First Time View on Human Metabolome Changes after a Single Intake of 3,4-Methylenedioxymethamphetamine in Healthy Placebo-Controlled Subjects

MDMA (3,4-methylenedioxymethamphetamine, “ecstasy”) is a widely used recreational psychoactive that releases monoamines and produces marked sympathomimetic and endocrine effects. Despite knowledge of its acute pharmacology and some animal metabolomic data, little is known about how MDMA acutely alters the human metabolome. Previous human work has been limited by forensic or retrospective sampling without controlled conditions, and metabolomics faces substantial challenges from interindividual and intraindividual variability arising from diet, genetics, lifestyle and sampling conditions. Boxler and colleagues analysed plasma from a randomized, double-blind, placebo-controlled crossover study to characterise for the first time acute changes in endogenous plasma metabolites after a single oral dose of MDMA (125 mg). Each participant served as their own control; the investigators used a targeted metabolomics kit (AbsoluteIDQ p180, Biocrates) to quantify up to 188 metabolites and aimed to detect time-dependent metabolite changes and potential biomarkers while accounting for interday variability.

Stored plasma samples from a previously conducted double-blind, placebo-controlled crossover clinical study were analysed. Sixteen healthy volunteers (eight male, eight female; age 20–27) participated in the original study, but metabolomics measurements were reported for 15 participants. Each subject completed sessions that included MDMA (125 mg orally, ~1.8 ± 0.2 mg/kg) and placebo, with a washout of at least 10 days. Blood was sampled 2 h before dosing (time point 0, baseline) and at 3 h (tp1), 8 h (tp2) and 24 h (tp3) after administration; additional pharmacokinetic samples existed but were not used here. Participants were observed during sessions and received a standardised meal at 12:30 pm. Plasma samples were stored at −80 °C for up to 2.5 years and underwent no more than two freeze–thaw cycles. A targeted metabolomics assay combining liquid chromatography–tandem mass spectrometry (UHPLC-MS/MS) and flow-injection analysis (FIA-MS/MS) was performed using the AbsoluteIDQ p180 kit. The kit allows quantification of up to 188 metabolites across amino acids, biogenic amines, acylcarnitines, hexoses, glycerophospholipids and sphingolipids. Samples were randomised across kit plates, 10 μL per sample were processed with stable isotope internal standards, derivatised, and analysed on a Thermo Fisher UHPLC coupled to a Sciex 5500 QTrap. Three kit plates (same batch) were used to measure 80 samples each; samples from the same participant were measured on the same plate and QC medium replicates were included. Data preprocessing normalised concentrations to the median QC medium per plate and log-transformed values to approach normality. Metabolites present in fewer than 80% of samples were removed; remaining missing values were imputed as half the minimum positive value. Nineteen metabolic ratios were also calculated. Baseline-corrected differences (tp1–tp0, tp2–tp0, tp3–tp0) were used to address interday variability. Statistical testing included paired t tests (p < 0.01 threshold, false discovery rate correction), two-way repeated measures ANOVA (within-subject) with FDR correction, and multivariate analyses (principal component analysis, PCA; partial least-squares discriminant analysis, PLS-DA). PLS-DA was validated with cross-model validation and permutation testing (1000 permutations); variable importance in projection (VIP) scores summarised metabolite contributions, with VIP ≥ 1.5 considered largely important.

From 188 assayed metabolites, 141 passed quality control and were analysed further (6 acylcarnitines, 21 amino acids, 13 biogenic amines, 88 glycerophospholipids, 12 sphingolipids, 1 sugar). Six metabolites (serine, 5-HT, spermine, symmetric dimethylarginine, PC aa C34:1, and SM C18:0) had analytical between-batch coefficients of variation >20% and were excluded. Interday and intraday QC repeatabilities were <20% for most metabolites, with SDMA and PC aa C34:1 showing slightly higher interday CVs. Two-way repeated measures ANOVA on baseline-corrected data identified nine metabolites that differed significantly between MDMA and placebo sessions (p < 0.05) and 23 metabolites with a session-by-time interaction; 88 metabolites changed across time independently of MDMA. The nine session-differing metabolites were PC aa C26:0, PC aa C38:1, PC aa C40:1, PC ae C30:2, PC ae C42:2, C18:1 (a free fatty acid), lysoPC a C26:1, lysoPC a C28:1, and the metabolic ratio methionine‑sulfoxide over methionine (Met‑SO/Met). Time-point analyses showed substantial interday variability. At baseline (tp0, 2 h before dosing) multiple metabolites (including several PCs, PCaes, Met and Met‑SO) already differed between sessions, indicating high interday individuality. At 3 h after dosing, paired t tests on uncorrected data suggested many differences (64 metabolites), but after baseline correction 23 metabolites remained significantly different between sessions, predominantly glycerophospholipids (numerous PC aa and PC ae species), asparagine and the Met‑SO/Met ratio. Multivariate PLS-DA did not produce significant separation at 3 h. At 8 h uncorrected data yielded 64 significant metabolites and PLS-DA showed apparent separation with R2 = 0.88 and Q2 = 0.75; permutation testing supported the model (p = 0.007). Important contributors (VIP >1.5) were mainly glycerophospholipids (multiple PC aa and PC ae species), several lysoPC a species, Met, Met‑SO and Met‑SO/Met. When baseline-corrected, 30 metabolites remained significant by paired testing (similar lipid species, C16 and C18:1 fatty acids, glycine, Met‑SO/Met, and Tyr/Phe), but permutation testing indicated overfitting (p = 0.893) despite cross-validation statistics suggesting a fit. At 24 h uncorrected paired tests showed significant changes in 42 metabolites; PCA detected one outlier that was removed. PLS-DA cross-validation (R2 = 0.54, Q2 = 0.29) and permutation testing (p = 0.005) suggested model validity for the uncorrected data, with VIP >1.5 for many glycerophospholipids and Met‑SO/Met. However, after baseline correction no significant differences were detected between sessions at 24 h and PLS-DA did not separate groups. Across analyses the most consistent finding was an increased Met‑SO/Met ratio after MDMA, interpreted as an index of systemic oxidative stress. Glycerophospholipids (diacyl- and acyl‑alkyl‑phosphatidylcholines) and certain lysophosphatidylcholines increased after MDMA in several analyses, suggesting altered lipid metabolism or lipoprotein changes. Tryptophan and serotonin (5‑HT) did not show consistent MDMA-specific changes in this targeted assay; the authors note methodological difficulties measuring 5‑HT in plasma due to platelet lysis. Overall, many amino acids and other metabolites varied with time independently of MDMA.

Boxler and colleagues interpret their principal finding—a rise in the Met‑SO/Met ratio after MDMA—as evidence of increased systemic oxidative stress following a single MDMA dose. The biochemical rationale offered is that methionine is oxidised by reactive oxygen species (ROS) to methionine sulfoxide, and the observed ratio change aligns with mechanistic pathways by which MDMA metabolism can generate redox‑active catecholamines and ROS. The authors note prior animal data and limited human retrospective findings and place their results as the first controlled human metabolomics profile after MDMA intake. The investigators also observed increased concentrations of multiple glycerophospholipids and some lyso‑phosphatidylcholines after MDMA, which they suggest may reflect altered plasma lipoprotein composition, cellular function or an upregulation of energy production to meet increased metabolic demands (for example cardiac or muscular β‑oxidation). They comment that lysoPC increases could be relevant to vascular risk because lysoPCs are components of oxidised LDL and affect endothelial permeability; however, causal links to MDMA‑related cardiovascular events remain speculative. Several important methodological limitations are acknowledged. The targeted Biocrates kit measures a defined panel and does not capture the full metabolome; many glycerophospholipids and sphingolipids are measured as sum species without fatty acid positional or stereo information. Storage up to 2.5 years and limited freeze–thaw cycles were judged acceptable but remain potential sources of variability. The sample size was modest (15 analysed participants), and the original clinical protocol did not strictly control pre-session sleep, activity or fasting; these factors and pronounced interday individuality substantially affected results. The authors emphasise that baseline (zero) samples are crucial: many apparent differences disappeared after baseline correction, and analyses without baseline adjustment risk false positives or overestimation. Methodologically, a combination of univariate and multivariate statistics was necessary; PCA was useful to detect outliers and PLS‑DA could separate groups in some uncorrected analyses but often showed overfitting once baseline variability was accounted for. The authors conclude that a single MDMA exposure produced modest, relatively nonspecific changes in the targeted metabolites, that Met‑SO/Met is the most robust MDMA‑associated signal in this data set, and that untargeted metabolomics or studies of repeated use might be required to identify more specific biomarkers. They also note ethical and practical constraints on performing controlled repeated‑dose studies with illicit drugs.

An increased Met‑SO/Met ratio indicates an overall rise in oxidative stress after a single MDMA dose. Many amino acids exhibited time‑dependent fluctuations unrelated to MDMA, emphasising the importance of baseline samples because interday individuality within subjects was high. A single intake did not strongly or specifically alter the targeted metabolite panel, and the metabolites that did change are involved in multiple pathways and thus are not suitable as specific biomarkers. The authors suggest that studies of repeated or frequent MDMA intake combined with untargeted metabolomics may better characterise metabolome changes.