Pharmacokinetics of N,N-dimethyltryptamine in Humans

Earlier clinical work and animal studies have suggested that psychedelic compounds can have therapeutic potential in major depressive disorder (MDD), with efficacy linked to dose-dependent receptor pharmacology and the subjective quality of the psychedelic experience. N,N-dimethyltryptamine (DMT) is the principal psychoactive constituent of ayahuasca, but prior human pharmacokinetic (PK) data have mainly come from oral ayahuasca studies in which harmala alkaloids inhibit monoamine oxidase (MAO) and thereby render DMT orally active. Injected DMT is extremely short acting owing to rapid oxidative deamination to indole-3-acetic acid (IAA), and earlier intravenous (IV) reports described very rapid peak levels and brief subjective effects but provided limited PK parameters. Good and colleagues set out to characterise, in detail, the in vitro metabolism and the clinical pharmacokinetics of DMT in healthy, psychedelic-naïve adults. The paper reports preclinical assays of physicochemical properties, MAO and cytochrome P450 (CYP) contributions to metabolism, followed by the Phase I component of a Phase I/IIa randomised, double-blind, placebo-controlled, parallel-group dose‑escalation trial (NCT04673383) in which escalating single doses of DMT fumarate (administered as a two‑phase 10‑min IV infusion) were evaluated for safety, tolerability and PK to inform later clinical development and infusion regimen design.

The preclinical work comprised a suite of validated in vitro assays using LC-MS/MS to inform metabolism and physiochemical properties. Human hepatocytes from ten donors were incubated with DMT fumarate at nominal concentrations (1–5 µM) with and without MAO‑A (clorgyline) and MAO‑B (deprenyl/selegiline) inhibitors; timepoints up to 60 min were sampled to estimate intrinsic clearance and half‑life. Human liver mitochondrial fractions (5‑donor pool) were assayed similarly to probe MAO activity specifically. CYP phenotyping used recombinant human CYP bactosomes for eight major isozymes with NADPH‑dependent incubations and time‑course sampling to derive isozyme-specific intrinsic clearance and half‑life estimates. Separate in vitro assays measured blood:plasma partitioning, plasma protein binding by rapid equilibrium dialysis, and lipophilicity (logD7.4) in octanol/buffer systems. Analytical methods and calibration ranges for DMT and IAA were reported for LC‑MS/MS instruments. The clinical study was a single‑centre, randomised, double‑blind, placebo‑controlled, parallel single ascending dose (SAD) design conducted under EMA and GCP standards. Eligible participants were aged ≥25 years, psychedelic‑naïve, BMI 18.0–30.9 kg/m2, medically and psychiatrically screened to exclude current or past DSM‑5 mental health disorders or family history of psychotic/bipolar disorder; the extraction does not report ethnic composition. The investigational product was a GMP‑manufactured fumarate salt of DMT (SPL026) formulated at 2.5 mg/ml free base; placebo matched excipients. Four sequential dose cohorts were planned, each randomised so six subjects received active and two received placebo, with sentinel dosing (1 active, 1 placebo) before dosing the remainder. Escalation depended on review of safety, tolerability and PK from prior cohorts. Study drug was delivered as a continuous 10‑min IV infusion split into two 5‑min phases via a single cannula with two syringe pumps: phase 1 (gentle rise to the fringe of psychedelic experience) and phase 2 (raise/maintain peak exposure). Dosing regimens were modelled from prior bolus IV data using a one‑compartment model fitted by iteratively reweighted non‑linear least squares; compartmental simulations were used to select four final infusion regimens intended to produce ascending peak exposures, including a target intended to elicit a 'breakthrough' experience. Pharmacokinetic blood sampling was intensive pre‑dose and up to 240 min post‑start of infusion with validated bioanalysis for DMT (calibration range down to 0.0619 ng/ml) and IAA. Noncompartmental PK parameters were calculated for evaluable subjects; dose proportionality of Cmax and AUClast was assessed with a power model fitted by REML mixed effects, and post‑hoc hierarchical regression assessed relationships of BMI and weight with dose‑normalised exposure. Missing concentrations below LLOQ were treated as zero if before Tmax or as missing otherwise; no imputation was performed.

In vitro experiments indicated different clearance behaviours across systems. In plated human hepatocytes (0.62 µM DMT), intrinsic clearance was 19.4 ± 0.8 µl/min/million cells with a half‑life of 98.9 ± 3.9 min. MAO inhibition in hepatocytes produced minimal change in intrinsic clearance at the tested concentrations. In contrast, human liver mitochondrial fractions showed high intrinsic clearance (175 µl/min/mg protein at 0.62 µM, half‑life 7.9 min) and a concentration‑dependent effect: at 3.1 µM, MAO‑A inhibition reduced intrinsic clearance by >90% and prolonged half‑life markedly, while MAO‑B inhibition had little effect. CYP phenotyping identified substantial intrinsic clearance via CYP2D6 (801 µl/min/nmol CYP, half‑life ~9 min) and a minor contribution from CYP2C19 (37 µl/min/nmol CYP, half‑life ~189 min); no appreciable turnover was detected for CYP1A2, 2B6, 2C8, 2C9, 3A4 or 3A5. Physicochemical assays returned mean logD7.4 = 0.15 (low lipophilicity), plasma unbound fraction ≈ 67.7% (i.e. most DMT unbound), and a blood:plasma ratio of 1.53 indicating notable partitioning into blood cells. Thirty‑two subjects were enrolled in the Phase I study (mean age 36.4 years, range 25–76; mean BMI 25.0 kg/m2; eight female, 25%). All subjects completed dosing; 13 subjects experienced drug‑related treatment‑emergent adverse events (drTEAEs), all categorised as mild or moderate and transient; no serious adverse events were recorded. Pharmacokinetic analyses excluded two subjects due to dosing errors and encountered missing samples in some cohorts (notably cohort 3), limiting parameter derivation for a small number of individuals. Across cohorts DMT was rapidly absorbed and eliminated. Median Tmax was approximately 10 min (near end of the 10‑min infusion). Cohort results reported were: cohort 1 (9 mg total, n=5 evaluable) mean Cmax 20.8 ng/ml (range 5.0–34.9), mean AUClast 349 ng·min/ml (range 71–705), mean t1/2 12.1 min (range 5.8–18.3); cohort 2 (12 mg, n=6) mean Cmax 30.6 ng/ml (range 12.7–62.3), AUClast 451 ng·min/ml (range 245–755), mean t1/2 9.5 min (range 6.0–17.0); cohort 3 (17 mg) sampling issues meant only Cmax (mean 72.1 ng/ml, range 16.2–126.0) and AUClast (mean 842 ng·min/ml) could be reported; cohort 4 (21.5 mg, n=6) mean Cmax 62.7 ng/ml (range approximately 29–107 reported) and AUClast 835 ng·min/ml (range 477–1052), mean t1/2 12.1 min (range 6.3–20.3). Individual concentrations at 5 min after a 6 mg administration (C5min) ranged from 3.32 to 43.0 ng/ml across cohorts and there were no significant differences between cohorts by one‑way ANOVA (F(3,20)=1.628, p=0.214). Exploratory IAA measures in eight subjects (cohorts 2 and 4) showed high baseline IAA (mean pre‑dose 316 ng/ml) and mean maximal plasma IAA concentrations of 1532.5 ng/ml (cohort 2) and 2182.5 ng/ml (cohort 4), recorded between 30–60 min after infusion start. Peak IAA concentrations exceeded DMT Cmax by roughly 30–100‑fold in the sampled subjects and IAA had not returned to baseline by 240 min in these cases. Dose proportionality assessment using a power model yielded slope (β) estimates for Cmax and AUClast of 1.58 (90% CI 0.84–2.33) and 1.35 (90% CI 0.65–2.04), respectively; sensitivity analysis excluding cohort 3 produced β estimates of 1.47 (0.74–2.20) for Cmax and 1.29 (0.59–2.00) for AUClast. As the 90% CIs encompassed 1.0, the authors report these findings as consistent with dose proportionality but caution that wide CIs and inter‑individual variability temper that conclusion. Post‑hoc hierarchical regression found no significant relationships between BMI or weight and dose‑normalised Cmax (BMI β=0.179, p=0.508; weight β=0.107, p=0.671) or C5min (BMI β=‑0.490, p=0.078; weight β=‑0.215, p=0.426) when controlling for age, sex and dose cohort.

Good and colleagues interpret the combined in vitro and clinical data as reinforcing a picture in which DMT is rapidly distributed and cleared in humans, with MAO‑A playing the primary role in systemic metabolism and CYP2D6 (and to a lesser extent CYP2C19) contributing under conditions where MAO‑A activity is limited. The contrasted behaviour between hepatocyte and mitochondrial fractions led the investigators to conclude that mitochondrial MAO‑A accounts for the high clearance seen in mitochondrial preparations and that MAO‑independent routes (CYPs) can nevertheless contribute to DMT breakdown in some physiological contexts. Low lipophilicity (logD7.4 ≈ 0.15) and a high unbound plasma fraction (>65%) corroborate the rapid availability of DMT for distribution and metabolism, while a blood:plasma ratio >1 suggests erythrocyte partitioning that should be considered in PK interpretation. Clinically, the infusion regimens produced rapid attainment of peak plasma levels near the end of the 10‑min infusion and a short terminal half‑life in the order of 9–12 min across doses. All regimens were tolerated without serious safety signals. The authors note substantial inter‑individual variability in Cmax and AUClast, consistent with prior reports, and propose that rapid clearance and sampling timing may amplify apparent differences at Tmax. Post‑hoc analyses did not identify BMI or weight as predictors of peak exposure, supporting the practicality of fixed infusion dosing rather than weight‑based dosing in this context. On metabolism, exploratory IAA data showed much higher plasma IAA concentrations than DMT and rapid increases in IAA after dosing, concordant with rapid oxidative deamination; however, the authors caution that high baseline and early fluctuations in IAA complicate interpretation. They also highlight that estimated blood clearance for the highest cohort substantially exceeded typical liver blood flow, suggesting extrahepatic clearance sites for IV DMT and a dominant role for MAO‑A expressed outside the liver. Important limitations acknowledged include missing PK samples in some cohorts, small cohort sizes typical of Phase I SAD studies, high inter‑individual variability, and limited IAA characterisation (sparse sampling, baseline fluctuations and the need for a more sensitive assay). The paper calls for larger studies and population PK/PK‑PD modelling, plus longer and more sensitive monitoring of IAA, to better define sources of variability and the full DMT‑IAA relationship. The authors suggest these data can inform design of IV infusion regimens for therapeutic development while clarifying metabolic pathways relevant to oral bioavailability and potential drug interactions.

This study provides a detailed characterisation of DMT pharmacokinetics following a slow 10‑min IV infusion in healthy adults, showing rapid attainment of peak plasma concentrations and rapid clearance largely mediated by MAO‑A. In vitro data identify CYP2D6 and CYP2C19 as potential contributors to DMT metabolism in MAO‑A‑sparse environments. The findings are presented as a basis for improved PK and metabolic models of DMT and to guide the design of IV infusion regimens for clinical development in mental health indications.