In Vivo Imaging of Cerebral Serotonin Transporter and Serotonin 2A Receptor Binding in MDMA and Hallucinogen Users

MDMA (3,4-methylenedioxymethamphetamine, “ecstasy”) and classical hallucinogens (for example, LSD and psilocybin) act on serotonergic neurotransmission, but their pharmacologic profiles differ. MDMA is a SERT substrate that induces serotonin release, inhibits reuptake and serotonin synthesis, and also has serotonin 2A receptor agonist activity. By contrast, hallucinogens are principally potent serotonin 2A receptor agonists and are not generally associated with long‑term serotonin depletion. Previous animal and human work has shown reduced cerebral serotonin levels and fewer SERT binding sites after moderate-to-heavy MDMA exposure, whereas the neurobiological consequences of recreational hallucinogen use are less well characterised. Erritzoe and colleagues set out to test whether recreational MDMA use and hallucinogen use have differential effects on presynaptic and postsynaptic serotonergic markers in vivo. Using PET imaging of the serotonin transporter (SERT) with [11C]DASB and of serotonin 2A receptors with [18F]altanserin, the study compared young adult users (subdivided into MDMA-preferring and hallucinogen-preferring groups) with age- and sex-matched nonusing controls. The investigators hypothesised that SERT binding would be reduced in MDMA users but not in hallucinogen-preferring users, and that serotonin 2A receptor binding would be reduced in both groups, with more pronounced SERT reductions after shorter abstinence periods.

Participants were recruited by advertisements, fliers and word of mouth and screened with structured interviews and substance-use questionnaires. Eligible individuals were 18–35 years old with at least 12 lifetime exposures to MDMA or hallucinogens and use within the year before scanning; users had to abstain for at least 7 days before PET to avoid acute receptor/transporter occupancy. Exclusion criteria included major psychiatric or neurologic disorders and prior use of antidepressants or antipsychotics. Controls were excluded if they reported more than 15 lifetime cannabis exposures or any other illicit drug use history. Written informed consent and local ethical approval were obtained. Twenty-four drug users (21 men, 3 women; mean age 24.6 years) and 21 controls (17 men, 4 women; mean age 24.0 years) were enrolled. Users were divided into MDMA-preferring users (MPUs, n=14) and hallucinogen-preferring users (HPUs, n=10) according to the ratio of lifetime exposures. Recent abstinence was confirmed with urine screens every 2–3 days for 7 days before scanning, and hair segments covering the prior ~3 months were analysed by GC–MS to corroborate self-reported MDMA use. All participants underwent MR imaging to allow precise VOI delineation and two PET scans: [11C]DASB for SERT (dynamic 90-minute acquisition) and [18F]altanserin for serotonin 2A receptors. PET scanning was performed on an 18-ring GE Advance scanner; for most participants the two scans were on the same day, though some had a median 5-day interval. Outcome measures were nondisplaceable binding potential (BP_ND) for [11C]DASB and binding potential of specific tracer binding (BP_P) for [18F]altanserin, using the cerebellum as a reference region. Due to technical issues with metabolite analysis, reliable [18F]altanserin quantification was available for 21 users and 20 controls. Volumes of interest were automatically delineated on individual MR images and included a volume-weighted neocortical composite (primary outcome), pallidostriatum, midbrain, amygdala and thalamus for SERT analyses. Statistical analysis used nonparametric tests for group comparisons where appropriate and one‑way ANOVA with Tukey correction for multiple comparisons. Potential confounders (age, BMI, daylight minutes on scan day, education, and recent use of other stimulants) were tested as covariates and were not retained in final models because they did not materially alter results. Dose–response and recovery relationships were explored using logarithmic transforms of lifetime tablet counts or lifetime hallucinogen exposures and with a log‑logistic four‑parameter model; time since last use was also modelled, with adjustment for lifetime MDMA intake when testing recovery of SERT binding.

Sample characteristics comprised 24 drug users (14 MPUs, 10 HPUs) and 21 controls. Hair testing confirmed self-reported recent MDMA use in 13 of 14 participants who reported MDMA within the prior ~3 months; urine tests documented at least 11 days’ abstinence before PET scans. Because of assay failure, [18F]altanserin data were available for 21 users and 20 controls. SERT binding (measured as BP_ND) differed by group in several regions. MDMA-preferring users had significantly lower SERT BP_ND than both HPUs and controls in the pallidostriatum (group F=7.89, P=.001), neocortex (F=23.05, P<.001) and amygdala (F=16.58, P<.001). Relative to controls, MPUs showed reductions of 56% in neocortex, 19% in pallidostriatum and 32% in amygdala. Regional neocortical reductions were large and heterogeneous (for example, 66% in sensorimotor cortex and 73% in occipital cortex). No significant group effect was seen in the midbrain; the thalamus showed a borderline group effect (F=2.6, P=.08). Inclusion of demographic and other drug‑use covariates did not materially change these findings. Dose–response analyses indicated a negative correlation between lifetime MDMA intake (log-transformed tablets) and SERT BP_ND in all regions except midbrain. A doubling of lifetime tablet consumption was associated with decreases in BP_ND of 0.021 in neocortex (95% CI, -0.029 to -0.014; P<.001), 0.062 in pallidostriatum (95% CI, -0.103 to -0.021; P=.005), 0.067 in thalamus (95% CI, -0.113 to -0.022; P=.006), and 0.075 in amygdala (95% CI, -0.106 to -0.045; P<.001). Time since last MDMA use (log2 transformed) was positively associated with pallidostriatal SERT BP_ND, and both lifetime intake and time since last use remained significant predictors in extended models except in the thalamus where the abstinence effect lost significance when lifetime use was included. For serotonin 2A receptor binding, the combined group of serotonin 2A agonist users (MPUs and HPUs together) showed a modest 9% decrease in neocortical BP_P compared with controls. However, this difference lost statistical significance when two control participants with very high BP_P values were excluded (Mann‑Whitney P=.10). Analysed separately, MPUs and HPUs did not differ significantly from controls in neocortical serotonin 2A binding (Kruskal‑Wallis P=.07). No consistent dose‑response or abstinence relationships for serotonin 2A binding were detected. Finally, within the user group there was a significant inverted U‑shaped (quadratic) relationship between neocortical serotonin 2A binding and pallidostriatal SERT binding (second‑order term P=.046). Overall, hallucinogen‑preferring users showed normal SERT binding, whereas MDMA‑preferring users exhibited pronounced reductions with a dose‑related pattern and partial subcortical recovery with longer abstinence.

Erritzoe and colleagues interpret the findings as evidence that MDMA use—but not primary hallucinogen use—is associated with marked reductions in cerebral SERT binding, particularly in cortical regions. The investigators argue that these changes are most plausibly explained by direct presynaptic actions of MDMA (and secondary serotonin depletion) rather than by serotonin 2A receptor agonism, because hallucinogen users (who share the 2A agonist effect) did not show reduced SERT binding. A dose–response relationship between lifetime MDMA exposure and regional SERT BP_ND supports a link between extent of use and SERT availability, and modelling of abstinence suggested partial recovery of subcortical SERT over months, whereas cortical SERT appeared the least likely to recover within the observed intervals. The authors place their results in the context of prior animal and human imaging studies that report cortical vulnerability to MDMA-related reductions in serotonergic markers and mixed evidence for recovery depending on brain region. They note that the observed SERT changes could reflect either axonal damage with subsequent reinnervation or adaptive downregulation of transporter expression; their cross‑sectional data do not allow distinguishing these mechanisms. Regarding serotonin 2A receptors, the modest cortical decrease in agonist users may reflect transient receptor downregulation after recent stimulation, compensatory changes related to altered synaptic serotonin, or could be an artefact influenced by outlying control values; subgroup analyses did not yield robust effects. Key limitations acknowledged by the investigators include the cross‑sectional design, which limits causal inference and cannot exclude preexisting differences that predispose to MDMA use. Self‑reported drug histories are imperfect, although hair and urine assays were used to corroborate recent use and abstinence. Methodological caveats include the use of a reference tissue model without arterial sampling for SERT quantification and relatively low cortical signal‑to‑noise for [11C]DASB, which increase variability. The modest sample size and the presence of two high‑value outliers in the serotonin 2A dataset reduce confidence in the small cortical 2A effect. Potential confounders such as tobacco, cocaine and amphetamine use were examined and did not account for the main findings. Finally, the authors recommend caution in interpreting decreased SERT availability as neurotoxicity because reductions in serotonergic markers can arise without neuronal loss. They propose that longitudinal and multimodal studies linking PET markers with functional and structural measures are needed to clarify the clinical significance and the reversibility of MDMA‑associated changes.