Despite a large body of research demonstrating that the popular psychoactive drug, (±)3,4-methylenedioxymethamphetamine (MDMA, “Ecstasy”), has the potential to destroy brain serotonin axon terminals (for reviews, see Green et al., 2003; Quinton and Yamamoto, 2006), recreational use of MDMA continues, and in recent years, numerous laboratories have begun testing various pharmacologic effects of MDMA in humans (for review, see Dumont and Verkes, 2006). Four clinical trials involving MDMA use are also underway (clinicaltrials.gov identifiers NCT00252174, NCT00090064, NCT00402298, and NCT00353938). MDMA use and abuse continue, at least in part, because of uncertainty regarding the clinical relevance of much of the animal research on MDMA-induced serotonin neurotoxicity. As we (Mechan et al., 2006) and others (Easton and Marsden, 2006) have discussed, this uncertainty stems from the fact that the majority of animal studies have used multiple high doses, have given these doses systemically rather than orally (as taken by humans), and, most often, have used rodents (rats and mice) that metabolize MDMA differently than primates (see Fig. 1) (Cho and Kumangai, 1994).

To begin bridging the gap between MDMA neurotoxicity studies in animals and human MDMA use patterns, we recently characterized the pharmacokinetic profile of oral doses of MDMA in nonhuman primates and compared results in squirrel monkeys to those in humans (Mechan et al., 2006). As might be expected, the biologic half-life (T_1/2) of MDMA in squirrel monkeys was shorter than in humans, due to the much smaller body mass of the squirrel monkey. However, other aspects of the pharmacokinetics and metabolism of MDMA in squirrel monkeys resembled those in humans, including the ratio of 3,4-methylenedioxyamphetamine (MDA) to MDMA (3–5/100). It is noteworthy that plasma concentrations of MDMA in squirrel monkeys that developed neurotoxicity were only two to three times higher than those that develop in humans given single 100 to 150-mg doses (Mechan et al., 2006), suggesting that the margin of safety of MDMA (at least with respect to brain serotonin neurotoxicity) is narrow.

Cytochrome P450 2D6 (CYP450 2D6) isoenzymes participates in the oxidative metabolism of MDMA (Tucker et al., 1994; Ramamoorthy et al., 2002). In particular, CYP450 2D6 isoenzymes are chiefly responsible for the demethylenation of MDMA to 3,4-dihydroxymethamphetamine (HHMA) (Fig. 1). Approximately 7 to 10% of Caucasians have deficient CYP450 2D6 activity (Ingelman-Sundberg et al., 2007). Whether or not such individuals are more (or less) susceptible to effects of MDMA has been a subject of discussion (see Yang et al., 2006).

An interesting and potentially very important feature of MDMA is its apparent tendency to accumulate in a nonlinear fashion (i.e., have nonlinear pharmacokinetics). The reason that this feature is potentially important is because seemingly small or trivial increases in dose could translate into large increases in plasma concentrations and, thus, a greater likelihood of toxicity (e.g., hyperthermia, serotonin neurotoxicity, etc.). We recently noted a tendency for nonlinear MDMA accumulation in squirrel monkeys (Mechan et al., 2006), and others have previously made similar observations in rats (Chu et al., 1996) and humans (de la Torre et al., 2000). However, our own studies in squirrel monkeys, like those of Chu et al. (1996) in rats, did not allow for accurate measurement of pharmacokinetic parameters, nor did they include measurements of its major metabolites HHMA and 4-hydroxy-3-methoxymethamphetamine (HMMA) (Mechan et al., 2006). The study by de la Torre et al. (2000) involved testing of different doses of MDMA in different individuals and thus left open the possibility that apparent nonlinear MDMA accumulation in humans was due to individual differences rather than nonlinear pharmacokinetics.

The purpose of the present studies was to determine if the pharmacokinetics of MDMA in nonhuman primates (squirrel monkeys) are nonlinear and, if they are, to identify plasma concentrations of MDMA at which nonlinear accumulation of MDMA occurs.

Materials and Methods

Animals. Six male adult squirrel monkeys (Saimiri sciureus) ranging in weight from 0.743 to 1.329 kg were used. Animals were housed in pairs (except during MDMA treatment when they were housed singly) in standard steel cages at an ambient temperature of 26 ± 3°C and 20 to 40% humidity, with free access to food and water. The colony room in which the animals were housed was maintained on a 14/10-h light/dark cycle (lights on: 7:00 AM). The facilities for housing and care of the animals are accredited by the American Association for the Assessment and Accreditation of Laboratory Animal Care. Animal care and experimental manipulations were approved by the Institutional Animal Care and Use Committee at the Johns Hopkins University School of Medicine and were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Drugs and Reagents. Racemic MDMA hydrochloride was obtained through the National Institute on Drug Abuse drug supply program (Rockville, MD). Racemic HHMA hydrochloride and methanolic solution of racemic MDMA hydrochloride and racemic MDA hydrochloride were purchased from Lipomed (Cambridge, MA). Methanolic solutions (1000 mg/l) of racemic HMMA and methanolic solutions (100 mg/l) of racemic MDMA-d₅ and MDA-d₅ were obtained from Cerilliant (Round Rock, TX). 4-Hydroxymethamphetamine (pholedrine), 4-methylcatechol, and EDTA were obtained from Sigma-Aldrich (St. Louis, MO). Sodium metabisulfite (SMBS) was obtained from E. Merck (Darmstadt, Germany). The authenticity of the MDMA, HHMA, HMMA, and MDA samples used in the present studies was confirmed using liquid chromatographic/mass spectrometric (LC/MS) methods.

MDMA Dosing. Each monkey received one of four doses of MDMA (on average, 6 weeks between doses) in random order. MDMA was given orally, and dose was expressed as the salt. For oral drug administration, animals were placed in a Plexiglas restraining chair (MED Associates, St. Albans, VT), and a number 8 French feeding tube (Popper & Sons, Inc., New York, NY) was inserted and used to administer the drug by gavage. Doses used for this experiment were calculated to be equivalent to 0.4, 0.8, 1.6, and 2.8 mg/kg doses in a human weighing 70 kg, using interspecies dose scaling methods (for discussion of these methods, including their limitations, see Mordenti and Chappell, 1989; Mahmood and Balian, 1996; Mahmood, 1999). Each monkey was administered a dose calculated using their individual weight at the time of administration. Animal equivalents of doses of MDMA used by humans were calculated using a standard allometric equation, shown below, where D is the dose in milligrams, W is the weight of the animal in kilograms, and 0.7 is a commonly used (empirically derived) exponent: D_human = D_animal (W_human/W_animal)^0.7 (Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

MDMA and its metabolism to HHMA, HMMA, MDA, and 3,4-dihydroxyamphetamine. In primates (humans and squirrel monkeys) ring-demethylenation predominates (as depicted by the larger arrow); in rodents (rats and mice), N-demethylation is more prominent (see Cho and Kumangai, 1994). Also shown are the CYP450 isoenzymes involved in these metabolic conversions in humans (see Kreth et al., 2000; Maurer et al., 2000).

Blood Sampling. For determination of plasma concentrations of MDMA (and its metabolites) and their pharmacokinetic profiles, blood was sampled at 0.75, 1.5, 3.0, 4.5, 6, 7, 9, 11, 23, and 25 h after MDMA administration. At each time point, approximately 0.7 ml of blood was collected as described previously (Mechan et al., 2006). Blood samples were dispensed into 5-ml Vacutainer7 hematology tubes containing 0.057 ml of 15% EDTA solution (BD Biosciences, Franklin Lakes, NJ) and stored on ice for up to 30 min until centrifuged. Samples were centrifuged at 1100g for 10 min at 4°C (Sorvall RC2-B; Kendro Laboratory Products, Newtown, CT). Plasma was withdrawn using a number 5 3/4 Pasteur pipette (Fisher Scientific, Pittsburgh, PA) and decanted into a 1.5-ml microcentrifuge tube, and SMBS (250 mM) was added at a volume of 30 μl/ml plasma to minimize oxidation of the compounds of interest. Plasma samples were stored at -20°C until assay.

Plasma Sample Preparation. Aliquots (100 μl) of squirrel monkey plasma were preserved with 20 μl of SMBS (250 mM) and 10 μl of EDTA (250 mM). After the addition of 100 μl of an aqueous solution of the racemic internal standards MDMA-d₅, MDA-d₅, and pholedrine (1.0 μg/ml each) and 300 μl of 0.5 M HCl, samples were mixed (15 s) on a rotary shaker and left at 100°C for 80 min to perform conjugate cleavage. After cooling to room temperature, 20 μl of 4-methylcatechol (1 mg/ml) was added, and samples were briefly vortexed. Perchloric acid (10 μl) then was added, and the samples were mixed again on a rotary shaker for 15 s to perform protein precipitation. The samples were centrifuged (16000g for 5 min), and the supernatant was transferred to autosampler vials. Aliquots (5 μl) were injected into the LC/MS system.

Determination of MDMA, HHMA, HMMA, and MDA Concentrations. Plasma concentrations of MDMA, HHMA, HMMA, and MDA were determined using a recently described LC/MS method modified to include acidic hydrolysis (Mueller et al., 2007). Values for HHMA and HMMA represent total free amounts (i.e., amounts measured after cleavage of sulfate and glucuronic acid conjugates).

Calculation of Pharmacokinetic Parameters. Peak plasma concentrations (C_max), times of peak plasma concentration (T_max), area under the concentration-time curve (AUC), and the elimination half-lives (T_1/2) were calculated using the software program WinNonlin (Pharsight Co., Mountain View, CA). At least three points in the declining portion of the curve were included in the calculation of T_1/2. A noncompartmental model with first-order output and elimination was used.

Statistics.C_max and AUC values for each analyte (MDMA, HHMA, HMMA, and MDA) were normalized (by dividing by the dose of MDMA administered) and then compared using repeated measures analysis of variance and subsequent Tukey's multiple comparison test. T_1/2 and T_max values at each dose were also compared using repeated measures analysis of variance and subsequent Tukey's multiple comparison test. Expected versus observed C_max and AUC values of MDMA and its major metabolites (HHMA and HMMA) at various MDMA doses were compared by means of paired t test. Expected values were calculated by multiplying the observed C_max or AUC values at the lowest dose (0.4 mg/kg) by the proportionate increase in dose. Statistical analyses were performed using Prism, Version 3.02 (GraphPad Software, Inc., San Diego, CA). Differences were considered significantly different if P < 0.05 (two-tailed).

Results

Results of pharmacokinetic studies in six squirrel monkeys, each of which received four different oral doses of MDMA in random order (with an average interval of 6 weeks between each dose), are shown in Table 2. As might be expected, absolute C_max and AUC values of MDMA increased with dose. To determine whether the observed increases were linear or nonlinear, the values (C_max and AUC) were normalized by dividing by the corresponding dose. If increases in C_max and AUC were linear, normalized C_max and AUC values would be expected to remain constant across dose (because, under linear conditions, plasma levels would rise in direct proportion to the dose). As can be seen in Table 2, normalized C_max and AUC values of MDMA did not adhere to this expectation. Instead, normalized C_max and AUC values of MDMA increased significantly with dose. T_max values of MDMA did not change as a function of dose. The T_1/2 of MDMA was significantly longer at the highest dose tested (2.8 mg/kg) (Table 2).

In sharp contrast to what was observed with MDMA, normalized C_max and AUC values of HHMA decreased significantly with dose (Fig. 2; Table 2). The T_max of HHMA did not change with dose. Similar to MDMA, the T_1/2 of HHMA was longer at the highest dose tested (2.8 mg/kg) (Table 2).

Results with HMMA paralleled those with HHMA (Fig. 2; Table 2). Results with MDA were unique in that some aspects paralleled those with HHMA and HMMA and others paralleled those with MDMA. In particular, as with normalized C_max values of HHMA and HMMA, normalized C_max values of MDA decreased significantly with dose (Table 2), whereas normalized AUC values of MDA (like those of MDMA) increased significantly with dose. The T_max and T_1/2 of MDA remained constant across dose.

Figure 3 shows the relative proportions of MDMA and its major metabolites at various times after administration of different oral doses of MDMA. At the lowest dose of MDMA tested (0.4 mg/kg), levels of MDMA were only one-half to one-sixth of those of HHMA and HMMA at comparable time points, respectively. As the dose of MDMA was increased, there was a clear shift in this pattern, with the relative proportion of MDMA increasing sharply, whereas levels of HMMA and HHMA remained relatively constant (Fig. 3).

It is noteworthy that once levels of approximately 100 ng/ml MDMA were achieved, plasma concentrations of HHMA and HMMA remained relatively constant, even though levels of MDMA rose sharply (Fig. 3).

Discussion

The present results are the first to provide unequivocal evidence of nonlinear pharmacokinetics for MDMA in nonhuman primates. Our findings in squirrel monkeys are in good agreement with those of Chu et al. (1996) in rats and speak to the species generality of nonlinear MDMA pharmacokinetics. Species generality of nonlinear MDMA pharmacokinetics is further demonstrated in findings in humans that were published (Kolbrich et al., 2008) while this article was under review. Unlike the previous study that had reported nonlinear pharmacokinetics of single doses of MDMA in humans (de la Torre et al., 2000), the study by Kolbrich et al. (2008) tested the same subjects at two different doses and thus eliminated the possibility that nonlinear MDMA accumulation might be due to individual differences rather than nonlinear pharmacokinetics. Taken together, these results indicate that the phenomenon of nonlinear MDMA pharmacokinetics has broad species generality.

In addition to documenting nonlinear pharmacokinetics of MDMA, we demonstrate that the pharmacokinetics of HHMA and HMMA, the two major phase I metabolites of MDMA in primates (Cho and Kumangai, 1994; de la Torre et al., 2004), are altered. In particular, despite disproportionate increases in plasma MDMA concentration, plasma concentrations of HHMA and HMMA remained relatively constant (Fig. 3), and once C_max and AUC values of HHMA and HMMA were dose-normalized, significant decreases in C_max and AUC values of HHMA and HMMA became evident (Table 2). This contrasts with dose-normalized pharmacokinetic parameters for MDMA, which increase with dose (Fig. 2). Considered together, these findings suggest that the nonlinear pharmacokinetics of MDMA are probably related to inhibition (or saturation) of the metabolic step that converts MDMA to HHMA: CYP450 2D6-mediated ring demethylenation (Tucker et al., 1994; Kreth et al., 2000; Maurer et al., 2000; Ramamoorthy et al., 2002) (Fig. 1), recognizing that the nomenclature for CYP450 2D6 enzymes is not necessarily the same for squirrel monkeys as humans. This suggestion is in keeping with observations that MDMA has the potential to inhibit CYP450 2D6 isoenzymes in vitro (Heydari et al., 2004; Yang et al., 2006; Van et al., 2007). Inhibition or saturation of CYP450 2D6-mediated ring demethylenation is not the only possible way that nonlinear MDMA pharmacokinetics could occur, although other metabolic mechanisms that may be involved remain to be identified.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Expected versus observed C_max (top) and AUC (bottom) values of MDMA and its major phase I metabolites, HHMA and HMMA (after conjugate cleavage), at various MDMA doses. Each dose, expressed as the human equivalent dose, was tested in the same six squirrel monkeys, as detailed in under Materials and Methods. Expected values were calculated by multiplying the observed C_max or AUC values at the lowest dose (0.4 mg/kg) by the proportionate increase in dose. Note that while observed C_max and AUC values of MDMA are greater than those expected, C_max and AUC values of HHMA and HMMA are less than those expected. Values are the mean ± S.D. *, significant difference at P < 0.05; **, significant difference at P < 0.01; ***, significant difference at P < 0.001 (paired t test).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Relative proportions of MDMA and its metabolites HHMA and HMMA (after conjugate cleavage) at various times after administration of different oral doses of MDMA to squirrel monkeys. Values are the mean ± S.D. (n = 6). Neither weighting functions nor statistical objectivity were used to decide on the best fit lines. Note that the time course and concentration of each compound is dependent on dose of MDMA administered and that, despite marked increases in plasma MDMA concentrations with dose, plasma concentrations of HHMA and HMMA remain relatively constant. In addition, note that once levels of approximately 100 ng/ml MDMA are achieved, plasma concentrations of HHMA and HMMA remain relatively constant, whereas MDMA concentrations continue to rise, suggesting inhibition or saturation of MDMA metabolism.

MDA, a relatively minor metabolite of MDMA in primates (Cho and Kumangai, 1994; de la Torre et al., 2004; Mechan et al., 2006), displayed unusual dose-related pharmacokinetic changes. In particular, while the normalized C_max of MDA decreased with dose, its normalized AUC increased with dose (Table 2). The decrease in normalized C_max of MDA parallels the decrease in normalized C_max of the other metabolites of MDMA (HHMA and HMMA) and may be due to decreased MDA formation. Whether this is due to impaired N-demethylation of MDMA to MDA remains to be determined. If it is, this would suggest that MDMA is relatively nonspecific in its ability to inactivate or saturate CYP450 enzyme systems responsible for MDMA metabolism, because separate and distinct CYP450 enzyme systems are believed to be responsible for N-demethylation and ring demethylenation (Kreth et al., 2000). In contrast to the decrease in normalized C_max, there was an increase in the normalized AUC of MDA with dose, suggesting that, over time, metabolism of MDA to 3,4-dihydroxyamphetamine (possibly by the same CYP450 enzyme system that converts MDMA to HHMA) is impaired. Assuming that this occurs, the accumulation of MDA would become nonlinear, much in the same way that the accumulation of MDMA becomes nonlinear. Additional studies are needed to investigate these possibilities.

The present results indicate that nonlinear MDMA accumulation occurs at plasma MDMA concentrations of 100 to 300 ng/ml and above. Once these plasma MDMA concentrations are achieved, plasma concentrations of HHMA and HMMA cease to increase (Fig. 3). Of note, plasma concentrations in the range of 100 to 300 ng/ml MDMA are the norm after doses of MDMA produce psychoactive effects in humans (Helmlin et al., 1996; de la Torre et al., 2000; Pacifici et al., 2001; Peters et al., 2003), suggesting that nonlinear MDMA accumulation takes place within the range of doses typically used by humans, either on the street or in the research laboratory.

In addition to demonstrating nonlinear MDMA pharmacokinetics and identifying plasma concentrations of MDMA at which nonlinear accumulation occurs, the present studies are the first reveal that the T_1/2 of MDMA in the squirrel monkey lengthens with dose. Although lengthening of the T_1/2 of MDMA was only significant at the highest dose tested (Table 2), there was a clear trend toward lengthening at lower doses as well. This was also the case for HHMA and HMMA (Table 2). A longer T_1/2 of MDMA at high dosage is noteworthy because it would effectively prolong the length of time that target sites are exposed to potentially toxic drug concentrations. Although direct confirmatory data are still needed, we suspect that nonlinear MDMA accumulation and T_1/2 prolongation are both related to impairment of MDMA demethylenation (Fig. 1). Consistent with this proposal is the observation that paroxetine, which impairs MDMA demethylenation (apparently by competing for CYP450 2D6 enzymes), increases both peak concentrations and T_1/2 of MDMA in humans (Segura et al., 2005).

The present results allow comment on the accuracy of interspecies dose scaling. Principles of interspecies dose scaling dictate that to arrive at comparable doses in animals and humans, it is important to take into account differences in body mass (see Mordenti and Chappell, 1989; Mahmood and Balian, 1996; Mahmood 1999). Therefore, as in past studies, we used interspecies dose scaling to calculate doses for squirrel monkeys that would be comparable to 0.4, 0.8, 1.6, and 2.8 mg/kg doses in humans. Comparison of plasma concentrations achieved in squirrel monkeys to those reported in humans (Helmlin et al., 1996; de la Torre et al., 2000; Peters et al., 2003; Kolbrich et al., 2008) reveals that, on average, peak plasma MDMA concentrations in squirrel monkeys were approximately 2-fold higher. This observation indicates that interspecies dose-scaling procedures, while useful for compensating for differences in body mass, are not perfect and underscores the importance of measuring actual drug plasma concentrations rather than relying solely on estimated dose equivalents.

Basic and clinical implications of the present findings remain to be determined. If the parent compound (MDMA) is chiefly responsible for pharmacological and toxic effects of MDMA (e.g., increased blood pressure, elevated body temperature, serotonin neurotoxicity), the present results suggest that seemingly small or trivial increases in dose could result in unexpected toxicities because of dose-disproportionate increases in plasma MDMA concentrations. Alternatively, if major metabolites (HHMA, HMMA) are chiefly responsible for the effects of MDMA, the present results suggest that MDMA accumulation per se is not a source of concern but that increases in dose would lead to increased exposure to potentially toxic levels of HHMA and HMMA due to T_1/2 prolongation of both of these metabolites (Table 2). However, if metabolite toxicity is a function of their C_max or AUC to a greater extent than duration of exposure (T_1/2), then the dose-dependent kinetics reported may not aggravate toxicities because the AUC of HHMA and HMMA did not increase and their C_max decreased (Table 2). With specific reference to brain serotonin neurotoxicity, it remains to be determined whether MDMA or one of its metabolites is primarily responsible. Some findings point to (but do not establish) the importance of the parent compound (dose dependence, high correlation between MDMA levels, and subsequent serotonin neurotoxicity) (see Mechan et al., 2006), whereas others suggest a possible role for metabolites (Monks et al., 2004; Erives et al., 2008). By exploring the relationship among MDMA, its metabolites, and serotonin neurotoxicity in the same animal in the context of nonlinear MDMA accumulation, it should be possible to begin discerning the relative importance of MDMA versus metabolites in the neurotoxic process.

In summary, the results of this study are the first to firmly establish nonlinear pharmacokinetics for MDMA in nonhuman primates (squirrel monkeys) and to show that the half-lives of MDMA and its major metabolites (HHMA and HMMA) increase with dose. Whether these dose-related pharmacokinetic changes (nonlinear accumulation and T_1/2 prolongation) influence the likelihood and severity of MDMA toxicities remains to be determined. It also remains to be determined whether, in the context of nonlinear pharmacokinetics, there is preferential metabolism of the S-(+) or R-(-)-enantiomer of MDMA. Of particular concern is the possibility that nonlinear MDMA pharmacokinetics, by causing disproportionate increases in plasma MDMA concentrations, narrows the already small gap that appears to exist between safe and neurotoxic doses of MDMA in primates. Additional studies are needed to explore the relationship between pharmacokinetic parameters of the parent drug (MDMA) and its metabolites (HHMA, HMMA) and brain serotonin neurotoxicity, to identify threshold neurotoxic doses (and associated plasma drug concentrations) of MDMA, and to assess the influence of nonlinear MDMA accumulation on the development of brain serotonin neurotoxicity.