicosapentanoylethanolamine (EPEA) did not change (Figure S1). Concerning monoacylglycerols, we detected a significant decrease in hippocampal 1/2-docosahexaenoylglycerol (1/2-DHG; Figure 2e), but not 1/2-linoleoylglycerol (1/2-LG), 1/2-arachidonoylglycerol (1/2-AG), 1/2-eicosapenaenoylglycerol (1/2-EPG) or 1/2 docosapentaenoylglycerol (1/2-DPG) (Figure S1).

FIGURE 2

Repeated lysergic acid diethylamide (LSD) administration (30 μg·kg⁻¹·day⁻¹ for 7 days, i.p.) modulates endocannabinoid and endocannabinoid-like mediators in the hippocampus. Lipid mediators were quantified using liquid chromatography with tandem mass spectrometry. Repeated LSD decreases hippocampal levels of the N-acylethanolamines (NAEs), (a) N-linoleoylethanolamine (LEA), (b) anandamide (N-arachidonoylethanolamine; AEA) and shows a strong trend towards decreasing (c) N-palmitoylethanolamine (PEA). Repeated LSD decreases hippocampal levels of (d) N-docosahexaenoylethanolamine (DHEA); (e) 1/2-docosahexaenoylglycerol (1/2-DHG); (f) prostaglandin D2 (PGD₂); (g) prostaglandin F2α (PGF_2α) and (h) thromboxane B2 (TXB2). Repeated LSD shows a strong trend towards decreasing (i) eicosapentaenoic acid (EPA). Data are represented as mean ± SEM if normally distributed or represented as box (fist quartile, median, third quartile) and whisker (minimum, maximum). Individual mice represented as dots. Statistical analysis was performed by two-tailed unpaired t test if normally distributed (all except g–i) or Mann–Whitney U-test if non-normally distributed (g–i). ^#P < 0.05 compared to vehicle (VEH)-treated mice

Similarly, LSD induced changes within a subset of hippocampal prostaglandins, which can be considered as metabolic products of 2-arachidonoylglycerol in the brain under certain conditions (Grabner et al., 2016; Nomura et al., 2011; Viader et al., 2015). Prostaglandin D2 (PGD₂; Figure 2f), prostaglandin F2α (PGF_2α; Figure 2g) and thromboxane B2 (TXB2, Figure 2h) levels significantly decreased in the group receiving repeated LSD compared to controls, whereas prostaglandin E2 (PGE₂) and 6-keto PGF_1α levels were unaffected (Figure S1).

With regard to long-chain polyunsaturated fatty acids, which can be considered as both ultimate biosynthetic precursors of polyunsaturated N-acylethanolamines (NAEs) and monoacylglycerols and products of their hydrolysis, eicosapentaenoic acid (EPA) showed a strong but non-significant trend towards being decreased in the hippocampus of LSD-treated mice (Figure 2i), whereas linoleic, arachidonic and docosahexaenoic (DHA) acids were unaffected (Figure S1).

Fewer alterations were observed in the prefrontal cortex of LSD-treated mice (Figure S2), suggesting that the hippocampus is a major site of action for LSD with respect to modulation of the eCBome. No overlap between the hippocampus and prefrontal cortex with respect to the mediators modified by LSD was observed. In the prefrontal cortex, only prostaglandin E1 (PGE₁) showed a trend towards being increased, without reaching statistical significance (Figure S2). No changes were found in any of the other mediators quantified in the prefrontal cortex (Figure S2 and Table S1).

3.3 Repeated LSD administration shifts the balance between the kynurenine and 5-HT pathways to the 5-HT pathway in the hippocampus

Although repeated LSD administration did not affect hippocampal tryptophan levels (Figure 3a), it selectively and significantly decreased kynurenine levels (Figure 3b) and the kynurenine/tryptophan ratio (Figure 3c). Moreover, repeated LSD did not affect the levels of 5-HT (Figure 3d) but was associated with a significant decrease in the kynurenine/5-HT ratio (Figure 3e). Lastly, the regimen did not affect hippocampal 5-HIAA levels (Figure 3f) or the 5-HIAA/5-HT ratio (Figure 3g) in the hippocampus. No differences were observed between VEH and repeated LSD administration concerning the prefrontal cortex levels of tryptophan, kynurenine, the kynurenine/tryptophan ratio, 5-HT, the kynurenine/5-HT ratio, 5-HIAA and the 5-HIAA/5-HT ratio (Figure S3).

FIGURE 3

Repeated lysergic acid diethylamide (LSD) administration (30 μg·kg⁻¹·day⁻¹ for 7 days, i.p.) shifts the balance between the kynurenine (Kyn) and 5-HT pathways to the 5-HT pathway in the hippocampus. Repeated LSD administration (a) does not affect tryptophan (Trp) levels in the hippocampus, but (b) it selectively decreases Kyn levels and (c) the Kyn/Trp ratio. The regimen does not affect hippocampal (d) 5-HT, but (e) it decreases the Kyn/5-HT ratio. The regimen does not affect (f) hippocampal 5-hydroxyindoleacetic acid (5-HIAA) levels, or (g) the 5-HIAA/5-HT ratio. Data are represented as mean ± SEM if normally distributed or represented as box (fist quartile, median, third quartile) and whisker (minimum, maximum). Individual mice represented as dots. Statistical analysis was performed by two-tailed unpaired t test if normally distributed (d, e, g) or Mann–Whitney U-test if non-normally distributed (a–c, f). ^#P < 0.05 compared to vehicle (VEH)-treated mice

3.4 Repeated LSD administration does not strongly modify the gut microbiome architecture

The global microbiota composition did not significantly differ between groups as also observed by non-metric multidimensional scaling (Figure S4). Despite this, repeated LSD (30 μg·kg⁻¹·day⁻¹ for 7 days, i.p.), but not repeated VEH (saline solution, i.p.), administration significantly decreased the gut microbiome Shannon alpha diversity (Figure 4a). Shannon diversity index is a measure of diversity that takes both evenness (distribution of abundances of taxonomic groups) and richness (number of the groups) into account; the higher the value of the index, the higher the diversity of species in a particular community (the whole microbiome) (Willis, 2019). The Firmicutes:Bacteroidetes (F/B) ratio significantly decreased over time in control mice, but not in LSD-treated mice (Figure 4b). The Firmicutes:Bacteroidetes ratio is an often-used marker of broad microbiome disruption and has been associated with the development of several pathological conditions, although the applicability of this simple measure has been brought into question (Magne et al., 2020).

FIGURE 4

Repeated lysergic acid diethylamide (LSD) decreases Shannon alpha diversity and prevents the Firmicutes:Bacteroidetes ratio decrease. (top) (a) Assessment of Shannon alpha diversity and (b) Firmicutes:Bacteroidetes ratio in vehicle- and LSD-treated (30 μg·kg⁻¹·day⁻¹ for 7 days) mice (n = 16 per group). Data are represented as mean ± SEM with individual mice shown as dots and statistically analysed using the Wilcoxon test. (bottom) (c) Dubosiella was significantly increased in mice receiving repeated LSD (30 μg·kg⁻¹·day⁻¹ for 7 days), but not in mice receiving vehicle. Similarly, (d) Ileibacterium, (e) the Rikenellaceae RC9 gut group and (f) Bifidobacterium were increased in LSD-treated, but not in control, mice. Statistical comparison within the group (c–f) was performed using the Wilcoxon test with false discovery rate correction for multiple testing, based on a significance threshold of P < 0.05. Differences in taxa levels between the groups at the end of the intervention, accounting for cage effects, were also compared using a linear mixed effects model. Data are represented as mean ± SEM if normally distributed or represented as box (fist quartile, median, third quartile) and whisker (minimum, maximum). Individual mice represented as dots. ^#P < 0.05 compared to baseline.

The microbiota composition significantly differed across time (permutational ANOVA; P = 0.007, pseudo-F = 4.245), although post hoc analysis indicated that this difference did not reach significance for the VEH (P(perm) = 0.1049, t = 1.53) or the LSD (P(perm) = 0.090, t = 1.55) groups. However, repeated LSD resulted in alterations in the relative abundance of four of the ~90 bacterial genera identified (only ~4.5%). Significantly increased levels of bacterial genera belonging to the Erysipelotrichaceae family (Dubosiella and Ileibacterium, respectively, Figure 4c and 4d), as well as the Rikenellaceae RC9 gut group of bacteria (Figure 4e) and Bifidobacterium (Figure 4f), were observed in LSD-treated mice, but not in VEH-treated mice. On the other hand, the levels of these taxa in the LSD group did not significantly differ from those of the VEH group at the end of the intervention (linear mixed effects model, accounting for cage effects).

3.5 Correlations between hippocampal eCBome shifts and behavioural phenotypes

We then investigated the possibility that specific behaviours might be associated to eCBome tone or gut microbiota composition by assessing the correlations between the level of hippocampal eCBome mediators and behavioural phenotypes (Figure 5) and between gut microbiota composition and behavioural phenotypes (Figure 6). Both treatment groups were included in this correlation analysis in order to investigate the possibility that specific behaviours might be mediated by the eCBome tone or gut microbiota composition. For example, certain behaviours are more or less accentuated with higher or lower brain concentrations of certain eCBome mediators and/or with higher or lower relative abundances of certain microbiota taxa. Therefore, we reasoned that a treatment that modifies (or does not modify) behaviour should also modify (or not modify) the other parameters. For the same reason, we included not only the behaviours that were altered by the LSD treatment but also other behaviours of interest that were not different between VEH- and LSD-treated mice.

FIGURE 5

Correlations between the hippocampal endocannabinoidome and behavioural phenotypes. 1.2.DHG, mono-docosahexaenoyl-glycerol 1 and 2 isomers; 1.2.EPG, mono-eicosapentaenoyl-glycerol 1(3) and 2 isomers; 1.2.LG, mono-linoleoyl-glycerol 1(3) and 2 isomers; 1.2.SDG, 1-arachidonoyl-d₅-rac-glycerol; 1.2-AG, 1/2-arachidonoylglycerol; 1.2-OG, 1-arachidonoyl-d₅-rac-glycerol; AA (arachidonic acid); anandamide (N-arachidonoylethanolamine; AEA); DHA, (docosahexaenoic acid); DHEA (N-docosahexaenoylethanolamine); DPA (n-3-docosapentaenoic acid); DPEA (N-docosahexaenoyl-ethanolamine); DSI (direct social interaction); EPA (eicosapentaenoic acid); EPEA (N-eicosapentanoylethanolamine); LA (linoleic acid); LEA (N-linoleoylethanolamine); OEA (N-oleoylethanolamine; PEA, N-palmitoylethanolamine; SDA, stearidonic acid; SEA, N-stearoylethanolamine); TCT (three chambers test of sociability); TintT (total interaction time in Phase 2 + Phase 3 of the TCT). *P < 0.05. The legend on the right shows the colour-coding of the correlations' r values: Blue colours indicate a positive correlation (r > 0); red colours indicate a negative correlation (r < 0).

We observed that the abundance of several hippocampal N-acylethanolamines (PEA, LEA, SEA and DHEA) was significantly inversely correlated with the social interaction time in the direct social interaction test. Moreover, anandamide, LEA and SEA levels were significantly negatively correlated with the social novelty index and significantly positively correlated with increased preference for the familiar mouse in Phase 3 of the three chambers test. The time spent rearing in the habituation phase of the three chambers test, which can be considered as an index of exploratory behaviour, was significantly positively correlated with the levels of PEA, LEA and SEA, whereas the number of grooming events during the habituation phase of the direct social interaction test was significantly positively correlated with the levels of OEA, LEA, SEA, DHEA, arachidonic acid, LA, n-3-docosapentaenoic acid (DPA), anandamide and PEA. Lastly, anandamide and LEA, SEA and 1/2-EPG were directly and significantly correlated with the preference for the familiar intruder mouse during Phase 3 of the three chambers test.

3.6 Correlations between gut microbiome shifts and behavioural phenotypes

Next, we assessed correlations between changes in gut microbiome composition and behavioural phenotypes to investigate the hypothesis that the gut microbiome changes observed following repeated LSD administration might play a causal role in the behavioural phenotypes observed (Figure 6). At the genus level, we found that the social interaction time in the direct social interaction test was positively correlated with the abundance of the genera LachnospiraceaUCG-006 and Ruminococcaceae UCG-003 significantly. The social preference index was negatively correlated with the abundance of the genera Bacteroides and Coprococcus 3 and positively correlated with Acetitomaculum and Helicobacter significantly. The social novelty index correlated positively with the abundance of the genus Anaerofustis significantly. Whereas the time spent interacting with a stranger mouse in Phase 2 of the three chambers test was negatively correlated with Bacteroides abundance, it was signifcanlty positively correlated with the abundance of Candidatus Arthromitus. The time spent interacting with a novel intruder mouse during Phase 3 of the three chambers test was significantly positively correlated with the abundance of Anaerofustis, ASF356, Candidatus Arthromitus, Dubosiella and Enterorhabdus. Similarly, the total interaction time with the familiar and novel intruder in Phase 3 of the three chambers test was also significantly positively correlated with the abundance of Anaerofustis, ASF356 and Enterorhabdus.

FIGURE 6

Correlations between gut microbiome composition shifts at the genus level and behavioural phenotypes. DSI (direct social interaction); TCT (three chambers test of sociability); TintT (total interaction time in Phase 2 + Phase 3 of the TCT). *P < 0.05. The legend on the right shows the colour-coding of the correlations' r values: Blue colours indicate a positive correlation (r > 0); red colours indicate a negative correlation (r < 0).

3.7 Correlations between gut microbiome and eCBome changes in the hippocampus

Lastly, we investigated correlations between gut microbiome changes observed and shifts in hippocampal eCBome mediators (Figure 7). The abundance of several bacterial families (Bacteroidaceae, Christensenellaceae, Clostridales vadinBB60 group, Enterobacteriaceae, Lachnospiraceae, Marinfilaceae and Muribaculaceae) was significantly positively correlated with hippocampal anandamide, whereas the abundance of Bacteroidaceae, Family_XIII, Lachnospiraceae and Marinfilaceae was significantly positively correlated with hippocampal LEA levels. EPEA was significantly negatively correlated with the abundance of Clostridaceae_1 and Deferribacteraceae and significantly positively correlated with Family_XIII, Helicobacteraceae, Mycoplasmataceae, Staphylococcaceae and Streptococcaceae. Noteworthy, hippocampal SDA levels were significantly negatively correlated with several bacterial families such as Bacteroidaceae, Bifidobacteriaceae, Eggerthellaceae, Muribaculaceae, Rikenellaceae and Saccharimonadaceae, whereas EPA levels were significantly positively correlated with the abundance of the families Clostridales vadinBB60 group, Eggerthellaceae, Marinfilaceae, Muribaculaceae, Mycoplasmataceae, Rikenellaceae and Saccharimonadaceae.

FIGURE 7

Correlations between gut microbiome changes at the family level and hippocampal endocannabinoidome shifts. *P < 0.05. The legend on the right shows the colour-coding of the correlations' r values: Blue colours indicate a positive correlation (r > 0); red colours indicate a negative correlation (r < 0).

4 DISCUSSION

In this study, we tested whether a repeated LSD regimen that elicits prosocial (de Gregorio et al., 2021) and anxiolytic (de Gregorio et al., 2022) outcomes affects the eCBome, 5-HT metabolism and the gut microbiome. LSD treatment decreased the hippocampal levels of the N-acylethanolamines LEA, anandamide, DHEA and 1/2-docosahexaenoylglycerol (1/2-DHG), as well as PGD₂, PGF_2α and thromboxane B2 (TXB2), a stable metabolite of thromboxane A2, while also decreasing kynurenine and the ratio kynurenine/tryptophan. LSD had less impact on microbiome, decreasing the Shannon alpha diversity, preventing the Firmicutes:Bacteroidetes ratio decrease observed in VEH-treated mice and increasing the abundance of the bacterial taxa Bifidobacterium, Ileibacterium, Dubosiella and Rikenellaceae RC9.

Ours is the first report describing the outcomes elicited by a psychedelic compound on the brain eCBome and gut microbiome and is in partial agreement with three recent reports on the modulation of the humans plasma eCBome following Ayahuasca (a psychoactive brew containing N,N-dimethyltryptamine and β-carbolines) ingestion, in which eCBome mediators increased acutely (90 min) and decreased afterwards (240 min) (dos Santos et al., 2018, 2022; Madrid-Gambin et al., 2022). Two main differences between the current study and these three clinical investigations consist in the drug and treatment timing (repeated LSD for 7 days vs. acute Ayahuasca), and the tissue under investigation (mouse brain vs. human plasma). We adopted a 7-day administration schedule based on the previous observation that acute administration of 30 μg·kg⁻¹, i.p., does not elicit prosocial or anxiolytic effects, whereas repeated dosing has prosocial (de Gregorio et al., 2021) and anxiolytic (de Gregorio et al., 2022) effects. Moreover, whereas acute administration decreases 5-HT neuronal firing in the dorsal raphé, repeated administration potentiates firing via the 5-HT_1A receptor desensitization (de Gregorio et al., 2022). In addition, this regimen modulates DNA methylation in the prefrontal cortex in genomic areas involved with neurotropic, neurotrophic and neuroplasticity signalling (Inserra et al., 2022).

Endocannabinoids are a class of arachidonic acid-derived lipophilic signalling molecules, which act through the cannabinoid receptors CB₁ and CB₂, G-coupled receptors playing a crucial role in neurodevelopment (Viveros et al., 2011), sociability (Folkes et al., 2020), neurogenesis (Fogaça et al., 2018) and stress (Hill et al., 2009). They are synthesized and released by postsynaptic neurons in response to increases in intracellular calcium levels or activation of metabotropic receptors to modulate postsynaptic depolarization (Ohno-Shosaku et al., 2001), by retrogradely inhibiting neurotransmitter release (Wilson & Nicoll, 2001). 5-HT_2A receptor agonists stimulate the formation and release of 2-arachidonyl-glycerol (2-AG) in vitro via phosphatidylinositol-specific phospholipase C activation (Parrish & Nichols, 2006). Given that LSD-induced 5-HT_2A activation activates phospholipase C-β (González-Maeso et al., 2007) leading to increased accumulation of inositol phosphates and intracellular calcium mobilization (Barnes & Sharp, 1999), it seems plausible that LSD might also acutely increase the production of endocannabinoids as observed for Ayahuasca (dos Santos et al., 2018, 2022; Madrid-Gambin et al., 2022) and other 5-HT_2A agonists (Parrish & Nichols, 2006).

Yet, we did not observe any increase in either anandamide or 2-AG levels following repeated LSD in either of the two brain regions analysed. Instead, anandamide and two of its congeners, LEA and DHEA, decreased in the hippocampus. LEA does not potently activate cannabinoid receptors but stimulates/desensitizes TRPV1 channels and antagonizes TRPV2 channels (Raboune et al., 2014; Schiano Moriello et al., 2018), while being up-regulated following acute brain inflammation (Raboune et al., 2014). These TRP channels are expressed in the brain and in the hippocampus in particular, where they appear to play various roles in affective and neuroinflammatory disorders (Cristino et al., 2020; Sawamura et al., 2017). DHEA, also known as synaptamide, exhibits anti-inflammatory and neurotrophic properties mediated by GPR110 (ADGRF1; Park et al., 2019) and weak agonist activity at anti-inflammatory cannabinoid CB₂ receptors (Paton et al., 2020; Yang et al., 2011). Also, the DHEA homologue, 1/2-docosahexaenoylglycerol (1/2-DHG), was decreased by LSD in the hippocampus, possibly indicating a reduction by LSD of the incorporation of the common precursor of these two metabolites, that is, docosahexaenoic acid (DHA), into the membrane phospholipids acting as precursors for N-acylethanolamines and monoacylglycerols (Di Marzo, 2018). Interestingly, the decrease in anandamide, PGD₂ and PGF_2α was paralleled by a decrease in kynurenine and the kynurenine/tryptophan ratio.

The kynurenine pathway of tryptophan metabolism plays a central role in the behavioural sequelae of inflammation, stress and oxidative stress. The first-rate limiting enzymes of this metabolic pathway, tryptophan 2,3-dioxygenase and indoleamine 2,3-dioxygenase, decrease the availability of tryptophan in the brain for the synthesis of 5-HT in favour of the increased formation of kynurenine and neurotoxic downstream metabolites (Comai et al., 2020). These data suggest that in the hippocampus, the decrease in anandamide and endocannabinoid-like lipids and kynurenine level might be the result of an anti-inflammatory effect elicited by LSD. Importantly, the gut microbiota has been shown to modulate the metabolism of tryptophan by the kynurenine pathway and kynurenines are important modulators of the gut–brain axis (Kennedy et al., 2017).

PGD₂, which we showed here for the first time to be decreased after repeated LSD treatment, is produced by mast cells and eosinophils, and signals through the DP₁ and DP₂ G-protein coupled receptors, which are expressed by different subsets of immune and vascular cells, and elicit different intracellular responses (Peebles, 2019). Also, the levels of PGF_2α, which is produced de novo by PGF synthases (PGFS) (Komoto et al., 2004; Peebles, 2019), were found to be decreased in the hipppocampus following repeated LSD, further suggesting an overall down-regulation of the hippocampal prostaglandin system and, potentially, inflammation. Future studies should assess whether the regimen employed here elicits anti-inflammatory effects by measuring cytokine levels, or microglial activation.

Concerning the correlations observed between hippocampal eCBome shifts and behavioural phenotypes, three main trends were observed: (i) Several N-acylethanolamines (PEA, LEA and SEA) were negatively correlated with the time spent interacting with a stranger mouse in the direct social interaction test and positively correlated with the time spent rearing and the number of grooming events during the direct social interaction test; (ii) a similar behaviour was observed in the ‘metabolite–product’ couples anandamide–arachidonic acid and LEA–LA in their positive correlation with grooming events in the direct social interaction test; and (iii), the N-acylethanolamines, anandamide, LEA and SEA, were all negatively correlated with the social novelty index during Phase 3 of the three chambers test and positively correlated with the preference for a familiar intruder mouse in the same phase. These observations reinforce the hypothesis that LSD-induced changes in hippocampal N-acylethanolamine levels may have contributed to, or have been the consequence of, LSD amelioration of sociability.

In this study, we found that a decrease in endocannabinoids was paralleled by an enhancement in social behaviour. This seems in contrast with previous literature because endocannabinoids have been positively linked to social behaviour (Sciolino et al., 2010; Wei et al., 2015; Wei et al., 2017) and are decreased in autism (Aran et al., 2019), even though the endocannabinoid-mediated social behaviour can have a biphasic action (Trezza & Vanderschuren, 2008). We have already demonstrated that LSD enhances social behaviour through 5-HT_2A, AMPA receptors and mTORC1 complex in excitatory neurotransmission (de Gregorio et al., 2021). Consequently, the decrease of endocannnabinoids in the hippocampus could be strictly linked to an anti-inflammatory effect and not necessarily to the outcomes elicited by LSD on social behaviour.

The current paper represents the first study to investigate putative microbiome changes elicited by the repeated administration of a psychedelic. The global composition of LSD-treated mice remained broadly unchanged compared to VEH-treated mice, despite the fact that the gut is rich in 5-HT_2A receptors (Fiorica-Howells et al., 2002) and 5-HT-producing bacteria (Yano et al., 2015), and that in rodents, 70%–80% of radioactively labelled LSD was found in the gut/feces 3 and 12 h post i.p. administration (Boyd et al., 1955). It is possible that the i.p., instead of oral, administration of LSD, although motivated by previous studies showing that this route is optimal to observe prosocial effects in mice (de Gregorio et al., 2021), may have prevented the observation of stronger effects on the gut microbiota. In fact, this administration route differs from clinical or naturalistic studies, in which LSD is usually administered orally and it is expected to have a different pharmacokinetic profile (Passie et al., 2008). Future translational studies should assess whether similar outcomes are achieved following oral administration. Despite this, four taxa, including members of the Erysipelotrichaceae family (Dubosiella and Ileibacterium), the Rikenellaceae RC9 gut group bacterium and Bifidobacterium, significantly increased in relative abundance following LSD treatment, although their relative abundance did not significantly differ from those of VEH-treated mice, suggesting that they may not contribute to the between-group phenotypic differences. It is interesting to note that the VEH-treated mice show a similar, though statistically insignificant, temporal trends in these same taxa. This suggests that LSD may be quickening temporal changes of the microbiome after their introduction into a new animal facility, even though animals were habituated to the testing facility for 2 weeks prior to the start of the protocol and there were no dietary alterations upon introduction to the testing facility. It was previously shown that introduction to a new facility can influence gut microbiota composition over several generations (Choo et al., 2017). For example, the abundance of specific genera including Bifidobacterium (which was increased in LSD- but not in VEH-treated mice) fluctuates over time upon introduction to a new facility (Choo et al., 2017). A similar effect might be responsible for the minor changes observed here in gut microbiota composition over time.

Post-treatment Bifidobacterium abundance was increased following the LSD regimen. Bifidobacterium abundance is decreased in the faeces of children diagnosed with autism spectrum disorder (Coretti et al., 2018; Wang et al., 2011), whereas Bifidobacterium- and Lactobacillus-based probiotics improve autism spectrum disorder and gastrointestinal symptoms by increasing Bifidobacteria (Shaaban et al., 2018) and GABA and decreasing inflammation (Kong et al., 2022). Therefore, the expansion of Bifidobacterium elicited by LSD could represent a mechanism of psychological and gastrointestinal therapeutic improvement in individuals with autism spectrum disorder (Markopoulos et al., 2021; Sigafoos et al., 2007). The Dubosiella genus was also increased by LSD. This genus is part of the family Erysipelotrichaceae and it is increased following stress (Westfall et al., 2021) and negatively associated with colonic and hippocampal inflammation (Fan et al., 2021), suggesting that LSD might have neuroprotective and anti-inflammatory effects. This hypothesis is also suggested by the decrease in generally proinflammatory PGF_2α (Michaud et al., 2014; Rajakariar et al., 2007) and PGD₂ (Xue et al., 2005) within the hippocampus and the increasing trend in the generally anti-inflammatory PGE₁ (Gezginci-Oktayoglu et al., 2016; Tate et al., 1988) in the prefrontal cortex. Erysipelotrichaceae is also decreased in the maternal immune activation model of autism along with the Ruminococcaceae family (Hsiao et al., 2013), a member of which, Ruminococcaceae_UCG-003, we found here to be positively associated with social interaction time in the direct social interaction test. Conversely, we found that Bacteroides abundance was negatively associated with several measures of social behaviour while being positively associated with empty cage preference. This latter finding is in line with reports that this genus is increased in the cecum of a genetic autism mouse model in association with altered social behaviour (Golubeva et al., 2017). These data also suggest the interesting possibility that LSD treatment of rodent models of autism spectrum disorder, or indeed patients, may induce changes in gut microbiota composition that revert it to that more representative of control animals or healthy humans, in conjunction with improvements in social behaviour (Markopoulos et al., 2021; Sigafoos et al., 2007).

LSD decreased gut microbial diversity (Shannon diversity) and prevented the decrease in the Firmicutes:Bacteroidetes ratio observed in VEH-treated mice. This latter could potentially be an effect of the stress induced by the multiple injections, exposure to the experimenter or the fact that the animals were exposed repeatedly to the behavioural phenotyping facility. LSD prevention of this trend may suggest an enteroprotective effect, which should be further explored. However, the finding that LSD decreased Shannon diversity should also be further explored to assess potential damaging effect of LSD over gut homeostasis.

The findings presented here should be interpreted in light of some limitations: (i) the study was correlative; therefore, we cannot attribute causality nor a sequence to the changes observed; (ii) endocannabinoids, endocannabinoid-like mediators and prostaglandins were quantified in the whole prefrontal cortex and hippocampus, thus not considering the different cell populations; (iii) we did not measure prefrontal cortex and hippocampal levels of downstream metabolites of kynurenine with neurotoxic or neuroprotective activities due to methodological limitations; and (iv), the study was performed in naïve wild-type mice; therefore, the findings cannot be generalized to mouse models of autism spectrum disorder. Validation studies in rodent models of autism spectrum disorder are warranted. In spite of these limitations, our study suggests for the first time that psychedelics can affect the hippocampal eCBome and 5-HT system while impacting the gut microbiome. Future studies should assess the translational value of the present findings and the mechanism of action by which LSD interacts with the eCBome and gut microbiota.

ACKNOWLEDGEMENTS

We wish to thank Dr. Fadil Dahhani for lipidomics sample processing. We also wish to thank Dr. Charlène Rosine Roussel for the help in calculating the Shannon alpha diversity index.

CONFLICTS OF INTEREST

D.D.G. is a consultant at Diamond Therapeutics Inc., Toronto, ON, Canada. D.D.G. and G.G. are inventors of a pending patent on the method of use of LSD.

AUTHOR CONTRIBUTIONS

Antonio Inserra: Conceptualization; data curation; formal analysis; funding acquisition; investigation; methodology; writing-original draft; writing-review and editing. Giada Giorgini: Data curation; formal analysis; investigation; visualization. Sebastien Lacroix: Data curation; formal analysis; visualization. Antonella Bertazzo: Formal analysis. Jocelyn Choo: Data curation; formal analysis; software; visualization; writing-original draft; writing-review and editing. Athanasios Markopolous: Investigation. Emily Grant: Investigation. Armita Abolghasemi: Investigation. Danilo De Gregorio: Funding acquisition; writing-original draft. Nicolas Flamand: Investigation. Geraint Rogers: Conceptualization; supervision; writing-original draft; writing-review and editing. Stefano Comai: Data curation; formal analysis; investigation; writing-original draft; writing-review and editing. Cristoforo Silvestri: Conceptualization; data curation; formal analysis; funding acquisition; investigation; writing-original draft; writing-review and editing. Gabriella Gobbi: Conceptualization; funding acquisition; investigation; methodology; project administration; resources; supervision; writing-original draft; writing-review and editing. Vincenzo Di Marzo: Conceptualization; funding acquisition; investigation; methodology; project administration; resources; supervision; writing-original draft; writing-review and editing.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design & Analysis and Animal Experimentation, and as recommended by funding agencies, publishers and other organizations engaged with supporting research.

Open Research

DATA AVAILABILITY STATEMENT

Data will be rendered available upon request to the corresponding authors.