Although its popularity has somewhat decreased in the last few years, ecstasy (MDMA) is still one of the most widely used illicit recreational drugs, especially used by young people 1 . Since in the late eighties studies have started to examine the effects of ecstasy on the brain, concern has risen about its potential neurotoxicity, especially to the axons of the serotonin (5-HT) cells of the brain. Most of this evidence has been derived from studies in animals, mainly in rats and non-human primates 2 . Also in humans, many studies have been performed in recent years that indicated that ecstasy use might be neurotoxic in humans too. Most of these studies showed differences between heavy ecstasy users and non-users. It was shown that heavy ecstasy users had lower serotonin transporter (SERT) densities, had higher levels of depression, were more impulsive and scored worse on neuropsychological memory tests than non-users 3-6 . Despite the increasing evidence that heavy ecstasy use is neurotoxic in humans many questions remained unanswered. Most of the evidence of neurotoxicity in humans has been derived from retrospective studies. Therefore, it remains possible that differences between ecstasy users and non-users were not caused by ecstasy, but were pre-existent and even predisposed people to start using ecstasy 7 . Moreover, there are many potential confounders like use of drugs other than ecstasy, gender, lifestyle, and SERT polymorphism that could influence the results of studies in ecstasy users 8 . Finally, it is unknown what the effects of low dose ecstasy use are 2,9-12 .

Studying the potential neurotoxic effects of ecstasy on the brain is relevant because many young people all over the world use this drug or will start experimenting with it. Well-conducted studies can probably help these young adults in making their own benefit-risk analysis and making their own choices. In addition, it can guide governments and social organizations in the development of their prevention messages. Moreover, in the last few years, interest in the potential beneficial effects of MDMA as an adjunct in psychotherapy to reduce anxiety, tension or agitation in patients with post-traumatic stress disorder or in last stage cancer, has increased and the first studies to evaluate these beneficial effects have already started 13,14 . In this regard, a careful and scientifically funded analysis should be made in order to decide whether potential benefits outweigh potential risks.

The aim of this thesis was to gain more insight in the effects of ecstasy use on the brain, especially regarding causality, course and clinical relevance while considering the most important potential confounders. Most of the studies in this thesis were part of the Netherlands XTC Toxicity (NeXT) study or studies on which the NeXT study was based. Part of the objectives of the NeXT study was addressed in this thesis. The objectives of the NeXT study included:

•  To study the causality of ecstasy use in observed brain pathology in humans;

•  To study the long-term course of brain pathology and related clinical characteristics in ecstasy users;

•  To study the clinical relevance of observed brain pathology in ecstasy users;

•  To study the dose-response characteristics of ecstasy use in the causation of brain pathology;

•  To study vulnerability and protective factors in the causation of brain pathology among ecstasy users;

•  To study potential neurotoxic consequences of ecstasy use in relation to the use of other drugs;

•  To study the presence of functional or structural damage to neurotransmitter systems other than serotonin following ecstasy exposure.

A combination of neuroimaging techniques, psychopathology questionnaires and neuropsychological tests were used as parameters of neurotoxicity. Neuroimaging included [ 123 I] -CIT single photon emission computed tomography (SPECT) measuring SERT densities; proton magnetic resonance spectroscopy ( 1 H-MRS) measuring neurometabolites 15-17 ; diffusion tensor imaging (DTI) measuring apparent diffusion coefficient (ADC) and fractional anisotropy (FA) of the diffusional motion of water molecules in the brain as indicators of axonal integrity 18,19 ; and perfusion weighted imaging (PWI) measuring regional relative cerebral blood volume (rrCBV) indicative of brain perfusion 20,21 . With these combined imaging techniques we measured both structural ( 1 H-MRS and DTI) and functional ([ 123 I]-CIT SPECT and PWI) aspects of neurotoxicity. Psychopathological assessment included the Beck Depression Inventory (BDI 22 ), the Barratt Impulsivity Scale (BIS 23 ) and the Spanning Behoefte Lijst (SBL 24-26 ), self-report questionnaires measuring symptoms of depression, impulsivity and sensation seeking, respectively 27 . Neuropsychological tests mainly included tests on various aspects of memory.

Part I: Introduction

In Chapter 1 a general introduction in the history, the effects and the potential neurotoxicity of ecstasy is given and the outline of the thesis is presented.

Most of the studies in this thesis are based on the NeXT study. In chapter 2 a detailed outline of the objectives and methods of the NeXT study is provided. The NeXT study is a combination of different approaches with three substudies: (1) a cross-sectional substudy among heavy ecstasy users and controls with variation in drug use, which should provide information about potential neurotoxic consequences of ecstasy in relation to other drugs, (2) a prospective cohort substudy in ecstasy-naive subjects with high risk for future ecstasy use, which should provide information on the causality and short-term course of ecstasy use and potential neurotoxicity, and (3) a retrospective cohort substudy in lifetime ecstasy users and matched controls of an existing epidemiological sample, which should provide information on the long-term course and outcome of ecstasy use in the general population. This chapter also gives an overview of the imaging techniques, psychopathology questionnaires, neurocognitive tests, and measurements of potentials confounders that were used. The final aim of the NeXT study, which also includes studies not described in this thesis, is to come to conclusions that can be used in prevention messages, clinical decision-making, and the development of an (inter)national ecstasy policy.

Part II: Use and validity of imaging techniques in ecstasy research

In part II we reviewed the existing literature on neuroimaging studies in human ecstasy users and further validated [ 123 I] -CIT SPECT, one of the most frequently applied imaging technique to study ecstasy neurotoxicity in humans.

In chapter 3 we reviewed the results of previous imaging studies in ecstasy users, in particular studies that used SPECT and positron emission tomography (PET) to measure SERT densities and 1 H-MRS to measure certain brain metabolites. Most of the PET and SPECT studies provided suggestive evidence that people who heavily use ecstasy are at risk of developing subcortical, and probably also cortical reductions in SERT densities. These effects seem to be dose-dependent and probably (partly) reversible. Moreover, females may be more vulnerable than males. From this review it seems that 1 H-MRS is a less sensitive technique for studying ecstasy's neurotoxic potential. The studies in this review were all retrospective and mainly included heavy ecstasy users, so we suggested that future ecstasy studies should address the effects of low dose ecstasy use and that longitudinal studies in human ecstasy users are needed to draw more definite conclusions on the causal role of ecstasy use in the observed differences in neuroimaging paramteres between ecstasy users and non-users.

In chapter 4 we assessed the v alidity of [ 123 I] -CIT SPECT in detecting MDMA-induced neurotoxicity in vivo in rats using a newly developed high-resolution pinhole SPECT system. We showed that both in vivo and ex vivo , thalamic, but not striatal, uptake ratios were reduced after MDMA treatment. This suggests that [ 123 I] -CIT SPECT is able to detect MDMA-induced loss of SERTs and therefore may be a promising technique to perform serial studies on MDMA-induced serotonergic neurotoxicity in living small animals.

Although [ 123 I] -CIT SPECT has already been used to assess SERT densities in the human brain, there was still discussion about its validity because [ 123 I] -CIT does not bind selectively to SERTs but also to dopamine transporters (DATs). In chapter 5 we aimed to investigate the validity of [ 123 I] -CIT SPECT to measure SERT densities in both SERT-rich and SERT-low areas in the living human brain using a double-blind, placebo-controlled, crossover design with the selective serotonin reuptake inhibitor (SSRI) citalopram. We report that citalopram reduced [ 123 I] -CIT binding ratios in SERT-rich midbrain and (hypo)thalamus. Binding ratios were also lower after citalopram in SERT-low cortical areas, but statistical significance was only reached in several cortical areas using a voxel-by-voxel analysis and not with a region of interest (ROI) analysis. In addition, we showed that citalopram increased binding ratios in the DAT-rich striatum. The results show that [ 123 I] -CIT SPECT is a valid technique to study SERT binding in vivo in human brain in SERT-rich areas. Although some evidence was provided that [ 123 I] -CIT SPECT may be used to measure SERTs in SERT-low cortical areas, these measurements must be interpreted with caution.

The fact that [ 123 I] -CIT does not selectively bind to SERTs but also to DATs is a disadvantage when one aims to study the serotonin system, like in ecstasy users. Therefore, a new radiotracer [ 123 I]ADAM has been developed with a high affinity for SERTs but not for other transporters like DATs, which makes it possible to asses SERTs more selectively. In chapter 6 we examined the optimal time course of [ 123 I]ADAM binding to central SERTs in young adults. The time of peak-specific [ 123 I]ADAM binding was highly variable among subjects, but specific binding in the SERT-rich (hypo)thalamus peaked within 5 h post injection (p.i.) in all subjects. Moreover, in this brain area, binding ratios of specific to nonspecific binding did not significantly change between 3 and 6 h p.i., and peaked 5 h p.i.. Therefore, we suggest that 5 h p.i. is an optimal time point for single-scan [ 123 I]ADAM SPECT studies in humans.

Part III: Retrospective studies in heavy ecstasy users

There is an ongoing discussion whether previously reported neurotoxic effects are caused by ecstasy, by other drugs or by the combinations of drugs 3,28-32 . Therefore, in chapter 7 we aimed to distinguish the specific/independent effects of ecstasy and the relative contributions of amphetamine, cocaine, and cannabis on the brain in a sample with variation in type and amount of drugs that were used with a combination of 1 H-MRS, DTI, PWI and [ 123 I] -CIT SPECT. Heavy ecstasy use showed no effect on 1 H-MRS brain metabolite ratios and DTI-derived ADC. However, heavy ecstasy use was associated with lower FA in the thalamus, higher rrCBV in the thalamus and temporal grey matter and lower [ 123 I] -CIT binding in the thalamus, frontal grey matter, and temporal grey matter. After adjusting for the use of drugs other than ecstasy and potential confounders (gender, verbal IQ, smoking) there was still a significant specific effect of ecstasy on the brain imaging parameters in the thalamus. Amphetamine and cocaine had a significant effect on some outcome parameters in brain areas other than the thalamus, but these findings were less consistent and converging than the robust findings associated with ecstasy use. Cannabis had no effect on any of the outcome parameters. This study therefore suggests strong converging evidence for a specific toxic effect of ecstasy on serotonergic axons in the thalamus with decreased [ 123 I] -CIT binding, probably reflecting damage to terminals of serotonergic axons, with a related decrease in FA due to axonal loss and increased rrCBV due to vasodilatation caused by sustained serotonin depletion.

Because serotonin is important for many neurocognitive and psychopathological processes, like memory and mood 33-36 we also looked at potential clinical consequences of ecstasy use. In chapter 8 we assessed the effects of ecstasy use on mood, measured with the BDI and the composite international diagnostic interview (CIDI) and focussed on its association with SERT densities, measured with [ 123 I] -CIT SPECT, dose, and gender. The prevalence of clinical depression assessed with CIDI did not differ between groups of moderate ecstasy users, heavy ecstasy users, former heavy ecstasy users, and drug-using but ecstasy-naive controls. However, BDI scores were higher in former heavy ecstasy users than in ecstasy-naive controls. Moreover, the total number of ecstasy tablets taken lifetime was associated with higher BDI scores for depressive mood. We did not find that depressed mood in ecstasy users was associated with gender or with a decrease in SERT density.

In chapter 9, the effect of moderate, heavy, and former ecstasy use on cognitive function was investigated. As females may be more vulnerable for the effects of ecstasy than males 37-39 and as SERTs are important in the regulation of synaptic serotonin transmission, we examined whether the effects of ecstasy on cognition, measured with neuropsychological tests, were different for females and males and for subjects with a different polymorphism in the serotonin transporter promoter gene region (5-HTTLPR). Heavy and former ecstasy users performed poorer on memory tasks than controls, while moderate ecstasy users did not. There was no difference between groups on reaction times or attention/ executive functioning. We also did not observe a significant effect of 5-HTTLPR or gender on test performance.

Part IV: Prospective studies in low dose ecstasy users

In part IV we describe the prospective substudies of the NeXT study, in which for the first time sustained effects of ecstasy were prospectively assessed in novel users. For this purpose a group of 188 ecstasy-naive subjects with an increased risk for future ecstasy user were assessed at baseline and followed during a period of about 18 months. In chapter 10, the first 30 incident ecstasy users were assessed with a combination of 1 H-MRS, DTI and PWI and self-report questionnaires on psychopathology before and quite soon after their first ecstasy use (mean of 1.8 ecstasy tablets). As brain metabolites and FA, parameters of structural neuronal damage, did not change after ecstasy use, we found no indications that incidental ecstasy use leads to extensive axonal damage. However, we did find sustained decreases in rrCBV in the thalamus, dorsolateral frontal cortex and superior parietal cortex, and a decrease in ADC in the thalamus after ecstasy use. This may indicate that even a low dose of ecstasy can induce prolonged vasoconstriction in some brain areas, although it is not known yet whether this effect is permanent. However, this has to be replicated in additional studies, because after correction for multiple comparisons only the rrCBV decrease in the dorsolateral frontal cortex remained significant. We also observed a small but significantly increase in impulsivity and small but significant decrease in depression scores after ecstasy use.

At the end of the follow-up period we assessed the effects of ecstasy use on the brain with SPECT and MR imaging parameters and psychopathology self-report questionnaires by comparing 59 incident ecstasy users (mean use of 6.0 tablets) with 56 persistent ecstasy-naive controls. Comparisons were corrected for baseline measurements. In chapter 11 we describe that compared to persistent ecstasy-naive subjects, novel, mainly, low dose, ecstasy users showed a decreased FA, increased ADC and decreased rrCBV in certain brain areas, mainly the basal ganglia. This suggests sustained vasoconstriction and probably damage to axons of brain neurons due to low dosages of ecstasy. Although we did not observe changes in SERT densities and neurometabolites, these results suggest that ecstasy even in low doses may have sustained effects on the brain.

In chapter 12 we assessed the relationship between ecstasy use and self-reported depression, impulsivity, and sensation seeking in (almost) the same prospective study group. We found that depression, impulsivity, and sensation seeking did not predict first time ecstasy use in this population of young adults with the intention to start using ecstasy. At the follow-up session, a significant effect of ecstasy use on the general and the disinhibition subscales of the sensation seeking scale were observed, while no effects of ecstasy use were found on depression and impulsivity scores. This suggest that low level ecstasy use does not seem to cause depression or impulsivity, although low level ecstasy use may increase (certain aspects of) sensation seeking.