Oxytocin-dependent reopening of a social reward learning critical period with MDMA
This mice study investigates the restoration/regulation of social reward learning for patients with long-term depression using MDMA. The findings point towards the potential of understanding the pathogenesis of neurodevelopmental diseases, which are defined by social impairments, and of disorders that are effected by social influence or are a consequence of social injury.
Authors
- Boyden, E.
- Dölen, G.
- Lewis, E. M.
Published
Abstract
A critical period is a developmental epoch during which the nervous system is expressly sensitive to specific environmental stimuli that are required for proper circuit organization and learning. Mechanistic characterization of critical periods has revealed an important role for exuberant brain plasticity during early development, and for constraints that are imposed on these mechanisms as the brain matures. In disease states, closure of critical periods limits the ability of the brain to adapt even when optimal conditions are restored. Thus, identification of manipulations that reopen critical periods has been a priority for translational neuroscience. Here we provide evidence that developmental regulation of oxytocin-mediated synaptic plasticity (long-term depression) in the nucleus accumbens establishes a critical period for social reward learning. Furthermore, we show that a single dose of (+/−)-3,4-methylendioxymethamphetamine (MDMA) reopens the critical period for social reward learning and leads to a metaplastic upregulation of oxytocin-dependent long-term depression. MDMA-induced reopening of this critical period requires activation of oxytocin receptors in the nucleus accumbens, and is recapitulated by stimulation of oxytocin terminals in the nucleus accumbens. These findings have important implications for understanding the pathogenesis of neurodevelopmental diseases that are characterized by social impairments and of disorders that respond to social influence or are the result of social injury.
Research Summary of 'Oxytocin-dependent reopening of a social reward learning critical period with MDMA'
Introduction
A critical period denotes a developmental window when the nervous system is highly sensitive to particular environmental stimuli required for appropriate circuit formation and learning. Earlier work has shown that exuberant plasticity during early development and subsequent maturation-imposed constraints shape these windows, and that once closed they can limit adaptive responses in disease. In this context, oxytocin (OT) signalling in the nucleus accumbens (NAc) has been implicated in juvenile social reward learning via a presynaptically expressed long-term depression (LTD) of excitatory synapses. Nardou and colleagues set out to test whether developmental regulation of OT-dependent synaptic plasticity establishes a critical period for social reward learning, and whether that period can be reopened. They examined the maturational trajectory of social conditioned place preference (social CPP) across mouse ages, tested age-dependent expression of OT-LTD in the NAc, and asked whether MDMA (±3,4-methylenedioxymethamphetamine) or direct stimulation of OT terminals can reinstate juvenile-like plasticity and behaviour in adults. The study aims to link cellular mechanisms in the NAc with behavioural critical-period regulation and to probe mechanistic requirements for any reopened window.
Methods
The study used C57BL/6 mice (both sexes bred in-house; most experiments in males) and newly generated oxytocin-2A-Flp knock-in (OT-2A-Flp KI) mice to enable molecular targeting of OT neurons. Behavioural testing was standardised to a controlled light phase and experimenters were blinded where subjective scoring was required. Sample sizes were estimated from prior work and mice were randomly assigned to groups; all procedures followed institutional animal care guidelines. Social conditioned place preference (social CPP) was the principal behavioural assay. In brief, after a pre-conditioning baseline trial in a two-zone chamber marked by distinct bedding cues, mice underwent 24 h social conditioning with cage-mates on one bedding and 24 h isolation conditioning on the other, followed immediately by a 30-min post-conditioning test. The assay indexes learned preference for a social-associated cue; analyses used both normalized (post/pre) and subtracted (post minus pre) preference scores. Exclusion criteria for behaviour were defined by extreme baseline preferences. Cocaine CPP used the same apparatus with drug (20, 10 or 5 mg/kg i.p.) paired to one bedding and saline to the other, with a 24-h post-test. Electrophysiology was carried out in parasagittal slices (250 µm) containing the NAc core. Whole-cell voltage-clamp recordings from medium spiny neurons (MSNs) measured evoked excitatory postsynaptic currents (EPSCs) at −70 mV, with GABAA receptors blocked by picrotoxin to isolate EPSCs. LTD induction tested bath application of OT (1 µM, 10 min), MDMA (2 µM, 10 min) or 5HTR1b agonist CP-93129 (2 µM, 10 min). Antagonists (OTR-A L-368,899; fluoxetine for SERT; SB203186 for 5HTR4; WAY-100635 for 5HTR1A) were used with pre-incubation to probe mechanism. Summary LTD was expressed as EPSC amplitude normalized to baseline; statistical comparisons used t-tests and ANOVA. For optogenetics, the authors constructed an AAV-CAG-Fon Chronos-TdTomato virus and injected it into the paraventricular nucleus (PVN) of OT-2A-Flp KI or wild-type mice to restrict Chronos expression to OT neurons. Chronos function was validated in PVN slices by current-clamp recordings and optical stimulation (470 nm, 5 ms pulses at 10–40 Hz). Optogenetic stimulation of OT terminals in the NAc was delivered via implanted fibre cannulas (5 ms pulses, 20 Hz, 20 mW) for 30 min and animals were tested for social CPP 48 h later. Miniature EPSCs (mEPSCs) were recorded 0–6 h after photostimulation in the presence of TTX to assess presynaptic changes. Stereotaxic procedures, virus production, Retrobead tracing, cannula implantation for local NAc microinjection, histology, immunohistochemistry and genotyping followed standard protocols described in the Methods. Statistical tests included two-tailed t-tests, repeated-measures ANOVA and MANOVA, with checks for normality and homogeneity of variance; P < 0.05 was considered significant (Bonferroni correction where stated).
Results
Behavioural maturation and critical period identification: Across male mice aged P21 to P112, social CPP magnitude showed a significant developmental decline (repeated-measures ANOVA F(13,311) = 3.119, P = 0.000217). Normalized and subtracted preference scores indicated a peak in social CPP at P42 with a decline thereafter and apparent closure of the critical period by P98 (normalized one-way ANOVA F(13,311) = 3.187, P = 0.000162; subtracted one-way ANOVA F(13,311) = 3.060, P = 0.000279). The maturational profile was largely similar in females and was not attributable to gross locomotor differences. Developmental regulation of OT-LTD and specificity: In acute NAc slices, bath application of OT (1 µM, 10 min) induced LTD of EPSCs in juvenile (P40) but substantially less LTD at P90, indicating an age-dependent decline in OT-LTD. Retrograde labelling showed no evidence for decreased oxytocinergic inputs to the NAc across maturation, consistent with downregulation of OT receptors rather than input loss. By contrast, 5HTR1b-mediated LTD induced by CP-93129 (2 µM) remained robust and unchanged between ages. Cocaine-conditioned place preference did not differ between P42 and P98, indicating that dopamine-dependent cocaine reward learning is maintained across development. MDMA effects on NAc synaptic plasticity: In adolescent NAc slices, bath-applied MDMA (2 µM, 10 min) induced significant LTD of EPSCs. This MDMA-induced LTD was blocked by pre-incubation with an oxytocin receptor antagonist (OTR-A L-368,899, 5 µM) and was occluded by prior induction of OT-LTD, implying overlapping mechanisms. MDMA-LTD was fully blocked by the SERT antagonist fluoxetine (10 µM), suggesting a role for serotonin transporter-mediated MDMA action. Pharmacological dissection showed blockade of MDMA-LTD by a 5HTR4 antagonist (SB203186, 10 µM) but not by 5HTR1A antagonism; these serotonin manipulations did not alter OT-LTD. Reopening the critical period in vivo: A single systemic MDMA injection (10 mg/kg i.p.) administered to adult mice at P96 reinstated social CPP when tested 48 h later at P98; saline-treated controls did not. The reinstatement required a social context, as MDMA given during social isolation failed to reinstate social CPP. Time-course experiments indicated reinstatement onset within 6 h, durability for at least 2 weeks, and return to baseline by 4 weeks. Cocaine pretreatment (20 mg/kg) did not reinstate social CPP, and MDMA pretreatment did not alter cocaine CPP magnitude. MDMA metaplasticity and localisation: Slices from MDMA-pretreated mice (48 h post) showed increased magnitude of OT-LTD relative to saline-pretreated controls, consistent with metaplastic upregulation of OT-dependent plasticity following MDMA. Systemic co-administration of OTR-A (L-368,899, 7.5 mg/kg) blocked MDMA-induced behavioural reinstatement; OTR-A alone had no effect. Local NAc microinjection of MDMA (11 µM) 48 h before testing also reinstated social CPP in adults, and co-delivery of OTR-A (500 ng) into the NAc blocked this effect, localising the requirement for OT receptors to the NAc. Optogenetic recapitulation: Using OT-2A-Flp KI mice expressing Chronos in OT neurons, optical stimulation reliably evoked action potentials in PVN OT neurons. Optical stimulation of OT terminals in the NAc (30 min, 5 ms pulses at 20 Hz) decreased mEPSC frequency but not amplitude in MSNs, an effect prevented by OTR-A, consistent with a presynaptic decrease in glutamate release. A single 30-min NAc OT-terminal photostimulation session produced robust social CPP in OT-2A-Flp KI mice but not in wild-type controls when tested 48 h later, demonstrating that targeted OT-terminal stimulation in the NAc is sufficient to reopen the critical period for social reward learning. The authors present a working model in extended data linking SERT and 5HT4-dependent OT release to metaplastic upregulation of OT-LTD.
Discussion
Nardou and colleagues interpret their results as the identification of a critical period for social reward learning in mice that peaks during adolescence and closes in adulthood, with a closure that corresponds to a developmental downregulation of OT-mediated synaptic plasticity in the nucleus accumbens. They argue that this downregulation reflects changes in OT receptor expression rather than loss of oxytocinergic inputs, since retrograde labelling did not show input decline while OT-LTD diminished with age. The preservation of 5HTR1b-LTD and cocaine CPP across development supports a selective regulation of OT signalling mechanisms as the relevant factor for the critical-period closure. Mechanistically, the authors present evidence that MDMA reopens the social-learning critical period by inducing OT release via SERT and 5HT4-dependent pathways, thereby producing a metaplastic increase in OT-LTD in the NAc that endures beyond the acute pharmacological effects. MDMA-induced reopening required OT receptors in the NAc and could be recapitulated by optogenetic stimulation of OT terminals, strengthening causal inference about local OT signalling. The investigators emphasise that MDMA is not known to bind OT receptors directly; rather, MDMA’s action at SERT and downstream serotonergic receptors on OT neurons drives OT release. The authors note parallels between their proposed metaplastic mechanism and other drug-induced receptor upregulations (for example, acetylcholine receptor upregulation after nicotine). They suggest that insights into reopening critical periods for social reward learning could inform therapeutic development. In particular, they link features of MDMA-induced reopening—rapid onset, durability beyond acute drug presence, and dependence on social context—to clinical findings from MDMA-assisted psychotherapy for PTSD, which has received FDA Breakthrough Therapy designation and is in Phase III trials, proposing that reopening of a social-reward-learning window might contribute to therapeutic effects by enhancing the therapeutic alliance. Limitations and uncertainties are acknowledged in specific respects: the authors state there is no evidence that MDMA binds OT receptors directly, and they characterise mechanistic steps primarily in mice and ex vivo slices. The extracted text does not provide an extended discussion of species generalisability or other translational caveats beyond the points above. The investigators propose that further mechanistic and translational work could refine therapeutic strategies that target OT signalling to treat disorders characterised by social deficits.
View full paper sections
SECTION
A critical period is a developmental epoch during which the nervous system is expressly sensitive to specific environmental stimuli that are required for proper circuit organization and learning. Mechanistic characterization of critical periods has revealed an important role for exuberant brain plasticity during early development, and for constraints that are imposed on these mechanisms as the brain matures. In disease states, closure of critical periods limits the ability of the brain to adapt even when optimal conditions are restored. Thus, identification of manipulations that reopen critical periods has been a priority for translational neuroscience. Here we provide evidence that developmental regulation of oxytocin-mediated synaptic plasticity (long-term depression) in the nucleus accumbens establishes a critical period for social reward learning. Furthermore, we show that a single dose of (+/-)-3,4-methylendioxymethamphetamine (MDMA) reopens the critical period for social reward learning and leads to a metaplastic upregulation of oxytocin-dependent long-term depression. MDMA-induced reopening of this critical period requires activation of oxytocin receptors in the nucleus accumbens, and is recapitulated by stimulation of oxytocin terminals in the nucleus accumbens. These findings have important implications for understanding the pathogenesis of neurodevelopmental diseases that are characterized by social impairments and of disorders that respond to social influence or are the result of social injury. In juvenile mice, oxytocin (OT) evokes a unique form of synaptic plasticity, long-term depression (LTD), in the nucleus accumbens (NAc), and this cellular mechanism is correlated with social reward learning in vivo. In this study, we began by examining the maturational profile of social conditioned place preference (social CPP; Fig.) to determine whether this behaviour is constrained by a critical period in mice. Ages were chosen to span milestones in brain development, including weaning (postnatal day 21, P21), onset of puberty (P28), sexual maturity (P42), early adulthood (P60), andThe Solomon H. Snyder Department of Neuroscience, Brain Science Institute, Johns Hopkins University School of Medicine, Baltimore, MD, USA.The Solomon H. Snyder Department of Neuroscience, Wendy Klag Institute, Johns Hopkins University School of Medicine, Baltimore, MD, USA.The Solomon H. Snyder Department of Neuroscience, Kavli Neuroscience Discovery Institute, Johns Hopkins University School of Medicine, Baltimore, MD, USA.Department of Brain and Cognitive Sciences, MIT, Cambridge, MA, USA.McGovern Institute, MIT, Cambridge, MA, USA.Department of Biological Engineering, Media Laboratory, Koch Institute, MIT, Cambridge, MA, USA. *e-mail: gul@jhu.edu of male mice from P21 to P112 reveal a significant developmental decline in the magnitude of social CPP (repeated measures ANOVA, F (13,311) = 3.119, P = 0.000217). See Extended Data Tablefor sample size and additional statistical comparisons. p, q, Comparison across ages using the normalized and subtracted social preference scores show that social CPP peaks at P42 and then declines until the closure of the critical period at P98 (normalized: one-way ANOVA, F (13,311) = 3.187, P = 0.000162; subtracted: one-way ANOVA, F (13,311) = 3.060, P = 0.000279). Data are presented as mean ± s.e.m. *P < 0.05; NS, comparisons not significant (P > 0.05).
LETTER RESEARCH
mature adulthood (P90) (Fig.). We compared age groups using normalized and subtracted social preference scores (Fig., q); these data show that, in male mice, the magnitude of social CPP peaks at P42 and declines thereafter until the close of the critical period at P98 (Extended Data Table). This maturational profile is largely recapitulated in females (Extended Data Fig.). Notably, locomotor behaviours are similar between adolescent and adult mice (Extended Data Fig.), so it is unlikely that the effects of maturation on social CPP are due to a general decline in activity across development. These studies provide a comprehensive profile of social reward learning across maturation, and identify a critical period that constrains this behaviour. To determine whether developmental changes in social reward learning behaviour correspond to a change in the expression of OT-dependent LTD (OT-LTD), we recorded excitatory postsynaptic currents (EPSCs) from medium spiny neurons (MSNs) in acute NAc slices prepared from male mice at P40 and P90. The magnitude of LTD induced by bath application of OT (1 µM, 10 min) was significantly decreased at P90 compared to P40 (Fig.). Retrograde labelling studies revealed no evidence for a maturational decline in the number of oxytocinergic inputs to the NAc (Extended Data Fig.). We also examined the magnitude of serotonin 1b receptor-mediated LTD (5HTR1b-LTD); bath application of the selective 5HTR1b agonist CP-93129 (2 µM, 10 min) induced robust LTD of EPSCs, the magnitude of which did not differ between ages (Fig.). Finally, we also examined cocaine-induced (20, 10 or 5 mg kg -1 ) CPP at P42 and P98, but found no difference in the magnitude of cocaine reward learning between ages (Fig.-j, Extended Data Fig.). Together, these data provide evidence that selective regulation of OT-mediated synaptic plasticity in the NAc is at least one mechanism that underlies the establishment of a critical period for social reward learning. We next investigated whether OT signalling mechanisms could be strategically targeted to reopen the critical period for social reward learning. OT does not cross the blood-brain barrier; consistent with this observation, intraperitoneal injection of OT did not reinstate social reward learning in adult mice (Extended Data Fig.). Notably, the phenethylamine MDMA has been shown to stimulate OT neurons, raising the possibility that the prosocial effects of MDMA 10 might be the consequence of OT activity. To characterize the synaptic effects of MDMA in the NAc, we recorded EPSCs from MSNs in acute slices prepared from adolescent male mice. Bath application of MDMA (2 µM, 10 min) induced significant LTD, which was blocked by pre-incubation with an OT receptor antagonist (OTR-A, L-368,899, 5 µM; Fig.), and occluded by OT-LTD (Extended Data Fig.), indicating that these forms of synaptic plasticity share overlapping mechanisms. Consistent with a dominant role for MDMA binding at the serotonin transporter (SERT)(but see ref.), a SERT antagonist (fluoxetine, 10 µM) completely blocked MDMA-induced LTD (Extended Data Fig.) but had no effect on the magnitude of OT-LTD (Extended Data Fig.). Stimulation of OT release by MDMA and 5-HT in the hypothalamus requires 5HT1A and 5HT4 receptors. In the NAc, MDMA-induced LTD was blocked by the 5HTR4 antagonist SB203186 (10 µM; Extended Data Fig.), but not by the 5HTR1A antagonist WAY-100635 (10 nM and 10 µM; data not shown); again, these manipulations had no effect on the magnitude of OT-LTD (Extended Data Fig.). Thus, MDMA induces synaptic plasticity in the NAc, and this plasticity requires activation of SERT, 5HTR4, and OTR. We next investigated the potential of MDMA to reopen the critical period for social reward learning. We administered a single intraperitoneal dose of MDMA (10 mg kg -1 ; determined empirically, Extended Data Fig.) to adult mice (P96) and assessed the magnitude of social CPP 48 h later. Pretreatment with MDMA, but not saline, induced the reinstatement of social CPP measured at P98 (Fig.). Further characterization of the effects of MDMA revealed that pretreatment with MDMA in a socially isolated setting failed to reinstate social reward learning in adulthood (Extended Data Fig.). Examination of the time course of these effects revealed that reinstatement of social CPP began within 6 h of MDMA administration, lasted at least 2 weeks, and returned to the levels seen in saline-pretreated mice by 4 weeks (Extended Data Fig.). Cocaine pretreatment (20 mg kg -1 ) did not reinstate social reward learning in adult mice (Fig.), and MDMA pretreatment had no effect on the magnitude of cocaine (20 mg kg -1 ) CPP in adult mice (Extended Data Fig.). Finally, we pretreated adult mice with MDMA and prepared acute slices containing the NAc 48 h later. The magnitude of OT-LTD was increased in slices prepared from MDMA-pretreated mice, compared with saline-pretreated control mice (Fig.). These data provide evidence that MDMA pretreatment leads to the reopening of the critical period for social reward learning by inducing a metaplasticreinstatement of OT-LTD in the NAc.
LETTER RESEARCH
Next, we next sought to strengthen the causal link between the effects of MDMA and OT signalling mechanisms. Co-administration of MDMA with an OTR-A that can cross the blood-brain barrier (L-368,899, 7.5 mg kg -1 ) prevented the reinstatement of social CPP in adulthood induced by MDMA pretreatment (Extended Data Fig.). Pretreatment with the OTR-A alone did not alter the magnitude of social CPP in adult mice (Extended Data Fig.). These effects can be localized to the NAc, as MDMA (11 µM) delivered directly to this brain region via bilateral cannulas 48 h before social CPP testing reinstated social reward learning in adults, and this effect was blocked by co-delivery of OTR-A (500 ng) to the NAc (Fig.). Next, to determine whether OT is sufficient to reopen the critical period for social reward learning, we generated and validated an OT-2A-Flp optimized knock-in (OT-2A-Flp KI) mouse (Fig., Extended Data Fig.). We also generated an AAV-CAG-Fon Chronos-TdTomato virus (Fig., Extended Data Fig.) to molecularly target OT neurons for expression of Chronos, a channelrhodopsin protein with enhanced stimulation kinetics. In acute slices of the paraventricular nucleus (PVN) prepared from OT-2A-Flp KI mice previously injected with AAV-CAG-Fon Chronos-TdTomato, optical stimulation across a range of frequencies (10-40 Hz) reliably evoked action potentials in molecularly specified OT neurons (Fig.). Consistent with our previous finding that OT-LTD is presynaptically expressed, pretreatment of acute slices from the NAc with optical stimulation (30 min, 5-ms pulse, 20 Hz, 15 mW) significantly decreased the frequency, but not the amplitude, of miniature EPSCs (mEPSCs) in MSNs treated with saline compared to those treated with OTR-A (Fig.). Finally, mice received a single 30-min session of NAc OT terminal stimulation (5-ms pulse, 20 Hz, 20 mW) and were tested for social CPP 48 h later, at P98. Pretreatment with optogenetic stimulation of OT terminals in the NAc led to the expression of robust social CPP in OT-2A-Flp KI mice but not in C57BL/6 wild-type mice (Fig.), indicating that this manipulation reopens the critical period for social reward learning. Together, these data provide direct evidence that MDMA-induced reopening of the critical period for social reward learning requires OTRs in the NAc and is recapitulated by stimulation of OT terminals in the NAc. Cognitive neuroscientists have speculated on the existence of a critical period for social behaviour. To our knowledge, these studies are the first to identify and characterize such a critical period by demonstrating that: (1) mice are maximally sensitive to social reward learning cues during adolescence (Fig.); (2) this sensitivity declines in adulthood (Fig.); and (3) these adaptations correspond to a maturational downregulation of OT-mediated synaptic plasticity in the NAc (Fig.). Moreover, the results of our retrograde labelling studies (Extended Data Fig.) are consistent with previous reports that OT receptors, rather than oxytocinergic inputs, are downregulated during development. As 5HTR1b-LTD (Fig.) is not developmentally downregulated, it is unlikely that a maturational decline in glutamatergic signalling mechanisms could account for decreased OT-LTD in adulthood. Moreover, because dopamine-dependentcocaine CPP behaviour (Fig., Extended Data Fig.) is maintained across development, these results indicate that the closure of the critical period for social reward learning requires selective regulation of OT signalling mechanisms. The current studies also provide the first direct evidence, to our knowledge, to link the prosocial effects of MDMAwith metaplastic changes in OT-LTD (Fig.), as well as synaptic and behavioural effects that require activation of OT receptors in the NAc (Fig.). We are aware of no evidence that MDMA binds to OT receptors. Instead, our results (Extended Data Fig.) are consistent with previous reports that MDMA binds to SERTand triggers OT releasethrough activation of 5HT4 receptors on OT neurons. These results add to growing evidencethat the 5HT4 receptor is a regulator of presynaptic neurotransmitter release. Extended Data Fig.shows our working model. Notably, our proposed mechanism of metaplastic upregulation of OT Letter reSeArCH signalling mechanisms parallels previous observations of acetylcholine receptor upregulation following nicotine administration. Finally, mechanistic insights into the reopening of the critical period for social reward learning may inform the development of novel therapeutic agents. Indeed, on the basis of its safety and efficacy profile, MDMA-assisted psychotherapy for post-traumatic stress disorder (PTSD) has been designated a 'Breakthrough Therapy' by the FDA, and phase III trials are currently underway. Despite the excitement over these clinical findings, the mechanisms that underlie the therapeutic efficacy of MDMA for PTSD are largely unknown. Some have proposed that these results are the consequence of the ability of MDMA to induce memory reconsolidation and fear extinction. Alternatively, our results indicate that the therapeutic effects of MDMA share a number of features with MDMA-induced reopening of the critical period for social reward learning, including: rapid onset (Extended Data Fig.); durability beyond the acute effects of the drug (Extended Data Fig.); and dependence on social setting (Extended Data Fig.). These similarities suggest that the therapeutic efficacy of MDMA might stem from its ability to reopen the critical period for social reward learning, thereby fostering the alliance between psychologist and patient during MDMA-assisted psychotherapy.
METHODS
Mice. Male and female C57BL/6 mice were bred in-house and weaned or delivered from Jackson Laboratory (stock # 00664) at 3 weeks of age. All experiments were performed in males except Extended Data Fig.. Oxytocin-2A-Flp optimized knock-in (OT-2A-Flp KI) mice were generated by Cyagen Biosciences (Santa Clara, CA) according to our design. Similar to the strategy employed by the oxytocin-IRES-Cre knock-in mousewe replaced the stop codon in exon 3 of the endogenous oxytocin gene (Oxt) with a 2A-Flp construct. The co-translational cleavage 2A peptide strategy (known to have more robust expression of its downstream gene than the internal ribsome entry site (IRES) strategy) was used to separate Flp and OT, because Flp is directed to the nucleus, while OT is directed to the cytoplasm. Flp-dependent GFP reporter mice (Stock # 32038; Swiss Webster background) were obtained from the Mutant Mouse Resource Center. Transgenic mice were bred in-house and were weaned at 3 weeks of age. All mice other than GFP reporter mice were kept on a C57BL/6 background and all mice were maintained on a 12 h:12 h natural light-dark cycle, starting at 8:30 a.m. with food and water provided ad libitum. All behavioural experiments were conducted during the same circadian period (8:30am to 8:30pm) in a dedicated, sound and odourcontrolled behavioural testing room, which is separated from the vivarium, and no other experiments were conducted simultaneously in the same room. Sample size was estimated based on previous work and published literature. Experimenters were blinded to the treatment condition when subjective criteria were used as a component of data analysis, and control and test conditions were interleaved for all experiments. Mice were randomly assigned to experimental and control groups. All procedures complied with the animal care standards set forth by the National Institutes of Health and were in accordance with protocols approved by the Johns Hopkins University Animal Care and Use Committee. Social CPP. The protocol for social CPP was adapted from previously published work. Upon weaning, animals were socially housed (4-5 same sex cage mates, unless otherwise specified) in a cage containing corncob bedding (Anderson Cob, 1/4" cob, Animal Specialties and Provisions) until the pre-determined age for social CPP testing. Each animal was used at only one behavioural time point. At the pre-determined age, animals were placed in an open field activity chamber (ENV-510, Med Associates) equipped with infrared beams and a software interface (Activity Monitor, Med Associates) to monitor the position of the mouse. The apparatus was partitioned into two equally sized zones using a clear Plexiglas wall, with a 5-cm diameter circular hole at the base; each zone contained one type of novel bedding (Alpha-Dri, Animal Specialties and Provisions or Kaytee Soft Granule, Petco). The amount of time spent freely exploring each zone was recorded during 30-min test sessions. After an initial pre-conditioning trial to establish baseline preference for the two sets of bedding cues, mice were assigned to receive social conditioning with cage mates for 24 h on one type of bedding, followed by 24 h of isolation conditioning (without cage mates) on the other bedding. To assure unbiased design, chamber assignments were counterbalanced for side and bedding cues. Immediately after the isolation conditioning, a 30-min post-conditioning trial was conducted to establish preference for the two conditioned cues. CPP is a learned association between a condition (for example, social, cocaine) and a cue (bedding). It does not require scent from other mice, as the bedding itself serves as the cue. This assay does not measure the quality of social interactions, but the magnitude of the reward learning is almost certainly linked to changes in the types of social interactions animals engage in as they mature (for example, decreased play behaviour, increased agonistic behaviour). Unlike drug rewards, which are non-contingently rewarding, natural rewards, including water, food, and social interactions, are state-dependent, and require that the animal be thirsty, hungry, or lonely. Consistent with this observation, previous studies have shown that in order for social CPP to be expressed, the animals must either (1) undergo 1 week of social isolation before conditioning or 2) experience the isolation phase of conditioning immediately before the post-conditioning trial. We elected to use the second protocol because it simplified our experimental design, especially for molecular manipulations that require 1 week for full expression of viruses. Exclusion criteria for this behaviour were strictly defined as a pre-conditioning preference score of >1.5 or <0.5. Animals were never excluded based on the quality of their social interactions; furthermore, because these animals have been housed together since weaning, aggressive interactions were extremely rare, even in adults. Pre-conditioning versus post-conditioning social preference scores were considered significant if paired student's t-test P values were <0.05. Comparisons between experimental conditions were made using both normalized social preference scores (time spent in social zone post divided by pre), and subtracted social preference scores (time spent in social zone post minus pre); these were considered significant if unpaired student's t-test (two conditions), or ANOVA (more than two conditions; Fig.and Extended Data Fig.) P values were <0.05, and Bonferonni corrected P values were <0.0036. Although the convention in the field is to refer to this assay as a measure of social reward, the assay actually measures the contribution of both social preference and isolation aversion. In pilot experiments, we corroborated these findings, and determined that comparisons between the socially-conditioned cue and the isolation-conditioned cue improve the robustness of the two-day protocol and therefore the scalability of this assay. All experiments were performed during the mouse rest period (light cycle), as pilot experiments revealed that social CPP is most robust if assayed during this period. Prior to intraperitoneal (i.p.) drug treatment experiments (MDMA, OT, oxytocin receptor antagonists (OTR-A or L-368,899 hydrochloride), cocaine, or saline), animals were habituated to the injection procedure by daily i.p. injections of saline in the home cage for 3 days. Pharmacological delivery schedules were counterbalanced for type of drug. Unless otherwise stated, for MDMA pretreatment experiments animals were tested 48 h after the injection to allow for complete clearance of MDMA and its metabolites. For the experiment testing involvement of the social context on the effect of MDMA, the injection of MDMA was followed by 3 h in either the isolate or social condition, after which time the animals were brought back to their home cages with their cage mates for 2 days before testing for social CPP. Cocaine CPP. For cocaine CPP, experiments were performed in the same open field activity chambers as social CPP using an identical configuration. After 3 days of habituation to i.p. saline injections in the home cage, the amount of time spent exploring each zone of the activity chamber during a 30-min pre-conditioning test was measured. After 24 h, mice were randomly assigned in a counterbalanced fashion to receive an i.p. injection of cocaine (20 mg/kg (Fig., Extended Data Fig.), 10 mg/kg (Extended Data Fig.) or 5 mg/kg (Extended Data Fig.)) immediately followed by 30 min conditioning on one bedding (Soft Granule or Alpha Dri). A second 30-min conditioning session was conducted 24 h later on the other bedding after an i.p. injection of saline (equal volume to cocaine). A 30 min post-conditioning test session was conducted 24 h later to determine each mouse's preference for the cocaine versus saline associated beddings. The time course of social CPP and cocaine CPP are slightly different, as social CPP is performed immediately after the non-social pairing, whereas cocaine CPP is performed 24 h after conditioning. Analyses were conducted and significance determined as for social CPP. Pharmacological delivery schedules were counterbalanced for the type of drug (that is, saline, cocaine or MDMA). Electrophysiology. Parasagittal slices (250 µm thick) containing the NAc core were prepared from C57BL/6 mice using standard procedures. In brief, after mice were anaesthetized with isoflurane and decapitated, brains were quickly removed and placed in ice-cold low-sodium, high-sucrose dissecting solution (228 mM sucrose, 26 mM NaHCO 3 , 11 mM glucose, 2.5 mM KCl, 1 mM NaH 2 PO 4 , 1 mM MgSO 4 , 0.5 mM CaCl 2 ). Slices were cut by adhering the two sagittal hemispheres containing the NAc core to the stage of a Leica vibroslicer. Slices were allowed to recover for a minimum of 60 min in a submerged holding chamber (∼25 °C) containing artificial cerebrospinal fluid (ACSF) consisting of 119 mM NaCl, 2.5 mM KCl, 2.5 mM CaCl 2 , 1.3 mM MgSO 4 , 1 mM NaH 2 PO 4 , 11 mM glucose and 26.2 mM NaHCO 3 . Slices were then removed from the holding chamber and placed into the recording chamber, where they were continuously perfused with oxygenated (95% O 2 , 5% CO 2 ) ACSF at 2 ml/min at 25 °C. To isolate EPSCs, picrotoxin (50 µM, Sigma) was added to the ACSF to block GABA A receptor-mediated inhibitory synaptic currents. Whole-cell voltage-clamp recordings from MSNs were obtained under visual control using a 40× objective. The NAc core was identified by the presence of the anterior commissure. Recordings were made with electrodes (2.5-3.5 MΩ) filled with 115 mM CsMeSO 4 , 20 mM CsCl, 10 mM HEPES, 0.6 mM EGTA, 2.5 mM MgCl, 10 mM Na-phosphocreatine, 4 mM Na-ATP, 0.3 mM Na-GTP, and 1 mM QX-314. Excitatory afferents were stimulated with a bipolar nichrome wire electrode placed at the border between the NAc core and cortex dorsal to the anterior commissure. Recordings were performed using a Multiclamp 700B (Molecular Devices), filtered at 2 kHz and digitized at 10 kHz. EPSCs were evoked at a frequency of 0.1 Hz while MSNs were voltage-clamped at -70 mV. Data acquisition and analysis were performed on-line using custom Igor Pro software. Input resistance and access resistance were monitored continuously and experiments were terminated with a change >15%. Summary LTD graphs were generated by averaging the peak amplitudes of individual EPSCs in 1-min bins (6 consecutive sweeps) and normalizing these to the mean value of EPSCs collected during the 10-min baseline immediately before the LTD-induction protocol. Individual experiments were then averaged together. As described previously, all experimental drugs (OT, 1 µM, Tocris Biosciences; CP-93129 dihydrochloride, 2 µM, Tocris Biosciences; MDMA, 2 µM, NIDA, Drug supply program) were bath applied for 10 min following the collection of baseline data to test their ability to induce LTD. For experiments examining the blockade of MDMA-LTD or OT-LTD (Fig., Extended Data Fig.), slices were pre-incubated in antagonist (OTR-A, 5 µM L-368,899 hydrochloride; serotonin transporter (SERT) antagonist, fluoxetine hydrochloride, 10 µM; 5HTR4 antagonist, SB203186 hydrochloride, 10 µM; 5HTR1a antagonist, WAY 100635 maleate, 10 nM or 10 µM, Tocris Biosciences) for at least 30 min before recording. For experiments examining the occlusion of MDMA-LTD, after stabilization of OT-LTD (at 60 min post induction), 2 µM MDMA was bath applied for 10 min. Comparisons between experimental manipulations were made using a twotailed, unpaired Students t-test with P < 0.05 considered significant. All values are reported as mean ± s.e.m. To confirm the functional efficiency of Chronos expressed in OT cells, coronal slices containing the PVN (250 µm) were prepared in a manner similar to that described above. After at least 60 min recovery, slices were transferred to the recording chamber and Chronos-expressing OT cells were visually targeted for recording based on TdTomato fluorescence. Recordings were conducted in current clamp using patch pipettes (2-4 MΩ) filled with internal solution containing 130 mM K-gluconate, 10 mM HEPES, 1 mM NaCl, 1 mM CaCl 2 , 10 mM EGTA, 1 mM MgCl, 2 mM Mg-ATP, 0.5 mM Na-GTP and 0.125% neurobiotin. As PVN OT neurons are often spontaneously active, a small amount of hyperpolarizing current (about -20 pA) was injected to eliminate spontaneous action potentials and facilitate analysis. Optical stimulation (470 nm; 5 ms; 20 pulses) was delivered through the 40× microscope objective at 10, 20 and 40 Hz. Data were acquired and analysed using the Recording Artist plugin in Igor Pro. For the miniature EPSC experiments, slices prepared from OT-2A-Flp KI mice injected with AAV-CAG-Fon Chronos-TdTomato were optically stimulated for 30 min through the 40× microscope objective (5 ms pulse, 20 Hz, 15 mW, 470 nm) in the presence or absence of OTR-A (5 µM, L-368,899 hydrochloride, Tocris Biosciences) before mEPSCs were recorded. Miniature EPSCs were collected from 0-6 h after photostimulation at a holding potential of -70 mV in the presence of tetrodotoxin (TTX) (0.5 µM, Tocris Biosciences) and picrotoxin (50 µM, Sigma). Two minutes after break-in, 30-s blocks of events (total of 200 events per cell) were acquired and analysed using custom Igor Pro software with threshold parameters set at 5 pA amplitude and <3 ms rise time. All events included in the final data analysis were verified visually. Virus generation. The pAAV-CAG-Fon Chronos-TdTomato construct (deposited on Addgene, accession number 105834) was assembled using seamless cloning by Epoch Life Science, cloning the Chronos-TdTomato gene into a pAAV-CAG-Fon backbone at the NheI and AscI sites. The gene is expressed under the CAG promoter and flanked by Flp recombination sequences FRT and F5 separated with a 50-bp spacer. The subsequent AAV virus was packaged in serotype 2/8 by the UNC vector core via co-transfection of AAV293 cells with pHelper, pAAV rep2/cap8 and pAAV-CAG-Fon Chronos-TdTomato at 1:1:1. The cell pellets were re-suspended in the AAV lysis buffer and freeze-thawed three times. Supernatants were collected and then gradient ultracentrifugation and dialysis were performed. Stereotaxic injection. Stereotaxic injection of AAV-CAG-Fon Chronos-TdTomato into the PVN of C57BL/6 wild-type WT and OT-2A-Flp KI mice was performed at P33 under general ketamine-medetomidine anaesthesia using a stereotaxic instrument (David Kopf). A small volume (∼1 µl) of concentrated virus solution was injected bilaterally into the PVN (bregma -0.6 mm; lateral 0.35 mm; ventral -5.1 mm), at a slow rate (100 nl/min) using a syringe pump (Harvard Apparatus, MA). The injection needle was withdrawn 10 min after the end of the infusion. Following virus injection, animals were housed as dyad pairs. Injection sites and viral infectivity were confirmed post-hoc by preparing sections (50 µm) containing the PVN. Robust TdTomato expression was observed throughout the rostral-caudal extent of the PVN in all animals. Stereotaxic injection of Retrobeads (RtBs) into the nucleus accumbens (NAc) was performed under general ketamine-medetomidine anaesthesia using a stereotaxic instrument (David Kopf). A small volume (∼30 nl) of diluted RtB solution (1:4) was injected unilaterally into the NAc core (bregma 1.54 mm; lateral 1.06 mm; ventral 4.1 mm), at a slow rate (20 nl/min) using a syringe pump (Harvard Apparatus, MA). The injection needle was withdrawn 10 min after the end of the infusion. Injection sites were confirmed post-hoc by preparing sections (20-50 µm) containing the NAc. Following RtB injection, the animals were returned to their home cages for one week before intracardial perfusion. Fibre implantation. Two months after virus injection, optic fibre cannulas were implanted bilaterally above the NAc (bregma 1.6 mm; lateral 1.1 mm; ventral -3.95 mm) under general ketamine-medetomidine anaesthesia using a stereotaxic instrument (David Kopf). The cannulas were made in-house using 1.25 mm diameter multimode ceramic ferrule, 200 µm fibre optic cable (NA 0.53) and epoxy. Cannulas were secured to the skull using miniature screws (thread size 00-90 × 1/16) and dental adhesive cement. Optogenetics. For optogenetic photostimulation, the fibre optic cannulas were connected to a 473-nm laser diode through a FC/PC adaptor and a fibre optic 1 × 2 rotary joint. Laser output was controlled using a pulse generator (Master-8), which delivered 5-ms light pulses at 20 Hz. Light output through the optical fibres was adjusted to 20 mW using a digital power-meter console. Each member of the dyad received 30 min simultaneous optical stimulation (5 ms pulse, 20 Hz, 20 mW) in a chamber divided by a perforated grid, which provided partial access to social cues (visual, olfactory, auditory), and the animals were tested for social CPP 48 h later, at P98. Following behavioural testing, viral expression, OT neuronal specificity, and fibre placement were confirmed for each animal (Extended Data Fig.). Notably, dyad housing alone had no effect on the magnitude of social CPP (Extended Data Fig.). Infusion cannula implantation and microinjection. Standard stereotaxic procedures were used to implant guide cannulas under ketamine-dexmedetomidine anaesthesia. In brief, at P91, double cannula guides (C235GS, 26GA, C/C distance 2.0 mm, 5 mm pedestal, cut 4 mm below pedestal, custom specified for mouse bilateral NAc coordinates, Plastics One Inc.) were implanted above the NAc of male mice following bilateral craniotomy (bregma 1.6 mm; lateral 1.0 mm) and attached to the skull using dental acrylic. A third hole was drilled a few mm posterior to the guide cannula holes; this hole was enlarged for placement of an anchor screw. After recovery, all animals were returned to their home cage and given free access to food and water. One week after cannula implantation, and 2 days before social CPP testing, mice were microinjected with MDMA (11 µM) with or without OTR-A (L-368,899, 500 ng) into the NAc. Immediately before microinjection, a 33-gauge injector cannula that protruded 1.6 mm below the guide cannula and was connected to a syringe pump (Harvard Apparatus, MA) was inserted into the guide cannula. Microinjections were delivered at a rate of 1 µl/min, with a total injection volume of 1 µl. Injector cannulas remained in place for an additional minute before being removed. After infusion the mice were put back into their home cage for 2 days before the social CPP. Cannula placement was verified by post-hoc histology. Histology and immunohistochemistry. After transcardial perfusion with phosphate buffered saline (PBS (pH 7.4, 1 M)) followed by 10% formalin, brains were post-fixed overnight in formalin at 4 °C and transferred to PBS. Coronal sections of the PVN (20 µm) and the NAc (50 µm) were cut in PBS using a vibratome. For experiments in which fluorescent colocalization was analysed, serial brain sections (20 µm thickness) were cut using a cryostat following cryoprotection of tissue with a 30% sucrose solution containing 0.01% sodium azide. Sections were mounted on slides in PBS and allowed to dry. Subsequently, sections were rinsed four times for 10 min each and blocked with 10% normal horse serum (NHS) and 2% bovine serum albumin (BSA) in PBS/0.5% triton-X100 (PBST). To label OT neurons immunohistochemically, the anti-OT-neurophysin antibody PS38 was used(gift from H. Gainer; 1:150). The PVN or NAc was stained for OT in 1% NHS in PBST at room temperature (RT) overnight. Sections were rinsed four times for 10 min each in PBS followed by 2 h secondary antibody stain using either Alexa 350 (RtB or Tdtomato colocalization; Life Technologies; goat anti-mouse; 1:200) or Alexa 647 (GFP colocalization; Life Technologies; goat anti-mouse; 1:1,000 in 1% NHS in PBST) at RT. Sections were then rinsed four times for 10 min at RT and stained with DAPI (diluted 1:5,000 in PBS) and the slides were coverslipped. Images were acquired using an EVOS or Olympus BX41 microscope with 4×, 10× and 40× objectives or a Zeiss LSM880 confocal microscope using a 63× objective. Images were analysed using ImageJ. Genotyping. Oxytocin-2A-Flp optimized knock-in mice (OT-2A-Flp KI) were genotyped before weaning using polymerase chain reaction (PCR) analysis of DNA isolated from tail snips. The wild-type allele was identified by a 622-bp PCR product and the mutant allele by a 308-bp PCR product using the following primers: mOxtR1: TCCGACAATTAGACACCAGTCAA; mOxtF1: CTACCTGAGCAGCTACATCAACAG; mOxtF2: AGGGCTTTGGGAAGTG TTAGGCT. The reactions were run under the following conditions: 94 °C × 3 min, (94 °C × 30 s, 60 °C × 35 s, 72 °C × 35 s) × 38 cycles, 72 °C × 5 min. For OT-2A-Flp::FonGFP mice, in addition to the genotyping for OT-2A-Flp, the wild-type allele was identified by a 603-bp PCR product, and the FonGFP allele by a 320-bp PCR product using the following primers; mutant: CCAGGCGGGCCATTTACCGTAAG; common: AAAGTCGCTCTGAGTTGTTAT; wild-type: GGAGCGGGAGAAATGGATATG. The reactions were run using the Gt(ROSA)26Sortm(CAG) protocol published by Jackson Laboratories. Statistics. Comparisons between experimental manipulations were made using a two-tailed Students t-test (paired or unpaired, as appropriate), ANOVA for repeated measures, and MANOVA for comparisons between multiple outcome measures, with P < 0.05 considered significant. Homogeneity of variance was tested using Levene's test for equality of variances and normality was tested using the Shapiro-Wilk test of normality using SPSS; post-hoc tests were carried out only if global measures were significant. Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this paper. 4,5,6,7,8,9,10,11,12,13,14,18,19,and 20 have associated raw data, which will be made available upon request.