Bioisosteric analogs of MDMA: Improving the pharmacological profile?

3,4-Methylenedioxymethamphetamine (MDMA) is a ring-substituted amphetamine with psychostimulant and prosocial effects that has re-emerged in clinical research, notably for post-traumatic stress disorder (PTSD) when combined with psychotherapy. MDMA's pharmacology includes interaction with serotonin (SERT), dopamine (DAT) and norepinephrine (NET) transporters, causing non-exocytotic efflux of these monoamines, and agonism at 5-HT2A/2B/2C receptors. However, MDMA is associated with acute and chronic adverse effects (cardiovascular, hyperthermia, hepatotoxicity) and complex, partly non-linear metabolism; the methylenedioxy group has been implicated in inhibition of CYP enzymes (notably CYP2D6) and in formation of catechol/quinone metabolites that can generate reactive oxygen species and may contribute to neurotoxicity in preclinical models. Sofia and colleagues set out to design and characterise three MDMA bioisosteres in which the methylenedioxyphenyl moiety is replaced by 2,1,3-benzoxadiazole, 2,1,3-benzothiadiazole, or 2,1,3-benzoselenadiazole, yielding ODMA, TDMA and SeDMA, respectively. The study aimed to compare these analogs with MDMA at key molecular targets (hSERT, hDAT, hNET; 5-HT2A/2B/2C receptors; human organic cation transporters (hOCT1-3); plasma membrane monoamine transporter (hPMAT)) and to characterise their in vitro hepatic metabolism, with the rationale that bioisosteric replacement might preserve therapeutic transporter activity while reducing CYP inhibition and metabolite-related toxicity.

The investigators performed a suite of in vitro assays using human cell systems and in silico modelling; complete procedural detail is reported in the Supplementary Information. Human embryonic kidney 293 (HEK293) cells stably expressing human isoforms of SERT, DAT, NET, OCT1-3 and PMAT were used for uptake-inhibition and release assays. Radiolabelled tracers were [3H]5-HT for hSERT and [3H]MPP+ for hDAT, hNET, hOCTs and hPMAT. Uptake-inhibition curves were generated from at least three independent cell culture preparations (n ≥ 3), fitted to sigmoidal dose–response models to obtain IC50 values; hDAT/hSERT ratios were computed as 1/(hDAT IC50) : 1/(hSERT IC50). Release assays employed monensin to distinguish transporter substrates from non-transported inhibitors, and area under the curve (AUC) from 8–18 min was used to quantify efflux. Statistical comparisons for release assays used mixed-effects models with Šidák correction. Electrophysiology included whole-cell patch-clamp recordings in HEK293 cells overexpressing hSERT or hDAT (Vh = −60 mV) and two-electrode voltage clamp recordings in Xenopus laevis oocytes injected with transporter cRNA, to detect substrate-elicited steady-state currents. Serotonin receptor activity (5-HT2A/2B/2C) was characterised using Gq-dissociation bioluminescence resonance energy transfer (BRET) assays in HEK293T cells, FLIPR-based calcium flux (Fluo-4), and a GCaMP6s reporter system; surface receptor expression was quantified by anti-FLAG ELISA. Molecular docking used the hSERT crystal structure and homology models of hDAT, with proteins prepared via molecular dynamics simulations and ligands docked with GOLD. Hepatic metabolism studies used pooled human liver microsomes (pHLM) and pooled human liver S9 fraction (pS9) incubations, analysed by liquid chromatography–high-resolution tandem mass spectrometry (LC‑HRMS/MS) to identify phase I and II metabolites and to perform isozyme mapping against a panel of CYPs and FMO3. Metabolic stability (substrate depletion) incubations with pHLM assessed in vitro half-lives (t1/2) and intrinsic clearance parameters; a 150 min cut-off was applied because enzyme activity declines after ~2 h. Interaction with hOCT1-3 and hPMAT was assessed by uptake-inhibition assays similar to those for the high-affinity transporters. Replication across assays was generally three or more independent preparations, with assays run in duplicate or triplicate as stated.

Uptake inhibition: MDMA and the three bioisosteres inhibited substrate uptake at hSERT, hDAT and hNET in HEK293 cells at low micromolar concentrations. Relative potencies were broadly similar to MDMA; ODMA was approximately 2-fold less potent at hSERT, whereas SeDMA showed inhibitory potency similar to MDMA. Calculated hDAT/hSERT ratios were low (<10) across compounds, an index the authors associate with lower predicted abuse liability. Release assays and electrophysiology: Using monensin-potentiated release assays, all compounds behaved as transporter substrates/releasers. They produced robust [3H]5-HT release via hSERT and potent [3H]MPP+ release via hNET, while eliciting more moderate release at hDAT. Monensin significantly potentiated efflux at hSERT (confirmed by mixed-effects modelling with Šidák correction), but potentiation at hNET was not significant. Electrophysiological recordings showed full-efficacy substrate profiles at hSERT (sustained inward currents during application) and partial-efficacy profiles at hDAT, consistent with stronger functional engagement of SERT than DAT. Molecular docking: Docking poses for MDMA and the analogs overlapped substantially with the natural substrates at both hSERT and hDAT. The positively charged amine oriented toward the transporter bundle domain, forming electrostatic and hydrogen-bond interactions with conserved residues (e.g. D98 and S438 at hSERT; D79 and F320 backbone at hDAT), while the aromatic ring engaged the scaffold domain. The similarity of interacting residues with those known to bind 5-HT and dopamine supports the substrate-like behaviour observed experimentally. 5-HT2 receptor activity: MDMA activated 5-HT2A, 5-HT2B and 5-HT2C receptors more potently than the bioisosteres. All three analogs (ODMA, TDMA, SeDMA) showed weaker potency and efficacy in Gq‑dissociation BRET assays and in calcium-flux assays; SeDMA was the weakest agonist. The bioisosteres were approximately 10-fold less potent than MDMA at 5-HT2B and 5-HT2C in these assays. Docking at 5-HT2 receptors indicated different, albeit only modestly distinct, binding poses across compounds and receptor subtypes. Hepatic metabolism and stability: LC‑HRMS/MS-based metabolite identification showed that N‑demethylation (and in some cases N‑hydroxylation) was the only metabolic transformation shared between MDMA and the three analogs. The bioisosteric replacements prevented demethylenation of the aromatic ring found with MDMA, thus precluding formation of catechol metabolites and subsequent phase II conjugates that are reported for MDMA. In metabolic stability assays with pHLM, TDMA had an in vitro half-life of 53 min and a scaled intrinsic clearance (CLint) of 11.2 mL/min/kg. For MDMA, ODMA and SeDMA, half-lives exceeded the 150 min assay cut-off, so clearance values could not be calculated under these conditions. Isozyme mapping and transporter interactions: Mapping indicated involvement of multiple CYP isozymes in phase I metabolism of the analogs, including CYP2D6 and CYP1A2 among others; the extracted text lists CYP1A2 and CYP2D6 involvement and notes additional CYPs were evaluated, but the detailed table is not fully present in the extraction. Interaction screening with low-affinity, high-capacity transporters showed that MDMA and its analogs interact with hOCT1-3 and hPMAT at low micromolar concentrations; overall, the analogs displayed weaker interactions with hOCT1, hOCT2 and hPMAT compared with MDMA. Full numerical IC50 values and isozyme details are reported in the paper’s tables and Supplementary Information.

Sofia and colleagues interpret their findings as demonstrating that bioisosteric replacement of MDMA's methylenedioxy ring preserves primary activity at high‑affinity monoamine transporters while reducing agonist activity at 5-HT2 receptor subtypes and altering hepatic metabolism. The preserved low micromolar interaction with hSERT, hDAT and hNET, together with a preference for full substrate activity at hSERT and partial activity at hDAT, aligns with a pharmacological profile that may retain prosocial and therapeutic effects but carry lower dopaminergic-driven reinforcing potential; the authors link the low hDAT/hSERT ratios to lower predicted abuse liability. Reduced potency at 5-HT2A receptor and notably weaker agonism at 5-HT2B and 5-HT2C receptors are highlighted as potentially favourable. Weaker 5-HT2A agonism could correspond to decreased hallucinogenic potential, and diminished 5-HT2B activation could reduce risk factors associated with valvular heart disease; the authors note, however, that cardiotoxic risk also depends on NET interaction and so further evaluation is required. The hepatic metabolism data are framed as especially important: because the analogues cannot undergo demethylenation, they do not form the catechol and phase II metabolites typical of MDMA that have been implicated in oxidative stress and some aspects of MDMA-related toxicity. The authors therefore suggest the bioisosteres may generate fewer reactive metabolites, though they emphasise this must be tested further. The discussion acknowledges limitations and outstanding questions. In vitro systems were used throughout, so in vivo pharmacokinetics, pharmacodynamics, toxicity and behavioural effects remain to be established. The study could not fully characterise clearance for compounds with half-lives beyond the 150 min assay window, and the detailed CYP isozyme contributions require further work to predict drug–drug interaction (DDI) risk and interindividual variability; the extracted text notes CYP2D6 and CYP1A2 involvement and indicates mapping across a panel of CYPs but the complete mapping details are reported in tables. The authors also raise the possibility that binding kinetics (e.g. slow on/off rates) might influence DAT partial efficacy and in vivo effects, recommending further kinetic and in vivo studies. Overall, the authors conclude that ODMA, TDMA and SeDMA warrant additional investigation as potential alternatives to MDMA that retain transporter-mediated pharmacology while showing reduced 5-HT2 agonism and altered hepatic metabolism, but they caution that in vivo and safety studies are necessary to confirm any risk‑reduction.