Introduction
Neuroscientists interested in studying pharmacological and toxicological effects of MDMA continue to conduct studies in rats, mice, and monkeys, and in vitro studies with human, rodent, and other cell lines. The bulk of the research published between March 2004 and January 2005 focused on MDMA neurotoxicity, metabolism, and pharmacology, while other studies focused on specific issues relating to MDMA pharmacology or toxicity.

This review will only examine research containing information that can be used in making estimates of the potential risks or benefits of MDMA to humans, and hence will not examine drug discrimination or the majority of behavioral pharmacology studies. However, the review will address some of this research if and when findings shed light on relevant research. People wishing to learn more about studies excluded from this review can learn more about drug discrimination and behavioral research in non-human animals in reviews published in 2003 (Cole and Sumnall 2003B; Green et al. 2003). Most of the studies published subsequent to the most recent update of the IB sought to clarify or further examine an issue, problem or hypothesis proposed in previously published reports. These include studies examining neurotoxicity and hyperthermia. In a few cases, researchers tried to replicate features of the dance event or nightclub environment to see whether these features altered the degree of MDMA neurotoxicity or hyperthermia. Though results from several recent publications question the accuracy of brain serotonin and other measures as indicators of MDMA neurotoxicity, none of these studies produced findings that significantly increase or decrease the estimated risk to human study participants in trials of MDMA. There are now even more studies clearly suggesting that variations in the enzyme CYP2D6 are of little significance in estimating risk of adverse events with MDMA, but evidence supporting this case has already been discussed in the IB itself and in all successive updates (see Baggott et al. 2001; Jerome and Baggott 2003; Jerome 2004). Studies in non-human animals continue to find reduced brain serotonin and other changes associated with damage to serotonin axons, and some studies detected effects on muscle tissue, the immune system and the liver. High ambient and body temperature continues to play a role in MDMA toxicity in non-human animal models. A number of studies sought to understand mechanisms of MDMA-induced hyperthermia, and include a review seeking to incorporate findings on hyperthermia from separate research programs in non-human animals. Studies published between March 2004 and January 2005 offer a number of interesting findings concerning MDMA, but do not alter the degree of estimated risks and benefits faced by humans exposed to MDMA in clinical trials.

Neurotoxicity
A majority of the published studies of MDMA in non-human animals are concerned with MDMA neurotoxicity. The history of MDMA neurotoxicity research in non-human animals has been addressed in the IB and in updates to the IB (Baggott et al. 2001; Jerome and Baggott 2003; Jerome 2004), and elsewhere (Cole and Sumnall 2003; Green et al. 2003). Researchers continue to study mechanisms of MDMA neurotoxicity, environmental factors that might exacerbate or attenuate it, and means of abating it. Some reports offer support for one of several models of MDMA neurotoxicity, with a number of studies seeking to establish the source of oxidative stress seen after MDMA. Several recently published studies call into question the use of reduced brain serotonin as an indicator of MDMA neurotoxicity, while others failed to find signs of neurotoxicity after lower doses of MDMA. If researchers continue to discover difficulties with using reduced brain serotonin as a measure of MDMA neurotoxicity, this may lead to questioning the significance of a large body of research findings in non-human animals. However, at present, findings questioning the accuracy of lower brain serotonin as a marker of MDMA neurotoxicity have not yet been replicated, and it is notable that other researchers have detected damage to serotonin neurons without relying on measures of brain serotonin, serotonin transporter sites, or levels of glial activation (Callahan et al. 2001). While future study findings may lead to a reconsideration of the significance of previous studies in non-human animals, reconsideration of this research would be premature at present.

A study in mice measured striatal dopamine and serotonin levels after central or peripheral administration of MDMA or the reactive MDMA metabolite HHMA (DHMA) (Escobedo et al. 2004). When assessed seven days after drug administration, the researchers found that peripherally administered MDMA reduced striatal dopamine and metabolites, but only at the highest dose of centrally administered MDMA, a dose far in excess of brain MDMA levels measured after peripheral administration. Peripherally injected HHMA failed to reduce striatal dopamine seven days later, but intrastriatal administrations at the highest dose tested reduced striatal dopamine. HHMA was detected in plasma, but not in brain, after MDMA administration. The researchers' failure to detect HHMA in brain indicator that HHMA does not cross the blood-brain barrier and does not arise during metabolism in the brain. Study findings fail to indict HHMA as a direct producer of MDMA neurotoxicity in mice, but suggest that HHMA may be metabolized into other compounds that are responsible for mouse MDMA neurotoxicity. Since mice are the only species so far to show dopamine neurotoxicity after MDMA, studies in mice are probably not relevant to estimates of human MDMA neurotoxicity, but the findings may be helpful in considering studies in other species that seek to separate the effects of MDMA from effects produced by its metabolites.

Three studies in rats generated interesting findings concerning the presence and significance of presumed indicators of MDMA neurotoxicity (Orio et al. 2004; Sanchez et al. 2004; Wang et al. 2004). Two of these studies were performed in the Dark Agouti rat strain (Orio et al. 2004; Sanchez et al. 2004), a rat strain believed to be more sensitive to the effects of MDMA than other strains, while the other study was performed in the more typical Sprague-Dawley strain. Orio and colleagues administered a single i.p. injection of 12.5 mg/kg MDMA to Dark Agouti rats housed in a comfortably warm (22° C. or 72° F) or a cool (4° C, or 39° F) environment, and then assessed microglial activation and levels of glial fibrillary acidic protein (GFAP) in frontal cortex and hypothalamus one to 24 hours later, and 7 days later (Orio et al. 2004), with increased GFAP levels considered indicative of neurotoxicity. The researchers found increased microglial activation in rats kept at warm and cold ambient temperatures, and failed to detect any changes in GFAP in rats from either condition. The cold environment prevented hyperthermia, but did not prevent microglial activation, or increased levels of proinflammatory cytokine IL-Beta. Orio and colleagues also determined that increased IL-Beta was not associated with hyperthermia, since an IL-Beta antagonist failed to attenuate MDMA-induced hyperthermia, and the serotonin uptake inhibitor fluoxetine reduced IL-Beta levels without reducing hyperthermia. The authors concluded that 12.5 mg/kg MDMA produced a stress response in the brain, but that it did not produce neurotoxicity. In a study in Sprague-Dawley rats, Wang and colleagues also assessed GFAP levels after administering three doses of 7.5 mg/kg MDMA in a six-hour period (Wang et al. 2004). Rather than use interspecies scaling, Wang and colleagues selected their dose with "effects scaling," which estimates dose equivalence on the basis of producing similar pharmacological effects, such as drug recognition in drug discrimination studies. Rats had lower brain serotonin levels after MDMA, but they did not have fewer brain serotonin transporter sites or increased GFAP levels. By contrast, the known serotonin neurotoxin 5,7-dihydroxytryptamine (5,7-DHT) decreased numbers of serotonin transporter sites and increased GFAP levels. Assuming that GFAP levels and number of serotonin transporter sites, are accurate measures of serotonin neurotoxicity, these findings suggest that reduced serotonin levels may be less indicative of neurotoxicity, and that at least at the doses used in this study, MDMA does not produce effects similar to those of a known serotonin neurotoxin. Lastly, Sanchez and colleagues assessed serotonin and dopamine levels in Dark Agouti rats given three doses of up to 6 mg/kg MDMA in a six-hour period to simulate human "binge" dosing, and compared effects of 4 mg/kg given in a 19° C (66° F) and a 30° C (86° F) environment. MDMA dose- dependently reduced brain serotonin and increased body temperature, but failed to reduce brain dopamine at either dose regimen or when given at either ambient temperature. This research was spurred on in part by findings of MDMA-induced dopamine toxicity in non- human primates that were later retracted when the researchers learned that they had administered the wrong drug (Ricaurte et al. 2002; Ricaurte et al. 2003).

If the results described above are replicated in future studies, they may indicate that lower brain serotonin may not be a sufficient indicator of damage to serotonin axons. If this is the case, then less weight may be given to studies that determined MDMA neurotoxicity solely on the basis of lower brain serotonin levels. Since each study assessed specific brain areas, such as cortex (Orio et al. 2004; Sanchez et al. 2004; Wang et al. 2004), hypothalamus (Orio et al. 2004) or hippocampus (Sanchez et al. 2004; Wang et al. 2004), it is possible that researchers would have detected fewer serotonin transporter sites or increased GFAP in other brain areas. It is also notable that researchers have used other methods of detecting MDMA neurotoxicity, such as reduced anterograde axonal transport after repeated doses of MDMA (Callahan et al. 2001), so recent research cannot be used to dismiss all MDMA neurotoxicity studies. Since one of the studies listed above found dose-dependent reductions in brain serotonin (Sanchez et al. 2004), and since the study of anterograde axonal transport relied on considerably higher doses of MDMA, it might also be the case that higher dose regimens could have affected serotonin transporter and GFAP levels. Nevertheless, these provocative findings should stimulate further research into evaluating the accuracy and sensitivity of MDMA neurotoxicity measures. Secondly, these studies suggest that MDMA effects on brain serotonin are dose-dependent. Controversy over interspecies scaling (de la Torre and Farre 2004) may lead to more research employing "effects scaling" or other means of calculating equivalent drug doses across species.

Researchers investigating the relationship between bioenergetic stress and MDMA neurotoxicity in neonatal and adult rats assessed glycogenolysis (glycogen consumption) in the left caudal quarter of rat brains one hour after a single s.c. administration of 20 mg/kg MDMA (Darvesh and Gudelsky 2004). The authors found that MDMA given at 17° C (approximately 63° F) was not associated with increased glycogenolysis in 21-day old or 70-day old rats, and that MDMA administered at 24° C (approximately 75° F) increased glycogenolysis in both groups of rats. However, when striatal serotonin was assessed seven days after MDMA administration, Darvesh and Gudelsky only detected reduced brain serotonin in adult rats kept at both 17° and 24° C, suggesting that glycogenolysis was not associated with MDMA-induced reductions in brain serotonin. If brain glycogenolysis is a good measure of bioenergetic stress, then these findings in turn suggest that bioenergetic stress does not play a prominent role in MDMA neurotoxicity. The possibility that reduced brain serotonin may not be a good indicator of damage to serotonin axons further clouds an interpretation of these findings, since in this case, glycogenolysis, but not reduced brain serotonin, may be a better marker of MDMA neurotoxicity.

Garcia-Osta and colleagues compared the effects of 10 mg/kg MDMA with 5 mg/kg para-chlorophenylalanine (PCPA), a compound that interferes with serotonin synthesis and depletes brain serotonin, in rats killed from 2 to 48 hours post-drug (Garcia-Osta et al. 2004). The researchers measured serotonin in frontal cortex and hippocampus, tryptophan hydroxylase (TPH) gene expression (TPH is an enzyme that transforms tryptophan into serotonin), and levels of a protein associated with neuronal activity. MDMA and PCPA both increased this protein and decreased TPH gene activity early after drug administration, but MDMA was associated with increased TPH activity 48 hours post-drug, while PCPA reduced TPH gene expression 48 hours later. MDMA produced a transient decline in frontal cortex serotonin that began returning to normal 48 hours later, and a lasting reduction in hippocampal serotonin that was present 48 hours later. By contrast, PCPA-associated reductions in frontal cortex and hippocampal serotonin levels were still detectable 48 hours later. These study results are in agreement with earlier reports of MDMA interfering with serotonin synthesis (e.g. Che et al. 1995; Johnson et al. 1992), but also suggest that MDMA does not permanently reduce TPH activity, and that reductions in TPH gene expression does not necessarily indicate MDMA neurotoxicity.

Sveen and colleagues incubated neonatal rat hippocampal slices for a week with 50 or 100 mcM MDMA (about 0.996 or 1.93 mg), and examined slices with the Fluoro-Jade stain (Sveen et al. 2004). They failed to find any signs of neurotoxicity. The lack of evidence for MDMA-associated neuronal damage implies that serotonin projections into the hippocampus or MDMA metabolites are needed to produce neurotoxicity . However, neonatal rats may be less sensitive to the neurotoxic effects of MDMA than adults (Broening et al. 1995; Broening et al. 2001; Colado et al. 1997, but see Meyer and Ali 2002). One previous in-vivo study using Fluoro-Jade staining in rat forebrain detected signs of neuronal loss after higher doses of MDMA in combination with hyperthermia (Schmued et al. 2002) while another study examining mouse striatum detected signs of oxidative stress, such as indications of DNA fragmentation, but no signs of cell death (Fornai et al. 2004).

Winsauer and colleagues compared the effects of intra-peritoneal (i.p.) and subcutaneous (s.c.) MDMA on brain neurotransmitter levels in rats (Winsauer et al. 2004), with i.p. dosing considered more similar to oral dosing, the most common route of administration in humans. Rats received twice-daily doses of saline or 10 mg/kg MDMA for four consecutive days, and the researchers measured monoamine levels in brain areas from rats killed 3, or 13 to 14 days after the last saline or MDMA administration. Both i.p. and s.c. MDMA reduced serotonin in most brain areas three days post-MDMA, but brain serotonin levels were normal in most brain areas 14 days later, though levels of the serotonin metabolite 5-HIAA was reduced in some brain areas. Neither i.p. nor s.c. MDMA altered brain norepinephrine levels, and s.c. MDMA failed to change dopamine or epinephrine levels either three or 14 days later. Winsauer and colleagues detected increased midbrain dopamine levels three days after i.p. MDMA, decreased hippocampal dopamine 13 to 14 days later, lower midbrain dopamine levels three days post-drug and lower hypothalamic dopamine 13 to 14 days later (Winsauer et al. 2004). It is notable that this is the first study of brain epinephrine levels after MDMA. Because changes in brain neurotransmitter may be secondary to changes in other neurotransmitter systems or may reflect changes in numbers active receptors for a neurotransmitter, the significance of these findings is not clear, and the authors do not make claims concerning the cause or causes of altered neurotransmitter levels. To date, no one has reported any other indicators of damage to other neurotransmitter systems after MDMA.

Two in vitro studies examined the role played by dopamine in MDMA serotonin neurotoxicity, either in connection with MDMA metabolites (Jones et al. 2004) or in production of oxidative stress through generation of hydrogen peroxide (Hrometz et al. 2004). In the first study, the researchers assessed inhibition of the serotonin transporter and reactive oxygen species levels (the direct cause of oxidative stress) in human cell cultures made to express the human serotonin or dopamine transporter and exposed to one of four compounds; MDMA, MDA, or two thioether compounds proposed as potential MDMA metabolites. Jones and colleagues found that all four compounds inhibited the serotonin transporter and produced reactive oxygen species in cells. The thioether compounds produced more reactive oxygen species than MDMA or MDA. Jones and colleagues also found that all four compounds, especially the thioether compounds, ushered dopamine into cells expressing the serotonin transporter. In the other study, Hrometz and colleagues measured hydrogen peroxide in serotonin-transporter expressing human cells after exposure to dopamine, MDMA, and the two combined (Hrometz et al. 2004). The combination of MDMA and dopamine was associated with a far greater increase in hydrogen peroxide than either compound individually, and MDMA alone did not increase hydrogen peroxide. The researchers used the SSRI fluoxetine and the MAO-B inhibitor L-deprenyl to investigate whether presence of the serotonin transporter or an aspect of dopamine metabolism were involved in hydrogen peroxide production. They found that both fluoxetine and L-deprenyl reduced, but did not eliminate, hydrogen peroxide production after the combination of MDMA and dopamine. None of the combinations described above reduced cell viability. Both the work of Jones and colleagues and that of Hrometz and colleagues imply that dopamine plays a role in MDMA-associated oxidative stress.

A study in mice found that a protein called metallothionein-1 (MT-1) protected the rodents from MDMA neurotoxicity (Xie et al. 2004). Gene expression associated with this protein increased after administration of 47 mg/kg MDMA, and MT-1 "knockout" mice had lower brain dopamine after receiving four doses of 30 mg/kg MDMA than "wild-type" mice. When MT-1 "knockout" mice were given supplemental MT-1 before MDMA, their brain dopamine levels were more similar to those of wild-type mice. Xie and colleagues also found that methamphetamine and MPTP increased MT-1 gene expression, suggesting that this protein is called on in response to various compounds with dopamine neurotoxicity. This study suggests that at least in mice, some neuroprotective genes are activated in response to repeated doses of MDMA. Previous investigations of the effects of environmental factors on MDMA neurotoxicity centered mainly around studies of ambient temperature (Malberg et al. 1996; Malberg et al. 1998; Malpass et al. 1999), and one study examined the effects of loud noise and MDMA on cardiac tissue (Gesi et al. 2002). Three studies published between March 2004 and January 2005 attempted to model specific features of common settings of MDMA use to study their effects on neurotoxicity (Darvesh and Gudelsky 2004; Gesi et al. 2004) or muscle tissue (Duarte et al. 2004). The first study will be discussed here, and the second will be addressed below in "Toxicity."

Darvesh and Gudelsky found that rats receiving a single s.c. injection of 20 mg/kg MDMA at 24° C (75° F) had lower striatal serotonin levels than rats that received the same dose of MDMA in a 17° C (63° F) environment, with brain serotonin measured 1 hour post-drug (Darvesh and Gudelsky 2004). It should be noted young rats did not exhibit reduced brain serotonin after receiving MDMA under either ambient temperature. Gesi and colleagues studied potential interactions between MDMA and continuously broadcast loud noise on striatal dopamine, tyrosine hydroxylase (an enzyme involved in dopamine synthesis), and GFAP in brain tissue from mice killed seven days after exposure. The researchers found that combining four injections of 7.5 mg/kg MDMA with six hours of white noise produced a greater decrease in striatal dopamine and a greater increase in GFAP than MDMA alone. The combination of prolonged loud noise and MDMA increased locomotion in an open field two hours and seven days post-treatment, possibly indicating that MDMA and noise-exposed mice were less anxious than saline-treated mice. This research team has previously examined the effects of MDMA and loud noise on heart tissue (Gesi et al. 2002A; Gesi et al. 2002B). Humans who use ecstasy at dance events and nightclubs differ from mice in that they choose to expose themselves to loud noise and likely find the experience enjoyable, while the mice may find loud noise to be stressful. In this case, it is possible that stress, and not loud noise, may increase MDMA neurotoxicity (Johnson et al. 2002, but see Johnson et al. 2004). Future rodent models of dance events may wish to train animals to associate a reward with loud noise before combining noise with MDMA. If these study results are replicated, then they might suggest another way in which people consuming ecstasy at a party or dance event may face a greater risk than people in clinical trials of MDMA. None of the MDMA neurotoxicity studies reviewed here present findings that require a change in the estimated risks and benefits to participants in trials of MDMA. If further research continues to find that brain serotonin and serotonin transporter levels are not reliable indicators of serotonin axon toxicity, as reported by Wang and colleagues, then a large body of research will need to be reconsidered. However, more research following along these lines is needed before concluding that earlier research overestimated MDMA neurotoxicity, and it should be noted that methods other than those described above have also detected indicators of damage or dysfunction in serotonin axons (Callahan et al. 2001). It is already recognized that higher ambient and body temperature play a role in MDMA neurotoxicity, and so recent findings do not lead to changes in the current understanding of MDMA neurotoxicity or effects on the serotonin system. As discussed in "Clinical Trials," MDMA produces a slight increase in body temperature when given in controlled settings, and clinical trials are conducted in comfortably warm environments, so that people will be very unlikely to experience the high ambient or body temperatures that exacerbate MDMA neurotoxicity in rodents. Findings of potentiated neurotoxicity when MDMA was combined with loud noise, and findings of increased hyperthermia and muscle damage when MDMA was combined with vigorous exercise, described below in "Toxicity") all point to the significance of setting in accentuating or attenuating drug-related risks. After considering current research findings, the risk of MDMA neurotoxicity in humans enrolled in clinical trials remains low.

Long-term behavioral effects of MDMA
Studies in non-human animals have examined anxiety, social interaction, and cognitive function subsequent to various doses of MDMA. Many, but not all, previous reports used MDMA in doses intended to reduce brain serotonin (for example Clemens et al. 2004; McGregor et al. 2003A; McGregor et al. 2003B; Sumnall et al. 2004B; Thompson et al. 2004). Some researchers used non-human animal studies to detect long-term effects potentially linked with ecstasy use, such as anxiety or impaired cognitive function. To date, the body of research into the long-term behavioral effects has been contradictory (Baggott et al. 2001; Green and McGregor. 2002; Jerome and Baggott 2003; Jerome 2004).

Seven studies addressing one or more behavioral effects of MDMA have been published since the last update of the literature review (Jerome 2004). Five studies assessed anxiety (Gesi et al. 2004; Ho et al. 2004; Piper and Meyer 2004; Sumnall et al. 2004B), and two examined changes in cognitive function (Piper and Meyer 2004; Winsauer et al. 2004). The researchers performed most of these studies in adult rats, though one study used "periadolescent" rats (Piper and Meyer 2004), and another used mice (Gesi et al. 2004). While fewer reports examining potential long-term effects of MDMA on anxiety in rats appeared this year than in past years, the issue remains of interest to some researchers (Ho et al. 2004; Piper and Meyer 2004; Sumnall et al. 2004B), as do the issues of sensitization to other drugs (Sumnall et al. 2004B). An investigation of the combined effects of loud noise and MDMA also assessed open field activity, usually treated as a measure of anxiety, seven days post-treatment (Gesi et al. 2004).

Ho and colleagues assessed the acute and long-term effects of a single 7.5 or 15 mg/kg dose of MDMA on Sprague-Dawley rats previously rated as low or high on trait anxiety on the basis of performance on the elevated plus-maze, with anxiety assessed again via plus-maze, open field and active avoidance. When tested 9 to 15 days later, MDMA did not increase or decrease anxiety, and the researchers did not detect an interaction between MDMA and trait anxiety. However, 7.5 mg/kg MDMA increased active avoidance in high-anxious rats and decreased it in low-anxious rats fifteen days post-MDMA, a finding that may indicate that MDMA accentuated differences in trait anxiety, or that MDMA facilitated fear-based learning in high-anxious rats and impaired fear-based learning in low-anxious rats. The researchers also detected a non-significant increase in immobility in the "forced swim" test, considered a model of "depression-like" behavior in rodents. Ho and colleagues' findings do not support the hypothesized interaction between MDMA and trait anxiety in rats proposed by Green and McGregor (2002), but it should be noted that Ho and colleagues performed a within-strain comparison rather than making cross-strain comparisons. It is interesting that a presumably non-neurotoxic dose of MDMA (7.5 mg/kg) accentuated differences in active avoidance seen in "low anxious" and "high anxious" rats, a sign that this change did not arise as a result of damage to serotonin axons. In another study examining the long-term effects of MDMA on plus- maze performance (Sumnall et al. 2004B), rats pre-treated with four injections of 10 mg/kg MDMA every 2 hours performed similarly to vehicle-treated rats. The researchers also found that MDMA and vehicle treated rats behaved similarly on the plus maze after heroin, ethanol, cocaine and MDMA, though there were differences relating to single behaviors, such as cocaine producing fewer head dips in MDMA-pretreated rats than in vehicle-pretreated rats, suggesting lower levels of anxiety. Though the MDMA dose regimen used by Sumnall et al. generally failed to alter the anxiety-related effects of other drugs, it did lower hippocampal serotonin levels and estimated numbers of serotonin transporter sites, suggesting that changed serotonin levels do not necessarily lead to changes in behavioral sensitivity to other drug. Two studies, one in rats and one in mice, found reduced anxiety after MDMA (Gesi et al. 2004; Piper and Meyer 2004). Piper and Meyer gave "periadolescent" two doses of 10 mg/kg MDMA every five days, administering the second dose 4 hours after the first dose (Piper and Meyer 2004). Rats received approximately seven pairs of doses for a month-long period. Approximately nine days after the final dose, these rats spent more time in the open arms of the elevated plus maze, a sign of reduced anxiety, without increased activity in the open field. Piper and Meyer detected reduction in estimated serotonin transporter binding in MDMA-treated rats that was slightly less extreme than seen after more traditional dose regimens used in MDMA neurotoxicity research, and an association between hippocampal serotonin transporter sites and reduced anxiety. In a study in mice discussed above in "Neurotoxicity", Gesi and colleagues found that mice that received a demonstrably neurotoxic regimen of MDMA in combination with prolonged loud noise demonstrated increased open field activity seven days post-drug (Gesi et al. 2004). When it was not administered along with loud noise, the same dose of MDMA altered striatal dopamine and increased markers of MDMA neurotoxicity, but did not affect open field activity.

Researchers have examined the long-term effects of MDMA on learning and memory in non-human animals, often with mixed results (see discussion in IB; Frederick et al. 1995; Frederick et al.1997; Seiden et al. 1993; Taffe et al. 2002; Taffe et al. 2003; Williams et al. 2003). Two recent publications tackled long-term effects of MDMA on learning and memory (Piper and Meyer 2004; Winsauer et al. 2004). In one study, Piper and Meyer detected impaired object recognition in "periadolescent" rats given two injections 10 mg/kg MDMA four hours apart every five days for a month (Piper and Meyer 2004). Though rats in this study had moderately reduced levels of serotonin transporter, the researchers failed to detect any relationship between numbers of serotonin transporter sites and impaired object recognition. These findings are similar to dissociations between indirect measures of serotonin transporter sites and memory and executive function in ecstasy users (see Buchert et al. 2003; Curran et al. 2003; Gijsman et al. 2002; Thomasius et al. 2003; Verkes et al. 2000).

Researchers examined the acute effects of the muscarinic antagonist (anti-cholinergic) scopolamine on learning and memory in rats before and after administering a regimen of 10 mg/kg MDMA given twice-daily on four consecutive days (Winsauer et al. 2004). As expected, scopolamine impaired acquisition and recall, and both MDMA regimens impaired acquisition and performance on the days MDMA was administered. Winsauer and colleagues found that the second MDMA regimen affected acquisition and recall to a greater degree than the first regimen. As described above in "Neurotoxicity", Winsauer and colleagues found that both i.p. and s.c. MDMA reduced serotonin in brain regions, including hippocampus, hypothalamus and (after s.c. dosing) cortex. However, contrary to expectations, the researchers found that MDMA attenuated scopolamine-induced impairment in acquisition and memory. The researchers offer several hypotheses for these findings, including the attenuation of serotonin-regulated inhibition of acetylcholine release, and the possibility that the combined effects of reduced serotonin and acetylcholine are different from reductions in each system alone. Though it is not clear that findings from a scopolamine challenge can be generalized to reduced cholinergic function in humans, as seen with Alzheimer's disease, these findings do not support claims that reductions in brain serotonin will further exacerbate impaired learning or memory seen with cholinergic deficits.

None of the studies referred to above employed comparable methods, and each focused on different long-term effects, but their findings suggest that the relationship between MDMA neurotoxicity and subsequent behavioral changes is complex and that presence or absence of changes in anxiety or learning after MDMA are not necessarily linked to damage to brain neurotransmitter systems. Similar dissociations can be seen when examining findings in ecstasy users. Findings from these studies in non-human animals do not argue for increasing or decreasing estimated risk to humans participating in clinical trials of MDMA.

Self-Administration
Previous research has found that rodents and non-human primates will self-administer MDMA and appear to find it rewarding (Fantegrossi et al. 2002; Fantegrossi et al. 2004; Robledo et al. 2004; Schenk et al. 2003; Wakonigg et al. 2004A). Two studies published between March, 2004 and January, 2005 examined MDMA self-administration (Daniela et al. 2004; Robledo et al. 2004B). In one study (Robledo et al. 2004B), both normal mice and knockout mice lacking the mu opioid receptor developed conditioned place preference, the tendency to prefer being in a location associated with a drug, in response to injections of 10 mg/kg MDMA. Unsurprisingly, these findings suggest that the rewarding properties of MDMA are more closely linked with dopamine than with opioids. Dopamine also played a role in the rewarding properties of MDMA for rats (Daniela et al. 2004). Rats stopped working for injections of 0.25 mg/kg MDMA when pre-treated with a D1 receptor antagonist, while rats worked even harder to receive 2 mg/kg MDMA when it was combined with the D1 antagonist, suggesting a role for D1 receptors in the rewarding properties of MDMA. These MDMA self-administration studies do not increase or decrease estimated abuse potential in humans. Up to approximately 6% of ecstasy users in a representative sample of Munich residents reporting ecstasy use were diagnosed with ecstasy abuse or dependence (Lieb et al. 2002), though higher figures have been reported in non-representative samples (Cottler et al. 2001; Topp et al. 2002). MDMA-naive participants did not report wishing to use MDMA outside the laboratory (Liechti et al. 2001).

Thermoregulation
Because hyperthermia is one of the most common serious adverse effects of ecstasy use in uncontrolled settings and because it might make MDMA neurotoxicity worse, researchers continue to perform studies on MDMA and hyperthermia in rodents and in vitro (Baggott et al. 2001; Jerome and Baggott 203; Jerome 2004). Ten studies published or located between March 2004 and January 2005 examined the thermoregulatory effects of MDMA. Researchers carried out studies in rodents (Bexis et al. 2004; Blessing et al. 2003; Duarte et al. 2004; Herin et al. 2004; Ho et al. 2004; Ootsuka et al. 2004; Rusyniak et al. 2004; Sprague et al. 2004), pigs (Rosa-Neto et al. 2004) and in vitro (Rusyniak et al. 2004). Most studies examined at least one model of MDMA-induced hyperthermia, while two studies in rodents examined the effects of environmental variables on MDMA- induced hyperthermia. In addition, a team of researchers offered their synthesis of the literature, focusing on the role of the sympathetic nervous system in MDMA-induced hyperthermia (Mills et al. 2004).

Researchers have investigated cutaneous vasoconstriction (constricted blood vessels near the skin), the sympathetic nervous system, specific neurotransmitter systems, mitochondrial metabolism, and hypothalamic activity as potential causes or factors involved in MDMA-induced hyperthermia. Blessing and colleagues reported that MDMA (6 mg/kg in rabbits, 10 mg/kg in rats) induced cutaneous vasoconstriction and hyperthermia, and that the atypical antipsychotic drugs clozapine and olanzapine reduced both effects (Blessing et al. 2003). An examination of sympathetic system activity in anesthetized rats and rabbits found that MDMA-induced hyperthermia was linked with the stimulation of spinal 5HT2A receptors (Ootsuka et al. 2004). Herin and associates reported that pretreatment with the serotonin 5HT2A receptor antagonist M100907 reduced hyperthermia in rats given 8 or 12 mg/kg S-(+)-MDMA (Herin et al. 2004), and the same 5HT2A antagonist reversed hyperthermia when given an hour after 12 mg/kg S- (+)-MDMA. Bexis and colleagues reported that administering the GABA(B) receptor agonist baclofen, but not the GABA(A) agonist muscimol, reduced hyperthermia in rats given 15 mg/kg MDMA (Bexis et al. 2004). Surprisingly, baclofen attenuated elevated body temperature without reducing MDMA-induced hyperactivity or elevated heart rate. Examining another transmitter system, Sprague and colleagues found the alpha1 receptor antagonist prazosin and the beta(3) antagonist SR59230A reduced MDMA-induced hyperthermia in rats (Sprague et al. 2004). As expected, researchers investigating the effects of MDMA on mitochondrial metabolism (discussed below) found that 40 mg/kg MDMA induced hyperthermia (Rusyniak et al. 2004), but found only slight indicators of mitochondrial dysregulation. An in vitro study performed by these researchers found that only extremely high concentrations of MDMA produced signs of dysregulation in mitochondrial metabolism. By contrast, the same research team reported that mice lacking a gene for uncoupling protein 3 (UCP3) were resistant to MDMA-induced hyperthermia (Mills et al. 2003). Sprague and colleagues also found that thyroid hormones are involved in methamphetamine-induced hyperthermia as well (Sprague et al. 2004), suggesting a common pathway for both types of drug-induced hyperthermia. Researchers in Denmark used PET imaging to scan the brains of anesthetized pigs given 1 mg/kg MDMA (Rosa-Neto et al. 2004), and found that increased cerebral blood flow in the hypothalamus was correlated with increased body temperature in individual pigs. It is notable that pig body temperature increased by 2 or 3 degrees C, whereas clinical trials using doses equal to or higher than 1 mg/kg produce only slight increases in human body temperature (see discussion below).

Researchers also examined potential impact of environmental factors and individual characteristics in MDMA-induced hyperthermia. Duarte and colleagues measured body temperature via radiotelemetry in a study of the effects of MDMA and strenuous exercise on rhabdomyolysis (muscle and organ damage) in mice (Duarte et al. 2004, see discussion below in "Toxicity"). Exercise increased body temperature, and 10 mg/kg MDMA, with or without exercise, further increased body temperature. In their study of more and less anxious rats described above in "Neurotoxicity", Ho and colleagues reported that 15 mg/kg MDMA, but not 7.5 mg/kg, increased body temperature in both less and more anxious rats, even though less anxious rats had higher baseline body temperatures (Ho et al. 2004).

Several investigations of MDMA neurotoxicity continue to support a role for high ambient temperature in MDMA-induced hyperthermia in rodents (Orio et al. 2004; Sanchez et al. 2004), and other researchers have studied the impact that other environmental factors, such as social interaction (Brown and Kiyatkin 2004) or ambient temperature (Darvesh and Gudelsky 2004) have on MDMA-induced hyperthermia. These studies found an association between higher ambient temperatures and higher body temperatures in rats given MDMA, and the relationship was detected in Dark Agouti and Long-Evans rats (Brown et al. 2004; Sanchez et al. 2004). One study found that interacting with an ovarectomized female rat also increased MDMA-induced hyperthermia, but not to the same degree as high ambient temperature (Brown and Kiyatkin 2004). These findings raise the possibility that people taking ecstasy in warm environments may be more likely to experience hyperthermia than people taking part in clinical trials of MDMA, placing them at greater risk for any potential MDMA neurotoxicity as well.

One of the research teams responsible for a number of rodent hyperthermia studies (e.g. Mills et al. 2003; Ootsuka et al. 2004; Sprague et al. 2004) reviewed research on MDMA- induced hyperthermia and offered a synthesis of the literature (Mills et al. 2004). Mills and colleagues posited that the sympathetic nervous system is involved in more than one mechanism behind MDMA-induced hyperthermia, with these mechanism including dysregulation of mitochondrial metabolism, vasoconstriction, and changes in monoamine release that affect brain areas that regulate thermoregulation. Mill and colleagues suggest that MDMA-induced hyperthermia can be reduced by using alpha or beta norepinephrine receptor antagonists.

Though each study used different measures and methods, an examination of these findings suggests that MDMA elevates body temperature through several potentially independent mechanisms. It is not clear how many of these mechanisms are present in humans, though Mills and colleagues note that unlike rodents, human adults do not possess brown adipose tissue, a source of non-shivering heat produced by uncoupling of mitochondrial metabolism. At doses used in controlled settings, MDMA produces a slight and sometimes undetectable increase in body temperature (see for instance Harris et al. 2002; Liechti and Vollenweider 2001; Mas et al. 1999; Tancer and Johanson 2003). Humans may be less sensitive than rodents are to the thermoregulatory effects of MDMA, or it may be that the effects will only be seen after the higher doses often used in rodent studies. Furthermore, clinical trials are not conducted in uncomfortably hot laboratories, and participants are not encouraged to exercise vigorously during clinical trials of MDMA. Participants undergoing MDMA-assisted psychotherapy will for the most part be sitting, reclining or lying down and will be involved in introspection. Hence risk of experiencing hyperthermia during a clinical trial of MDMA appears to be minimal.

Developmental Toxicity
Research on the potential developmental toxicity of MDMA has produced a number of contradictory findings (see Broening et al. 1995; Broening et al. 2001; Koprich et al. 2003; Meyer et al. 2002; Williams et al. 2003). In general, researchers found an absence of developmental effects when they gave MDMA prior to postnatal day 1, but effects appeared when they gave MDMA on or after postnatal day 10 (PND10), a period believed to be analogous to the third trimester of pregnancy in humans. Rats exposed to high and frequent doses of MDMA during this period exhibited impaired performance on measures of spatial memory (Williams et al. 2003). Some researchers have proposed that MDMA does not produce developmental effects before PND10 because it does not induce hyperthermia during this period (Aguirre et al. 1995, but see Meyer and Ali 2002). A few researchers also found increased anxiety and reduced social interactions after rats were given MDMA during a period after postnatal day 28, referred to by some as rat "periadolescence" (Bull et al. 2003; Bull et al. 2004; Fone et al. 2002). Three studies published between March, 2004 and January, 2005 investigated the developmental effects of MDMA, two in neonatal rats and one in slightly older rats. In one study, neonatal rats received 10 mg/kg MDMA twice-daily (the second dose four hours after the first) on postnatal days 1 to 4 (Meyer et al. 2004), and in the other study, rats received 5, 10 or 10 mg/kg twice-daily MDMA from PND11 to PND20 (Vorhees et al. 2004). Meyer and colleagues found that MDMA exposure immediately after birth reduced hippocampal serotonin when assessed 25 days later, but did not reduce forebrain serotonin levels. The researchers reported that MDMA increased signs of apoptotic (programmed cell death) activity a day afterwards. Numbers of hippocampal serotonin transporter sites were reduced after MDMA 25 and 60 days post-drug, while forebrain serotonin transporter sites were reduced 60 days later. When Meyer and colleagues examined brain serotonin transporter binding 9 months post-drug, fewer serotonin transporter sites were detected in some areas (visual and somatosensory cortex, caudate- putamen, and nucleus accumbens), but not in others. Inducing hyperthermia through a warm incubator had little effect on brain serotonin, serotonin transporter sites, or signs of apoptosis. Meyer and colleagues did not assess learning, memory or locomotion. Earlier studies failed to find changes in brain serotonin after giving MDMA to similarly aged rats (Broening et al. 1995; Broening et al. 2001), possibly as a result of using different dose regimens. In the second study, Vorhees and colleagues administered MDMA from PND11 and PND20. They found an association between MDMA and impaired performance on the Morris water maze, but not in the Barnes maze, a dry version of the water maze wherein a goal box is hidden in one of several holes in a brightly lit, and thus aversive, arena (Vorhees et al. 2004). However, water maze performance in MDMA- treated rats was only impaired when this task was presented before the Barnes maze. Vorhees and colleagues did not assess brain serotonin or serotonin transporter sites, though an earlier report by the same researchers detected lower levels of serotonin and dopamine in MDMA-treated rats (Koprich et al. 2003B). In contrast with this study, previous research detected an unqualified impairment in spatial memory after the same dosing schedule (Williams et al. 2003).

These and other studies in rats suggest that MDMA administered in the third trimester of pregnancy could alter developing brain organization, and that intense prenatal exposure could subsequently impair memory or learning. Several studies point to one or more "critical periods" of increased developmental toxicity, with most studies centering on the third trimester of pregnancy, though at least one study found developmental toxicity in rats born to mothers exposed to MDMA (Koprich et al. 2003B, but see Kelly et al. 2002). As described in "Clinical Trials," a case-control comparison examining a specific heart defect failed to establish a link between ecstasy use and the defect, largely owing to the extremely low level of ecstasy exposure seen in their sample (Bateman et al. 2004). It is possible that it is difficult to establish developmental toxicity in humans because most women discontinue or curtail use upon learning they are pregnant (Ho et al. 2001). Since at least some studies detect developmental toxicity after MDMA, women who are pregnant or lactating should continue to be excluded from clinical trials of MDMA. As described above in "Neurotoxicity," Darvesh and Gudelsky compared the effects of a single s.c. 20 mg/kg dose of MDMA on striatal serotonin 7 days post-drug in 21-day old and 70-day old rats (Darvesh and Gudelsky 2004). The authors detected reduced striatal serotonin only in adult rats. Somewhat surprisingly, these findings suggest that MDMA administered during a period occurring slightly later than the period described above failed to affect brain serotonin. However, the authors did not assess learning, memory or any other long-term effects in these rats.

Piper and Meyer administered MDMA during the period between PND35 and PND60, using a schedule of two 10 mg/kg doses spaced four hours apart, and with injections given every five days across this month-long period (Piper and Meyer 2004). As discussed above in "Long Term Behavioral Effects," rats receiving MDMA displayed less anxiety and impaired object recognition afterwards, effects not always seen in adult rats. However, few studies in adult rats have employed the dose regimen that Piper and Meyer used, so it is possible that these effects might be the result of dose regimen, and not developmental stage. As noted above, only reduced anxiety was related to changes in brain serotonin. Study findings suggest that long-term effects of MDMA may be age- dependent, but are preliminary at present. These findings raise the possibility that adults and adolescents may face risks of slightly different long-term effects, but do not increase or decrease the potential risk estimates for people taking part in clinical trials of MDMA.

Toxicity
Three studies of MDMA toxicity have been published between March, 2004 and January 2005. In one study, a team of researchers in Portugal first described in "Thermoregulation" examined the possible effects of strenuous exercise and MDMA on soleus (calf) muscle tissue by giving mice a 10 mg/kg injection of MDMA, placing them in a treadwheel set to spin at 75% of the maximum running speed for a mouse, or combining the two treatments (Duarte et al. 2004). When given alone, both strenuous exercise and MDMA were associated with signs of muscle damage, but soleus muscle fibers exhibited the most damage after a combination of exercise and MDMA, with signs of muscle damage apparent 1.5 hours after exercise and still visible 24 and 48 hours later.

Both this study and the work of Gesi and colleagues, described in "Neurotoxicity", suggest that adverse events seen after ecstasy use may due at least in part to aspects of the setting where ecstasy is most frequently used. It also remains true that humans, unlike the mice in these studies, voluntarily exercise after taking MDMA, so it is possible these findings might be due at least in part to stress, and not just to the combination of exercise and MDMA. Participants in clinical trials of MDMA will not be engaged in strenuous exercise, and so risk of muscle damage should be extremely minimal.

In a study investigating the effects of MDMA on the liver, rats received a single intra- gastric dose of 5 to 40 mg/kg MDMA, or 14 daily i.g. doses of 5, 10 or 20 mg/kg MDMA (Ninkovic et al. 2004, obtained abstract only). The researchers found increased superoxide dismutase, decreased glutathione, and increased lipid peroxidation after a single dose of MDMA, and dose-dependent increases in lipid peroxidation. Chronic MDMA also dose-dependently increased lipid peroxidation and reduced glutathione, but only the highest dose regimen (20 mg/kg) increased superoxide dismutase. Study findings indicate that oxidative stress can occur in the liver after MDMA. The lower rate of lipid peroxidation and attenuated decrease in glutathione after repeated dosing may indicate tolerance to MDMA after repeated doses, or these effects may be a sign of inhibited response to oxidative stress after repeated dosing. Previous investigations have found associations between usually high doses of MDMA and signs of hepatotoxicity, with body temperature playing a role in degree of toxicity (Beitia et al. 2000; Carvalho et al. 2002; Carvalho et al. 2003; see also Baggott et al. 2001; Jerome 2004). It is unclear whether the findings from this study are relevant to hepatotoxic potential in humans. They do not increase or decrease estimated risk of liver toxicity for study participants in clinical trials of MDMA.

Connor and associates sought to uncover the cause or causes for the decrease in the immune system stimulating and proinflammatory cytokine tumor necrosis factor alpha (TNFAlpha) and increase in the immunosuppressive and anti-inflammatory cytokine interleukin-10 (IL10) in rats given MDMA (Connor et al. 2004). The authors first confirmed that MDMA did produce these effects and demonstrated that an increase in IL- 10 was not causing the reduction in TNFAlpha. The researchers then manipulated levels of stress hormones, sympathetic system activity, and activity at adrenergic (norepinephrine-related) receptors. They concluded that reduction in TNFAlpha might arise as a result of stimulating the sympathetic nervous system or from norepinephrine release, and that MDMA increased IL-10 levels through stimulating beta adrenergic receptors. The immunological effects of MDMA are already known in humans (Pacifici et al. 2000; Pacifici et al. 2002; Pacifici et al. 2004) and rats (Connor et al. 2000). Surprisingly, Connor and colleagues did not examine the effects of serotonin release on TNFAlpha or IL-10 levels, even though a study in humans suggests that preventing serotonin release through paroxetine nearly eliminated IL-10 increase (Pacifici et al. 2004). Since this report studied mechanisms for the immunological effects of MDMA without reporting any new effects, these findings do not alter estimated risk for humans taking part in clinical trials of MDMA.

In addition to these studies, a team of Dutch researchers published a report on a possible mutation of the CYP2D6 enzyme, known to play a role in the metabolism of MDMA (Keizers et al. 2004). The researchers created a mutant version of the enzyme and genetically modified bacteria (e. coli) to express the new enzyme. Keizers and colleagues compared the "wild-type" and mutant enzyme on known CYP2D6 substrates (compounds the enzyme helps metabolize), including MDMA. The mutant version of CYP2D6 produced the familiar MDMA metabolites DHMA (HHMA) and MDA, but it also hydroxylated MDMA, producing what the authors referred to as N-OH-MDMA. These findings are of uncertain relevance, as there is no record of this mutation ever appearing in humans, and to date, the unusual metabolites produced by the potential mutant form of CYP2D6 have only been detected in horse urine (Damasia 2003, cited in Keizers et al. 2004). An examination of research into well-known variants of CYP2D6 found in humans suggests that these variants are not associated with adverse events (Gilhooly and Daly 2002), and that higher doses of MDMA may impede MDMA metabolism in most people (de la Torre and Farre 2004).

Concluding Remarks
Research reports published between March, 2004 and January 2005 examined the pharmacology and toxicity of MDMA in non-human animals and in vitro, with the bulk of the research testing or elaborating on models of MDMA neurotoxicity. Recent studies also investigated long-term behavioral effects of MDMA, potential developmental toxicity, effects on thermoregulation, and toxicity in other organs and systems, such as the liver and the immune system. If replicated, some findings that call into question indicators of MDMA neurotoxicity in other rodent studies may lead to a re-examination of previous research. However, at present, these findings remain preliminary. Other research studies have found changes in anxiety, learning and memory. It is notable that not all changes were associated with lower brain serotonin or other signs of MDMA neurotoxicity. Studies continue to support the possibility of developmental toxicity of MDMA in rats. Finally, one study examined oxidative stress in the livers of rats given MDMA, and another explored the cause or causes of immunological changes associated with MDMA. None of the studies call for an increase or decrease in estimated risk for people taking part in clinical trials of MDMA.