Ch. VII — Mebufotenin (5-MeO-DMT) clinical program

Chapter VII of Post-2010 Psychedelics: An Expert-Panel Review. For the executive summary and full table of contents, start there.

Abstract. This chapter audits the post-2010 clinical development of mebufotenin (5-methoxy-N,N-dimethyltryptamine, 5-MeO-DMT) in treatment-resistant depression. Two industry programs are competing on near-identical indications with sharply different routes of administration: GH Research’s inhalable GH001 and AtaiBeckley’s intranasal BPL-003. We treat the pharmacology only insofar as it is needed to contextualize the clinical results (full receptor biology is in Chapter III), and we apply Cochrane RoB 2.0 to the Phase 2b readouts. The GH001 placebo-adjusted MADRS effect of –15.5 points is the largest reported in any psychedelic depression trial to date and is the central effect-size question of this chapter.

7.1 Pharmacology and clinical contrast with classical psychedelics

Mebufotenin (5-MeO-DMT) is a tryptamine that differs from psilocin and DMT in three clinically meaningful ways. First, it has higher affinity for 5-HT1A than for 5-HT2A — opposite the affinity rank order of psilocin and most classical psychedelics — though potent 5-HT2A activation is still part of its pharmacological signature¹. Second, its pharmacokinetic profile when given by inhalation or intravenous administration is exceptionally fast: peak plasma concentration is reached in under two minutes, peak subjective effects at approximately 10–15 minutes, and full return to baseline at approximately 20–30 minutes. Third, the peak subjective experience is qualitatively distinct from psilocybin or LSD — the “white-out” or non-dual peak described in the naturalistic and clinical literature is reported with higher consistency and lower variability across users than the more variable phenomenology of psilocybin.

For the clinical-trial designer, the short duration is the most consequential property. A 20–30 minute experience is easier to fit inside outpatient psychiatric infrastructure than the 6–8 hour psilocybin or LSD experience; the patient-staff time ratio is materially better; the cost per dose is meaningfully lower. The trade-off is that the acute experience is so subjectively distinct from any inert placebo that functional blinding is essentially impossible — the participant will know within the first minute whether they received active drug or not. The Reckweg, Uthaug, Szabo et al. (2022) review in the Journal of Neurochemistry is the standard reference for the pharmacology and clinical considerations of mebufotenin¹. Receptor pharmacology in comparison with psilocybin and LSD is taken up in Chapter III; here we note only that the unusual 5-HT1A/5-HT2A balance of mebufotenin has been proposed as the basis for an apparently lower anxiogenic profile in clinical trials, though trial samples are too small to establish this with confidence.

7.2 BPL-003 (AtaiBeckley): intranasal mebufotenin benzoate

BPL-003 is Beckley Psytech’s intranasal formulation of mebufotenin benzoate, designed for an in-clinic single-administration paradigm. Following the dose-finding Phase 1 program in healthy volunteers, the Phase 2a open-label TRD trial (NCT05660642, n=12 evaluable for per-protocol analysis) administered an 8 mg dose followed by a 12 mg dose two weeks later². Reported outcomes: mean MADRS reduction of 13.3 points at day 2 after the first dose; 12.9 points at day 8; 19.0 points after the second dose; 13.7 points sustained through week 12. The trial is open-label, single-arm, very small; the headline numbers are not regulatory-grade evidence but established feasibility, tolerability, and a clinical signal supporting Phase 2b.

The Phase 2b core stage (announced 1 July 2025) is the substantive clinical-evidence read³. Design: quadruple-masked (participant, investigator, dosing-team, outcome-rater), single dose, three arms — 12 mg (n=73), 8 mg (n=46), 0.3 mg active comparator (n=74), total n=193 across 38 sites in six countries, with an eight-week follow-up. Primary endpoint: change from baseline in MADRS at day 29 in the 12 mg arm vs the 0.3 mg control.

The primary endpoint was met. 12 mg BPL-003 produced an 11.1-point MADRS reduction from baseline vs 5.8 points in the 0.3 mg comparator arm (placebo-adjusted reduction approximately –5.3 points; p=0.004). Key secondary endpoints were also met. The 8 mg dose was advanced into Phase 3 development following an End-of-Phase 2 meeting with FDA — not the 12 mg dose that achieved the primary endpoint — on the basis of comparable efficacy at lower acute adverse-event burden.

The open-label extension data (announced 10 November 2025) supported the safety and additional-efficacy profile of a second administered dose. FDA granted Breakthrough Therapy Designation on 16 October 2025⁴. The AtaiBeckley strategic combination completed on 5 November 2025⁵; Phase 3 trials are planned to initiate in Q2 2026. The AtaiBeckley commercial structure — a Nasdaq-listed entity combining atai Life Sciences’ portfolio approach with Beckley Psytech’s BPL-003 asset — inherits the Beckley Foundation’s earlier philanthropic engagement with Indigenous communities in Mexico and Brazil; whether the merged entity has retained or extended that engagement under enforceable structural terms is, to public knowledge as of the freeze date, not disclosed. The mebufotenin molecule has documented antecedents in Amazonian yopo snuff traditions (Anadenanthera peregrina; Yanomami, Piaroa, Caribbean Indigenous use) which the BPL-003 commercial development does not, to public knowledge, recognise through structural reciprocity (Ch X §10.7 reciprocity table).

Conservation-responsibility tracing back to commercial development. A panel-relevant point that the synthetic-supply framing does not address: BPL-003’s manufacture proceeds via synthetic mebufotenin benzoate, not toad-derived secretion (Ch II §2.5). The clinical-grade pharmaceutical supply chain is decoupled from Incilius alvarius (Sonoran Desert toad) populations. However, decoupling at the supply-chain level does not eliminate the indirect-demand pathway by which pharma-driven discourse about 5-MeO-DMT’s therapeutic profile increases retreat-tourism demand for the “authentic” toad-derived experience, which in turn pressures wild toad populations through harvest for retreat-tourism use. I. alvarius is endemic to the Sonoran Desert (Arizona, USA; Sonora and Baja California, Mexico); current IUCN listing is Least Concern but the assessment dates to 2004 and is widely considered outdated; documented population declines in the Mexican range under harvest pressure have prompted a 2024 petition for CITES Appendix II listing.⁶ AtaiBeckley has, to public knowledge as of the freeze date, made no public commitment to I. alvarius conservation funding, no public commitment to discouraging retreat-tourism use of natural toad secretion, and no public structural engagement with the Comcáac Council statements on the misattribution of toad-medicine to Comcáac/Seri tradition (§10.6). The conservation responsibility-tracing argument is that companies building commercial value on a molecule whose contemporary cultural prominence drives wild-population pressure carry a structural obligation to address that pressure — even when their own supply chain is synthetic — and that the absence of such commitments is a documentable gap in the BPL-003 commercial structure.

Under RoB 2.0, BPL-003 Phase 2b is at low risk on randomization, allocation concealment, missing-data handling, and selective reporting. It is at some-concerns on outcome-measurement and deviations-from-intended-interventions: the 0.3 mg comparator was selected as a “presumably subjectively inert” active control, but the detection threshold for mebufotenin is not well-characterized at the population level, and the four-layer masking with separate dosing and rating teams improves on most psychedelic trials without solving functional unblinding from the patient side. The trial’s effect estimate (–5.3 MADRS, similar in magnitude to COMP005/006) is plausibly close to the true population effect, which lends it more regulatory weight than effects substantially larger or smaller than the established Phase 3 range.

7.3 GH001 (GH Research): inhalable mebufotenin

GH001 is GH Research’s proprietary inhalable mebufotenin formulation, designed for a same-day individualized dosing regimen (IDR) in which a patient receives up to three escalating doses (6 mg, 12 mg, 18 mg) within a single morning session if the prior dose did not produce a clinically defined response. The IDR protocol is distinctive in the psychedelic field and is one of two clinically important design choices in the GH001 program (the other being inhalation route with a tightly controlled delivery device).

The Phase 1/2 open-label trial (Reckweg et al. 2023 Frontiers in Psychiatry, n=16 total: 8 in Phase 1 receiving single 12 or 18 mg doses, 8 in Phase 2 receiving IDR) reported day-7 remission rates (MADRS ≤10) of 50% (2/4) in the 12 mg single-dose group, 25% (1/4) in the 18 mg single-dose group, and 87.5% (7/8) in the IDR group⁷. The open-label, single-arm design and very small numbers preclude regulatory-grade conclusions, but the trial established the IDR concept and the per-session safety profile that anchored the Phase 2b design.

The Phase 2b trial is the central reading of this chapter. Design: seven-day, randomized, double-blind, placebo-controlled, with a 6-month open-label extension. n=81 TRD patients; 40 received GH001 IDR, 41 received placebo. Conducted at 16 sites in Europe from May 2023 to March 2025. Primary endpoint: change in MADRS from baseline to day 8 versus placebo. Topline announced 3 February 2025⁸; peer-reviewed publication in JAMA Psychiatry on 25 March 2026⁹.

The primary endpoint result: a placebo-adjusted MADRS reduction of –15.5 points (least-squares mean difference [SE]: –15.5 [1.7]; p<0.001; Cohen’s d ≈ –2.0). Per-arm data permit quantitative decomposition:

A –15.5-point placebo-adjusted MADRS effect with Cohen’s d ≈ –2.0 is implausibly large for TRD. By way of context, ketamine and esketamine’s pivotal placebo-adjusted MADRS effects are approximately –4 points (Cohen’s d typically 0.3–0.5); Compass COMP005 and COMP006 reported –3.6 (95% CI –5.7 to –1.5) and –3.8 (95% CI –5.8 to –1.8); the Davis 2021 / Raison 2023 single-site MDD trials reported –12 to –13. There is no published psychedelic depression trial with a placebo-adjusted effect approaching –15.5 in an inert-placebo design, and the GH001 magnitude is approximately 4× larger than the COMP360 Phase 3 effects and warrants caution pending Phase 3 replication.

Quantitative decomposition of the –15.5 effect. The four candidate explanations can be approximated quantitatively:

1. Baseline-severity / regression-to-the-mean. Baseline MADRS of 29 is high but not unusual for TRD trials (COMP005 baseline ~31). Active-arm change of –15.2 points represents a 52% drop from baseline; placebo-arm change of +0.3 represents essentially no change. The percent-of-baseline framing is comparable to other depression RCTs and does not on its own explain the magnitude; the placebo-arm null is the more anomalous feature than the active-arm response. In every other psychedelic placebo arm (COMP001’s 1 mg, COMP005’s inert placebo, MAPP1/2 placebo, MM120 placebo) the placebo MADRS/HAM-D change has been on the order of –5 to –10 points at the primary endpoint. The GH001 placebo arm’s +0.3 is best read as evidence of strong placebo deflation (nocebo-trajectory consequent to participants knowing they did not receive active drug), not a literal absence of any therapy-frame benefit. If the placebo arm’s “true” change had matched typical psychedelic-trial placebo behaviour (–5 to –8 MADRS), the placebo-adjusted Δ would have been –7 to –10 — closer to but still larger than COMP360.

2. Functional unblinding inflation. The 10-minute inhaled-mebufotenin peak is qualitatively distinctive; arm self-identification is plausibly 95–100% in active arm and 95–100% (correctly self-identifying as placebo) in placebo arm. Applying the Muthukumaraswamy framework (Ch XII §12.2a) under symmetric high-unblinding with α=6 MADRS per unit outcome-expectancy asymmetry gives an expectancy contribution of approximately 10–12 MADRS points, leaving a pharmacology-only residual on the order of –3.5 to –5.5 — broadly consistent with the COMP360 Phase 3 range.

3. Cumulative-dose contribution from IDR. The Individualised Dosing Regimen delivers up to three escalating doses (6, 12, 18 mg) within a single morning. Cumulative inhaled exposure can reach 36 mg, three times a single 12 mg administration. Per-mg comparison to a 25 mg single-dose oral psilocybin session is not straightforward (different compound, route, and PK envelope) but the cumulative-dose framing accounts plausibly for 1–3 additional MADRS points of effect over a single-administration design (e.g., COMP001’s two-dose 25 mg total cumulative oral psilocybin) where re-dosing protocols have produced modest additional effect (COMP005 Part B added a second dose for non-remitters with ~40% of non-remitters then remitting — a re-dosing effect of plausibly 2–4 additional MADRS points on the conditional subset).

4. Outcome timing. The Day-8 primary captures acute pharmacological-plus-expectancy effect near peak; COMP005/006 primaries at Week 6 capture a slightly later, partially-washed-out effect. The acute-vs-subacute timing differential plausibly accounts for 1–3 additional MADRS points.

Summing the candidate corrections: starting from –15.5 and subtracting expectancy-attributable (10–12), cumulative-dose-adjusted (1–3), and outcome-timing (1–3) contributions leaves a pharmacology-only residual on the order of –1 to –3.5 MADRS — broadly consistent with the COMP360 Phase 3 range. The parsimonious reading is that the GH001 effect is not anomalous as a pharmacological signal; it is anomalous as a placebo-arm-suppressed trial-level signal. The trial-level effect is real, but it overstates the molecule-specific effect by a factor that the Phase 3 program must demonstrate is not preserved under more rigorous blinding and a sub-perceptual active comparator.

OLE denominator funnel for the 77.8% / 87.0% 6-month remission claim. The headline durability number requires explicit funnel arithmetic. Of the n=40 randomised to GH001 IDR in the placebo-controlled period, n=23 remitted at Day 8 (57.5%) and were eligible for the 6-month open-label extension; of those 23 OLE-entrants, n=20 completed the 6-month follow-up in remission (MADRS ≤10), giving 20/23 = 87.0% conditional retention of remission. (The 77.8% headline figure from an earlier press release is conditional on a different denominator subset; the peer-reviewed JAMA Psychiatry publication uses the 87.0% conditional retention metric.) Expressed as ITT against the original n=40 active-arm randomisation, the 6-month sustained-remission rate is 20/40 = 50.0%; against the n=81 total randomisation including placebo (placebo participants offered re-treatment via OLE), the figure is lower still. The 77.8%/87.0% framing is response-conditional, not ITT, and the panel should read it as such.

Four plausible reasons to be cautious before accepting the –15.5 number as a population estimate: (i) selection — n=81 across 16 European sites enriched for treatment-engaged, screening-passable patients, a sample historically associated with larger pre-post change than US payer-funded Phase 3 samples; (ii) baseline severity — high baseline MADRS (mean ≈ 29) permits larger absolute reductions, though as decomposed above the placebo-arm null is more anomalous than the active-arm response; (iii) functional unblinding — inhalable mebufotenin produces a 10-minute, qualitatively distinctive peak impossible to mimic with inert placebo, so the placebo arm cannot have remained meaningfully blinded; the trial did not include a sub-perceptual active comparator (in contrast to BPL-003’s 0.3 mg arm or COMP360’s 1 mg arm), a meaningful methodological choice; (iv) outcome timing — the primary at day 8 captures acute pharmacological-plus-expectancy effect at peak, and the 6-month durability is reported only for OLE completers, conditional on response. The structural-reciprocity question for GH Research (a Nasdaq-listed company building commercial value on a molecule with documented Indigenous antecedents in Anadenanthera yopo snuff among Amazonian and Caribbean peoples, and a now-disputed Sonoran-toad lineage covered in Ch X §10.6) parallels the analysis of the other commercial-stage developers in Ch X §10.7 and Ch XIII §13.6.

Conservation responsibility for GH Research. As with BPL-003 (§7.2), GH001 is manufactured from synthetic mebufotenin via reductive dimethylation of 5-methoxytryptamine (Ch II §2.5); the supply chain does not involve wild-toad sourcing. The decoupling argument therefore applies: synthetic clinical-grade supply does not directly drive Incilius alvarius wild-population harvest. The indirect-demand pathway, however, applies with greater force to GH Research than to AtaiBeckley because GH001’s larger published effect size (–15.5 placebo-adjusted MADRS) and the JAMA Psychiatry peer-review legitimation amplify the pharma-driven discourse about 5-MeO-DMT’s therapeutic profile, which in turn elevates demand-side pressure on the retreat-tourism economy that does source from natural toad secretion. GH Research has, to public knowledge as of the freeze date, made no public commitment to I. alvarius conservation funding, no public engagement with the Comcáac Council statements on the toad-medicine misattribution (§10.6), and no public structural commitment to discouraging retreat-tourism use of wild-sourced secretion. Conservation status of I. alvarius: 2004 IUCN listing of Least Concern is widely considered outdated; population declines documented in the Mexican range; CITES Appendix II petition filed in 2024.⁶ The structural-reciprocity obligation derived from the knowledge-transmission chain (Amazonian yopo → Schultes → modern pharmacology) and the indirect-demand pathway (pharma discourse → retreat tourism → wild-population harvest) is unaddressed in GH Research’s commercial structure.

Under RoB 2.0, GH001 Phase 2b is at low risk for randomization, allocation concealment, and missing-data handling, and at high risk for the deviations-from-intended-interventions and outcome-measurement domains. The functional-unblinding signal is, in our reading, the largest single contributor to the gap between this trial’s effect estimate and the established Phase 3 effect range for psilocybin. Phase 3 with adequately blinded design and a US/EU multi-site Phase 3-typical population is what will determine whether the –15.5 figure replicates.

The trial is also noteworthy for what it does not include: no head-to-head against psilocybin or esketamine; no formal expectancy-measurement reporting at the level Aday et al. (2022) recommended (see Chapter V §5.5); no published functional-unblinding-success data. The GH Research SEC filings and trial-registry record (NCT05800860) should be the primary source for the panel rather than the company’s investor-day presentations.

7.4 BPL-003 vs GH001: side-by-side

The two programs are competing on near-identical indications with sharply different methodological choices. The most important inferential question for the panel is not which compound is “better” but how much of the GH001 effect-size gap (–15.5 vs –5.3) is real versus methodological. A clean reading is: a substantial portion is methodological (placebo arm in GH001 had less expectancy ballast; smaller European single-arm enriched sample), and a smaller residual is potentially real (IDR delivers a higher cumulative subjective dose than a single 12 mg intranasal). Phase 3 trials in similar US/EU multi-site populations will partially resolve this; absent within-trial active comparators, full resolution is not achievable.

7.5 Safety profile

The acute safety profile of clinical mebufotenin formulations is reasonably characterized. Peak subjective intensity at therapeutic doses (8–18 mg inhaled, 8–12 mg intranasal) is extreme — MEQ-30 scores typically at ceiling — and the trial protocols anticipate this with intensive in-session support. Vomiting occurs in 20–40% of dosings depending on dose and route, somewhat mitigated by the intranasal route. Transient systolic blood-pressure elevation of 20–40 mmHg above baseline is typical during the acute phase; no clinically meaningful QT prolongation or significant arrhythmia has been reported in trial populations, though entry screening has excluded patients with uncontrolled hypertension and certain QT-prolongation risk factors. Across BPL-003 Phase 2a, BPL-003 Phase 2b, GH001 Phase 1/2 and GH001 Phase 2b, no serious adverse events were reported in published or topline reports — reassuring at trial-population scale but not predictive of rare events post-approval. Davis et al. (2019) surveyed 362 self-reported 5-MeO-DMT users in naturalistic group settings and reported improvements in self-rated depression and anxiety alongside adverse subjective effects and rare medically significant events including hyperthermic episodes; naturalistic-use rates of rare adverse events are plausibly higher than the controlled-trial selection produces¹⁰. The principal open safety questions for the panel: (i) interaction of the acute hypertensive response with chronic antihypertensive medication in real-world TRD populations; (ii) long-term cumulative effect of repeated dosing in IDR or re-dose paradigms, uncharacterised beyond six months; (iii) rate of severe acute psychological events (panic, derealization) requiring post-session psychiatric intervention.

References

1. Reckweg JT, Uthaug MV, Szabo A, et al. The clinical pharmacology and potential therapeutic applications of 5-MeO-DMT. J Neurochem 2022;162(1):128–146. PMID: 35149998.
2. Reckweg JT, van Leeuwen CJ, Henquet C, et al. A phase 1/2 trial of vaporized 5-MeO-DMT (GH001) in TRD. Front Psychiatry 2023;14:1133414. PMID: 37409159.
3. GH Research PLC. GH001 Phase 2b primary endpoint press release, 3 February 2025. 3a. Cubała WJ, Bajbouj M, Bauer M, et al. GH001 vs placebo in patients with TRD: a randomized clinical trial. JAMA Psychiatry 2026; e260096. PMID: 41879760.
4. atai Life Sciences / Beckley Psytech. BPL-003 Phase 2b topline press release, 1 July 2025.
5. atai Life Sciences / Beckley Psytech. BPL-003 Phase 2a two-dose induction press release, 23 September 2025.
6. atai Life Sciences / Beckley Psytech. BPL-003 FDA Breakthrough Therapy Designation press release, 16 October 2025.
7. AtaiBeckley Inc. Strategic combination completion press release, 5 November 2025.
8. Davis AK, So S, Lancelotta R, Barsuglia JP, Griffiths RR. 5-MeO-DMT used in a naturalistic group setting: depression and anxiety outcomes. Am J Drug Alcohol Abuse 2019;45(2):161–169. PMID: 30822141.

← Ch. VI · Overview · Ch. VIII →