This paper outlines the formulation development, in vitro, and in vivo testing of transdermal drug-in-adhesive DMT patches using various adhesives and permeation enhancers. In vivo behavioural and pharmacokinetic studies performed with lead patch formulation (F5) in male and female Swiss Webster mice showed that transdermal administration provided consistent, extended drug release at a non-hallucinogenic dose, with a 77% bioavailability compared to IV at two dosages.
There is considerable evidence from the literature that psychedelics, such as N,N-dimethyltryptamine (DMT), are safe and effective treatments for depression. However, clinical administration to induce psychedelic effects and expensive psychotherapy-assisted treatments likely limit accessibility to the average patient. There is emerging evidence that DMT promotes positive behavioral changes in vivo at sub-hallucinogenic dosages, and depending on the target indication, subjecting patients to high, bolus dosages may not be necessary. Due to rapid metabolic degradation, achieving target levels of DMT in subjects is difficult, requiring IV administration, which poses risks to patients during the intense hallucinogenic and subjective drug effects. The chemical and physical properties of DMT make it an excellent candidate for non-invasive, transdermal delivery platforms. This paper outlines the formulation development, in vitro, and in vivo testing of transdermal drug-in-adhesive DMT patches using various adhesives and permeation enhancers. In vivo behavioral and pharmacokinetic studies were performed with lead patch formulation (F5) in male and female Swiss Webster mice, and resulting DMT levels in plasma and brain samples were quantified using LC/MS/MS. Notable differences were seen in female versus male mice during IV administration; however, transdermal administration provided consistent, extended drug release at a non-hallucinogenic dose. The IV half-life of DMT was extended by 20-fold with administration of the transdermal delivery system at sub-hallucinogenic plasma concentrations not exceeding 60 ng/mL. Results of a translational head twitch assay (a surrogate for hallucinogenic effects in non-human organisms) were consistent with absence of hallucinations at low plasma levels achieved with our TDDS. Despite the reported low bioavailability of DMT, the non-invasive transdermal DMT patch F5 afforded an impressive 77 % bioavailability compared to IV at two dosages. This unique transdermal delivery option has the potential to provide an out-patient treatment option for ailments not requiring higher, bolus doses and is especially intriguing for therapeutic indications requiring non-hallucinogenic alternatives.
Papers cited by this study that are also in Blossom
Alper, K., Cange, J., Sah, R. et al. · Frontiers in Pharmacology (2023)
Barker, S., McIlhenny, E. H., Strassman, R. J. · Drug Testing and Analysis (2012)
Cameron, L. P., Benson, C. J., Defelice, B. C. et al. · ACS Chemical Neuroscience (2019)
Cameron, L. P. · ACS Chemical Neuroscience (2018)
Psychedelic tryptamines such as N,N-dimethyltryptamine (DMT) are being reconsidered as therapeutic agents for psychiatric and neurological disorders, but their clinical use is constrained by delivery challenges. DMT produces rapid and intense psychedelic effects via 5-HT2A receptor agonism and is rapidly degraded by monoamine oxidases, making oral dosing ineffective; consequently most clinical and research administrations rely on intravenous, intramuscular, inhalation, or insufflation routes that produce high, transient peak concentrations and require supervised settings. Preclinical work suggests that sub-hallucinogenic exposures of DMT may still drive neuroplastic changes relevant to therapeutic outcomes, and the head-twitch response (HTR) in rodents has been used as a proxy to define hallucinogenic versus sub-hallucinogenic plasma levels. This study aimed to design, formulate, and evaluate single-layer drug-in-adhesive (DIA) transdermal patches delivering DMT freebase, with the objective of producing sustained peripheral exposure while keeping peak plasma concentrations below a proposed 60 ng/mL threshold for psychedelic subjective effects. Witowski and colleagues report in vitro screening (Franz cell diffusion) of six formulations varying adhesives, solvents and permeation enhancers, followed by GLP manufacture of a lead formulation (F5) and in vivo pharmacokinetic and HTR testing in male and female Swiss-Webster mice to assess pharmacokinetics, brain exposure, and hallucinogenic-like activity.
Formulation development screened six DIA patch formulations (F1–F6) using DMT freebase dissolved in ethanol or ethyl acetate, combined with three common transdermal adhesives (acrylate, silicone, polyisobutylene) and, in one formulation, the permeation enhancer isopropyl myristate. Non-GLP material was used for initial screening; GLP-grade DMT (98.5% purity) was used to manufacture patches for in vivo testing. Coating and drying procedures are described (150–250 μm coat thickness, oven drying at 70 °C for 2 h), and GLP patches were produced at a university facility. Some procedural parameters in the extraction text (e.g. specific stirrer rpm) are not clearly reported in the extracted material. In vitro release was assessed using Franz diffusion cells fitted with Strat-M membranes and a 10% ethanol/water receiver to aid solubility at 37 °C with stirring. Single-replicate Franz cell experiments sampled multiple timepoints to calculate flux and cumulative release; the methods state linear regression of concentration versus time was used to compute flux. Patch content and recovery were evaluated via solvent extraction, sonication, evaporation and HPLC quantification; recovery from the DIA matrix was reported as 92% and the HPLC method had a lower limit of quantification (LLOQ) of 20 ng/mL. In vivo work used male and female Swiss-Webster mice (6–8 weeks, 30–40 g). Hair was removed from the dorsal neck area and patches were applied under brief isoflurane anaesthesia and secured with jackets. Animals received either intravenous (IV) DMT or an active patch, with the other manipulation being vehicle to control handling. Doses tested were nominally 1 mg/kg and 5 mg/kg (extraction truncates some numeric labels but these doses are consistently described in the results). Blood sampling for IV cohorts occurred as early as 5 min and up to 240 min or more; sampling for transdermal cohorts began at 60 min and included later timepoints up to 1440 min. Brain tissue was collected at specified early timepoints and flash-frozen for LC-MS/MS analysis. The extracted text contains some truncated timing details for brain collections. Plasma and brain DMT concentrations were quantified by LC/MS/MS using validated calibration ranges and triplicate injections; assay performance criteria (calibration R ≥ 0.995, QC within ±15%, triplicate %RSD ≤ 15%) are reported. Pharmacokinetic parameters were computed from log-linear elimination slopes; AUCs were calculated using appropriate t = 0 values for IV and TD plots, and relative bioavailability (F) was computed as the dose-normalised AUC ratio. HTR was measured in magnet-implanted mice using automated magnetometer cages; mice were habituated, given treatments, and head-twitch counts were recorded and analysed relative to saline-treated controls. Statistical testing used GraphPad Prism with p < 0.05 as the significance threshold.
Formulation screening showed acrylate adhesive (Duro-Tak 4098) provided the highest DMT loading and flux among the adhesives tested; ethanol was the preferred solvent for DMT solubilisation. Six formulations (F1–F6) were characterised for coat weight, thickness and DMT content; F5 (acrylate adhesive, ethanol, 10.5% w/w DMT) was selected as the lead non-enhanced formulation and F6 (with isopropyl myristate, 16.1% w/w DMT) gave higher loading. Franz cell diffusion demonstrated largely linear, zero-order release over 72 h (R2 > 0.98) for all formulations except F4 (polyisobutylene), and cumulative delivery over 72 h was 520 μg/cm2 for F5 and 775 μg/cm2 for F6. F5 flux was stable at 1 week, 1 month and 2 months post-manufacture (flux range reported in the extracted text: 6.54 ± 2.73 μg/(cm2·hr)). In vivo pharmacokinetics revealed stark differences between IV and transdermal administration. For 5 mg/kg IV dosing, peak plasma concentrations (Cmax) occurred at the 5-min timepoint and averaged 597 ng/mL in males and 1437 ng/mL in females. At 1 mg/kg IV, Cmax averaged 372 ng/mL (males) and 528 ng/mL (females). The F5 transdermal patch produced much lower peak plasma levels and delayed Tmax: at 5 mg/kg, male Cmax = 37.0 ng/mL (1 h) and female Cmax = 54.8 ng/mL (4 h); at 1 mg/kg, male Cmax = 8.0 ng/mL (1 h) and female Cmax = 15.7 ng/mL (1 h). Apparent elimination half-lives were markedly prolonged with the patch: IV t1/2 ≈ 10.8 ± 2.5 min (males) and 10.0 ± 0.1 min (females), versus TD t1/2 ≈ 273 ± 10.5 min (males) and 210 ± 14.6 min (females). Relative plasma bioavailability following transdermal delivery was reported as approximately 77% compared to IV; the precise SD for the value in the results text was not clearly extracted but a value of 77 ± 10% is reported elsewhere in the manuscript. Brain concentrations mirrored peripheral exposure patterns. After 5 mg/kg IV, brain Cmax at 15 min averaged 4503 ng/g (males) and 8080 ng/g (females). At 1 mg/kg IV, brain Cmax were substantially lower (353 ng/g males, 250 ng/g females). For transdermal dosing, brain DMT was detectable above the LLOQ only at the 1-hour timepoint for the 5 mg/kg dose and those concentrations were significantly lower than after 5 mg/kg IV (Two-Way ANOVA: F Route [1,20] = 18.03, p = 0.0004). Brain concentrations were below the LLOQ at all timepoints after 1 mg/kg transdermal dosing. In the HTR assay, IV DMT produced a transient but robust increase in head-twitch counts, with an 18-fold elevation relative to IV saline during the 0–10 min bin (Sidak's test p = 0.0214) and time-dependent variation overall (Two-Way RM ANOVA: Route×Time F(11,99) = 3.69, p = 0.0002). The transdermal cohort did not show significant HTR changes compared with TD vehicle up to 1 h post-administration (Two-Way RM ANOVA: Route×Time F(11,242) = 0.66, p = 0.77). The investigators also observed sex-dependent differences in systemic exposure, with females generally achieving higher Cmax and AUC values than males across routes and doses. Finally, the authors note a discrepancy between in vitro 72-h sustained release and in vivo detectability of DMT, which in mice extended only to about 8 h, likely reflecting species-specific cutaneous and systemic metabolism.
Witowski and colleagues interpret these findings as the first demonstration that DMT can be delivered effectively via a non-invasive single-layer DIA transdermal patch, achieving prolonged systemic exposure, substantial relative bioavailability, and greatly reduced peak concentrations compared with IV dosing. The lead formulation (F5) produced a roughly 20-fold apparent extension in plasma half-life and maintained peripheral concentrations that, at the doses tested, did not exceed the authors' target threshold of 60 ng/mL thought to associate with subjective psychedelic effects. As a consequence, the patch did not induce significant HTR in mice, whereas IV DMT produced a rapid, transient increase in HTR consistent with hallucinogenic-like activity in this preclinical model. The discussion situates these results against earlier work suggesting that sub-hallucinogenic DMT exposures can still engage neuroplastic pathways; the researchers emphasise that transdermal delivery could enable outpatient, lower-risk regimens that avoid intense subjective effects and the need for continuous clinical supervision. They also highlight practical formulation insights: acrylate adhesives and ethanol solvent supported higher DMT loading and flux, and the addition of an enhancer (isopropyl myristate) further increased loading and permeation. Limitations acknowledged by the authors include species differences in metabolism and skin permeability, the faster drug clearance typical of rodents relative to larger mammals and humans, and discrepancies between in vitro Franz cell persistence and in vivo duration of exposure. Sex differences in PK were noted and the authors recommend that future clinical work monitor male–female differences. They also recognise that the current study does not establish whether the achieved exposure profiles will produce therapeutic efficacy or whether subjective psychedelic experiences are necessary for clinical benefit. The investigators call for further studies in larger animals and humans to determine translatability, to explore higher or enhancer-containing formulations that might elicit HTR via transdermal routes, and to evaluate neural and behavioural endpoints relevant to therapeutic indications.
The authors conclude that a low-dose transdermal DMT product is feasible and that their optimized DIA patch (F5) provides sustained systemic exposure with high relative bioavailability, a greatly extended apparent half-life, and peak plasma concentrations below the authors' proposed hallucinogenic threshold, thereby avoiding the HTR in mice. They state this approach could enable non-hallucinogenic, take-home dosing paradigms for psychiatric or neurological indications, while recommending additional PK/PD studies in larger species and humans to establish therapeutic potential and translatability.
Psychedelic compounds, largely overlooked as therapeutic approaches by industrialized nations since the 1960s, have emerged again as promising drug classes for psychiatricand addiction disorders. The tryptamine or "classic psychedelic" compound class is being evaluated in several late-stage clinical trials (Dshowing sustained efficacy in patients after only one or two doses. The doses selected in these trials are typically designed for maximum psychedelic subjective effects which require the use of therapists to observe patients within a clinical setting. This is a similar model to the now approved Spravato (esketamine) for treatment-resistant depression;however, the methods of drug delivery to patients have not been optimized for non-orally active psychedelic drugs or those with low bioavailability. One of these drugs, N,N-dimethyltryptamine (DMT), can produce rapid, intense, and short-lived psychedelic effects via serotonin 2A receptor (5-HT 2A R) agonism. Despite its potent nature, DMT is rapidly degraded by monoamine oxidase (MAO) enzymes limiting any potential of oral dosing. The use of ayahuasca, a plant-based aqueous extract consisting of naturally derived DMT and MAO inhibitors, has been used for ritualistic, spiritual, and therapeuticpurposes. Pure DMT formulations must be administered via intravenous, intramuscular, inhalation, and insufflation methods to reduce gastrointestinal first-pass effects. Definitive biological and ecological roles of DMT are unknown, but research shows it is ubiquitously produced throughout natureand has been found in the brains of ratsand in human biological fluids. Given its structural similarities to serotonin and the amino acid tryptamine, its biosynthetic precursor, DMT is also hypothesized to be the endogenous ligand to 5-HT 2A. Despite these intriguing developments, DMT remains a schedule 1 drug in the United States and a highly controlled substance in most countries worldwide. The therapeutic promise of DMT is currently being assessed in several FDA clinical trials including a phase I study (NCT05559931) with a focus on stroke and traumatic brain injury (TBI) and a phase I/IIa study for major depressive disorder (NCT04673383). To date, most clinical DMT studies have incorporated individualized therapy alongside higher doses of 0.2-0.4 mg/kg, which are sufficient to induce psychedelic and hallucinogenic effects. Studies incorporating this approach include the use of ayahuasca for addiction. It is presumed that drug-induced neuroplasticity, increasing dendritic spines and formation of new neural pathways contribute to the therapeutic effects of these drugs and there has also been increasing use of preclinical models to study the impact of DMT on these potential mechanisms, both together with, and apart from, hallucinogenic effects. Given the challenges in assessing altered states of consciousness in non-human species, the head-twitch response assay (HTR) has been used as a predictive pre-clinical method to define hallucinogenic and sub-hallucinogenic doses, as the plasma levels of drug required to induce head twitch in rodents has been demonstrated to be strongly correlated to the amount of drug required to induce subjective effects of psychedelic compounds in humans. In this assay, mice will repeatedly shake or twitch their head under the influence of strong 5-HT 2A agonists, like DMT, which can be accurately measured in real-time using an automated magnetometer approach to count head twitches. Sub-hallucinogenic doses (i.e., 10 % of hallucinogenic dose) of DMT produced similar synaptic signatures measured in rat brains to that of a larger, psychedelic dose. Furthermore, preclinical data suggests sub-hallucinogenic doses can be effective for ischemic stroke, mood disorders such as anxiety, and Alzheimer's disease. A functional study showed that using an antagonist to block DMT's effects at the sigma-1 receptor, a transmembrane protein located on the endoplasmic reticulum, abolished neuroplasticity, but 5-HT 2A R antagonism still achieved DMT-induced neuroplasticity. These results indicate there are multiple cellular pathways affected by DMT and that higher doses of DMT and 5-HT 2A R mediated hallucinations are not solely responsible for therapeutic effects. Therefore, non-invasive low-dose DMT dosage forms could offer a new paradigm to achieve therapeutic effects with fewer side effects such as hallucinations. Transdermal drug delivery systems (TDDS) of psychedelics have only been postulated with no peer-reviewed literature supporting viable technology in this field despite the positive application of TDDS in psychiatry with Emsam (selegiline) and Secuado (asenapine)as well as Adlarity (donepezil) for Alzheimer's disease. The physical properties of DMT (small molecule, low molecular weight, low boiling point, and suitable lipophilicity) make it an attractive candidate for transdermal delivery. Given that transdermal delivery is a direct route to the bloodstream without exposure to first-pass metabolism, it could offer a non-invasive alternative to intravenous administration. The major challenge to the druggability of DMT is its short half-life of about six minutes. Therefore, a drug delivery system must be designed to overcome the kinetics of metabolism to drive steady-state drug plasma concentrations with fewer fluctuations and blunting of maximum concentration (C max ) to avoid hallucinations. This study outlines the design and development of single layer drugin-adhesive (DIA) patches containing DMT freebase. Various formulations were trialed modifying adhesives, permeation enhancers, and concentrations of DMT to maximize drug loading and release characteristics. Franz cell diffusion was used as an initial tool to assess DMT release and flux. An optimized patch formulation was reproduced under good laboratory practices (GLP) and tested in both male and female mice for pharmacokinetic (PK) measurement of blood and brain concentrations of DMT, while the head-twitch response assay (HTR) was utilized to measure hallucinogenic potential. The goal of the study was to maintain plasma concentrations below that of 60 ng/mL, a concentration that has been clinically defined to cause subjective hallucinogenic and psychedelic effects in humans.
DMT freebase (non-GLP) was synthesized for formulation screening and in vitro Franz cell release testing using the Speeter-Anthony method. DMT freebase (GLP) utilized in patches for dose-response in mice was purchased from Organix Inc., Boston, MA, USA with a purity of 98.5 %., and formic acid (FA) were purchased from Fisher Scientific, Hampton, NH, USA. Duro-Tak 4098 and were provided by Henkel Corporation, Dusseldorf, Germany. Bio PSA 7-4302 and 4202 were provided by DuPont Inc., Missuagua, ON, Canada. Isopropyl myristate was provided by Croda International Plc, Edison, NJ, USA. Scotchpak 9709 and 9733 were provided by 3M Corporation, Saint Paul, MN, USA. Mylar packaging was purchased from Impak Corporation, Sebastian, FL, USA.
A total of six formulations were trialed to optimize drug loading in the DIA matrix. To make the DIA formulations, non-GLP DMT was dissolved in either ethanol or ethyl acetate before the addition of adhesive and/or permeation enhancers. Once combined, an overhead stirrer (IKA RW20, Wilmington, NC, USA) was used to blend the formulation at rpm and allowed to degas. The formulations were placed on the siliconized side of a Scotchpak 9709 liner and coated using a thin film applicator (150 mm, MTI Corporation, Richmond, CA, USA) at a thickness of 150-250 μm. Formulations were dried in an oven (Quincy Lab 10-180, Burr Ridge, IL, USA) for 2 h at 70 • C before lamination onto a Scotchpak 9733 backing. Patches were cut to size based on desired dose, heat sealed within Mylar packaging, and stored at room temperature until use. The patches used for the in vivo studies were manufactured under Phase 1 GLP conditions at the Zeeh Pharmaceutical Experiment Station at the University of Wisconsin, Madison. The GLP DMT freebase from Organix was placed under stability and used to create two patch dosages at 0.4 mg/cm 2 using the same procedure for non-GLP DMT patches.
The release liner was removed, and patches folded in half onto itself with adhesive sides touching. The patch was cut into four pieces, placed into a vial, and 10 mL of ethyl acetate added to cover the material for thorough extraction. Vials were submerged in a sonicator (Branson 2510, Emerson Electric, Round Rock, TX, USA) to extract the DIA from the patch backing into the solvent for 15 min. Once complete, the DIAless backing pieces were removed, and the resulting ethyl acetate extract was dried under an inert stream of nitrogen. The dried residue was reconstituted with 5 mL of methanol and sonicated for another 15 min. The resulting slurry was transferred into an Eppendorf tube and placed into a centrifuge (Mini Centrifuge, Fisher Scientific, Hampton, NH, USA) for 10 min at 6,000 rpm. The supernatant was pipetted into a 2 mL amber glass autosampler vial along with a tryptamine (50 μg/mL) internal standard.
To determine the rate and quantity of non-GLP DMT released from patches, Franz cell apparatuses (10 mL unjacketed, PermeGear, Hellertown, PA, USA) were equipped with Strat-M membranes (25 mm, Millipore Sigma, Darmstadt, Germany). In our experience, these membranes provide a more consistent platform to analyze different formulations without changes in thickness and follicle density from excised skin. Inconsistencies in these factors can significantly alter drug flux parameters necessitating greater replicates along with extra processing techniques (i.e. dermatome and stratum corneum removal). Strat-M membranes have excellent correlations to human skin across a number of different drugs further validating their usage. Due to low aqueous solubility of freebase DMT, 10 % ethanol in water was used in the receiving well which was incubated in a 37 • C water bath to simulate human biological temperatures with continuous stirring at 500 rpm. Patches were stored at room temperature for at least one week post-production to equilibrate before testing. Single replicates (n = 1) from each formulation were analyzed at multiple sampling times (0.5, 1,whereby the entire contents of the receiving well were emptied into a vial and fresh 10 % ethanol solution was added to the receiving well. Each sample was homogenized with a vortexer (Standard Vortex Mixer, Fisher Scientific, Fisher Scientific, Hampton, NH, USA) before pipetted into a 2 mL amber glass autosampler vial along with a tryptamine (50 μg/mL) internal standard.
An HPLC method was adoptedon an Agilent 1200 system to quantify DMT from patch formulations and Franz cell time points. A binary solvent system with water and ACN, each spiked with 0.1 % FA, was employed at a flow rate of 1.3 mL/min initially at 5 % ACN, ramping to 40 % ACN at 3.5 min, then up to 100 % ACN at 6 min and held for 1 additional minute. The stationary phase column, C18 InfinityLab Poroshell 120 (4.6 × 30 mm x 2.7 mm), was maintained at 35 • C. The sample injection volume was 5 µL and quantification was done at 280 nm. Samples were analyzed using Agilent ChemStation software whereby DMT peak areas were first normalized to a peak ratio of an external standard tryptamine vs. DMT at 50 μg/mL. The normalized DMT peak area was then divided by the peak area of tryptamine internal standard and multiplied by its concentration of 50 μg/mL. Studies to determine recovery of DMT from the patch DIA were assessed by spiking 50 μg/mL DMT into the patch extraction protocol of Section 2.3. The DIA patch matrix did retain a portion of DMT with recovery determined to be 92 %. Reproducibility of DMT peak areas compared to the standard was 99.8 %. The lower limit of quantification for DMT was 20 ng/mL. Flux (μg/hr) was calculated from Franz cell (4.91 cm 2 active area) studies as the linear regression of the DMT concentration vs. time plot.
Animal experiments were approved by the University of Wisconsin, Madison Animal Care and Use Committee (IACUC). All procedures were followed in compliance with the Research Animal Resources and Compliance (RARC) guidelines. All mice (Swiss-Webster; male and female; 6-8 weeks; 30-40 g; Charles River Laboratories, Wilmington, MA) were handled for 7 days prior to experimentation to acclimate to both the experimenter and vivarium. Mice were housed in groups of two or three under a 12h reverse light/dark cycle; temperature was maintained between 22 and 24 • C. All food (LabDiet) and water (Inno-Vive) were available to the mice ad libitum unless stated otherwise. For patch application, animals were initially anesthetized under a nose cone with 5 % isoflurane in oxygen flowing at 2 L/min, with maintenance anesthesia levels at 1.5-2 % isoflurane. Electric clippers (Fisher Scientific, Pittsburg, PA) were used to shave the dorsal neck area between the ears and shoulder blades. Veet hair removal cream (Andwin Scientific Industrial, Simi Valley, CA) was applied on the shaved portion and removed after 5 min using 70 % ethanol. Patches were cut to size prior to application. to deliver either 1 or 5 mg/kg of DMT. Patches containing either GLP DMT or vehicle (Psilera Inc., Tampa, FL) were placed on the hairless area and secured to the animals with jackets (Lomir Biomedical Inc, Malone, NY). All mice were placed in a recovery chamber until they were fully awake before administering an IV tail vein injection of either DMT or vehicle and remained in their home cage for 60 min prior to experimentation. While all animals received both manipulations, an injection and patch, to control for differences in handling arising from these manipulations, active drug was only delivered through one of these routes for each group. Post-administration blood samples were collected at 5, 10, 15, 30, 60, 240, 480, or 1440 min using microcentrifuge tubes. Following collection, the samples were then centrifuged at 10,000 rpm (11,292 g) for 10 min at 4 • C. The plasma fraction was separated and stored in the dark at -80 • C until LC-MS/MS analysis. Brain tissues were collected post-administration at 15 or min and were flash frozen in liquid N 2 and stored in the dark at -80 • C until LC-MS/MS analysis. Tissue extracts for DMT quantitation were prepared by placing weighed brains in a 2-ml screw-cap tube with 1.4 mm ceramic beads , adding 2 volumes of 150 mM ammonium bicarbonate per gram of tissue and homogenizing in an Omni Bead Mill Elite homogenizer.
For LC/MS/MS analysis plasma samples were prepared by precipitation and filtration. Briefly, 50 µl of plasma was precipitated with 150 µl ACN/1 % formic acid containing internal standard (6-methoxy DMT, Sigma Aldrich, St. Louis MO) in a 96-well format Sirocco plate (Waters Corp. Milford, MA) according to the manufacturer's protocol. Samples were pushed through the plate to a 2-ml receiver plate containing 800 µl of water. A similar protocol was used for preparation of brain extracts except that instead of a simple filtration plate, brain samples were processed through a Waters Ostro plate to reduce phospholipid content of the extracts. For both brain and plasma, calibration curves (1 ng/mL-2000 ng/mL) and QC samples for DMT were prepared in blank matrix and processed with unknown samples. After processing samples were briefly vortexed then analyzed using a liquid chromatography tandem mass spectrometry (LC/MS/MS). Sample (2 µl) were injected onto a Phenomenex Kinetex Phenyl-Hexyl 2.1 × 100 mm column packed with 1.7 µm particles using a Waters Acquity UPLC system (Waters, Milford MA). The column was held at 35 • C and the flow-rate was 0.375 ml/min. Solvent A was water/0.1 % FA and solvent B was ACN/0.1 % FA. Analytes were eluted from the column with an increasing gradient of ACN from 5-40 % B in 1.75 min then to % B in 0.35 min with a 0.4 minute hold at 95 % then a return to 5 % B in 0.25 min. Eluate from the column was analyzed in positive ion mode using a QTrap 5500 hybrid triple quadrupole mass spectrometer (SCIEX, Framingham MA) operating in multiple-reaction-mode (MRM) under conditions optimized for detection of the analyte and internal standard. For DMT the transitions were: parent ion 189.1/product ions 58, 144.1 and 91.1. The transitions for 6-methoxy DMT were: parent 219.1/ product ions 77, 130.1 and 174.1. All transitions had a 50 msec dwell time. Triplicate injections of samples, calibrators and QCs were used for quantitative analysis allowing calculation of mean and standard deviation. The mean area under the curve (AUC) of the analyte relative to ISTD was used to construct a quadratic fit for the calibration curve in MultiQuant software (SCIEX, Framingham, MA). Calibrators were excluded from quantitation models if their calculated concentrations were >15 % different from theoretical. For all calibrators, samples, or QCs if the calculated %RSD of the triplicate injections was >15 % samples were not considered valid. All calibration curves had R-values ≥ 0.995. For all assays QC samples at each concentration fell within 15 % of theoretical concentrations. The lower limit of detection was based on a signal to noise value of three (determined in MultiQuant). The lower limit of quantitation for each assay was based on either a signal to noise value of ten, or set at a DMT concentration equal to the lowest concentration calibrator, whichever was higher. The upper limit of quantitation for each assay was set at the highest calibrator meeting the ±15 % of theoretical metric.
To assess the dose-dependent effects of DMT on HTR, male and female Swiss Webster mice, 6-8 weeks, were separated into two cohorts (cohort 1: DMT IV + vehicle patch; cohort 2: vehicle IV + DMT patch) and compared to animals receiving Saline IV + vehicle patch. All mice underwent a magnet implantation procedure and HTR analysis using procedures in previously reported methods. Briefly, in this protocol, mice were anesthetized with isoflurane, as described above, and a nickel magnet (Eokoaiee multi-use fridge magnets; 4 × 2 mm) was cemented (DentalWorld, Bahadurgarh, India) to the skull attached to an anchor screw. All animals were allowed to recover for at least 5-7 days prior to any behavioral experimentation. On the experimental day, mice were placed in a magnetometer cage (15.24 cm height × 15.24 cm diameter) containing ~300 rotations of 30-gauge copper wire (Essex, Fort Wayne, IN). All mice were initially analyzed for 60 min with jackets applied for baseline activity. After the initial habituation period, mice were administered the active treatment and placed immediately back into the magnetometer cage for another 60 min trial period. All signals were transmitted through a Molecular Devices (Digidata 1440A) digitizer and analyzed with Clampex software. The sampling rate used in this study was set at 3003 Hz. All data analysis was performed with MATLAB software and Prism GraphPad. HTR was assessed at each timepoint as a percent of the corresponding saline response.
Elimination rate constants and half-lives were calculated from the slope of the best fit line for Ln (Concentration) over time. Y-intercepts were calculated for IV plots from this best fit. AUCs are reported for t = 0 -terminal collection, using the best fit Y-intercept value at t = 0 for IV plots, and 0 at t = 0 for transdermal (TD) plots. Relative bioavailability (F) was calculated as: F = (AUC A / Dose A ) / (AUC B / Dose B ). All statistical analyses were performed using GraphPad Prism, version 10 (San Diego, CA). In all cases, statistical significance threshold was set as p < 0.05. All data presented as Mean ± SEM.
Formulation screenings were undertaken with DMT freebase and the three most common types of transdermal adhesives (acrylate, polyisobutylene, and silicone) and isopropyl myristate, a known permeation enhancer. The formulations F1-F6 are listed in Table. The aim of screening was to maximize DMT concentrations in the adhesive-solvent system for in vitro drug release comparison. Ethanol was determined as the ideal solubilizer for DMT and was compatible across all adhesive systems. Ethyl acetate, used in F1, required the addition of ethanol to adequately solubilize DMT. Further formulations F2-F6 used ethanol as the sole solubilizer with greater DMT solvation capacity and increased drug loading. No crystallization was observed in any formulation (F1-F6). Solubility ranking of the adhesives (highest to lowest) are as follows: acrylate (Duro-Tak 4098) > silicone (Bio-PSA) > polyisobutylene (Duro-Tak 6098). The use of isopropyl myristate greatly increased solubility of DMT in the acrylate (Duro-Tak 4098) adhesive matrix. While DMT has a relatively low boiling point, care was taken to not evaporate DMT from the adhesive during the DIA drying process but did not pose problems for scale-up manufacturing.
After drying the DIA laminate, each patch from formulations F1-F6 were weighed, thickness assessed, and average DMT content (dry w/w %) determined. The averages of these values are reported in Table. After removal of the outer edges of the laminate, which results in uneven coating due to edge effect, the intra-batch results for patches were consistent with coat weight variability rarely exceeding 10 %. In light of this, one patch from each formulation was quantitatively tested within the average thickness and coat weight. Polyisobutylene adhesive (F4) had the lowest DMT content, thickness, and DIA mass as its solubility hindered formulation. The characteristics of silicone adhesive formulation (F3) were similar to that of acrylate (F2), though higher DMT concentrations were achieved with F2. High drug loading of DMT is achievable with the acrylate adhesive, making Duro-Tak 4098 the ideal candidate for further formulation screening. The highest drug concentration without additives was F5, achieving 10.5 % w/w DMT. The addition of isopropyl myristate, a solubility and permeation enhancer, in F6 aided DMT dry weight concentrations up to 16.1 % w/w.
Within two weeks of production and storage, each formulation was subjected to Franz cell diffusion studies to quantify in vitro release of DMT (Fig.). All formulations were able to achieve linear, zero-order DMT delivery over 72 h (R 2 > 0.98), except F4, under the tested conditions. The multiple timepoints for sampling (n = 8) add replicates for the flux calculations for a single Franz cell diffusion analysis. This is congruent with prior experiments with the Strat-M membranes (unpublished data). Unsurprisingly, higher DMT concentrations tended to increase overall DMT permeation through the Strat-M membrane into the receiving well. Though there was one notable exception with F1 (2.4 % DMT w/w) having slightly greater diffusion than F3 (3.9 % w/w). This demonstrates that acrylate is a superior adhesive for drug delivery over silicone or polyisobutylene systems. Optimized formulation F5 (10.5 % DMT w/w) was able to deliver 520 μg/cm 2 over 72 h without any permeation enhancers. Only F6 (16.1 % w/w), with the addition of isopropyl myristate, bettered F5 with 775 μg/cm 2 cumulative DMT delivery. The flux values for all formulations F1-F6 are listed in Table. This initial dataset confirms that transdermal delivery systems offer a plausible and novel administration route for DMT. Based on the in vitro diffusion data, it was determined that formulation F5 would be trialed in vivo to test the hypothesis that a transdermal delivery system, without permeation enhancers, could offer sustained DMT delivery. Patches from the F5 batch were subjected to Franz cell diffusion at one week, one month, and two months to determine stability. All of the resulting fluxes were within the range of F5 (6.54 ± 2.73 μg/(cm 2 *hr) -1 ) listed in Table, indicating suitable stability for in vivo testing.
To quantify concentrations of DMT present in plasma and brain samples following delivery through both intravenous and transdermal routes, male and female mice were administered DMT doses of either or 5 mg/kg. In these animals, samples were collected at 5-, 10-, 15-, 30-, 60-, and 240-min post-drug administration for IV DMT, and at 60-, 240-, 480-, and 1440-min post-drug administration for TD DMT. Precision and accuracy for the LC/MS/MS method was acceptable and reproducible (Table, Fig.) across runs. Regarding the pharmacokinetic profile for the IV infusion, this route exhibited rapid and robust maximum plasma concentrations (C max ) of 597 ng/mL for males and 1437 ng/mL for females with a 5 mg/kg dose (Fig.). All concentrations administered through IV infusions peaked at the initial 5-minute time point, irrespective of sex, and demonstrate a dose dependent plasma exposure profile. For the lower doses of DMT (1 mg/kg) administered to males, the average C max reached 372 ng/mL, while the same dose in females achieved a concentration of 528 ng/mL. The transdermal DMT patch F5, on the other hand, demonstrated a sustained release profile, resulting in a prolonged time to peak concentration (Fig.), and much lower maximum concentrations overall. The average C max for male animals reached 37.0 ng/mL at 5 mg/kg and 8.0 ng/mL at 1 mg/kg, both occurring at the 1-hour time point. Interestingly, female animals achieved a C max of 54.8 ng/mL at the 4-hour time point with a 5 mg/kg dose. Average C max for the 1 mg/kg dose was achieved at 1-hour post administration, peaking at 15.7 ng/mL. In addition to the DMT patch F5 blunting the peak concentrations of DMT achieved, these results reveal distinct prolongation of the apparent half-life (t 1/2 ) by the transdermal delivery system, as calculated using the transformed concentration-time profiles of plasma samples from the 1 mg/kg and 5 mg/kg doses. For the IV cohort, the average t 1/2 across was calculated to be 10.8 ± 2.5 min for males and 10.0 ± 0.1 min for females, whereas for the transdermal cohort, the t 1/2 was calculated to be 273 ± 10.5 min for males and 210 ± 14.6 min for females. Despite the quite different plasma concentration profiles demonstrated by the transdermal patch versus IV administration, the relative plasma bioavailability of DMT following transdermal delivery remains 77 ± % of that for IV administration across the 1 mg/kg and 5 mg/kg doses (Fig.). Concentrations of DMT were next assessed in brain samples taken from animals dosed with either IV or transdermal DMT (Fig.). In animals given IV DMT, brain concentrations of DMT were consistently found to be above the LLOQ only at the 15-minute time point. Interestingly, at 15 min, the 5 mg/kg IV dose induced the highest difference in brain DMT levels compared to plasma, with male and female brain samples reaching an average C max of 4503 and 8080 ng/g, respectively. In contrast, at a dose of 1 mg/kg IV, brain DMT concentrations were lower than those found in plasma; males achieved an average C max of 353 ng/g, and female animals reached a C max of 250 ng/g. In animals given transdermal DMT, brain concentrations of DMT were consistently found to be above the LLOQ only at the 1-hour time point, reaching ng/g in both male and female animals administered 5 mg/kg transdermally. The brain concentrations following 5 mg/kg transdermal DMT were significantly lower than those from 5 mg/kg IV DMT administration (Two-Way ANOVA: F Route [1,20] = 18.03, p = 0.0004). Brain concentrations of DMT were below the LLOQ at all timepoints for all animals given 1 mg/kg DMT transdermally. Overall, these data demonstrate that the transdermal preparation yielded a pharmacokinetic profile with dramatically reduced peak DMT concentrations in the periphery and CNS as compared to IV administration, while drug exposure is maintained at similar levels to IV due to a substantially prolonged half-life (Table).
To measure acute unconditioned responses to DMT mice were intravenously or transdermally dosed at 5 mg/kg (or given saline/ vehicle) and assessed for HTR(Fig.). The IV cohort demonstrated significant variation in HTR responding by time (Two-Way RM ANOVA: F RoutexTime (11, 99) = 3.69, p = 0.0002), including a significant 18-fold increase in head twitches relative to animals receiving IV saline at 0-10 min post-administration (Sidak's Multiple Comparisons, p = 0.0214). This robust increase was observed transiently with only a 2.4 and 1.2-fold increase in HTR's relative to animals receiving IV saline within the next two chronological bins. In contrast, the transdermal cohort did not show any significant variation in HTR's, as compared to animals receiving a TD vehicle patch, within 1hour post-drug administration at the same dose (Two-Way RM ANOVA: F RoutexTime (11, 242) = 0.66, p = 0.77).
This work herein demonstrates the first evidence of transdermal delivery of DMT, a notoriously low bioavailability and short half-life drug using oral administration. Despite increasing evidence of DMT's therapeutic potential, drug delivery for low or non-orally bioavailable psychedelic drugs have not been optimized. The development of a TDDS could offer a non-invasive and low-dose delivery option for DMT, inducing fewer hallucinations, and circumventing a side effect that requires therapists to continuously observe patients in clinical settings. A low-dose DMT patch can also reduce abuse potential or product misuse relative to the current high, psychedelic dose administration of DMT including intravenous or inhalation. The initial formulation strategy was to determine solvent systems which could maximize DMT drug loading into the DIA matrix. It was theorized that high drug loading of DMT would be necessary to increase flux and overcome the rapid metabolism of DMT. A solvent screening was employed to test DMT solubility in ethanol and ethyl acetate, two solvents which are both generally regarded as safe for topical use and are compatible with the adhesive systems. The first formulation (F1) was trialed using both solvents in a 1:1 ratio resulting in a cloudy formulation with a modest DMT concentration (2.4 %, dry w/w %). An optimized formulation (F2) was then tested using only ethanol in the same acrylate (Duro-Tak 4098) adhesive achieving a clear appearance with a much higher DMT concentration (6.4 %). Based on these initial results, ethanol was chosen for further screenings as the superior solvent with DMT solubility of ~400 mg/mL thereby reducing solvent evaporation and increasing DMT drug loading. The next steps were to determine the optimal adhesive system to increase DMT drug loading. Nearly all commercial transdermal patches contain either acrylate, polyisobutylene, or silicone-based adhesives so a representative from each adhesive was trialed using ethanol as a drug solubilizer. Polyisobutylene (Duro-Tak 6098, F4) had significant solubility issues with the ethanolic DMT solution, leading to a cloudy formulation with the lowest DMT concentration (0.5 %). Compared to F4, silicone (Bio-PSA, F3) achieved a higher DMT concentration (3.9 %) with a cloudy appearance post-drying. Comparing the three adhesive types, acrylate-based F2 had the highest DMT concentrations with a clear formulation. Therefore, acrylate was chosen as the ideal adhesive for further analysis. DMT concentration was scalable from F2 into F5 achieving a 10.5 % DMT concentration with only DMT, ethanol, and the acrylate adhesive. Lastly, the addition of isopropyl myristate in F6 increased the DMT concentration to 16.1 %. Franz cell diffusion assays, a common tool for transdermal formulation development, were used to test DMT drug flux for each formulation. A direct correlation was observed between DMT drug loading and flux and all formulations F1-F6 displayed the ability to permeate DMT through a skin-like membrane, with notable differences seen with adhesive types. When comparing adhesive types directly, acrylate (F2) displayed greater flux, though not significant, over silicone (F3) while both had significantly greater DMT fluxes over polyisobutylene (F4). This correlation could be driven by polarity with acrylate being the most polar, followed by silicone, then polyisobutylene being the most nonpolar. Increasing DMT concentration of F5 had a 2-fold greater impact on flux than F2 without the use of any permeation enhancers. Isopropyl myristate expectedly increased the flux in F6 over F5. DMT patch F5 was chosen for in vivo screening to show the applicability of a simple DIA formulation, without the use of permeation enhancers, to deliver DMT transdermally. Furthermore, the drug loading (10.5 % DMT) and DMT flux of 6.54 ± 2.73 μg/(cm 2 *hr) -1 of F5 in Franz cell testing were theorized to show meaningful plasma concentrations in mice without hallucinogenic effects, measured by the HTR. Compared to IV administration, DMT patch F5 elicited a dramatically lower C max in both plasma and brain samples of Swiss-Webster mice. Additionally, while all DMT was cleared from the system within 1 h of IV administration, the DMT patch F5 resulted in measurable concentrations of peripheral DMT for at least 8 h, with a 20-fold increase in the apparent half-life of DMT at doses of 1 to 5 mg/kg. Interestingly, there were notable sex differences observed in drug exposure, with DMT reaching higher plasma concentrations at the 5 mg/kg and 1 mg/kg doses in female animals as compared to males, regardless of route of administration. Human clinical trials would be wise to monitor PK differences in male and female patients especially with high bolus dosages. Previous studies have shown variations in DMT effects in male versus female rats that correlate with weightas well as efficacy differences in mice with the similar 4-substituted tryptamine psilocybin. Future studies to address the source of this difference should be undertaken. The rapid onset of DMT patch F5 is shown through detectable plasma concentrations in vivo at the first timepoint of 60 min. Peak plasma concentrations for DMT patch F5 were also detected at the first timepoint at both 1 mg/kg and 5 mg/kg. While initial drug onset and peak plasma concentration may be higher at timepoints prior to 1 h with TDDS, it is unlikely they approach levels seen in IV administration. This is especially true given the distinct lack of HTR response from the transdermal treatment group up through 1 h. There were notable differences between the in vitro Franz cell diffusion and the in vivo results for patch F5. The patches continued to show continuous delivery of DMT up to 72 h, however, the in vivo results showed no detectable DMT after 8 h. Presumably the lack of enzymes to metabolize DMT in the Franz cell led to these prolongated results in vitro. Rodents are known to be rapid metabolizers of drugs, often more than humans or pigs, the ideal species for transdermal drug development. In spite of this, Franz cell diffusion represents a good screening tool for transdermal formulation development but durability of drug effects can only be accurately monitored in vivo. Transdermal delivery is common to overcome the first-pass effect for drugs with low oral bioavailability as is the case with DMT. The plasma AUC for both IV and transdermal DMT were compared to show an apparent 77 ± 10 % bioavailability for DMT patch F5 across 1 and 5 mg/ kg doses in males and females. The PK results obtained with DMT patch F5 are notably higher than expected given the lack of permeation enhancers or MAOIs to reduce drug metabolism. This is an incredibly efficient delivery method of DMT with high bioavailability for a drug which is commonly administered therapeutically in non-ideal delivery formats (intravenous, smoked or inhaled, or aqueous phytoid extracts of ayahuasca). It is yet to be seen if the level of drug exposure we achieve relates to DMT's efficacy or if the subjective psychedelic experience, a C max driven effect, leads to enduring therapeutic effects. A 39-day study of non-hallucinogenic 1 mg/kg subcutaneous DMT doses every third day did provide anxiolytic properties of fear extinction in both male and female Sprague Dawley rats; however, only females benefitted from antidepressant effects via the forced swim test. Our dataset also supports the key differences in rodent gender, with notably higher C max and AUCs observed in females at 5 mg/kg. During the HTR assay, intravenously administered DMT showed an increase in head twitches relative to saline within the first 10 min of drug administration. This is consistent with DMT IV administration leading to hallucinogenic-like activity in rodent models. In contrast, DMT patch F5 did not show significant increases in HTR relative to saline, suggesting that use of TDDS to avoid a rapid rise in plasma and brain DMT concentrations are also sufficient to avoid induction of hallucinogenic-like activity in mice. The average peak plasma concentrations of the 5 mg/kg DMT patch F5 did not exceed 60 ng/mL, a proposed threshold for the psychedelic or hallucinogenic effects of DMT. Taken together these characteristics address possible limitations of DMT administration including the opportunity for unsupervised or take-home dosing. Future studies could assess higher doses to see if the HTR could be elicited through TDDS or if permeation-enhanced formulations such as F6 can exceed DMT's psychedelic threshold. Increased neural activity and dendritic densities are correlated with neuroplasticity, a proposed mechanism of action for DMT and other psychedelic 5-HT 2A agonists. Previous studies did not show increased dendritic spine densities with repeated 1 mg/kg dosesdespite showing similar spontaneous excitatory postsynaptic currents for both 10 mg/kg and 1 mg/kg doses. However, a single 10 mg/kg dose did achieve psychedelic effects via HTR and increased spine density. It is still unclear whether lower repeated doses, often referred to as microdosing, can lead to therapeutic effects in animals and humans. The screening efforts landed on a novel TDDS formulation which was able to deliver a sustained dose of DMT through in vitro and in vivo models. This work provides evidence supporting the design and development of a single layer DIA system containing DMT freebase to achieve long-lasting peripheral exposure to DMT without inducing hallucinogenic-like activity in rodent models, a condition that was achieved at a 5 mg/kg dose. While the results are quite surprising, the physical properties and lipophilicity of DMT make it an attractive drug for TDDS. These results demonstrate the adequacy of DMT as a transdermal drug candidate without using enhancers or techniques such as microneedles to bypass the stratum corneum, the outermost skin layer that can be impermeable for some drugs. Despite the promise of microneedles to aid the delivery of drugs, such as ketamine, challenges remain to manufacture these products reproducibly under strict regulatory guidelines, and to date, only a few microneedle transdermal patches have been approved by the FDA. The Franz cell and in vivo results of DMT patch F5 clearly validate the non-invasive approach of DMT TDDS. Psychedelic compounds remain a promising class of drugs for a wide variety of neurological disorders. However, challenges remain to treat large numbers of patients with medical observation and limited options are available for patients unable or unwilling to undergo a psychedelic treatment. New drug delivery methods can reduce peak drug exposures to reduce or eliminate hallucinogenic side effects and improve overall patient compliance. Notably, many neuropsychiatric researchers still debate whether or not subjective effects like hallucinations are necessary for efficacy. For the first time, this research demonstrates a highly effective TDDS for DMT to extend half-life, reduce peak drug concentrations likely to limit hallucinogenic-like effects correlated with a rodent HTR. Further work dosing larger animals and humans will allow a better understanding of the translatability and therapeutic potential of low-dose DMT regimens. Non-or sub-hallucinogenic delivery systems capable of enabling neuroplasticity would expand patient populations beyond psychiatry to include neurodegenerative disorders. Convenient TDDS of DMT could significantly reduce medical burdens and provide new tools for physicians and patient treatments.
A low-dose DMT product can offer new solutions for psychiatric and neurological disorders without inpatient dosing and supervision required for higher, psychedelic DMT doses. The primary goal was to demonstrate the viability of TDDS of DMT given general bioavailability concerns and rapid plasma clearance. The study was successful in producing optimized transdermal patch formulations of DMT with high bioavailability. The initial in vitro Franz cell diffusion screening provided examples of DMT in several commercially available adhesives and patch materials with increasing drug concentrations and drug fluxes with or without permeation enhancers. The optimized DMT patch F5 had suitable drug loading and flux to advance into in vivo studies in mice. Brain concentrations and pharmacokinetics of DMT along with the HTR were monitored to understand PK/PD drug effects with this novel TDDS. The IV half-life of DMT was extended by 20-fold with administration of the TDDS while plasma concentrations did not exceed 60 ng/mL, eliminating the HTR. Research to establish the therapeutic potential and neural effects of this delivery method along with PK profiling in further species will provide better value to this initial discovery of transdermally administered DMT.
Create a free account to open full-text PDFs.
D’Souza, D. C., Syed, S. A., Safi-Aghdam, H. et al. · Neuropsychopharmacology (2022)
Frecska, E., Bokor, P., Winkelman, M. J. · Frontiers in Pharmacology (2016)
Gallimore, A. R., Strassman, R. J. · Frontiers in Pharmacology (2016)
Gasser, P., Holstein, D., Michel, Y. et al. · Journal of Nervous and Mental Disease (2014)
Good, M., Joel, Z., Benway, T. et al. · European Journal of Pharmacology (2022)
Goodwin, G. M., Aaronson, S. T., Alvarez, O. et al. · New England Journal of Medicine (2022)
Goodwin, G. M. · Journal of Affective Disorders (2023)
James, E., Erritzoe, D., Benway, T. et al. · Frontiers in Psychiatry (2024)
Jiménez, J. H., Bouso, J. C. · Journal of Psychopharmacology (2022)
Kiilerich, K. F., Lorenz, J., Scharff, M. B. et al. · Molecular Psychiatry (2023)
Johnson, M. W., Gründer, G., Betzler, F. et al. · Current Addiction Reports (2021)
Ly, C., Greb, A. C., Cameron, L. P. et al. · Cell Reports (2018)
Nichols, D. E., Walter, H. · Pharmacopsychiatry (2020)
Palhano-Fontes, F., Barreto, D., Onias, H. et al. · Psychological Medicine (2018)
Polito, V., Stevenson, R. J. · PLOS ONE (2019)
Szabó, Í., Varga, V. É., Dvorácskó, S. et al. · Neuropharmacology (2021)
Thomas, G., Lucas, P., Rielle Capler, N. et al. · Current Drug Abuse Reviews (2013)
Vargas, M. V., Dunlap, L. E., Dong, C. et al. · Science (2023)