A Phase 1 single ascending dose study of pure oral harmine in healthy volunteers

Abstract

Background:

Harmine is a component of the hallucinogenic brew, Ayahuasca, which also contains the psychoactive compound, N, N-dimethyltryptamine. Whether pharmaceutical-grade harmine hydrochloride (HCl) has psychoactive effects, the doses at which these might occur, and the dose-response relationship to side effects and safety in humans are unknown.

Methods:

We conducted a Phase 1, open-label single ascending dose trial in healthy adults with normal body mass index and no prior psychiatric illness. The primary goal was to determine the maximum tolerated dose (MTD) of oral pharmaceutical-grade harmine HCl and to characterize safety and tolerability. A secondary goal was to ascertain whether any oral dose has psychoactive effects.

Results:

Thirty-four adult participants, aged 18–55 years, were screened for study eligibility. Twenty-five participants met eligibility criteria and were randomized to a single dose of 100, 200, 300, or 500 mg of harmine HCl, respectively, using a continuous reassessment method. The most common adverse events (AEs) observed were gastrointestinal and/or neurological, dose-related, and of mild to moderate severity. The MTD was determined to be between 100 and 200 mg and is weight-based, with 90% of those participants receiving >2.7 mg/kg experiencing a dose-limiting toxicity. No serious AEs of harmine HCl were identified.

Conclusions:

Harmine HCl can be orally administered to healthy participants in doses <2.7 mg/kg with minimal or no AEs. Doses >2.7 mg/kg are associated with vomiting, drowsiness, and limited psychoactivity. This study is the first to systematically characterize the psychoactive effects of pharmaceutical quality harmine in healthy participants.

Introduction

Harmine is one of several β-carboline alkaloids that have been ingested by humans for more than a millennium, and likely longer, as the herbal remedy, Banisteriopsis cappi, and as a component of the hallucinogenic brew Ayahuasca (Callaway et al., 1999; Djamshidian et al., 2016; Miller et al., 2019; Riba et al., 2003; Socha et al., 2022). Harmine is also a constituent in “Syrian rue,” made from the plant Peganum harmala, the seeds of which have been consumed by humans in the Middle East for millennia for a variety of conditions and diseases, and from which harmine derives its name (Moloudizargari et al., 2013). Harmine remains in wide use as one of two components of Ayahuasca in traditional religious services in Brazil. Ayahuasca in various formulations is also widely consumed as a recreational drug and folk therapeutic and is increasingly popular among individuals seeking spiritual and mind-altering experiences, including in the United States.

Ayahuasca is traditionally prepared as a decoction from Psychotria viridis leaves, which contain N, N-dimethyltryptamine (DMT), a potent hallucinogen, and Banisteriopsis caapi vines, which contain harmine and other structurally related β-carboline alkaloids, including harmaline and tetrahydroharmine (THH; Callaway et al., 1999; Riba et al., 2003). DMT is inactive orally and must be administered parenterally or nasally to achieve robust psychoactive effects. The inability of orally administered DMT to achieve psychoactive effects has been attributed to rapid metabolic degradation by monoamine oxidases (MAOs) in the intestine and liver, which prevent DMT from achieving meaningful circulating levels. As a reversible monoamine oxidase inhibitor (MAOI), harmine reduces the degradation of orally administered DMT, thus creating a permissive drug-drug interaction role for the hallucinatory and other psychoactive effects of DMT as a component of Ayahuasca (Callaway et al., 1999; Riba et al., 2003). This concept has recently been demonstrated for the combination of purified DMT and synthesized harmine in a controlled clinical trial and coined “pharmahuasca” (Aicher et al., 2023). The beneficial effects of Ayahuasca are described as mood-enhancing, mind-altering, and euphoric, generating pleasant visual and auditory hallucinations, and sustained peace of mind. Indeed, several recent small but well-controlled clinical trials found significant antidepressant and anxiolytic effects of Ayahuasca (Dos Santos et al., 2021; Palhano-Fontes et al., 2019; Rocha et al., 2021; Sanches et al., 2016).

Syrian rue, in contrast, does not contain DMT and is thought to contain primarily β-carboline alkaloids, including harmine, harmaline, and THH. P. harmala has been used as an herbal remedy for pain, hypertension, diabetes, and anxiety, and as an abortifacient (Moloudizargari et al., 2013).

Pharmacologically, harmine is a reversible inhibitor of MAO-A but not MAO-B, inhibits CYP3A4 and CYP2D6 and has also been demonstrated to inhibit multiple kinases, including the dual-specificity tyrosine-(Y)-phosphorylation-regulated kinase 1A (DYRK1A), cyclin-dependent kinases, and acetylcholinesterase (Zhang et al., 2020). Consistent with its preclinical neuroprotective effects (Zhang et al., 2020), harmine was sold commercially by Merck, beginning in the 1920s as a remedy for Parkinson’s disease and Parkinsonian tremor (Djamshidian et al., 2016), only to be replaced in the mid-1900s by l-1,3 dihydroxyphenylalanine (l-DOPA) and subsequent more effective dopaminergic preparations. Interest in harmine as a potential therapeutic agent has resurfaced recently, with preclinical studies providing strong evidence that harmine may possess the ability to induce proliferation of insulin-producing β-cells in the pancreas (Wang et al., 2015), consistent with P. harmala’s traditional use as an antidiabetic agent.

The adverse effects (AEs) of both Ayahuasca and P. harmala include nausea, vomiting, diarrhea, deep drowsiness, and in some cases bradycardia and hypotension (Brierley and Davidson, 2012; Callaway et al., 1999; Moloudizargari et al., 2013; Moshiri et al., 2013; Pennes and Hoch, 1957; Riba et al., 2003). While Brazilian law regulates the plants that can be used to prepare Ayahuasca, there is considerable variation in the method of preparation and the resultant concentrations of DMT and β-carboline alkaloids. Previously reported dosages of harmine as a component of Ayahuasca are in the range of 25–250 mg, with the most noted AEs being nausea and vomiting. Of course, the other β-carbolines in Ayahuasca and P. harmala preparations may contribute to their psychoactive as well as unpleasant effects. It also remains unclear whether harmine itself, perhaps via its MAO activity, has an intrinsic psychoactive ability (Brierley and Davidson, 2012). Intoxication case reports indicate that hallucinations, vomiting, tremor, ataxia, confusion, dizziness, hypotension, bradycardia, and agitation occur with large doses of β-carboline alkaloids (Frison et al., 2008; Moshiri et al., 2013), although no case reports exist for intoxication with pure harmine.

Given the recent positive clinical and preclinical trials, there is considerable interest in developing harmine for therapeutic use. Accordingly, its safety and toxicity are of considerable public health relevance, yet safety studies of pure harmine in humans are limited. In 1970, 35–45 mg (0.5 mg/kg) of harmine was administered intravenously to five healthy male volunteers (Slotkin et al., 1970). While it was reported that no psychedelic effects were observed, all five subjects did experience subjective symptoms, including bradycardia, trouble focusing the eyes, tingling, hypotension, cold extremities, and light-headedness, lasting 45 min. They also reported persistent drowsiness in four out of five subjects beyond 45 min. Pennes and Hoch administered harmine (20–960 mg) orally to 11 patients with schizophrenia (Pennes and Hoch, 1957). While no beneficial effects or hallucinations characteristic of Ayahuasca ingestion were reported, some subjects that received doses higher than 300 mg experienced nausea, vomiting, tremors, and altered sensations, including “waviness” of the environment, “sinking” sensation of the body, vibrations, and numbness. In the same study, harmine was also administered intravenously. Five of 11 subjects receiving 150–300 mg intravenous harmine did experience visual hallucinations and all experienced declines in systolic blood pressure (SBP) and heart rate. While collectively these studies suggest that harmine has many limiting side effects and may have intrinsic psychoactive properties at high doses, there is the caveat that these studies were conducted many decades ago, and the purity of the harmine preparation is unclear. Most recently, an oral disintegrating preparation of 150–200 mg synthetic harmine hydrochloride (HCl) was administered to 10 healthy male subjects in combination with placebo. In this study, no vomiting and no psychedelic effects were observed, although a close look at the data reveals that a mild blissful state did occur in the harmine-alone arm (Aicher et al., 2023). Together these studies suggest that harmine is likely well-tolerated at lower doses, but limited at higher doses by AEs, including some psychoactivity.

This Phase 1 Single-Rising Dose Study had two broad goals. The primary endpoint was the identification of the maximum tolerable dose (MTD) of oral pharmaceutical-grade synthesized harmine HCl in healthy normal humans. Note that throughout the manuscript, all doses reported are of harmine HCl (salt form). Secondary exploratory endpoints included defining whether harmine per se has intrinsic psychoactive properties and assessing the general safety and tolerability of a single oral dose of harmine HCl in healthy normal humans. This Phase 1 study was an open-label study without a placebo, thus every participant knowingly received harmine HCl. To ensure safety and to minimize participant exposure to unnecessarily high doses of harmine HCl, we utilized a continual reassessment method (CRM; O’Quigley et al., 1990) to inform the dose to be given to the next participant based on prior information. We selected a potential range of doses (100–1200 mg) based on several considerations. First, using the standard human dose conversion method of Nair (Nair and Jacob, 2016), we converted the highest non-toxic mouse dose (20 mg/kg; Li et al., 2018) and the highest non-toxic rat dose (15 mg/kg; Reus et al., 2010) to human equivalent dose, 1.62 and 2.4 mg/kg, respectively. For a 70 kg human male, these would translate to 144 and 215 mg, respectively. Oral bioavailability of harmine is likely between 5% and 10%. Thus, the oral equivalent to the 300 mg IV harmine dose employed by Pennes and Hoch ranges lies between 3000 and 6000 mg (approximately 40–90 mg/kg). Given that all subjects experienced AEs at these doses, we sought to conservatively set our high end of the range at less than half of the dose used by Pennes and Hoch, or 1200 mg (approximately 15–20 mg/kg).

We find that: (1) pure harmine HCl has very limited intrinsic psychoactive effects at tolerated doses; (2) the MTD in normal humans for a single dose is between 100 and 200 mg, or 2.7 mg/kg; and, (3) even at the higher doses of harmine HCl tested, no moderate-to-severe cardiovascular or metabolic AEs were observed. Nausea, vomiting, and drowsiness are common AEs of pure harmine HCl and are dose dependent.

Methods

Study participants

Healthy volunteers were recruited from the greater New York City metropolitan area by media advertisements or as referrals from the Screening Protocol at the Depression and Anxiety Center for Discovery and Treatment at the Icahn School of Medicine at Mount Sinai in New York, NY, USA, to evaluate the safety and tolerability of oral harmine HCl (NCT05526430). Eligible participants were healthy adults (aged 18–55 years), body mass index (BMI) of 19–30, with no active, unstable medical conditions, and no prior psychiatric history. We excluded children under 18 and adults over 55 due to concerns regarding neurodevelopmental and neurocognitive effects, respectively. Participants agreed to use contraception. Additionally, participants were required to have a level of understanding of English sufficient to complete examinations and consent.

Screening

To rule out pre-existing mental illness, participants were assessed using the Mini International Neuropsychiatric Interview (Sheehan et al., 1998). Participants were excluded if there was the presence of or a significant history of psychiatric illness requiring medication. To rule out pre-existing medical illness, all participants received a physical examination, electrocardiogram (EKG), urinalysis, and blood-based laboratory measures including complete blood count with differential, complete metabolic panel, and thyroid function panel. Exclusion criteria included the presence of significant medical disease, neurocognitive disease, neurological disease, prior or current psychiatric disease, urine toxicology positive for illicit drugs, concomitant medications with primary central nervous system (CNS) or cardiovascular effects, history of human immunodeficiency virus, hepatitis B or C, significant EKG abnormalities, heart rate >120 or <60, SBP outside the range of 100–140 mmHg, or diastolic blood pressure (DBP) outside the range of 60–90 mmHg. Women could not be pregnant or breastfeeding. Potential participants with abnormalities of clinical significance were referred to their primary physicians for further work-up and were excluded from the study.

Study design

Harmine HCl salt was administered in an open-label, single-rising dose-escalation design, using the CRM to inform the dose to be given to the next participant based on prior information. The CRM used a one-parameter logistic regression model to estimate the relationship between dose and dose-limiting toxicity (DLT) risk to inform decisions on dosing with a target DLT rate of 25%. It first starts with a selected target DLT rate and a mathematical model for the relationship between dose and toxicity, the prior dose-toxicity curve. After a participant experiences a DLT or not, the dose-toxicity curve is re-fit incorporating the latest outcome. At every step, the next patient is assigned the dose estimated to be nearest to the MTD. To minimize the exposure of participants to potentially toxic doses, we employed some modifications to the CRM, including starting with the lowest dose available and not skipping dose levels when escalating. Additionally, we included an early stopping rule that the study would end when either the maximum number of allowable patients treated per protocol was reached or would end early in the event that the probability of the next 10 patients being given the same dose level exceeds 90%, regardless of DLT outcomes observed.

The study design included a total of seven possible doses in a maximum of 40 participants, defined per protocol: 100, 200, 300, 500, 700, 900, and 1200 mg. Each participant received a single oral dose of harmine HCl on the treatment day and was observed for 8 h with continuous medical monitoring and followed up at 24 h. Note that doses >500 mg were not used due to the stopping rule above.

The primary endpoint of the trial was establishing the MTD of oral harmine HCl in healthy participants. The secondary endpoints were characterization of the psychiatric effects of oral harmine in healthy participants and identification of potential adverse events (AEs).

Approvals

The study was approved in advance by the Icahn School of Medicine Institutional Review Board and was registered with ClinicalTrials.gov: (NCT05526430). The study was supported by the National Institutes of Health (R01 DK128242).

Harmine HCl capsule drug substance preparation

Harmine HCl was prepared from a commercially available harmine base, recrystallized, purified, and packaged into gelatin capsules containing 100 mg each under Good Manufacturing Practice (GMP) conditions compliant with Food and Drug Administration (FDA) requirements at the independent Institute for Therapeutics, Discovery, and Development at the University of Minnesota College of Pharmacy. Periodic mass spectroscopy and high-performance liquid chromatography (HPLC) stability studies after manufacture and then at intervals out to 2 years confirmed stability and >99.9% purity during the entire duration of the study. The pharmaceutical-grade harmine HCl capsules were shipped to and remained under the control of, the Mount Sinai Investigational Drug Service. It was stored at 4°C and was dispensed on a day-of-visit, per-dose basis.

FDA Investigational New Drug (IND) submission and review

An application (IND#: 148461) was submitted to the FDA for review on May 20, 2020 and a letter indicating the study may proceed was received from the FDA on July 1, 2021.

Study protocol

Participants were requested to arrive on the study day having followed a low tyramine diet for 72 h (O’Halloran et al., 2004), having ingested nothing by mouth except water after midnight, and fasting on the day of study. On the day of the study, participants arrived at the Icahn School of Medicine Clinical Research Unit at 7:30 AM, an antecubital intravenous line was placed, and then baseline blood and urine, and assessments were collected. Participants were allowed to sit in a hospital bed or a reclining chair in a dimly lit hospital room, ad libitum for the 8 h of the study. No music was provided for participants, although they were allowed to bring their own reading or digital entertainment. The harmine HCl capsule(s) was/were administered at approximately 9:00 AM to achieve the desired dose for the subject. Blood was drawn at the times shown in Supplemental Figure 1. A physician and study coordinator were in attendance throughout the study. Participants were discharged at approximately 6:00 PM following a final medical safety assessment and returned on the following morning at approximately 9:00 AM for follow-up medical assessment and final blood and urine collection. Participants were offered a snack at approximately 11:00 AM and a standard 350–450 calorie meal (65%–67% carbohydrate, 16%–18% protein, and 14%–18% fat) at approximately 1:00 PM, as indicated in Figure 4 and Supplemental Figure 1.

Figure 4.

VAS of subjective states after administration of a single oral dose of harmine HCl. (a–j) Ten subjective states were assessed by self-report at repeated intervals after a single oral dose of harmine. Time “0” is baseline prior to dosing. (a) Nausea, (b) Hunger, (c) Feeling High/Intoxicated, (d) Drowsiness, (e) Anxiety, (f) Depressed mood, (g) Happy mood, (h) Excitement, (i) Feeling of control, (j) Vividness of image. All data are presented as the mean ± SEM. Three subjects receiving 300mg and the single subject receiving 500mg are grouped as “>200mg.”

Psychiatric assessment

A combination of clinician-administered and patient self-report instruments were administered at baseline and at regular intervals after harmine administration as shown in the Figures to determine the psychoactive effects of pure harmine. A psychiatric mental status exam was administered by a licensed psychiatrist at screening, prior to study drug administration, prior to discharge after study drug administration, and at the exit visit (24 h after study drug administration).

Profile of Mood States Bipolar Scale (POMS-Bi).

Mood and potential psychoactive effects of harmine HCl were assessed using the POMS-Bi, a 72-item psychological self-report rating scale used to assess transient, distinct mood states. It is also a validated instrument for identifying the effects of drug treatments (O’Halloran et al., 2004). Items are rated on a four-point scale from 0 “much unlike this” to 3 “much like this.” It includes six bipolar scales: composed-anxious, agreeable-hostile, elated-depressed, confident-unsure, energetic-tired, and clearheaded-confused. A maximum of score 36 can be earned on each subscale, with higher scores associated with better mood.

Brief Psychiatric Rating Scale (BPRS).

The BPRS is a clinician-administered scale that captures acute behavioral changes throughout treatment (Overall and Gorham, 1962). The scale includes 16 items, each aimed at assessing components of psychosis, including four items assessing symptoms of psychosis, for example, conceptual disorganization, hallucinatory behavior, suspiciousness, and unusual thought content. Another three items assess negative symptoms of psychosis, for example, blunted affect, emotional withdrawal, and motor retardation. The remaining items assess activation and hostility by capturing tension, mannerisms and posturing, uncooperativeness, and grandiosity. Items that assess guilt, anxiety, depressed mood, and somatic concerns are also present. Each item is rated on a scale from 1 (not present) to 7 (severe), and the item scores are summed to give a total score (range 16–112). This measure served as the primary measure of psychotic-like reactions related to the study drug.

Visual Analog Scale (VAS).

The VASs are self-report scales used to assess subjective state changes (bond and Lader, 1974). They are 100-mm horizontal lines marked proportionately to the perceived intensity of the subjective experience (0 = not at all, to 10 = extremely) for the following states: anxious, depressed, drowsy, high, hungry, nauseous, control/dominance, happiness/pleasure, excitement/arousal, and vividness of image.

Perceived Stress Scale (PSS).

The PSS is a 10-item self-report scale that measures the perception of stress (Cohen et al., 1983). Each item is rated on a five-point scale ranging from never (0) to almost always (4). Positively worded items are reverse scored, and the ratings are summed, with higher scores indicating more perceived stress. PSS-10 scores are obtained by reversing the scores on the four positive items: For example, 0 = 4, 1 = 3, 2 = 2, etc., and then summing across all 10 items. Items 4, 5, 7, and 8 are the positively stated items.

Patient Rated Inventory of Side Effects (PRISE).

The PRISE is a self-report Adverse Event Checklist used to qualify side effects by identifying and evaluating the tolerability of each symptom (Rush et al., 2004). This was administered at repeated intervals after harmine administration to capture subjective side effects not outwardly observable by the study team.

Columbia Suicide Severity Rating Scale (C-SSRS).

The C-SSRS is a clinician-administered suicidal ideation and behavior rating scale used to evaluate suicide risk (Posner et al., 2011) and was used to monitor for the emergence of any suicidal thinking or behavior after harmine administration.

Cardiovascular and general safety monitoring

Participants were visually observed continuously during the 8-h treatment visit and vital signs assessed every half hour for the first 4 h and then hourly for the remainder of the visit or when participants reported symptoms suggestive of cardiac source.

Metabolic blood testing

Since harmine has been suggested to affect blood glucose and is metabolized primarily through the gastrointestinal epithelium and liver (Brierley and Davidson, 2012; Callaway et al., 1999; Riba et al., 2003), blood was obtained for metabolic laboratory assessments at repeated intervals as shown in the Figures. Assays were performed in the Clinical Chemistry Laboratory of the Mount Sinai Hospital.

Statistical analyses

The mean, standard deviation, and the range or frequency (n) and percent (%) were each reported as appropriate for all patient demographics, vital signs, lab values, and psychological rating scales using SPSS (IBM, Armonk, New York, USA) and/or SAS Version 9.4 software (SAS Institute, Cary, North Carolina, USA). Additionally, all measures that were plotted visualize the mean ± standard error of the mean.

Results

Participant demographics

Participant enrollment is summarized in the CONSORT Diagram in Figure 1. Between September 13, 2022, and June 30, 2023, a total of 34 participants signed consent. Of those, seven participants failed screening because of abnormal screening labs or vital signs outside the normal range and were excluded. Twenty-seven participants were assigned to receive a single, oral dose of harmine HCl. Of these, two participants withdrew on or before the day of study due to the inability to attend the dosing visit, before receiving harmine HCl. Thus, a total of 25 participants were dosed with harmine HCl according to the CRM protocol. Overall, 10 participants received 100 mg, 10 received 200 mg, 4 received 300 mg, and 1 received 500 mg. All participants who received harmine HCl completed all study procedures and are included in this analysis. Participant demographics are summarized in Table 1.

Figure 1.

CONSORT (Consolidated Standards of Reporting Trials) diagram for the study.

Table 1.

Participant demographics.

Study participants (number)	n = 25
Median age, years (range)	28 (21–45)
Male, n (%)	13 (52)
Female, n (%)	12 (48)
Weight, kg (range)	66.6 (51.3–100.5)
Body mass index, kg/m2 (range)	25.2 (19.5–30)
Race, n (%)
White/Caucasian	12 (48)
Asian	7 (28)
Black/African American	3 (12)
American Indian/Alaskan Native	1 (4)
More than 1	2 (8)
Ethnicity, n (%)
Hispanic or Latino	6 (24)
Not Hispanic or Latino	18 (72)
Unknown/Did not disclose	1 (4)
Education level, n (%)
Some college	1 (4)
Graduated 2-year college/trade school	1 (4)
Graduated 4-year college	4 (16)
Some graduate/professional school	10 (40)
Completed graduate/professional school	9 (36)
Marital status, n (%)
Married/living with someone as if married	9 (36)
Never married	16 (64)

Harmine HCl tolerability, safety, and determination of the MTD

The main study goal was to define the MTD for harmine HCl in healthy people. A DLT was defined a priori as (1) a serious AE with causality at least “Possibly” related to study drug; or (2) a non-serious AE rated as at least moderate and “Possibly” related to study drug; or the experience of any of the following psychiatric symptoms after study drug administration; (3) visual hallucinations; (4) humming or buzzing noises; or (5) sensations of sinking, body vibrations or waviness of the environment; or a significant change in vital signs within 6 h of dosing, including (6) symptomatic hypotension or >20% decrease in SBP from baseline and absolute SBP < 90; (7) symptomatic hypertension or >20% increase in SBP or DBP from baseline and absolute SBP > 170 or DBP > 95; or (8) new onset tachycardia (heart rate >100 bpm) and >20% increase from baseline or symptomatic bradycardia (<60 bpm) and >20% decrease from baseline.

DLTs as a function of harmine HCl dose are displayed in Figure 2 and detailed in Table 2. The predominant DLTs from oral harmine HCl were in categories 2–4, and included vomiting which occurred in a dose-related fashion: 0 of 10 (0%) participants receiving 100 mg, 4 of 10 (40%) participants receiving 200 mg, 3 of 4 (75%) participants receiving 300 mg and in the single participant (100%) receiving 500 mg. Vomiting occurred approximately 60–90 min after dosing and participants typically vomited again 30–60 min later before resolving by 3 h after harmine regardless of administration of anti-nausea medication (metoclopramide). Despite the presence of nausea and vomiting, there was little effect of harmine HCl on hunger (Figure 3). Notably, DLTs were also dose-related on a weight basis: none of 15 participants receiving <2.7 mg/kg experienced a DLT, whereas all 10 participants receiving >2.8 mg/kg experienced a DLT. We concluded that the MTD of oral harmine HCl lies between 100 and 200 mg, or at 2.7 mg/kg. Based on this data, we determined that we would not dose any new participant with more than 200 mg and that we reached a 100% probability that no patients should be dosed at a higher dose regardless of DLT. Therefore, we closed the study to further enrollment.

Figure 2.

AEs and DLTs for a single oral dose of harmine HCl (HCl) are weight-based. (a) The number of AEs experienced per participant as a function of mg/kg harmine HCl. (b) The number of DLTs experienced per participant as a function of mg/kg harmine HCl. Note that no DLTs occurred below 2.7 mg/kg. Three subjects receiving 300 mg and the single subject receiving 500 mg are grouped as “>200 mg.”

Table 2.

Dose-limiting toxicity.

Number of participants (%)	100 mg	200 mg	>200 mg	Overall
	n = 10	n = 10	n = 5	n = 25
DLT Criteriona
1. A non-serious adverse event rated as moderate and at least "possibly" related to the study drugb	0 (0)	4 (40)	5 (100)	9 (36)
2. Sensation of sinking, body vibrations, or waviness of the environment	0 (0)	2 (20)	1 (20)	3 (12)
3. Humming/buzzing noises	0 (0)	0 (0)	2 (40)	2 (8)
4. New onset tachycardia and >20% increase in heart rate (HR) from baseline within 6 h of administration of study drug	0 (0)	0 (0)	1 (20)	1 (4)
5. Visual hallucinations	0 (0)	1(10)	0 (0)	1 (4)
6. >20% decrease in systolic blood pressure (SBP) from baseline and SBP < 90 mmHg (symptomatic hypotension) within 6 h of administration of study drug	0 (0)	0 (0)	1 (20)	1 (4)
7. A serious adverse event with causality of at least "possibly" related to the study drug	0 (0)	0 (0)	0 (0)	0 (0)
8. >20% increase in SBP or diastolic blood pressure (DBP) from baseline and SBP > 170 or DBP > 95 (symptomatic hypertension) within 6 h of administration of study drug	0 (0)	0 (0)	0 (0)	0 (0)

^aParticipants may meet criteria for more than one DLT.

^bThe DLT meeting this definition in all participants was vomiting.

Figure 3.

Assessment of cardiovascular effects of a single dose of oral harmine HCl in healthy participants. (a) Heart rate and (b) blood pressure were monitored at repeated intervals after administration of harmine. No significant hypertensive or hypotensive effects were observed, nor was any change in heart rate observed after administration of harmine HCl, even at the highest doses. Time “0” is baseline prior to dosing. All data are presented as the mean ± SEM. Three subjects receiving 300 mg and the single subject receiving 500 mg are grouped as “>200 mg.”

No serious AEs were reported at any dose, and no participant withdrew from the study due to an AE. However, non-serious AEs were reported in 48% of participants (Figure 2; Tables 2–4). All were mild to moderate in severity. As with DLTs, the most frequently reported AEs were gastrointestinal (nausea and vomiting) and CNS-related (drowsiness, dizziness, impaired concentration, and feeling intoxicated/high). AEs were typically mild, with onset typically within 30–60 min of dosing and lasting less than 90 min overall. At the 100 mg dose, only 2 of 10 participants experienced AEs, and in each case they were mild. Only 2 of 15 (13%) participants experienced mild AEs at doses below 2.7 mg/kg. Overall, AEs were more common in women than men (9 vs 3), which may have reflected the lower weight in female versus male participants. There were no apparent differences among ethnic groups in the frequency of AEs or DLTs, although the study was underpowered to draw firm conclusions on this parameter.

Table 4.

PRISE reporting.^a

Number (%)	100 mg	200 mg	>200 mg	Overall
	n = 10	n = 10	n = 5	n = 25
Decreased energy	3 (30)	8 (80)	5 (100)	16 (64)
Fatigue	4 (40)	9 (90)	3 (60)	16 (64)
Dizziness	3 (30)	5 (50)	5 (100)	13 (52)
Nausea/vomiting	0 (0)	5 (50)	5 (100)	10 (40)
Dizziness on standing	1 (10)	3 (30)	5 (100)	9 (36)
Poor concentration	2 (20)	3 (30)	3 (60)	8 (32)
Headache	2 (20)	4 (40)	1 (20)	7 (28)
Poor coordination	1 (10)	3 (30)	2 (40)	6 (24)
Anxiety	1 (10)	1 (10)	2 (40)	4 (16)
Restlessness	2 (20)	1 (10)	1 (20)	4 (16)
Palpitations	1 (10)	2 (20)	1 (20)	4 (16)
Dry mouth	2 (20)	1 (10)	1 (20)	4 (16)
Sleeping too much	0 (0)	2 (20)	1 (20)	3 (12)
Tremors	0 (0)	3 (30)	0 (0)	3 (12)
Blurred vision	0 (0)	3 (30)	0 (0)	3 (12)
Ringing in ears	0 (0)	0 (0)	2 (40)	2 (8)
Increased perspiration	0 (0)	1 (10)	1 (20)	2 (8)
General malaise	0 (0)	1 (10)	0 (0)	1 (4)
Difficulty sleeping	0 (0)	1 (10)	0 (0)	1 (4)
Constipation	1 (10)	0 (0)	0 (0)	1 (4)

The PRISE occurring in each dose group.

^aOnly new onset or worsening symptoms were included. Participants could report more than one AE.

No participants at any dose reported frank visual hallucinations, although one participant who received 300 mg (5.6 mg/kg) reported sensory disturbances, including confusion about whether her socks were visually on her feet or not and auditory hallucinations of her “voice moving faster” and possible buzzing, and another participant who received 200 mg (3.8 mg/kg) reported that the outline of objects appeared fuzzy like a television screen with static. Three participants reported experiencing giddiness, two at 300 mg (5.6 and 3.1 mg/kg) and one at 200 mg (3.8 mg/kg), respectively, although the study team only noted outward signs of this behavioral effect in one participant receiving 300 mg (3.1 mg/kg). One participant who received 100 mg (1.5 mg/kg) reported a feeling of “heaviness” but not “sinking.”

With respect to cardiovascular outcomes, mean systolic and DBP and heart rate were normal and unaffected (Figure 3). No participant became hypertensive, but one participant in the 300 mg group (5.6 mg/kg) became hypotensive (lowest BP = 85/48) and heart rate increased more than 20% from baseline (max heart rate = 91), attributed to vomiting, while another participant receiving the 300 mg dose (3.1 mg/kg) experienced mild transient (10 min) hypotension (BP 96/53) attributed to a vasovagal reaction during a blood draw. Both participants required intravenous fluids. Two participants in the 300 mg group (4 mg/kg, 5.6 mg/kg) reported subjective sensation of palpitations prior to vomiting, with no objective changes in blood pressure or heart rate or changes on EKG.

No participant had a seizure or severe depression of mental status, although several participants experienced drowsiness and/or impaired concentration (see Assessment of Psychoactive Effects below). No subject became hypoglycemic or hyperinsulinemic, and none displayed liver function test abnormalities (Supplemental Figure 1). No other safety issues emerged in the study.

Assessment of psychoactive effects

Given the well-documented psychoactive experiences associated with ingestion of Ayahuasca (Aicher et al., 2023; Brierley and Davidson, 2012; Callaway et al., 1999; Djamshidian et al., 2016; Dornbierer et al., 2023; Dos Santos et al., 2021; Palhano-Fontes et al., 2019; Riba et al., 2001, 2003; Rocha et al., 2021; Rossi et al., 2023; Sanches et al., 2016), and limited human data on pure harmine consumption (Dornbierer et al., 2023; Pennes and Hoch, 1957; Slotkin et al., 1970), we designed this study to characterize in detail the psychoactive effects of harmine as a secondary endpoint. We utilized the VAS to assess several subjective domains of physical and mental states (0 = not at all, 10 = extremely), the PSS to assess stress associated with the experience, the C-SSRS to evaluate suicidality, the BPRS to capture acute behavioral changes, the PRISE to identify and evaluate tolerability of AEs, and the POMS-Bi to assess transient mood states, including those induced by substances. These were repeated at frequent intervals as shown in Figure 5.

Figure 5.

POMS-Bi in healthy participants given a single oral dose of harmine HCl. The POMS-Bi was administered at repeated intervals after dose administration to assess drug-induced changes in six mood states. Scores range from 0 to 36, with higher scores being associated with positive mood states on each of the six subscales. Time “0” is baseline prior to dosing. (a) Calm-anxious subscale, (b) Confident-unsure subscale, (c) Elated-depressed subscale, (d) Energetic-tired subscale, (e) Agreeable-hostile subscale, (f) Clear-confused subscale. All data are presented as the mean ± SEM. Three subjects receiving 300 mg and the single subject receiving 500 mg are grouped as “>200 mg.”

As mentioned in the DLT section above, no participants reported or were observed to experience frank hallucinations or other significant psychoactive effects associated with ingesting Ayahuasca. No gross changes in mental status exam were observed in any patient, although notable drowsiness was observed in one participant receiving 100 mg harmine HCl (1.3 mg/kg) and three participants receiving 200 mg (3.8, 3.4, and 2.9 mg/kg; Table 3). Participants remained easily arousable with peak drowsiness at 2 h after harmine HCl administration and resolving within 5 h. Drowsiness associated with administration was also captured on the VAS (Figure 3), POMS-Bi (energetic-tired, Figure 5), and PRISE (fatigue and sleeping too much, Table 4). Three participants receiving 200 mg (3.8, 3.4, and 2.9 mg/kg) reported impaired concentration. Likewise, impaired concentration was captured on the POMS-Bi (clear-confused, Figure 4), and PRISE (Table 4), peaking at 2 h after administration and resolving by 6 h. The impairment was mild and did not prevent participants from following directions from the study team or completing assessments.

Table 3.

Adverse events.

Number (%)	100 mg	200 mg	>200 mg	Overall
	n = 10	n = 10	n = 5	n = 25
Any AEa	2 (20)	5 (50)	5 (100)	12 (48)
Death	0 (0)	0 (0)	0 (0)	0 (0)
AE severity
Serious	0 (0)	0 (0)	0 (0)	0 (0)
Moderate	0 (0)	4 (40)	3 (60)	7 (28)
Mild	2 (20)	3 (30)	2 (40)	7 (28)
Vomiting/emesis	0 (0)	3 (30)	5 (100)	8 (32)
Dizziness	2 (20)	1 (10)	1 (20)	4 (16)
Impaired concentration/confusion	1 (10)	3 (20)	0 (0)	4 (16)
Drowsiness	0 (0)	3 (30)	0 (0)	3 (12)
Visual illusion	0 (0)	1 (10)	0 (0)	1 (4)
Auditory hallucination	0 (0)	0 (0)	1 (20)	1 (4)
Giddiness	0 (0)	0 (0)	1 (20)	1 (4)
Hypotension	0 (0)	0 (0)	1 (20)	1 (4)
Vasovagal	0 (0)	0 (0)	1 (20)	1 (4)
Tingling	0 (0)	1 (10)	0 (0)	1 (4)
Numbness	1 (10)	0 (0)	0 (0)	1 (4)
Fatigue	1 (10)	0 (0)	0 (0)	1 (4)
Nausea w/o vomiting	0 (0)	1 (10)	0 (0)	1 (4)
Heaviness	1 (10)	0 (0)	0 (0)	1 (4)

AEs occurring in each dose group. AEs were coded by Medical Dictionary for Regulatory Activities (MedDRA) Version 21.0 Preferred Terms (International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH)).

^aThe number of participants experiencing any AE. Some participants experienced multiple AEs.

Symptoms of drowsiness and impaired concentration were accompanied by a dose-dependent report of feeling “high” on the VAS, with peak effect 1 h after harmine HCl administration and resolved by 3 h (Figure 4). There appeared to be little relationship to dose administration and reported symptoms of anxiety, control, depression, happiness, vividness of image or excitement on the VAS (Figure 4) or on calm-anxious, elated-depressed, agreeable-hostile domains on the POMS-Bi (Figure 5); however, one participant receiving 300 mg (3.1 mg/kg) experienced transient (5 min) giddiness with outbursts of laughter about 30 min after administration of harmine HCl. Despite this minimal effect on mood, none of the participants who experienced effects from harmine reported it to be a pleasant experience at the exit visit.

On the BPRS, two participants receiving the 200 mg dose were noted to have altered behavior compared to screening, with one participant (3.6 mg/kg) being anxious before administration of harmine HCl and the other (3.8 mg/kg) having slower speech 2 h after administration of the dose (Supplemental Figure 2). No effect of harmine HCl was observed on the PSS or C-SSRS scales (not shown). No participants required medication for psychiatric reasons during the course of harmine treatment.

Discussion

Despite its long and extensive ethnobotanical history, there is currently little clarity on the important issue of harmine’s human psychoactivity, safety, and tolerability, since studies using pure, high-quality harmine are limited in healthy participants (Dornbierer et al., 2023). Further, the available studies of harmine alone tend to utilize a narrow range of doses, so it is unknown what the upper tolerable limit is. Because Ayahuasca and Syrian rue have psychoactive properties, and because of the ubiquity of DYRK1A expression in the body, it has been assumed by some that harmine will have psychoactive and other off-target AEs. Therefore, one primary goal of the study was to clarify whether harmine alone, at comparable doses to those used in Ayahuasca or higher, causes hallucinations, as assessed by a team of experienced psychiatric investigators.

In this Phase 1 single-rising dose trial in healthy participants receiving oral harmine HCl, we found that overall, harmine HCl was well-tolerated. We determined that the MTD is between 100 and 200 mg and that the MTD is weight-based, with an apparent cutoff of 2.7 mg/kg. There were no serious AEs at any dose studied. The predominant non-serious DLT observed at doses higher than 2.7 mg/kg was nausea and vomiting. The most common non-serious AEs at all doses other than nausea and vomiting, were drowsiness, followed in frequency by impaired concentration and dizziness. No cardiovascular effects were observed. Participants tolerated harmine HCl with few psychoactive effects, even at the higher doses.

The CRM model employed here was designed to include up to 40 participants at doses of up to 1200 mg. We selected our initial dose range for this study based on prior human studies with Ayahuasca (Brito-da-Costa et al., 2020; Callaway et al., 1999; Dos Santos et al., 2021; Palhano-Fontes et al., 2019; Riba et al., 2001, 2003; Rocha et al., 2021; Sanches et al., 2016) or earlier, less rigorous clinical studies with harmine (Pennes and Hoch, 1957; Slotkin et al., 1970). We used the CRM for dose escalation rather than the conventional “3 + 3 + 3…” dose-escalation model because it offers the potential to reach and define the MTD more rapidly (O’Quigley et al., 1990). Thus, as shown in Table 1, it quickly became clear that doses of 100 and 200 mg were reasonably well-tolerated, whereas doses of 300 and 500 were poorly tolerated. Although the CRM algorithm may have recommended doses of 300 mg, after the fourth participant received 300 mg, we determined that from a medical standpoint, no further participants should be dosed above 200 mg, due to the predicted 25% risk of AE/DLT and/or discomfort in participants receiving doses greater than 200 mg. Thus, the study was terminated when 10 participants at 100 and 200 mg doses had been enrolled. All 10 participants who received 100 mg tolerated harmine HCl remained well, with no DLTs and only 2 had mild AEs of dizziness, drowsiness, and/or impaired concentration. At the 200 mg dose, 6 of 10 experienced DLTs, with AEs that were considered mild or moderate, typically nausea, vomiting, dizziness, and drowsiness. At 300 mg dose, all four participants experienced moderate AEs and two had DLTs. Only one participant received 500 mg, and experienced nausea and vomiting, considered a DLT, and an AE of drowsiness. No participant experienced a severe AE.

As the study progressed, we appreciated that although our protocol required normal participants with a normal BMI (19–30 kg/m²), some of the participants were thin and others were at the high end of the normal BMI range. Retrospective analysis of the AEs and DLTs on a mg/kg body weight basis revealed that the weight-based cutoff for AEs and DLTs appeared to be 2.7 mg/kg, with no participant below that cutoff experiencing a DLT and only two experiencing mild AEs (Figure 2). Among the participants in the 200 mg group with a DLT, all were women with smaller body mass and received doses above 2.7 mg/kg.

The major dose-limiting AEs of harmine HCl administration were nausea and vomiting. The mechanisms for nausea and vomiting are uncertain and may include direct effects of harmine on the CNS and/or on the intestine via its MAO-like or serotoninergic properties. Nausea and vomiting are common AEs among Ayahuasca users (Brito-da-Costa et al., 2020), and our data suggest that harmine contributes to vomiting associated with its use. In contrast, more recent reports with Ayahuasca note doses of harmine (approximately 25–130 mg, 0.85–1.86 mg/kg) that are lower than would be expected to be associated with vomiting (Aicher et al., 2023; Dornbierer et al., 2023; Dos Santos et al., 2021; Palhano-Fontes et al., 2019; Rocha et al., 2021; Rossi et al., 2023; Sanches et al., 2016) based on our findings here. However, if one considers the total dose of β-carbolines (180–230 mg, 2.56–3.3 mg/kg) ingested, the reported percentage of participants that experienced vomiting (33%–57%) is similar to our findings. Notably, a recent study used an orally disintegrating form of harmine, which allows for buccal absorption and bypasses the gastrointenstinal tract (Dornbierer et al., 2023). In that study, no vomiting was observed, regardless of harmine or DMT dose. It should be noted that both the DMT and harmine used in the study were pure pharmaceutical grade, rather than the traditional decoction of Ayahuasca, which contains numerous alkaloids. In contrast, diarrhea, which is also common following Ayahuasca use, was not reported in any of the 25 study participants, all of whom returned for their 24-h follow-up visit, suggesting that other alkaloids present in Ayahuasca are responsible for diarrhea.

Cardiovascular AEs were rare, and none were serious. Mean systolic or DBP and heart rate were normal. Importantly, 2 of 25 participants did experience brief hypotension: one experienced a vasovagal reaction during a blood draw and one during an episode of vomiting. Both were transient and self-limited and not attributed to the direct effects of harmine itself.

The current report supports the premise that harmine prevents the degradation of DMT in Ayahuasca to enable hallucinations. Specifically, we conclude that harmine per se does not have intrinsic psychoactive hallucinatory activity at doses comparable to those commonly used in religious, spiritual, or recreational settings, where ~252 mg of harmine is ingested as described by Callaway et al. (1999), and at doses of <2.7 mg/kg in the current study. At doses >2.7 mg/kg, some participants were noted to have mild alterations in visual and tactile perception, but no frank hallucinations were reported, nor did the team of experienced psychiatrists note outward appearances or behavior consistent with hallucination. While reports of intoxication with larger amounts of β-carbolines include frank hallucinations (Moshiri et al., 2013), it is likely that there is an additive effect of harmine, harmaline, and THH. Although there was limited psychoactivity in terms of hallucinations, drowsiness, and a sense of being intoxicated were reported by many participants, even those receiving the lowest dose. The mechanisms for drowsiness resemble the effects observed for other MAOI-class drugs (Rapaport, 2007) and may likely act via this mechanism in the CNS.

Overall, these studies make the following points: (1) harmine HCl can be safely ingested orally by healthy humans in a single dose of up to 200 mg without serious AEs or toxicities; (2) future studies with 50 mg dose increments, targeting the 100–300 mg range more finely, would more precisely define the MTD; (3) a weight-based mg/kg cutoff of 2.7 mg/kg will likely prove more useful in future studies rather than an absolute mg/person cutoff; (4) the most common AE of oral harmine HCl is nausea and vomiting; (5) harmine does not appear to have intrinsic hallucinatory effects at doses up to 300 mg; (6) adverse cardiovascular effects are unlikely at doses below 300 mg or 2.7 mg/kg; (7) diarrhea is not a common harmine AE; (8) multi-day studies in animals and humans are required to assess longer-term safety and tolerability; (9) future studies should assess absorption, distribution, metabolism, excretion, and pharmacokinetics of pure harmine HCl in humans and should consider route of administration: intravenous versus oral versus oral disintegrating tablet.

Strengths and limitations

This is the first study to our knowledge to assess the safety and tolerability of oral administration of GMP harmine HCl in healthy participants across a range of doses (100–500 mg). In line with prior reports (Callaway et al., 1999), we found that the MTD of harmine lies between 100 and 200 mg (with a more accurate and useful MTD being weight-based at 2.7 mg/kg), and the primary DLT being vomiting.

There are several limitations that should be considered. Our observational study design was not powered to detect statistically significant differences between groups, nor did we include a placebo group. Although 100 and 200 mg doses of harmine HCl were generally well-tolerated, there was a wide range of responses, suggesting individual differences in pharmacokinetics that warrant further investigation. This is consistent with a recent study of “pharmahuasca,” where there was considerable individual variability in harmine absorption and metabolism (Aicher et al., 2023).

Additionally, although a secondary endpoint was assessment of the psychoactive properties of harmine alone, we did not include a scale that directly assesses psychedelic responses other than the BPRS. Future studies should employ a scale such as the Clinician-Administered Dissociative States Scale (Bremner et al., 1998) or the Hallucinogenic Rating Scale (Strassman et al., 1994) to more accurately capture the acute psychoactive properties of harmine and distinguish them from those of DMT. Based on our findings here, it is likely that significant psychoactive effects may occur at doses above those associated with vomiting (>300 mg), thus the study should be designed to anticipate such effects of oral harmine administration.

Conclusions

We conclude that a single oral dose of harmine HCl below 200 mg or 2.7 mg/kg is generally well-tolerated in healthy young adults and that harmine itself is not the principal psychoactive component of Ayahuasca. Harmine, as part of the total β-carboline and alkaloid load in Ayahuasca, likely contributes to the vomiting, and perhaps some of the sedation, associated with Ayahuasca ingestion, but does not likely contribute to diarrhea. Future studies with smaller dose increments, longer-term dosing, and determination of pharmacokinetic profiles are warranted to fully understand the safety and tolerability of oral dosing of harmine HCl.