History, pharmacology and therapeutic mechanisms of 3,4-methylenedioxymethamphetamine (MDMA)

Abstract

The illicit drug 3,4-methylenedioxymethamphetamine (MDMA) has recently shown promising efficacy as an adjunct to psychotherapy for posttraumatic stress disorder (PTSD), although the underlying mechanisms are poorly defined. In this review, we contextualize the emergence of MDMA-assisted psychotherapy (MDMA-AT) within 20th century psychiatry and trace its journey from assisting with military interrogation, through prohibition and to clinical use. We outline three core domains of MDMA’s subjective effects—prosocial behaviour, reduced threat perception and euphoria—and explore how each may relate to its therapeutic efficacy and abuse liability. Drawing from clinical, behavioural, and pharmacological studies, we highlight the central role of 5-HT (serotonin) in mediating the effects of MDMA, while identifying key gaps in our understanding of its mechanism of action. We also assess how preclinical models capture therapeutic-relevant processes, discuss the limitations of existing data (including sex biases and an assumed role for therapeutic alliance) and suggest strategies to unravel the neurobiological basis of MDMA’s therapeutic effects. Clarifying these mechanisms will be critical for optimizing future clinical protocols, mitigating risks and guiding the clinical development of safer, mechanistically informed MDMA-like therapeutic agents.

1 |. INTRODUCTION

The synthetic phenethylamine 3,4-methylenedioxymethamphetamine (midomafetamine, MDMA) has a rich and somewhat apocryphal history marked by an uneventful discovery of shelved chemical matter (Freudenmann et al., 2006) that was later explored as an adjunct for military interrogation (Passie & Benzenhöfer, 2018) and psychotherapy (Shulgin & Shulgin, 1990), prior to strict and enduring prohibition by the United States and world governments. Over the last 40 years, steadfast proponents for the therapeutic use of MDMA led a grass roots effort to fund several small-to-moderate-sized clinical trials seeking legitimization of the drug through approval by the US Food and Drug Administration (FDA). These trials indicated that MDMA may be an effective adjunct to psychotherapy for the treatment of post-traumatic stress disorder (PTSD), a debilitating emotional disorder caused by traumatic experience that is characterized by (1) intrusive symptoms (e.g., recurrent distressing memories), (2) arousal symptoms (e.g., exaggerated startle response and hypervigilance), (3) avoidance of trauma cues and (4) negative alterations in cognition and mood (e.g., overgeneralized memory, anhedonia and social withdrawal).

In this review, we discuss three overarching themes associated with the subjective effects of MDMA based on decades of human observational and psychological studies: (1) prosocial effects described as ‘closeness’ or ‘empathy’, (2) reduced threat perception and (3) euphoria. We examine how these frequently reported subjective effects of MDMA are reflected in animal models and align with in vitro pharmacology and neural circuit mechanisms. We also summarize the current state of MDMA clinical development and identify important gaps in the current research landscape that, if addressed, could enhance our understanding of neural mechanisms underlying trauma recovery.

2 |. EMERGENCE OF MDMA AS AN ADJUNCT TO PSYCHOTHERAPY

The idea that pharmacological agents may be useful adjuncts for psychotherapy began in the 1930s and represented a paradigm shift in psychiatry. Reports that pharmacological shock (i.e., drug-induced convulsive seizures) could treat early-stage schizophrenia and depression (Bennett, 1939; Finkelman et al., 1938) provided strong support for a biological origin of mental illness (Grinker & Mclean, 1940). Pentylenetetrazol-induced convulsions soon became a popular method of pharmacological shock and were thought to unblock pathways in the cortex, facilitating verbal communication and releasing repressed psychological content (Grinker & Mclean, 1940). Similar observations popularized the use of sodium pentothal-assisted psychotherapy for treating PTSD-like symptoms in World War II veterans (Heath & Sherman, 1944). Psychiatrists then began experimenting with drug combinations, including alternating sodium amytal and amphetamine treatments, to render patients more accessible to psychoanalysis. By the 1940s, methamphetamine was considered an improvement over amphetamine (Levine et al., 1948) for treating depressed and psychotic patients, with clinicians citing greater euphoria, verbalization and disclosure of previously unobtainable emotional material (Jonas & Parkway, 1954; Simon & Taube, 1946), a therapeutic approach related to ‘narcosynthesis’ and ‘abreaction’.

Cardiovascular safety concerns associated with methamphetamine highlighted a need for improved therapeutic agents. It had been previously suggested that the psychedelic drug mescaline may serve as an adjuvant to reassure psychiatric patients that their feelings of ‘unreality’ were reversible, thus making them susceptible to psychotherapy (Guttmann, 1936). However, mescaline was minimally explored as an adjunct to psychotherapy, on account of its association with overt hallucinations (Cattell, 1954) and categorization as an endogenous ‘psychotogen’ that caused mental illness (Guttmann, 1936; Osmond & Smythies, 1952; Stockings, 1940). Researchers therefore aimed to combine molecular features of methamphetamine and mescaline to identify an improved adjunct to psychotherapy that lacked the cardiac risk and thought disturbances associated with methamphetamine and mescaline, respectively. These efforts led to the identification of 3,4-methylenedioxyamphetamine (MDA), an apparently effective facilitator of psychotherapy (Naranjo et al., 1967), and predecessor to MDMA.

In the late 1960s, MDA (and soon after, MDMA) was new to the field of psychiatry, however both drugs had already been explored by the pharmaceutical industry (Freudenmann et al., 2006) and US military (see (Passie & Benzenhöfer, 2018) for an extensive review). While mescaline was found to be too psychotogenic to assist interrogation, military researchers synthesized mescaline derivatives, including MDA and MDMA, hoping to retain its putative mind-altering properties with minimal psychotogenic effects (Figure 1). A series of derivatives were then shared with the New York State Psychiatric Institute whereupon researchers hastily administered them to humans without informed consent, resulting in one death after a high dose of MDA (Passie & Benzenhöfer, 2018). The compounds then underwent animal toxicology tests, where it was discovered that derivatives harboring a 3,4-methylenedioxy motif (e.g., MDA and MDMA) posed the greatest risk for toxicity (Hardman et al., 1973), potentially explaining the lack of government investigation into these derivatives, save for one preclinical report (Alles & Fairchild, 1962). There is no direct evidence that the US military administered MDMA to humans.

FIGURE 1.

Structures of example phenethylamine derivatives associated with distinct subjective effects (chiral carbons denoted by *).

The first administration of MDMA to humans is unknown, although it was identified in a 1970 Chicago drug seizure (Gaston & Rasmussen, 1972) and gradually made its way to the west coast, where it was introduced to the chemist Alexander Shulgin (Shulgin & Shulgin, 1990). Shulgin, along with David Nichols, published the first psychopharmacological account of MDMA in humans, describing it as, ‘… an easily controlled altered state of consciousness with emotional and sensual overtones’ (Shulgin & Nichols, 1978). Although no mention was made of psychotherapy, Shulgin is credited with introducing MDMA to the underground psychiatric community in 1976 (Shulgin & Shulgin, 1990). Thereafter, the administration of MDMA for therapeutic purposes was largely performed in the therapist’s home to increase trust (Greer & Tolbert, 1990), and perceived efficacy was disseminated through word-of-mouth rather than academic publications.

Greer, Tolbert and Downing published the first broad, observational case series’—totaling 50 subjects—on the effects of MDMA administered during therapy (Greer & Tolbert, 1986) or at a socially immersive, scenic gathering (Downing, 1986). In addition to an almost unanimous sense of benefit, clients described feeling a greater sense of intimacy with others, greater receptivity to compliments and criticism, enhanced communication and self-confidence, the release of repressed emotions, and euphoria. A potentially important and therapeutically-relevant observation was that MDMA caused seemingly less perceptual distortions compared to other purported psychotherapeutic adjuncts and did not affect cognitive functioning (Grinspoon & Bakalar, 1986). More recent psychopharmacology studies using quantitative rating scales confirmed many of these early qualitative impressions. For example, MDMA elicits a sense of well-being, emotional sensitivity and euphoria (Peroutka et al., 1988; Vollenweider et al., 1998), concordant with its designation as an entactogen, derived from Latin and Greek roots to express compounds that produce a touching within (Nichols, 1986). Moreover, the perceptual distortions elicited by MDMA are milder than that of lysergic acid diethylamide (LSD) (Holze et al., 2020), although MDMA does impair sustained attention, visual–spatial memory and encoding of specific emotional information (Doss et al., 2024).

3 |. CLINICAL DEVELOPMENT OF MDMA

Research into the therapeutic use of MDMA halted abruptly after its prohibition in 1985 and classification as a drug with high abuse potential and no recognized medical use. For the next three decades, most academic work focused on MDMA’s potential for drug abuse and neurotoxicity (Figure 2) (McCann & Ricaurte, 2014). In direct response to the scheduling of MDMA, the non-profit Multidisciplinary Association for Psychedelic Studies (MAPS) was founded to obtain recognized medical use for MDMA in the eyes of the FDA. To garner public support, MAPS sought a patient population that could evoke compassion from the public; based on the opinion of contemporary psychiatrists that MDMA reduced the fear response associated with emotionally threatening information, the candidate populations were identified as those with PTSD or facing terminal illness (; Doblin, 2002). As patients with PTSD were likely to be healthier and taking fewer medications than those with terminal illness, the development of MDMA assisted psychotherapy (MDMA-AT) has focused on PTSD. Thus, the use of MDMA-AT for PTSD was a politically strategic decision, designed to oppose cultural biases and associated bureaucratic barriers that prevented the progression of clinical trials involving MDMA (Doblin, 2002).

FIGURE 2.

PubMed-indexed citations mentioning MDMA in the context of drug abuse and toxicity (red circles) or therapeutic potential (purple squares).

MAPS sponsored over a dozen small Phase II clinical trials assessing the efficacy and safety of MDMA-AT for the treatment of PTSD (Bahji et al., 2020), wherein patients received one to three drug treatments, bookended by psychotherapy sessions. With the exception of one study (Oehen et al., 2013), these small trials reported large reductions in PTSD symptoms, high rates of remission (56%–100%) and efficacy persisting beyond 12 months (Bahji et al., 2020). Two Phase III trials also reported efficacy for MDMA-AT compared to non-assisted psychotherapy for PTSD (n = ~200 across both trials) (Mitchell et al., 2021; Mitchell et al., 2023). Patients treated with MDMA-AT over 18 weeks received three preparation therapy sessions and nine integration sessions interspersed with three experimental drug sessions. Phase III trials reported rapid and enduring reductions in PTSD symptoms, depression symptoms and disability following MDMA-AT, with ~70% of patients treated with MDMA-AT no longer meeting diagnostic criteria for PTSD at study termination, compared to ~40% of patients treated with psychotherapy alone. However, in August of 2024, the FDA declined to approve MDMA-AT for PTSD, most likely influenced by highly publicized critiques, biased outcome reporting by therapists, an episode of therapist sexual misconduct and concerns over adverse event reporting (Colcott et al., 2024; Mustafa et al., 2024).

In addition to PTSD, MDMA-AT has been explored in small clinical trials for the treatment of alcohol use disorder and social anxiety in autism spectrum disorder (Danforth et al., 2018; Sessa et al., 2021), with promising, though preliminary, suggestions of efficacy. Taken together, the clinical trial results suggest that MDMA-AT may represent a novel, transdiagnostic therapeutic modality not just for PTSD, but other brain disorders as well. However, more research is needed to understand how it compares to available treatment options and to what extent expectations contribute to the observed therapeutic effects (Szigeti & Heifets, 2024). Trials with rigorous reporting practices in larger patient populations across a range of doses are needed to more clearly understand the risks associated with MDMA in a therapeutic context. In the meantime, efforts are already underway to develop next generation MDMA-like drugs with ‘entactogenic’ properties, but without substantial abuse potential or toxicity. This drug design effort requires an understanding of mechanism(s), particularly how MDMA produces its putative therapeutic effects. In sections below, we review what is known about MDMA’s mechanism of action, from receptor pharmacology to behaviour, across species.

4 |. MOLECULAR PHARMACOLOGY

Embedded in the molecular structure of MDMA resides a chiral carbon atom, positioned α to the amine (Figure 1), which allows the molecule to adopt one of two three-dimensional arrangements that are non-superimposable mirror images of each other (i.e., enantiomers). Despite this handedness, most studies have used the racemic mixture of (±)-MDMA for historical reasons as well as technical and economic challenges associated with chiral separation. Thus, specific enantiomers in this review will only be mentioned for relevant distinctions.

Early biochemical studies of MDMA suggested that, despite its structural similarity to psychedelic phenethylamines such as mescaline and 2,5-dimethoxy-4-methylamphetamine (DOM, Figure 1), its unique insensitivity to N-methylation (from MDA to MDMA) and stereoselectivity indicated a mechanism distinct from activation of 5-HT (serotonin) receptors (Cheng et al., 1974; Glennon, 1979). Indeed, MDMA blocks reuptake of 5-HT, noradrenaline and dopamine by competing as a substrate for the respective monoamine transporter proteins (SERT, NET and DAT) located on the nerve terminal (Verrico et al., 2007), a process driven by ion gradients of sodium, chloride and potassium (Rudnick & Wall, 1992; Schmidt et al., 2022). Under typical conditions, active transport of monoamines into the cytosol of the corresponding presynaptic neuron is followed by packaging into storage vesicles by vesicular monoamine transporters (VMATs) where they await exocytosis via activity-dependent release. Once inside the cell, MDMA serves as a substrate for VMAT2 (Erickson et al., 1996), whereupon it inhibits monoamine uptake and disrupts vesicular storage, leading to their displacement into the cytosol and synaptic release through SERT, NET or DAT (Rudnick & Wall, 1992).

The specific release of 5-HT, noradrenaline and dopamine is regulated primarily by the potency of MDMA to act as a substrate at SERT, NET and DAT, as VMATs are expressed in all aminergic neuron types and package monoamines with little recognition bias (Ye et al., 2024). Unlike structurally related psychostimulants such as amphetamine and methamphetamine (Figure 1), which primarily release noradrenaline and dopamine over 5-HT via an identical mechanism, MDMA preferentially releases 5-HT and noradrenaline (Rothman et al., 2001) (Table 1). However, the selectivity by which MDMA releases monoamines is species-dependent. For example, in rat brain synaptosomes, MDMA preferentially releases 5-HT with modest selectivity (<10-fold) over noradrenaline and dopamine (Rothman et al., 2001; Setola et al., 2003). In contrast, studies using cloned human transporters in heterologous cells indicate that MDMA preferentially inhibits the release of noradrenaline with modest selectivity over 5-HT and dopamine (Han & Gu, 2006; Verrico et al., 2007). These species differences are most prominent for SERT, as MDMA more potently inhibits rodent over human SERT, with no species differences at NET and DAT (Han & Gu, 2006).

TABLE 1.

Comparative in vitro potencies of (±)-MDMA and its enantiomers at monoamine transporters (EC₅₀ of release, nM) and receptors (EC₅₀, IC₅₀ or K_i, nM).

	(±)-MDMA	(S)-MDMA	(R)-MDMA	References
SERT	72a,g,i	74a,g,i	340a,g,i	Setola et al. (2003)
NET	110a,g,i	136a,g,i	560a,g,i
DAT	278a,g,i	142a,g,i	3700a,g,i
VMAT1	19,500b	-	-	Erickson et al. (1996)
VMAT2	6900b	-	-	Erickson et al. (1996) and Partilla et al. (2006)
	19,500c	22,800c	31,100c
5-HT1A	6850e,g,i,j	>10,000e,g,i,j	4200e,g,i,j	Lyon et al. (1986)
5-HT2A	8300e,g,i,k	>10,000e,g,i,k	3310e,g,i,k
5-HT2B	2000d,f,h	6000d,f,h	900d,f,h	Setola et al., 2003
5-HT2C	>10,000e,f,h,k	2600d,g,h	5400d,g,h	Setola et al., 2003 and Nash et al. (1994)
α 2A	2532e,f,h,j	-	-	Setola et al. (2003)
α 2B	1785e,f,h,j	-	-
α 2C	1123e,f,h,j	-	-
TA1 (rat)	370e,g,h,k	-	-	Simmler et al. (2013)
TA1 (mouse)	2400e,g,h,k	-	-

Note: Potency is reported as the ^aEC₅₀ for monoamine release, IC₅₀ for ^b[³H]5-HT or ^c[³H]dopamine uptake, ^dEC₅₀ for receptor activation or ^eK_i, with priority given to studies and assay formats that internally compared (±)-MDMA with the individual enantiomers. Values were derived from ^fhuman or ^grodent proteins expressed in ^hheterologous cell systems or ⁱnative tissue. ^jAgonist or ^kantagonist radioligand is noted for K_i values. ‘-’ denotes ligand affinities that are, to the best knowledge of the authors, not reported in the literature. The in vitro results above represent an abbreviated panel of target activities, a more extensive list of values for (±)-MDMA is available via the NIMH Psychoactive Drug Screening Program database, https://pdspdb.unc.edu/pdspWeb/.

Irrespective of species, the monoamine releasing effects of (±)-MDMA are predominantly driven by (S)-(+)-MDMA (Table 1), and although the proportion of 5-HT to dopamine release at a given concentration of (R)-(−)-MDMA may be greater compared to (S)-(+)-MDMA, it is considerably less potent (5–36-fold depending on species and transporter) (Verrico et al., 2007). In addition to its action at monoamine transporters, MDMA has affinity, albeit low, at aminergic G protein coupled receptor binding sites including 5-HT_1A and 5-HT₂-type (5-HT_2A, 5-HT_2B and 5-HT_2C), where it acts as a partial agonist, and α₂-adrenoceptors (Eshleman et al., 2014; Lyon et al., 1986; Nash et al., 1994; Setola et al., 2003). Interestingly, (R)-(−)-MDMA binds with modest stereoselectivity to 5-HT_1A, 5-HT_2A, and 5-HT_2B receptors, while the reverse is true for 5-HT_2C receptors (Table 1). Given the low potency of MDMA at aminergic G protein-couped receptors, little evidence supports its direct engagement in vivo, except perhaps in the case of 5-HT_2B, α₂-adrenoceptors, trace amine (TA₁) receptors or L-type calcium channels where individual enantiomers have determined or presumed sub-micromolar affinity (Setola et al., 2003). We note that the low potency of MDMA to act as a substrate at VMAT2 may represent a special physiologically relevant case, as SERT, NET and DAT may enrich the intracellular concentration of MDMA. Although MDMA has additional targets involved in monoamine release (Simmler et al., 2016), and the precise mechanisms of MDMA uptake and release are currently unknown due to a lack of structural biology data, the present molecular pharmacology data suggests that MDMA primarily elicits monoamine release, causing indirect activation of 5-HT, adrenoceptors and dopamine receptors (Figure 3a).

FIGURE 3.

Molecular, behavioural and brain-region specific mechanisms of MDMA, and their potential relationship to human clinical presentations. (a) MDMA (red glow) binds monoamine transporters and VMAT and, with lower affinity, monoamine receptors, grouped (top to bottom) as 5-HT and dopamine receptors and adrenoceptors. Lower affinity is reflected by lower red glow intensity. (b) Major preclinical behavioural models used to investigate MDMA span sociability (blue box), fear learning (red box), drug reward (green box) and toxic effects (grey box). Contributions of receptors to each domain is indicated by an arrow, colour-coded by behavioural domain. Brain region specific mechanisms are indicated below each behavioural category by a red dot on a rodent brain hemisection. PVN = paraventricular nucleus (5-HT_1A receptor stimulates oxytocin release from oxytocinergic cell bodies); NAc = nucleus accumbens; BLA = basolateral amygdala; ILA = infralimbic area; VTA = ventral tegmental area. Toxic effect icons represent serotonergic terminal degeneration and cardiac effects included acute haemodynamic alternations and valvular hypertrophy associated with long term 5-HT_2B receptor exposure. (c) Proposed clinical relevance of preclinical models. PTSD = posttraumatic stress disorder; ASD = autism spectrum disorder; SUD = substance use disorder; CHF = congestive heart failure. Created with BioRender.com.

5 |. PSYCHOPHARMACOLOGY

5.1 |. Social behaviour

MDMA increases self-reported ratings of sociability (Bedi et al., 2009), enhances use of social and sexual words during interpersonal communication (Baggott et al., 2015) and differentially modulates some subjective and physiological drug responses depending on the drug-state of other humans present (Kirkpatrick & de Wit, 2015). Although MDMA increases self-reported empathy, task-based measures of empathy are not uniformly increased by MDMA (Bedi et al., 2010). Empathy measures are usually categorized as ‘emotional’ (the degree of evoked arousal or concern in the observer) or ‘cognitive’ (accurately labeling observed emotions). In healthy human volunteers, MDMA enhanced emotional empathy for positive emotional situations (Hysek et al., 2014; Kuypers et al., 2014), and neural responses to happy faces in the nucleus accumbens (NAc) (Bedi et al., 2009). Interestingly, MDMA generally impaired cognitive empathy, specifically the identification of negatively valanced faces and sounds (see Section 5.4).

The prosocial effects of MDMA have been modelled in animals by direct social interaction, social approach, social reward learning and social transfer of affective states. MDMA reportedly increased measures of direct social contact, such as investigation, adjacent lying and social sniffing in rodents (Adam et al., 2024; Curry et al., 2018; Curry et al., 2019; Morley et al., 2005; Ramos et al., 2013; Ramos et al., 2015; Ramos et al., 2016; Thompson et al., 2007; Thompson et al., 2009) (Figure 3b) as well as huddling and affiliative vocalization in squirrel monkeys (Pitts et al., 2017). It is not clear whether direct social contact reflects an affective state versus a change in the quality of sensory perception—in humans, MDMA (but not methamphetamine) enhanced the pleasantness of touch (Bershad et al., 2019)—which may explain, for example, huddling in squirrel monkeys. While some measures of direct social contact have been repeatedly measured within individual research groups, we note that attempts to replicate findings across labs have not been successful in rats (Maldonado & Navarro, 2001; Navarro et al., 2004; Smith, Schmidt, et al., 2021) possibly due to MDMA preexposure (Thompson et al., 2008), handling (Ramos et al., 2016) or arena familiarity (Morley & McGregor, 2000). Direct social contact has also proven unreliable in wild-type (WT) mice, with some reports showing that multiple doses of MDMA were needed to increase contact (Curry et al., 2018; Curry et al., 2019), while a single dose reduced (Maldonado & Navarro, 2001) or did not change contact (Walsh et al., 2021). However, MDMA did increase social contact in mouse models of autism spectrum disorder that display social deficits (Walsh et al., 2021).

The prosocial effects of MDMA may be more consistently observed in social approach assays not requiring direct contact (Esaki et al., 2023; Heifets et al., 2019; Kuteykin-Teplyakov & Maldonado, 2014; Ramos et al., 2016; Walsh et al., 2021). Using these models, we and others have shown that local infusions of MDMA in the NAc (Heifets et al., 2019) or basolateral amygdala (BLA) (Esaki et al., 2023) were sufficient to enhance social approach in mice. More recently, MDMA enhanced empathy-like behaviour in a social transfer of pain assay, where an uninjured mouse, treated with MDMA, developed prolonged hyperalgesia (pain-like) following brief exposure to an MDMA-naive mouse in pain (Rein et al., 2024). MDMA also enhanced the social transfer of morphine analgesia, a positive affective state, albeit less durably (Rein et al., 2024). Notably, in this report, the transfer of affective state was not linked to the duration of direct physical contact between the mice and could be mimicked by optogenetic stimulation of dorsal raphe inputs to the NAc.

Finally, MDMA may enhance the sensitivity of rodents to social reward learning measured with social-conditioned place preference (social CPP). This assay differs from measures of social interaction, approach and state transfer in that it measures the value assigned to an opportunity for social contact, linking reward theory to social approach (Panksepp & Lahvis, 2007). Ramos and colleagues modified a CPP protocol often used to measure drug preference (drug paired with Context A and vehicle with Context B) such that drug (i.e., MDMA) pairing sessions were now performed in the presence of conspecific animal or a novel object (Ramos et al., 2016). Rats expressed a preference for the MDMA-conspecific paired context to a greater extent than the MDMA-novel object paired context, but did not prefer a context paired with only MDMA, suggesting that the rewarding effects of MDMA are susceptible to social facilitation. Notably, similar effects have been observed with subthreshold doses of methamphetamine or cocaine (Reyna et al., 2021; Thiel et al., 2008), suggesting a role for catecholamines. Moreover, Nardou and colleagues found similar results using a distinct social CPP and MDMA administration procedure. Mice were administered MDMA in a social context and 48 h later exposed to bedding associated with social or isolate conditions. In two reports from this group (Nardou et al., 2019; Nardou et al., 2023), social CPP, not observed in saline pretreated adult mice (but see Harda et al., 2022), was restored after systemic administration with MDMA, or local infusion of MDMA into the NAc. Notably, these effects were dose-specific.

5.2 |. Serotonergic mechanisms of MDMA-elicited sociability

Across laboratories and behavioural assays, the prosocial effects of MDMA seem to require 5-HT receptor activation, likely triggered indirectly via MDMA-mediated 5-HT efflux. In humans, the SSRI citalopram attenuates a broad range of MDMA effects, including extroversion (Liechti, Baumann, et al., 2000). Indeed, we found that genetic deletion of SERT or pharmacological blockade with the SSRI escitalopram blocked MDMA-induced social approach behaviour in mice (Heifets et al., 2019). Another study reported that MDMA-induced social approach was quantitatively lower following pretreatment with citalopram, but did not reach significance (Esaki et al., 2023), probably due to floating baseline measurements and a distinct arena design.

With respect to individual 5-HT receptors, several studies have demonstrated a role for 5-HT₁-type receptors in MDMA-elicited sociability (Table 2 and Figure 3a,b). For example, systemically administered WAY 100635, a potent and selective 5-HT_1A receptor antagonist (Forster et al., 1995; Martel et al., 2007), blocked MDMA-elicited social interaction in rats (Morley et al., 2005) and approach in mice (Kuteykin-Teplyakov & Maldonado, 2014), with the latter effect being reproduced by local infusion (~4 mM) into the mouse BLA (Esaki et al., 2023). Similarly, local infusion of the 5-HT_1B antagonist NAS 181 (~2 mM) into the NAc medial shell blocked MDMA-elicited social approach (Heifets et al., 2019) and social reward learning (Dölen et al., 2013). Although these data partly conflict with reports that the 5-HT_1B antagonist GR 55562 or the partial agonist SB 216641 did not affect MDMA-elicited social interaction or approach following systemic administration (Esaki et al., 2023; Morley et al., 2005), it is possible that the partial agonist efficacy of SB 216641 (see Price et al., 1997) was insufficient for functional antagonism, and the high concentration of drug locally infused engaged off-target receptors. Future studies should carefully titrate the concentration of tool compounds to limit off-target activities. Nevertheless, these findings are generally consistent with data showing that other 5-HT_1A/1B receptor agonists enhance sociability in mice (Armstrong et al., 2020; Thompson et al., 2007; Walsh et al., 2021). However, coadministration of MDMA with WAY 100635 paradoxically augmented huddling without affecting affiliative vocalizations in squirrel monkeys (Pitts et al., 2017), and pindolol, a 5-HT_1A receptor antagonist showed no effect on emotional empathy or self-reported friendliness in humans (Kuypers et al., 2014; van Wel et al., 2012). Together, these studies support a role for BLA and NAc 5-HT₁-type receptors in a subset of prosocial effects of MDMA, at least in rodents.

TABLE 2.

Animal models and mechanisms associated with discrete components of the subjective MDMA experience in humans.

Component	Study	Behaviour	Species	Strain	Dose	Mechanism	Brain region	Effect	Duration
Sociability	Walsh et al. (2021) a	Social contact	Mouse (♂/♀)	C57BL/6Jc	7.5 mg·kg−1 (i.p.)	-	-	↑	≤1 h
	Morley and McGregor (2000) Cornish et al. (2003) Morley et al. (2005) Thompson et al. (2007) Ramos et al. (2013) Ramos et al. (2015) Ramos et al. (2016) Adam et al. (2024) d		Rat (♂)	Long Evansb, Wistarb	5–10 mg·kg−1 (i.p.)	5-HT1A, 5-HT2B/5-HT2C, V1A	-	↑
	Pitts et al. (2017)		NHP (♂)	Squirrelb	0.3–1 mg·kg−1 (i.m.)	5-HT2A	-	↑
	Kuteykin-Teplyakov and Maldonado (2014) Heifets et al. (2019)a Walsh et al. (2022)a Esaki et al. (2023)	Social approach	Mouse (♂/♀)	C57BL/6Jb,c, ICRb	3–15 mg·kg−1 (i.p.)	SERT, 5-HT1A, 5-HT1B	NAc, BLA	↑	≤1 h
	Rein et al. (2024) c	Social transfer	Mouse (♂/♀)	C57BL/6Jb,c	7.5 mg·kg−1 (i.p.)	5-HT	NAc	↑	≤48 h
	Nardou et al. (2019) Nardou et al. (2023)	Social reward	Mouse (♂)	C57BL/6Jb	10 mg·kg−1 (i.p.)	OT, β-arrestin 2	NAc	↑	2 weeks
	Ramos et al. (2015)		Rat (♂)	Long Evansb	5 mg·kg−1 (i.p.)	-	-	↑	4 weeks
Threat perception	Morley and McGregor (2000)	Predator odour avoidance	Rat (♂)	Wistarb	5 mg·kg−1 (i.p.)	-	-	↑	<1 h
	Arluk et al. (2022)	Predator scent stress	Rat (♂)	Wistarb	5 mg·kg−1 (i.p.)	5-HT1A, 5-HT2A, glucocorticoid receptors	-	↓	7–8 days
	Young et al. (2015) Young et al. (2017)	CS + retrieval	Mouse (♂)	C57BL/6Jb	5.6–7.8 mg·kg−1 (i.p.)	-		↓	<1 h
	Hake et al. (2019)		Rat (♂)	Long Evansb	3–5 mg·kg−1 (i.p.)	-	-	↓
	Young et al. (2015) Young et al. (2017)	Extinction learning	Mouse (♂)	C57BL/6Jb	7.8 mg·kg−1 (i.p.)	-	-	→	-
	Hake et al. (2019)		Rat (♂)	Long Evansb	10 mg·kg−1 (i.p.)	-	-	↓	24 h
	Young et al. (2015) Young et al. (2017)	Extinction retention	Mouse (♂)	C57BL/6Jb	5.6–7.8 mg·kg−1 (i.p.)	SERT, BDNF, 5-HT2A	ILA, BLA	↑	1–10 days
	Hake et al. (2019)		Rat (♂)	Long Evansb	3–10 mg·kg−1 (i.p.)	-	-	↓	24 h
	Young et al. (2015)	Reconsolidation	Mouse (♂)	C57BL/6Jb	7.8 mg·kg−1 (i.p.)	-	-	→	-
	Hake et al. (2019)		Rat (♂)	Long Evansb	5 mg·kg−1 (i.p.)	-	-	↓	8 days
Euphoria	Orejarena et al. (2011)	Self-administration	Mouse (♂)	C57BL/6Jb	0.125 mg·kg−1 (i.v.)	5-HT2A		↑	≤1 h
	Shin et al. (2008)		Rat (♂)	Wistarb	30 mM (i.e.)	D1-like, D2-like	NAc medial shell, NAc core, olfactory tubercle [medial part]	↑
	Fantegrossi et al. (2002)		NHP (♂)	Rhesusb	0.03–0.1 mg·kg−1 (i.v)	5-HT2A		↑
	Doly et al. (2009) a	Conditioned place preference	Mouse (♂/♀)	129/SvPASb	10 mg·kg−1 (i.p.)	5-HT2B		↑	24 h
	Heifets et al. (2019) Pomrenze et al. (2025)		Mouse (♂)	C57BL/6Jb	10–15 mg·kg−1 (i.p.)	DAT, D2-like	NAc medial shell	↑
	Pomrenze et al. (2025)		Mouse (♂/♀)	C57BL/6Jb	5 mg·kg−1 (i.p.)e	SERT, 5-HT2Ce	NAc medial shell	→/↑e
	Vidal-Infer et al. (2012))		Mouse (♂)	OF-1b	10 mg·kg−1 (i.p.)	D1-like, D2-like, DAT		↑

^aNo sex difference reported.

^bEffect observed in WT animals.

^cEffect observed in autism spectrum disorder models.

^dSex difference reported.

^eSubthreshold dose of MDMA that effectuated CPP upon pharmacological challenge.

There is some evidence, albeit scant, that 5-HT₂-type receptors also modulate the prosocial effects of MDMA. For example, pretreatment with the putatively selective 5-HT_2A receptor antagonist ketanserin (see Casey et al., 2022) decreased self-reported friendless in a small human study (van Wel et al., 2012), and the 5-HT_2A receptor antagonist M100,907 blocked the effect of MDMA on huddling in squirrel monkeys (Pitts et al., 2017). However, in mice, neither ketanserin nor M100,907 appreciably reduced MDMA-elicited social approach (Esaki et al., 2023), social interaction (Morley et al., 2005) or social reward learning (Nardou et al., 2023), and while M100,907 blocked the increased social contact following repeat MDMA administration, this effect could not be mimicked by a 5-HT_2A agonist (Curry et al., 2019). In contrast, the dual 5-HT_2B/5-HT_2C receptor antagonist SB 206553 abolished MDMA-evoked social interaction in rats (Morley et al., 2005). As the 5-HT_2C receptor antagonist SB 242084 increased social interaction in rats (Kennett et al., 1997), this latter finding might be due to an interesting and understudied role of 5-HT_2B receptors in regulating MDMA-evoked 5-HT release. For example, two studies from one lab show that the genetic ablation of the 5-HT_2B receptor, or pharmacological blockade of 5-HT_2B receptors with RS 127445, completely blocked MDMA-evoked 5-HT and dopamine release, in a manner that (paradoxically) appears to be upstream of SERT and DAT (Doly et al., 2008; Doly et al., 2009).

5.3 |. Neuropeptides and MDMA-elicited sociability

Despite the substantial literature linking oxytocin to social behaviour across species (Froemke & Young, 2021), the role for oxytocin release and oxytocin receptors in MDMA-evoked social behaviour is less clear. The most compelling evidence in humans supporting a causal relationship between MDMA, oxytocin release and enhanced sociability comes from a study in patients with central diabetes insipidus, in whom release of the hormone arginine vasopressin (AVP) from the CNS is severely impaired. The investigators used MDMA to test whether these patients showed a similar impairment in the release of oxytocin, which is closely related to AVP. Not only was oxytocin release impaired, but nearly all MDMA’s subjective effects were strongly attenuated (with the possible exception of ‘stimulation’). Interpreting this result is complicated by the low ratings for openness, talkativeness or desire for social contact obtained in the healthy comparison group given MDMA. There may also be a potentially important contribution of other neurohormonal deficits (such as reduced AVP release) present in these patients: AVP itself can improve social deficits in children with autism (Parker et al., 2019). Other human studies (Kirkpatrick, Francis, et al., 2014; Kirkpatrick, Lee, et al., 2014; Kuypers et al., 2014) comparing intranasal oxytocin administration and oral MDMA have found an inconsistent effect of oxytocin on prosocial feelings; when these subjective effects were present, oxytocin and MDMA appeared to affect different domains of social behaviour, with oxytocin improving cognitive empathy, but not preference for social interaction, while MDMA had an opposite set of effects (Kamilar-Britt & Bedi, 2015). There have not yet been human studies testing MDMA with selective oxytocin receptor antagonists that cross the blood brain barrier.

Animal studies have made use of selective antagonists for oxytocin receptors in studying the prosocial effects of MDMA, particularly important as oxytocin can activate both oxytocin receptors and a receptor for AVP, the vasopressin 1A (V_1A) receptor. Overall, these studies reveal a limited role for oxytocin and oxytocin receptors in some MDMA-evoked prosocial behaviuors, and again raise the possibility that AVP may play a role in the human studies noted above. In rats, MDMA-elicited direct social interactions (e.g., adjacent lying and sniffing) are mimicked by oxytocin and AVP (Ramos et al., 2013), blocked by a non-selective oxytocin/AVP receptor antagonist (tocinoic acid) (Thompson et al., 2007) and a selective V_1A receptor antagonist (SR 49059), but not blocked by the selective oxytocin receptor antagonist cligosiban (Ramos et al., 2013). The same research group found, in rats, that neither oxytocin or V_1A receptors were required for a different MDMA-induced prosocial behaviour, social approach (Ramos et al., 2016). Similar results were obtained by our group in mice, where we found that MDMA-enhanced social approach was not blocked by peripheral or central administration of oxytocin receptor antagonists, nor by genetic deletion of oxytocin receptors from SERT-expressing neurons (Heifets et al., 2019). However, oxytocin and its receptors may be required for MDMA-facilitated social reward learning. Nardou et al. (2019) found that a centrally administered oxytocin receptor antagonist blocked the ability of MDMA to restore a developmentally regulated form of social CPP (Nardou et al., 2019. A rat study using a more traditional drug CPP paradigm found that oxytocin or MDMA, but not AVP, could enhance CPP when paired with a social stimulus, to a greater extent than pairing drug with a novel object (Ramos et al., 2015), an effect that may involve 5-HT_1A receptor-mediated activation of oxytocin neurons in the paraventricular nucleus of the hypothalamus (PVN) (Hunt et al., 2011). Thus, the accumulated data suggest a specific role for oxytocin and potentially oxytocin receptors in MDMA-induced social reward learning, but not social approach or direct social interaction (Table 2).

5.4 |. Threat perception

Early adopters of MDMA-AT, drawing from personal experience and clinical observation, believed it could attenuate feelings of defensiveness and fear arising in patients accessing negative emotional content (Greer & Tolbert, 1990; Grinspoon & Bakalar, 1986). To what extent disclosure of sensitive emotional content during therapy constitutes ‘fear’ is unknown, but the idea has pervaded research into the therapeutic mechanisms of MDMA (Figure 3b). One preeminent view is that PTSD reflects an aberrant associative learning process. For example, acquisition of fear conditioning is the learning process whereupon an aversive unconditioned stimulus (US), such as air-puff or footshock, is paired with an otherwise neutral stimulus, thereafter eliciting a conditioned stimulus (CS) response. Following acquisition of the CS-US contingency, mammals display enhanced CS-appropriate responses (e.g., increased skin conductance in humans, defensive freezing in rodents). If the CS is repeatedly presented without a reinforcing US, the expectancy-outcome mismatch leads to a new memory that inhibits the conditioned behavioural response (i.e., extinction). Although the neurobiology underlying fear conditioning and extinction is one of the most well-studied areas of neuroscience owing to its tractability, little is known about how MDMA affects these processes.

In two studies of healthy humans, MDMA did not, on average, impact the rate of fear extinction learning measured 24–48 h later (Maples-Keller et al., 2022; Vizeli et al., 2022), although an exploratory analysis hinted at improved extinction retention (i.e., the ability to recall extinction memories) (Maples-Keller et al., 2022). Moreover, MDMA impaired subjects’ ability to discriminate between neutral and conditioned stimuli (Vizeli et al., 2022), suggesting that MDMA may alter the perception of aversive cues. The same authors showed that MDMA impaired the identification of negative emotional faces, and facilitated misclassification of fearful faces as happy. When combined with reports that MDMA blunted emotional responses to aversive autobiographical memories (Carhart-Harris et al., 2014), and increased the likeability of negative emotional faces (Zhang et al., 2025), these data indicate that MDMA diminishes the salience, and may reverse the valence, of negative emotional content.

The effects of MDMA on fear extinction in mice are similar to those in humans, albeit both effects are admittedly weak. In an auditory fear conditioning paradigm, MDMA acutely suppressed defensive freezing to the CS, potentially due to confounding hyperlocomotion (Young et al., 2015). However, MDMA-treated mice displayed less freezing 1–10 days after administration, long after drug excretion, indicative of improved extinction retention (Young et al., 2015). Similar effects were demonstrated in a fear-potentiated startle paradigm (Young et al., 2017). In contrast, a study in rats found that MDMA impaired auditory fear extinction (Hake et al., 2019) (i.e., significantly increased freezing), without affecting contextual fear extinction. However, these authors reported that MDMA administration immediately after context retrieval decreased freezing up to 8 days later, suggesting that MDMA may disrupt the original memory trace (i.e., reconsolidation). These conflicting reports may represent species discrepancies similar to work showing that chronic SSRI treatment can facilitate or impair fear extinction in mice (Karpova et al., 2011) and rats (Burghardt et al., 2013), respectively.

Only one rodent study, in rats, examined the enduring effect of MDMA in a non-classical threat model. Predator-scent stress (exposure to soiled cat litter) was used to induce lasting PTSD-like behavioural alterations, and when MDMA administration was paired with a situational reminder (fresh, unused litter), it caused a lasting reduction (7–8 days) in anxiety-like behaviours, startle response and freezing (Arluk et al., 2022).

5.5 |. Neural mechanisms of MDMA-impaired threat perception

Similar to its effects on sociability, activation of specific 5-HT receptors are necessary for MDMA-impaired threat responses (Table 2 and Figure 3b). For example, acute inhibition of SERT, and to a smaller extent NET, but not DAT, blocked MDMA-mediated extinction retention in mice (Young et al., 2017). Likewise, chronic administration of citalopram, as well as acute M100,907, also blocked MDMA-enhanced extinction retention. Interestingly, this effect was not mimicked by fenfluramine, suggesting that 5-HT efflux alone is insufficient to enhance extinction retention. It has been proposed that the acute hyperlocomotive and fear extinction effects of MDMA can be dissociated by low dose (S)- and high dose (R)-MDMA, respectively (Curry et al., 2018); however, this activity appears at odds with findings from racemic MDMA, which has minimal enantiomeric excess and leads one to assume (R)-MDMA at one half the equivalent of (±)-MDMA should be sufficient. Pharmacological challenge with ketanserin, or pindolol, but not SB 242084, also blocked the ability of MDMA to alleviate predator scent stress-induced anxiety and freezing behaviours in rats (Arluk et al., 2022), suggesting that 5-HT_2A, and 5-HT_1A (or β₂), but not 5-HT_2C, receptors were necessary. The same authors also demonstrated that MDMA increased urinary corticosterone in a 5-HT_2A, and 5-HT_1A, dependent manner that paralleled the observed reduction in anxiogenic and freezing behaviours, consistent with the role of 5-HT and glucocorticoids in corticosterone release and memory consolidation, respectively (Meyer et al., 1984; Wolf, 2008). Notably, the necessity of 5-HT_2A receptors to effectuate MDMA-impaired threat detection in rodents is mirrored by human studies showing that MDMA reduced vigilance and avoidance to negative sounds in a 5-HT_2A receptor-dependent manner (Kuypers et al., 2018; Liechti, Saur, et al., 2000).

Impaired fear extinction is a core feature of PTSD and the basis for exposure-based psychotherapy treatments (Sep et al., 2023). Human neuroimaging studies indicate that remembering negative emotional content as well as recollecting traumatic memories enhance amygdala activity that correlates with symptom severity (Sharot et al., 2004; Shin et al., 2004; Zhu et al., 2017). In addition to storing fear memories (Gale et al., 2004; Nader et al., 2000), the amygdala also promotes long-term extinction memories (Zhang et al., 2020) and receives feedback from the infralimbic area (ILA) and anterior cingulate cortex (ACC) to facilitate extinction (Etkin et al., 2011), a mechanism impaired in PTSD (Rauch et al., 2006). These dual roles for the amygdala may explain the somewhat paradoxical findings from human neuroimaging wherein MDMA decreased amygdala activity (Bedi et al., 2009; Zhang et al., 2025), while findings in mice indicate that MDMA increased amygdala activity (Young et al., 2015). In a recent neuroimaging study, MDMA decreased amygdala reactivity to threat and increased functional connectivity between the amygdala and subgenual ACC in participants defined by subthreshold PTSD and high amygdala activity at baseline (Zhang et al., 2025), suggesting direct modulation of this circuit by MDMA at least in certain patient populations. These findings are paralleled by preclinical work showing that MDMA infusion into the BLA or ILA was sufficient to enhance fear extinction retention in mice (Young et al., 2015).

In aggregate, the data presented above support the notion that MDMA impairs the detection of external threat and enduringly changes the behavioural expression of aversive memories. However, we note that only a few human studies have explicitly examined the effects of MDMA on threatening memories, all in healthy, predominantly male volunteers. How these effects might translate to individuals with PTSD is unknown since a reduction in symptom scores could be driven by component processes other than avoidance. Accordingly, learned fear is only one component of PTSD, and the only one addressed by classical fear conditioning models which have been the primary method of measuring MDMA-elicited fear memory modulation. Moreover, the interpretation of threat-activated responses such as freezing in rodents as “fear” may be a misnomer (LeDoux, 2014), and it is unclear what variants of fear are (non)specifically modulated by MDMA. Future work using models that incorporate other or more components of PTSD is necessary to understand the extent to which MDMA modulates behavioural and neural processes relevant to PTSD.

5.6 |. Euphoria

Since the emergence of MDMA in popular culture, it is best known as a highly pleasurable recreational drug (i.e., ‘ecstasy’), with some data suggesting that >20% of users may meet criteria for substance use disorder (Haug et al., 2024). The rewarding properties of MDMA are widely represented in psychometric rating scales measuring ‘elation’, ‘heightened mood’, ‘drug-liking’ and ‘happiness’ (Liechti, Saur, et al., 2000b; van Wel et al., 2012). Consistent with its abuse potential in humans, non-human primates and rodents self-administer MDMA (Fantegrossi et al., 2002; Schenk et al., 2003; Trigo et al., 2006). As we discuss below, the rewarding properties of MDMA are governed by many of the same neuropharmacological mechanisms that promote sociability and memory modulation (Figure 3b). While these latter effects have historically been the focus of MDMA’s therapeutic effects, it is unknown if they are inextricably linked to abuse potential in humans.

5.7 |. Monoamines and MDMA-elicited reward

The reinforcing properties of addictive drugs are linked to their ability to drive midbrain dopamine release. Consistent with this notion, the dopamine D₂-like receptor antagonist haloperidol blunted MDMA-evoked positive mood in humans (Liechti & Vollenweider, 2000). These effects are mirrored in animal models (Table 2). For example, the dopamine D₁-like receptor antagonist SCH 23390, and the D₂-like antagonist eticlopride, attenuated aspects of MDMA self-administration in rats (Daniela et al., 2004; van de Wetering & Schenk, 2017). Similarly, Scheme 23390, haloperidol, as well as DAT inhibitors blocked MDMA-evoked CPP in rodents (Vidal-Infer et al., 2012). Local infusion of MDMA into the medial, but not lateral, subregions of the rat NAc facilitated intracranial self-administration through D₁- and D₂-like receptors (Shin et al., 2008), indicating a key role for mesolimbic dopamine release in MDMA self-administration. Interestingly, local striatal infusion of fluoxetine or the 5-HT_2A antagonist M100,907 diminished MDMA-evoked striatal dopamine release (Schmidt et al., 1994), indicative of a modulatory role via serotoninergic mechanisms (Figure 3b).

Indeed, several studies suggest an important role for 5-HT receptors in MDMA-elicited euphoria/reward. In humans, pretreatment with citalopram, fluoxetine, or ketanserin attenuated MDMA-evoked elation and positive mood (Kuypers et al., 2018; Liechti, Baumann, et al., 2000; Liechti & Vollenweider, 2000; van Wel et al., 2012), suggesting that 5-HT efflux and/or 5-HT_2A receptor activation promotes euphoria. Preclinical studies corroborate this interpretation, as MDMA-self administration was attenuated in non-human primates following pretreatment with the 5-HT_2A antagonist M100,907 (Fantegrossi et al., 2002), and in mice lacking the 5-HT_2A receptor (Orejarena et al., 2011).

In mice, 5-HT_2B receptors appear necessary for CPP induced by low doses of MDMA (Doly et al., 2009) via a mechanism involving 5-HT_2B receptor-dependent dopamine release. Recent findings from our lab show that genetic deletion of SERT, or pharmacological blockade of NAc SERT or 5-HT_2C receptors, enhances CPP and NAc dopamine release evoked by low dose MDMA (Pomrenze et al., 2025). These results suggest that direct 5-HT_2B receptor activation by MDMA or its metabolite, MDA (Setola et al., 2003), may promote dopamine release and reward in the absence of SERT, while 5-HT efflux—through activation of 5-HT_2C receptors—can suppress these effects in wild-type mice. Taken together, these findings indicate that serotonergic mechanisms both facilitate, and constrain, the hedonic rewarding properties of MDMA, probably through indirect modulation of mesolimbic dopamine signalling (Table 2 and Figure 3b).

The abuse liability of MDMA has undoubtedly raised concerns relevant to its therapeutic development (Mustafa et al., 2024), especially because PTSD is associated with an increased risk of substance abuse disorders (Goldstein et al., 2016). To this point, it has not been formally examined whether the same euphoria-enhancing properties of MDMA reported in recreational users or healthy volunteers also occur during MDMA-AT. The substantially different internal and external contexts of drug administration across these populations are a notable distinction. In fact, the largest MDMA-AT trials to date (Mitchell et al., 2021; Mitchell et al., 2023) found no positive association with substance abuse, and MDMA elicits more emotional than euphoric responses in some individuals with a history of trauma (Zhang et al., 2025).

From a medicinal chemistry perspective, the therapeutic-like effects of MDMA on sociability and memory modulation may be favoured over the rewarding effects, which are often viewed as a side effect (Alberto-Silva et al., 2024; Curry et al., 2018; Mayer et al., 2023). Consistent with the in vitro potency of MDMA-mediated monoamine release, preclinical studies suggest low doses of MDMA primarily enhance sociability and extinction retention via serotonergic mechanisms, whereas high doses drive reward-like behaviour via dopaminergic mechansims (Heifets et al., 2019; Rein et al., 2024; Young et al., 2017). However, the therapeutic-like and reward-like behavioural effects of MDMA are not entirely separable by monoamines. For example, 5-HT release is insufficient to enhance extinction retention (Young et al., 2017), serotonergic mechanisms promote MDMA-elicited euphoria, and Phase II clinical trials for MDMA-AT experimented with low doses (30–125 mg) (Bahji et al., 2020) of MDMA whereas Phase III trials settled on a high dose regimen (180 mg cumulative within-session) (Mitchell et al., 2021; Mitchell et al., 2023), presumably owing to greater efficacy. Thus, present data indicate that MDMA-elicited euphoria comprises myriad neural and behavioural processes, including some involved in sociability and threat detection. For example, ambient temperature modulates MDMA-, but not cocaine-elicited social contact, reward (Cornish et al., 2003) and monoamine release (O’Shea et al., 2005).

6 |. GAPS IN KNOWLEDGE AND TRANSLATION

6.1 |. Neurotoxicity and safety concerns

The extent to which MDMA poses a public health risk is hotly debated (Mithoefer et al., 2003). A plethora of research demonstrates that frequent high doses of MDMA can elicit structural and functional neurotoxicity in 5-HT and dopamine neurons, most likely through monoamine-derived reactive oxygen species (see Costa & Gołembiowska, 2022, for detailed review; Figure 3a–c). However, meta-analyses and systematic reviews of human neuroimaging data on brain alterations in MDMA users have yielded conflicting results (Mueller et al., 2016; Roberts et al., 2016; Roberts et al., 2018), probably due to heterogeneous consumption, concomitant recreational polydrug use, and preexisting attributes of MDMA users that may crop up in observational trials. To be sure, there is some risk of neurotoxicity associated with MDMA, just as there is with the pharmacologically similar FDA-approved drug fenfluramine, a 5-HT releasing agent known to elicit long-lasting decrements in 5-HT axonal density across multiple species, even humans (McCann et al., 1997). A notable distinction in the context of these pharmacotherapies is the frequency of administration. Fenfluramine is administered up to twice daily and indefinitely, whereas all extant clinical trials of MDMA-AT use ≤3 doses spread over several weeks. Highlighting this discrepancy is not to minimize the risks associated with MDMA, but to place it within the cost–benefit context wherein all neurotherapeutic modalities exist. Indeed, MDMA-enhanced heart rate, blood pressure, and body temperature may pose greater risk in some patients (Hysek et al., 2011; Hysek et al., 2013). Moreover, a recent study associated MDMA with an increased risk of all-cause mortality when combined with psychiatric drugs (Cohen et al., 2021). Insofar as readily observable somatic and psychic side-effects of MDMA are concerned, these effects are largely acute and effectively managed in a medical setting (Mitchell et al., 2021; Mitchell et al., 2023). To the best of our knowledge, however, the extent of neurotoxicity (if any) elicited by MDMA within a clinically relevant dosing schedule is presently unclear and warrants further investigation, ideally using a within-subjects clinical trial in healthy volunteers.

6.2 |. The role of psychotherapy in MDMA-AT

A frequently cited mechanism for the apparent clinical efficacy of MDMA is its ability to facilitate therapeutic alliance (Bahji et al., 2020; Maples-Keller et al., 2022; Mitchell et al., 2021; Mitchell et al., 2023; Nardou et al., 2019), that is, a relational bond between the patient and therapist. Although ‘alliance’ is a widely accepted predictor of treatment outcomes in psychotherapy (Howard et al., 2022), it is a nebulous psychological construct that may relate to basic research on sociability and threat perception. Moreover, the extent to which psychotherapy contributes to, or is necessary for, the therapeutic efficacy of MDMA-AT is unclear. All clinical trials of MDMA to date have combined drug administration with an intensive psychotherapy regimen, likely to be due to early experimental and ongoing underground practices, and to minimize risks to trial participants. A recent self-report survey found that the majority of perceived long-term positive emotional effects of MDMA were independent of the environmental setting of drug administration (Elmer et al., 2024). In the same survey, perceived emotional benefits from MDMA were associated with the users’ underlying motivation to obtain self-insight, rather than, say, to feel euphoric or energized. This putative role of intention may harmonize with the psychotherapy, as a recent clinical study demonstrated that fenfluramine-evoked 5-HT release, a pharmacodynamic property shared with MDMA, reduced participant sensitivity to aversive outcomes, enhanced impulse control during aversive interference, and enhanced recall of verbally encoded information (Colwell et al., 2024). Similarly, MDMA increased the likability of implicit threat (i.e., presentation of angry faces) in humans with high amygdala activity (Zhang et al., 2025), which could be useful for engaging with emotional content in psychotherapy. Furthermore, as a broad spectrum of psychotherapy approaches effectively treat PTSD symptoms (Bradley et al., 2005), it is conceivable that the approach used in MDMA-AT trials may be effective in its own right, but additional clinical trials are needed to parse the contribution of psychotherapy and MDMA factors, as well as their interaction, on PTSD symptoms.

While psychotherapy cannot be modelled in laboratory animals, rodent models of social transfer of affective states may serve as a translational proxy for relevant components processes such as interpersonal emotional exchange. For example, a bystander mouse exposed to a demonstrator undergoing pain, analgesia or contextual fear conditioning displays behavioural phenotypes consistent with the experience of the demonstrator (Smith et al., 2021). Follow on studies showed that bystander administration of systemic or intra-NAc MDMA enhanced the social transfer of pain and analgesia, which could not be mimicked by methamphetamine (Rein et al., 2024). It is unknown, however, whether the effect generalizes to social transfer of fear, is mimicked by other effective adjuncts such as chronic SSRIs, or whether MDMA administration to the demonstrator affects the efficiency of social transfer.

6.3 |. Pharmacological specificity

Controlled laboratory studies indicate that MDMA’s effects on sociability, threat perception, and reward are interconnected by its ability to effectuate serotonergic neurotransmission, consistent with its preferential monoamine release profile (Table 1 and Figure 3). A major strength of some clinical and preclinical studies has been the direct comparison of MDMA to other psychoactive drugs (Bershad et al., 2016, Cornish et al., 2003; Heifets et al., 2019; Nardou et al., 2019; Rein et al., 2024). In clinical laboratory studies, head-to-head comparisons have found both overlapping and distinct subjective effects of MDMA which vary depending on the choice comparator and dose (Dolder et al., 2018; Holze et al., 2020; Kirkpatrick et al., 2012). However, therapeutic trials have rarely—if ever—included an ‘active placebo’ that mimics some of the subjective components of MDMA via a dissimilar pharmacological mechanism (e.g., methamphetamine, see ‘Emergence of MDMA as an adjunct to psychotherapy’). As a result, it remains unclear to what extent MDMA’s therapeutic effects are pharmacologically unique as opposed to being broadly attributable to general alterations in affect or perception.

The majority of mechanistic clinical and preclinical pharmacology research with MDMA has focused on its role in releasing 5-HT. The contribution of serotonergic mechanisms to MDMA-elicited sociability, memory modulation and reward are thus well studied, as is the dopaminergic contribution to reward. However, catecholamines have well described roles in sociability (Gunaydin et al., 2014; Solié et al., 2022) and fear extinction (Bierwirth & Stockhorst, 2022; Salinas-Hernández et al., 2023) that have been underappreciated in mechanistic MDMA studies. Notably, two clinical studies support a role for noradrenaline, through NET (Hysek et al., 2011) and α₁-adrenoceptors (Hysek et al., 2013), in the euphoric and emotional effects of MDMA, which remain largely unexplored in preclinical research. While we acknowledge that MDMA exhibits a distinct rank order of transporter-mediated monoamine release (NET ≥ SERT > DAT) compared to amphetamine or methamphetamine (NET > DAT ≫ SERT), direct release of noradrenaline (Rothman et al., 2001) and indirect 5-HT-dependent dopamine release (Schmidt et al., 1994) are not negligible. Therefore, the use of ‘active placebo’ conditions in clinical and preclinical assays is necessary to avoid inappropriately perpetuating the oversimplified idea that 5-HT release is therapeutic and catecholamine release is a mere side effect. Tool compounds useful for this purpose include fenfluramine and methamphetamine, which possess partially overlapping physicochemical, pharmacodynamic, and subjective properties with MDMA, while effectively isolating the role of preferential 5-HT or catecholamine release. Appropriate dosing of control compounds needs to be empirically determined.

6.4 |. Sex differences

There is also the issue that nearly all extant rodent studies investigated the behavioural effects of MDMA exclusively in male subjects, and the majority of participants in many clinical studies are male. This bias is problematic from a therapeutic development and translational neuroscience perspective because PTSD is about twice as prevalent in women (Schein et al., 2021). Moreover, it is unclear whether sex differences observed in MDMA-treated humans (Hysek et al., 2014; Mitchell et al., 2023) are reflected in animal behavioural models where MDMA more effectively releases 5-HT and dopamine in females (Lazenka et al., 2017). Only a few rodent studies reported sociability effects of MDMA in both sexes, with conflicting results (Adam et al., 2024; Heifets et al., 2019; Walsh et al., 2021), and none have reported the effects of MDMA on fear extinction in females.

6.5 |. Future therapeutic use of MDMA

The limited validity of animal models to infer human psychiatric processes is a well-acknowledged criticism of preclinical neuroscience, the foundation of which rests on the assumption that neuroanatomical relationships and function are conserved across species. There is still considerable uncertainty as to which (if any) preclinical effects of MDMA are most relevant to its apparently enduring therapeutic efficacy in humans, as a role for the prosocial and fear extinction effects of MDMA, observed across species, is widely assumed based largely on translational metrics bearing face validity. Delineating the contribution of these effects will require precise mapping of the neurophysiological activities of MDMA onto behaviourally relevant cell types and/or neural circuits in animals to generate new, precise and testable hypotheses for clinical research using molecules or interventions designed to mimic or optimize aspects of response to MDMA-AT.

The therapeutic potential, and future, of MDMA is presently at a crossroads. Promising clinical trial results for MDMA-AT have accelerated research interest and swayed public perception. Indeed, the Australian Therapeutic Goods Administration approved the use of MDMA for treatment of PTSD in 2023. However, the US FDA and several advisory committees voted against the approval of MDMA-AT for PTSD in the summer of 2024 owing to legitimate concerns about clinical trial conduct and composition, data integrity, and long-term abuse liability (Mustafa et al., 2024). These decisions highlight a need for more rigorous research aimed at understanding the contribution of MDMA, psychotherapy, and their combination to the treatment of PTSD. Whether MDMA will ever demonstrate sufficient efficacy to overcome social norms and gain widespread approval is unknown. At the very least, further research will determine whether MDMA-activated neural pathways, or monoamine efflux mediators, can effectively treat symptoms of PTSD, and if so, will provide a foundation for novel drug discovery and development programmes.

6.6 |. 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 2023/24 (Alexander, Christopoulos et al., 2023; Alexander, Fabbro et al., 2023).

Abbreviations:

5-HT

5-hydroxytryptamine (serotonin)

ACC

anterior cingulate cortex

AVP

arginine vasopressin

BDNF

brain-derived neurotrophic factor

BLA

basolateral amygdala

CHF

congestive heart failure

CPP

conditioned place preference

CS

conditioned stimulus

DAT

dopamine transporter

DOM

2,5-dimethoxy-4-methylamphetamine

FDA

U.S. Food and Drug Administration

GPCR

G protein–coupled receptor

ILA

infralimbic area

LSD

lysergic acid diethylamide

MAPS

multidisciplinary association for psychedelic studies

MDA

3,4-methylenedioxyamphetamine

MDMA

3,4-methylenedioxymethamphetamine (midomafetamine)

MDMA-AT

MDMA-assisted psychotherapy (MDMA-assisted therapy)

NAc

nucleus accumbens

NET

norepinephrine transporter

NHP

non-human primate

OT

oxytocin

OTR

oxytocin receptor

PTSD

posttraumatic stress disorder

PVN

paraventricular nucleus (of the hypothalamus)

SERT

serotonin transporter

SSRI

selective serotonin reuptake inhibitor

SUD

substance use disorder

TA1

trace amine receptor type 1

US

unconditioned stimulus

V1A

vasopressin 1A receptor

VMAT

vesicular monoamine transporter

VMAT1 / VMAT2

vesicular monoamine transporter 1 / 2

VTA

ventral tegmental area

WT

wild type