Abstract
Alcohol use disorder (AUD) is a highly prevalent disorder with limited therapeutic options. The central amygdala (CeA) is a critical brain region as dysregulation within the CeA and the corticotropin-releasing factor (CRF) system are associated with AUD pathology. CeA CRF1 receptors regulate alcohol drinking and have served as a therapeutic target in alcohol treatment. One emerging potential therapeutic for AUD is psilocybin. Psilocybin has been shown to decrease drinking in some clinical studies; however, the effects are variable and the underlying mechanisms are poorly understood. Psilocybin engages many brain regions, including the CeA, and may produce therapeutic effects on drinking through interactions with CeA CRF1 neurons. The current study explores the effects of psilocin, the active metabolite of psilocybin, on voluntary ethanol drinking and CeA CRF1 activity to understand potential mechanisms underlying the therapeutic effects of psilocin. Psilocin acutely decreased ethanol consumption in mice exposed to two different models of chronic ethanol exposure without producing changes in locomotor behavior. Psilocin increased CeA activation and decreased relative CRF1 activation in CeA subregions from ethanol-naive female CRF1:GFP mice. These results were also observed in chronic ethanol-exposed mice at 24 and 72 h withdrawal timepoints. Psilocin increased corticosterone at 24 h withdrawal but not at 72 h withdrawal. Collectively, these results demonstrate that psilocin engages CeA circuitry and decreases relative CRF1 activation, in parallel with acute reductions in drinking. These results contribute to our understanding of the mechanisms underlying the actions of psilocin and inform the interpretation of therapeutic effects in clinical studies.
Significance Statement
Alcohol is one of the most commonly used substances that dysregulate brain regions involved in emotional processing and stress. An important regulator of the stress response is the neuropeptide known as corticotropin-releasing factor (CRF). Alcohol can dysregulate brain regions through the engagement of corticotropin-releasing factor receptor 1 (CRF1)-containing neurons and thus promote continued alcohol use. Although alcohol use disorder (AUD) is a highly prevalent condition, few treatment options are available. Psilocybin, a psychedelic prodrug that is broken down into the active metabolite, psilocin, has emerged as a potential treatment for AUD in recent studies. The current study explores the effects of psilocin on alcohol drinking and central amygdala CRF1-containing neurons in female mice to better understand potential therapeutic mechanisms.
Introduction
Alcohol use disorder (AUD) is a complex condition characterized by an inability to limit drinking despite negative consequences (Flores-Bonilla and Richardson, 2020). AUD remains a major public health concern as a leading cause of death that lacks widely effective treatment options (Pilar et al., 2020). Alcohol alters neurotransmission in reward- and stress-related brain regions that may contribute to AUD (Agoglia et al., 2020a). One site of alcohol-associated plasticity and dysregulation is the central amygdala (CeA). The CeA coordinates emotional processing with implications in stress, anxiety, and AUD (Agoglia et al., 2020a). The CeA is implicated in behaviors associated with the withdrawal/negative affect stage of AUD (Gilpin et al., 2015; Centanni et al., 2019), and alcohol dysregulates CeA synaptic transmission after acute and chronic exposure (Roberto et al., 2003, 2004a,b; Agoglia and Herman, 2018). The CeA contains subnuclei including medial (CeAm) and lateral (CeAl) subregions that are implicated in fear conditioning (Ciocchi et al., 2010; Li et al., 2013) and AUD (Melkumyan and Silberman, 2022). The CeAl projects to the CeAm mainly through local GABAergic circuits (Melkumyan and Silberman, 2022). Previous work reports differential effects of alcohol on CeA subregions, contributing to the idea that CeA microcircuitry is involved in AUD (Melkumyan and Silberman, 2022).
The CeA contains many cell types differentially impacted by stress and alcohol, including corticotropin-releasing factor receptor 1 (CRF1) neurons. CRF is a neuropeptide that regulates the stress response and modulates CeA synaptic transmission (Nie et al., 2009; Roberto et al., 2010; Agoglia et al., 2022). The CRF1 receptor has a functional role in alcohol dependence and is a focus for AUD therapeutic strategies (Nie et al., 2009; Roberto et al., 2010; Pomrenze et al., 2017; Agoglia et al., 2022). Previous studies in rodents report increased amygdala CRF and CRF1 expression after chronic ethanol and CRF1 antagonists reversed alcohol-induced CeA GABA plasticity and reduced drinking in mice (Agoglia and Herman, 2018). Although preclinical studies have demonstrated a functional role for the CRF1 receptor in alcohol self-administration, clinical studies using CRF1 antagonists in AUD patients have been unsuccessful (Kwako et al., 2015; Schwandt et al., 2016).
Despite its prevalence, AUD treatments are limited and more effective therapeutics are needed. One potential treatment is psilocybin, a prodrug that is dephosphorylated into the active metabolite, psilocin. Preclinical studies report variable impacts of psilocybin on alcohol drinking with some studies showing no effect on drinking or relapse (Benvenuti et al., 2023) while others show sex-specific reductions in drinking and ethanol preference (Alper et al., 2023). Clinical studies investigating the effects of psilocybin report increased abstinence in participants with AUD that were administered psilocybin paired with psychotherapy (Bogenschutz et al., 2015), and similar results were seen in a subsequent double-blind, randomized clinical trial where AUD patients received two doses of psilocybin alongside psychotherapy (Bogenschutz et al., 2022). However, a more recent study found reduced alcohol craving, but limited long-term effects of a single dose of psilocybin (Rieser et al., 2025). Although clinical research provides some evidence of psilocybin reducing AUD symptoms, the neurobiological mechanisms underlying these effects are unclear. Psilocybin/psilocin can engage many brain regions, including the CeA, leading to region-specific neuroplasticity (Shao et al., 2021; Davoudian et al., 2023; Effinger et al., 2023; Aboharb et al., 2025). We examined the effects of psilocin on CeA activation and on CRF1 neuron activation in the context of overall CeA activity in naive or chronic ethanol-exposed female CRF1:GFP mice. Female mice were used for this study since they typically drink more than males (Salazar and Centanni, 2024), allowing the examination of psilocin effects in a high-drinking population. We also examined the effects of psilocin on drinking in female mice with a history of chronic ethanol vapor exposure and voluntary drinking or chronic voluntary drinking alone. An improved understanding of how psilocin impacts CeA CRF1 activity following chronic ethanol exposure increases our understanding of the mechanisms underlying the therapeutic potential of psilocin for AUD.
Materials and Methods
Animals
Adult female CRF1:GFP mice (n = 90, 2–8 months, 18–30 g) were bred and housed in a temperature- and humidity-controlled 12 h reverse light/dark (7 A.M. lights off) facility with ad libitum food (RMH-3000) and water access. Mice in the ethanol-naive cohort (n = 10) or the locomotor cohort (n = 20) were group housed prior to treatment. Mice in the ethanol drinking cohorts (n = 60) were individually housed for two-bottle choice (2BC) drinking. All experimental procedures were approved by the University of North Carolina Institutional Animal Care and Use Committee.
Drugs
Psilocin (Cayman Chemical) was dissolved in 2% glacial acetic acid and brought up to a 1 mg/ml dilution in 0.9% saline. Mice were subcutaneously injected with 2 ml/kg of either 2 mg/kg psilocin or vehicle (2% glacial acetic acid in 0.9% saline).
Chronic intermittent ethanol vapor-two-bottle choice drinking
Mice completed a modified chronic intermittent ethanol vapor-two-bottle choice (CIE-2BC) paradigm adapted from a study that used an established model of ethanol dependence (Smith et al., 2020). Mice underwent 2 weeks of baseline 2BC drinking followed by a week of CIE (16 h ethanol vapor, Mon–Thurs) and alternated between 2BC and CIE weeks for another 4 weeks. Mice were individually housed for at least 5 d prior to the start of drinking for habituation. Mice were given 2 h access (Mon–Fri) to two bottles, one containing 20% ethanol solution made from 95% ethanol (Pharmco-Aaper) and one containing water. One hour intake was recorded to compare any immediate consumption to the total 2 h duration and determine whether mice were front-loading during ethanol access. Alcohol intake was determined by visual inspection of ethanol volume according to meniscus on serological pipettes. Bottle placements were alternated to prevent side preference. Before being placed in the CIE chambers (LJARI), mice were given a loading dose (intraperitoneal injection, 10 ml/kg) of 68 mg/kg pyrazole (Sigma-Aldrich), 0.8–2.0 g/kg ethanol, in 0.9% saline to maintain intoxication level during ethanol vapor exposure. AIR mice received a control loading dose (68 mg/kg pyrazole in 0.9% saline). Blood was collected at least twice a week during CIE via a tail nick to monitor blood alcohol concentrations (BACs), which were measured using an Analox AM1 Analyzer. Chamber flow rate and loading doses were monitored and adjusted to maintain BACs between 100 and 300 mg/dl throughout the study. If the BACs measured showed signs of tolerance (e.g., a significant decrease from previous measurements using the same parameters), flow rate was decreased and the loading dose (0.8–2 g/kg ethanol) was adjusted to increase BACs and maintain target levels in following sessions. For individual mice, if the BAC measurement showed signs of intolerance to a change in parameters, such as a reading of >350 mg/dl, the ethanol in the loading dose would be removed for the following session to avoid severe intoxication.
Locomotor testing
Percent of time in center (thigmotaxis) and spontaneous locomotor activity were recorded in a computer-controlled open field chamber (41 cm × 41 cm × 30 cm; Versa Max, AccuScan Instruments). In this chamber, multiple values were measured including total distance traveled in center, total distance traveled, thigmotaxis, and time course data of locomotion. The x–y–z spatial positioning was recorded through three sets of infrared photo beams placed around the exterior of the acrylic box. Mice were subcutaneously injected with either vehicle or psilocin (2 mg/kg) before habituating in the room with the machines on for 90 min. Once habituation was complete, the mice were placed into the boxes and allowed to move freely for 10 min with overhead lights and fans running to provide white noise. Boxes were wiped clean between subjects and cleaned with 2% acetic acid in water at the conclusion of all trials.
Corticosterone assay and analysis
Blood was collected using 1.5 ml heparinized microcentrifuge tubes (Sarstedt AG & Co. KG), and each sample was immediately centrifuged. Serum was collected in 0.2 ml PCR tubes (Eppendorf) and stored at −20°C. All serum samples were prepared and analyzed using the CORT ELISA kit (Arbor Assays, catalog #K014-H5) following the protocol provided.
Immunohistochemistry
Mice were anesthetized by isoflurane and perfused with 1× phosphate-buffered saline (PBS) then 4% paraformaldehyde (PFA) in PBS. Brains were postfixed in 4% PFA before being transferred to 30% sucrose in PBS at 4°C. Using a microtome (HM450, Thermo Fisher), brains were cut into 40 μm sections and stored in 0.01% sodium azide/PBS at 4°C. Slices containing the CeA were washed in PBS (two times for 10 min), incubated in 50% methanol/PBS (30 min), washed in 3% hydrogen peroxide (5 min), and then incubated in blocking solution (0.3% Triton X-100/PBS; Thermo Fisher), 1% bovine serum albumin (BSA; Sigma) for an hour. Slices were incubated in rabbit anti-cFos primary antibody (1:3,000, Millipore Sigma; ABE457) in blocking solution for 48 h at 4°C. Slices were then washed with Tris, NaCl, Triton X-100 (TNT) buffer (10 min) and washed in Tris, NaCl, blocking reagent (TNB; PerkinElmer) for 30 min, followed by a 30 min incubation in horseradish peroxidase (HRP; 1:200, Abcam ab6721), and four times for 5 min washes in TNT. Slices then incubated in Cy3 (1:50) TSA amplification diluent (Akoya Biosciences, NEL744001KT) for 10 min at room temperature to amplify the fluorescence signal. Slices were washed in TNT buffer (two times for 10 min), PBS (three times for 10 min), 0.3% Triton X-100/PBS (30 min), and PBS (two times for 10 min) before being blocked in 0.3% Triton/10% normal donkey serum/1%w/v BSA/PBS and left overnight in primary antibody (chicken anti-GFP). Slices were washed in PBS (three times for 10 min) before being shaken in the secondary antibody (donkey anti-chicken). After two final 10 min PBS washes, slices were mounted onto slides using Vectashield DAPI with hard-set (Vector Labs; H1500-10). Fluorescence was imaged using a Keyence BZ-X800 microscope: Cy3 (c-Fos) 555 nm excitation, 570 nm emission, green fluorescent protein (GFP, CRFR1) 488 nm excitation, 510 nm emission, DAPI 350 nm excitation, 461 nm emission. Regions of interest (ROIs) were confirmed at a lower magnification (4×) and imaged at higher magnification (20×) for quantification. ROIs were outlined by an investigator blinded to experimental group. Quantification of cell counts was performed by manual detection using ImageJ software by an investigator blinded to experimental group.
Statistical analysis
Statistical analysis and data visualization were performed using Prism 10.0 (GraphPad). Data were analyzed and compared using either unpaired t test or two-way ANOVA with Sidak's multiple comparisons with repeated measures as appropriate. All data are expressed as mean ± SEM with p < 0.05 as the threshold for statistical significance.
Results
Impact of psilocin on drinking following chronic intermittent ethanol vapor exposure and ethanol drinking
Mice were administered 2 mg/kg psilocin or vehicle 72 h into withdrawal from 7 weeks of CIE-2BC to examine the impact of psilocin on drinking behavior (Fig. 1A). Ethanol intake remained relatively stable throughout the drinking paradigm, including 1 and 2 h timepoints (Fig. 1B). Mice maintained high levels of intoxication during each week of ethanol vapor exposure (Fig. 1C). Psilocin decreased ethanol consumption in the session immediately following injection with a main effect of treatment and time (two-way ANOVA, treatment: ***p = 0.0003, F(1,18) = 19.33, time: #p < 0.0001, F(1,18) = 30.82, treatment × time: *p = 0.0248, F(1,18) = 5.994; Fig. 1D). Water intake was unchanged in the first hour however the psilocin group increased water intake in the second hour (two-way ANOVA, *p = 0.0217, F(1,18) = 4.970; Fig. 1E). The session following injection remained the only time when psilocin decreased drinking at the 1 h timepoint (two-way ANOVA, time: #p < 0.0001, F(6,108) = 15.32, treatment × time: **p = 0.0081, F(6,108) = 3.074, multiple comparisons: 8.1 Veh vs Psi: **p = 0.0072; Fig. 1F). Similarly at the 2 h timepoint, psilocin only decreased ethanol consumption during the drinking session following injection (two-way ANOVA, time: #p < 0.0001, F(6,108) = 9.742, treatment × time: *p = 0.0284, F(6,108) = 2.465, multiple comparisons: 8.1 Veh vs Psi: ***p = 0.0007; Fig. 1G). There was no difference in water intake at either 1 h (two-way ANOVA: p = 0.8586, F(6,108) = 0.4284; Fig. 1H) or 2 h timepoints (two-way ANOVA: p = 0.3319, F(6,107) = 1.163; Fig. 1I). There was no difference in corticosterone levels 1 week after vehicle or psilocin injection (unpaired t test: 0.4078, t = 0.8476, df = 18; Fig. S2A). There was no difference in CeA cFos expression, including medial and lateral regions, 1 week after vehicle or psilocin injection (Fig. S2B–E). There was also no difference in the proportion of CRF1 neurons in the active CeA population, including medial and lateral regions, 1 week after vehicle or psilocin injection (Fig. S2F–I).
A, Experimental timeline outlining treatment 72 h following the last CIE session relative to postinjection 2BC drinking. B, 1 and 2 h ethanol intake for each week of 2BC drinking, n = 20. C, Blood alcohol concentrations throughout each week of CIE, n = 20. D, 1 and 2 h ethanol intake in psilocin and vehicle exposed mice during the first 2BC drinking session following injection, Vehicle: n = 10; Psilocin: n = 10. Two-way ANOVA, treatment: ***p = 0.0003, F(1,18) = 19.33, time: #p < 0.0001, F(1,18) = 30.82, treatment × time: *p = 0.0248, F(1,18) = 5.994. E, 1 and 2 h water intake in psilocin and vehicle groups during the first 2BC drinking session following injection, Vehicle: n = 10; Psilocin: n = 10. Two-way ANOVA, *p = 0.0217, F(1,18) = 4.970. F, 1 h ethanol intake including the last 2BC session before treatment, during treatment, and for five sessions following injection, Vehicle: n = 10; Psilocin: n = 10. Two-way ANOVA, time: #p < 0.0001, F(6,108) = 15.32, treatment × time: **p = 0.0081, F(6,108) = 3.074, multiple comparisons: 8.1 Veh versus Psi: **p = 0.0072. G, 2 h ethanol intake including the last 2BC session before treatment, during treatment, and for 5 sessions following injection, Vehicle: n = 10; Psilocin: n = 10. Two-way ANOVA, time: #p < 0.0001, F(6,108) = 9.742, treatment × time: *p = 0.0284, F(6,108) = 9.742, multiple comparisons: 8.1 Veh versus Psi: ***p = 0.0007. H, 1 h water intake including the last 2BC session before treatment, during treatment, and for five sessions following injection, Vehicle: n = 10; Psilocin: n = 10. Two-way ANOVA: p = 0.8586, F(6,108) = 0.4284. I, 2 h water intake including the last 2BC session before treatment, during treatment, and for five sessions following injection, Vehicle: n = 10. Psilocin: n = 10. Two-way ANOVA: p = 0.3319, F(6,107) = 1.163.
A separate cohort of mice exposed to chronic intermittent ethanol vapor and drinking (CIE-2BC) or air and drinking (AIR-2BC) were injected with 2 mg/kg psilocin 2 weeks after vehicle injection (Fig. 2A). A within-subject design was used for vehicle and psilocin injections rather than splitting the CIE-2BC and AIR-2BC groups into treatment groups yielding five mice per treatment instead of 10. This design allowed for drinking cohorts (Figs. 1, 2) to have the same sample sizes for each treatment group while maintaining consistent psilocin injection timelines (7 weeks of CIE-2BC or AIR-2BC). Baseline levels of ethanol intake were established at 1 and 2 h timepoints (Fig. 2B). Mice were split into AIR and CIE groups based on average baseline ethanol intake (unpaired t test: p = 0.8758, t = 0.1586, df = 18; Fig. 2C). Ethanol intake was monitored in AIR and CIE groups to examine any group-specific changes in consumption (Fig. 2D). BACs were monitored during CIE exposure to ensure blood alcohol levels remained high and stable across weeks (Fig. 2E). One and two hours ethanol intake was reduced in CIE mice immediately following psilocin injection (two-way ANOVA, treatment: *p = 0.0313, F(1,15) = 5.643; Fig. 2F). One and two hours ethanol intake was also reduced in AIR mice immediately following psilocin injection (two-way ANOVA, treatment: **p = 0.0013, F(1,18) = 14.34, time: #p < 0.0001, F(11,17) = 45.32; Fig. 2G). One hour (two-way ANOVA, time: #p < 0.0001, F(4.193,73.37) = 14.22; Fig. 2H) and 2 h ethanol intake (two-way ANOVA, time: #p < 0.0001, F(4.845,84.79) = 7.803; Fig. 2I) had significant effects of time across drinking sessions in AIR and CIE mice, with no significant effects of treatment. Water intake was unaffected at 1 h (two-way ANOVA: p = 0.8709, F(6,103) = 0.4101; Fig. 5J) but there were main effects of time on 2 h water intake (two-way ANOVA, time: #p = 0.0481, F(3.904,68.32) = 2.552; Fig. 2K). Collectively, these data demonstrate that psilocin acutely reduced drinking in both AIR and CIE mice as compared with within-subject vehicle injection, but there were no significant differences between groups by ethanol exposure history.
A, Experimental timeline outlining drinking paradigm relative to within-subject vehicle and psilocin injections. B, 1 and 2 h baseline ethanol consumption during the first 2 weeks of two-bottle choice drinking (2BC), n = 20. C, Splitting of air (AIR) and ethanol vapor exposed (CIE) mice based off of baseline ethanol intake, AIR: n = 10; CIE: n = 10. Unpaired t test: p = 0.8758, t = 0.1586, df = 18. D, 2 h ethanol intake in AIR and CIE mice for subsequent weeks of 2BC drinking sessions, AIR: n = 10; CIE: n = 10. E, Blood alcohol concentrations measured for each week of ethanol vapor exposure, CIE: n = 10. F, Within-subject comparison of ethanol intake following vehicle and psilocin injection in CIE-exposed mice, AIR: n = 10; CIE: n = 10. Two-way ANOVA: *p = 0.0313, F(1,15) = 5.643. G, Within-subject comparison of ethanol intake following vehicle and psilocin injection in AIR control mice, AIR: n = 10; CIE: n = 10. Two-way ANOVA: treatment: **p = 0.0013, F(1,18) = 14.34, time: #p < 0.0001, F(11,17) = 45.32. H, 1 h ethanol intake in CIE and AIR mice including last 2BC session before psilocin injection, AIR: n = 10; CIE: n = 10. Two-way ANOVA: time: #p < 0.0001, F(4.193,73.37) = 14.22. I, 2 h ethanol intake in CIE and AIR mice including last 2BC session before psilocin injection, AIR: n = 10; CIE: n = 10. Two-way ANOVA: #p < 0.0001, F(4.845,84.79) = 7.803. J, 1 h water intake in CIE and AIR mice including last 2BC session before psilocin injection, AIR: n = 10; CIE: n = 10. Two-way ANOVA: p = 0.8709, F(6,103) = 0.4101. K, 2 h water intake in CIE and AIR mice including last 2BC session before psilocin injection, AIR: n = 10; CIE: n = 10. Two-way ANOVA: time: #p = 0.0481, F(3.904,68.32) = 2.552.
Effects of psilocin on CeA activity and locomotion in naive female CRF1:GFP mice
To examine how psilocin impacts CeA activity, mice were injected with 2 mg/kg psilocin or vehicle 90 min prior to brain collection (Fig. 3A). This dose of psilocin (2 mg/kg) was chosen as it is a full active dose that has been shown to elicit significant head twitch response in mice (Halberstadt et al., 2011). CeA activation was assessed in the medial and lateral subregions (Fig. 3B, CeAm, CeAl). We performed immunohistochemistry to label c-Fos, an immediate early gene marker for neuronal activation, in ethanol-naive female CRF1:GFP mice following psilocin or vehicle (Fig. 3C). Psilocin increased CeA c-Fos expression compared with vehicle (unpaired t test, **p = 0.0026, t = 4.553, df = 7; Fig. 3D). There was no significant difference in CeAm c-Fos between psilocin and vehicle (unpaired t test: p = 0.2864, t = 1.154, df = 7; Fig. 3E); however, psilocin increased CeAl c-Fos expression (unpaired t test, ***p = 0.0003, t = 6.453, df = 7; Fig. 3F). To assess changes in the proportion of CRF1 neurons in the total activated population, we measured GFP as a marker for CRF1+ neurons and quantified the number of CRF1+ c-Fos+ neurons out of total c-Fos+ neurons following psilocin or vehicle (Fig. 3G). Psilocin lowered the proportion of CRF1 neurons in the active CeA population (unpaired t test, **p = 0.0099, t = 3.504, df = 7; Fig. 3H). In addition, psilocin lowered the proportion of CRF1 neurons in the active CeAm population (unpaired t test, *p = 0.0190, t = 3.036, df = 7; Fig. 3I) and CeAl population (unpaired t test, *p = 0.0320, t = 2.671, df = 7; Fig. 3J). Locomotor testing was completed to determine whether psilocin induced sedation or hyperlocomotion which could contribute to changes in drinking. In a separate cohort of mice, we measured locomotion and found no difference in total distance traveled (unpaired t test: p = 0.5354, t = 0.6318, df = 18; Fig. 3K), distance traveled by minute (two-way ANOVA: p = 0.1170, F(9,162) = 1.608; Fig. 3L), or time spent in center between vehicle and psilocin (unpaired t test: p = 0.6409, t = 0.4744, df = 18; Fig. 3M).
A, Experimental timeline outlining time of treatment relative to brain collection for analysis. B, Representative image of c-Fos expression (red) in the central amygdala (CeA) and anatomical outline of the medial (CeAm) and lateral central amygdala (CeAl). C, Representative image of CeA c-Fos expression (red) in psilocin and vehicle treated mice. Scale bar, 50 μm. D, CeA c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: **p = 0.0026, t = 4.553, df = 7. E, CeAm c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: p = 0.2864, t = 1.154, df = 7. F, CeAl c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: ***p = 0.0003, t = 6.453, df = 7. G, Representative image of CeA c-Fos expression (red) and green fluorescent protein expression (GFP, green) in psilocin and vehicle treated mice. Scale bar, 50 μm. H, Proportion of CRF1 neurons in the CeA c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: **p = 0.0099, t = 3.504, df = 7. I, Proportion of CRF1 neurons in the CeAm c-Fos+ population in female psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4 mice. Unpaired t test: *p = 0.0190, t = 3.036, df = 7. J, Proportion of CRF1 neurons in the CeAl c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: *p = 0.0320, t = 2.671, df = 7. K, Distance traveled during locomotor testing in vehicle and psilocin groups, Vehicle: n = 10; Psilocin: n = 10. Unpaired t test: p = 0.5354, t = 0.6318, df = 18. L, Distance traveled by minute during locomotor testing in vehicle and psilocin groups, Vehicle: n = 10; Psilocin: n = 10. Two-way ANOVA: p = 0.1170, F(9,162) = 1.608. M, Time spent in center during locomotor testing in vehicle and psilocin groups, Vehicle: n = 10; Psilocin: n = 10. Unpaired t test: p = 0.6409, t = 0.4744, df = 18.
Psilocin effects on CeA activity 24 h into withdrawal from chronic intermittent ethanol vapor exposure and drinking
To assess how psilocin changes CeA activity following chronic ethanol exposure and drinking, mice were injected with 2 mg/kg psilocin or vehicle 24 h into withdrawal (Fig. 4A) from 7 weeks of CIE-2BC (Fig. S1A). Mice displayed stable ethanol and water drinking (Fig. S1B,C), ethanol preference (Fig. S1D), and intoxication levels during CIE-2BC (Fig. S1E,F). Psilocin increased corticosterone compared with vehicle (unpaired t test, **p = 0.0083, t = 3.481, df = 8; Fig. 4B). Neuronal activation was assessed in CeAm and CeAl (Fig. 4C), and c-Fos expression was examined 24 h into withdrawal (Fig. 4D). Psilocin increased CeA c-Fos compared with vehicle (unpaired t test, *p = 0.013, t = 3.182, df = 8; Fig. 4E). There was no difference in CeAm c-Fos expression between groups (unpaired t test: p = 0.6200, t = 0.5157, df = 8; Fig. 4F); however, psilocin increased CeAl c-Fos (unpaired t test, *p = 0.0104, t = 3.328, df = 8; Fig. 4G). The number of CRF1+ c-Fos+ neurons out of total c-Fos+ neurons was quantified to assess changes in the proportion of CRF1 neurons in the total activated population following psilocin or vehicle (Fig. 4H). Psilocin lowered the proportion of CRF1 neurons in the active CeA population (unpaired t test, *p = 0.0249, t = 2.755, df = 8; Fig. 4I). There was no difference in the proportion of CRF1 neurons in the active CeAm (unpaired t test: p = 0.1000, t = 1.860, df = 8; Fig. 4J) or CeAl population between groups (unpaired t test: p = 0.1345, t = 1.665, df = 8; Fig. 4K). Together, psilocin increased CeA and CeAl activity and decreased CRF1 neurons in the active CeA population. The decrease in active CeA CRF1 neurons occurred in parallel with increased serum corticosterone levels.
A, Experimental timeline outlining treatment 24 h following the last CIE session relative to brain and blood collection. B, Corticosterone concentrations compared between vehicle and psilocin groups, Vehicle: n = 5; Psilocin: n = 5. Unpaired t test: **p = 0.0083, t = 3.481, df = 8. C, Representative image of a coronal section including the medial and lateral regions of the central amygdala (CeAm, CeAl). D, Representative images of CeA c-Fos expression (red) in vehicle and psilocin exposed mice. Scale bar, 50 μm. E, CeA c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 5. Unpaired t test: *p = 0.013, t = 3.182, df = 8. F, CeAm c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 5. Unpaired t test: p = 0.6200, t = 0.5157, df = 8. G, CeAl c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 5. Unpaired t test: *p = 0.0104, t = 3.328, df = 8. H, Representative image of CeA c-Fos expression (red) and green fluorescent protein expression (GFP, green) in psilocin and vehicle exposed mice. Scale bar, 50 μm. I, Proportion of CRF1 neurons in the CeA c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 5. Unpaired t test: *p = 0.0249, t = 2.755, df = 8. J, Proportion of CRF1 neurons in the CeAm c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 5. Unpaired t test: p = 0.1000, t = 1.860, df = 8. K, Proportion of CRF1 neurons in the CeAl c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 5. Unpaired t test: p = 0.1345, t = 1.665, df = 8.
Psilocin effects on CeA activation 72 h into withdrawal from chronic intermittent ethanol vapor exposure and drinking
To assess how psilocin changes CeA activity following chronic ethanol exposure after more prolonged withdrawal, mice were injected with 2 mg/kg psilocin or vehicle 72 h into withdrawal from 7 weeks of CIE-2BC (Fig. 5A). Mice maintained stable ethanol intake and intoxication levels throughout CIE-2BC (Fig. S1). There was no difference in corticosterone between groups (unpaired t test: p = 0.7788, t = 0.2919, df = 7; Fig. 5B). Neuronal activation was assessed in the CeAm and CeAl (Fig. 5C), and c-Fos expression was examined 72 h into withdrawal (Fig. 5D). Psilocin increased CeA c-Fos expression compared with vehicle (unpaired t test, **p = 0.0043, t = 4.150, df = 7; Fig. 5E). There was no difference in CeAm c-Fos between groups (unpaired t test: p = 0.2535, t = 1.244, df = 7; Fig. 5F); however, psilocin increased CeAl c-Fos expression (unpaired t test, **p = 0.0041, t = 4.188, df = 7; Fig. 5G). The number of CRF1+ c-Fos+ neurons out of total c-Fos+ neurons was quantified to assess changes in the proportion of CRF1 neurons in the total activated population following psilocin or vehicle (Fig. 5H). Psilocin lowered the proportion of CRF1 neurons in the active CeA population (unpaired t test, **p = 0.0033, t = 4.370, df = 7; Fig. 5I). Psilocin also lowered the proportion of CRF1 neurons in the active CeAm (unpaired t test, *p = 0.0377, t = 2.557, df = 7; Fig. 5J) and CeAl population (unpaired t test, *p = 0.0140, t = 3.252, df = 7; Fig. 5K).
A, Experimental timeline outlining treatment 72 h following the last CIE session relative to brain and blood collection. B, Corticosterone concentrations compared between vehicle and psilocin groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: p = 0.7788, t = 0.2919, df = 7. C, Representative image of a coronal section including the medial and lateral regions of the central amygdala (CeAm, CeAl). D, Representative images of CeA c-Fos expression (red) in vehicle and psilocin exposed mice. Scale bar, 50 μm. E, CeA c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: **p = 0.0043, t = 4.150, df = 7. F, CeAm c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5. Psilocin: n = 4. Unpaired t test: p = 0.2535, t = 1.244, df = 7. G, CeAl c-Fos expression in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: **p = 0.0041, t = 4.188, df = 7. H, Representative image of CeA c-Fos expression (red) and green fluorescent protein expression (GFP, green) in psilocin and vehicle exposed mice. Scale bar, 50 μm. I, Proportion of CRF1 neurons in the CeA c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: **p = 0.0033, t = 4.370, df = 7. J, Proportion of CRF1 neurons in the CeAm c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: *p = 0.0377, t = 2.557, df = 7. K, Proportion of CRF1 neurons in the CeAl c-Fos+ population in psilocin and vehicle groups, Vehicle: n = 5; Psilocin: n = 4. Unpaired t test: *p = 0.0140, t = 3.252, df = 7.
Discussion
The current study examined how psilocin impacted voluntary drinking and subregion-specific CeA CRF1 neuron activation in female CRF1:GFP mice. Psilocin acutely decreased ethanol intake in mice following chronic ethanol vapor and chronic drinking or after chronic drinking alone. In ethanol-naive mice, psilocin increased CeA activation but reduced CRF1 neuron activation relative to overall CeA activity. In ethanol-exposed mice, psilocin increased CeA activation in lateral CeA (CeAl), but not medial CeA (CeAm). During withdrawal, psilocin increased activation in CeAl and reduced the proportion of CRF1 neurons within the activated CeAm and CeAl populations. Psilocin increased corticosterone at 24 h, but not 72 h into withdrawal. These results reveal a potential interaction between psilocin and CeA CRF1 neurons that may play a part in reducing AUD pathology.
Psilocybin increased abstinence and decreased heavy drinking in AUD patients (Bogenschutz et al., 2015, 2022). Psilocybin has been shown to decrease drinking in preclinical (Alper et al., 2023) and clinical studies (Bogenschutz et al., 2015, 2022); however, the mechanisms underlying these effects are unclear. We used chronic intermittent ethanol vapor-two-bottle choice drinking (CIE-2BC) and 2BC alone (AIR-2BC) to explore whether ethanol history severity impacted the effects of psilocin. We used a shortened CIE-2BC paradigm, which did not elicit escalation of drinking (Smith et al., 2020) but reflects a history of heavy ethanol exposure. It is important to note that the AIR-2BC group drank more than the CIE-2BC group on the day of vehicle injection potentially because the AIR-2BC group was compensating for the week in which they were not given access to ethanol. The AIR-2BC group consistently drank more than the CIE-2BC group in the first 2BC session following a week of vapor exposure as demonstrated in Figure 2D, sessions 4.1 and 6.1. Psilocin acutely reduced drinking following both CIE-2BC and AIR-2BC, suggesting that psilocin acutely reduces drinking across levels of alcohol drinking exposure. A recent study reported that a single dose of psilocybin paired with psychotherapy acutely reduced alcohol craving but did not significantly change heavy drinking days or abstinence at 4 week or 6 month follow-ups (Rieser et al., 2025). Like our findings, psilocybin seemed to have an acute effect; however, a single dose was not sufficient to produce long-lasting effects on drinking.
The CeA is implicated in AUD, specifically alcohol-associated plasticity that contributes to withdrawal-associated negative affect. One study in C57BL/6J mice measured CeA activation 26 and 74 h after chronic ethanol vapor exposure and reported increased CeA cFos expression at 26 h withdrawal (Smith et al., 2020). Psilocybin has been shown to engage the CeA as well as other brain regions (Davoudian et al., 2023; Effinger et al., 2023; Aboharb et al., 2025). We hypothesized that psilocin may be acting on CeA microcircuitry to indirectly impact CRF1 neurons and reduce alcohol consumption. We observed that psilocin increased CeA activation, driven by CeAl but not CeAm, and decreased the proportion of CRF1 within the c-Fos+ population in both subregions. CeA microcircuitry is involved in ethanol effects, as well as fear conditioning and pain (Ciocchi et al., 2010; Li et al., 2013; Jiang et al., 2014; Kiritoshi et al., 2024). Acute ethanol increased firing in CeAm CRF1+ neurons that project to the bed nucleus of the stria terminalis in CRF1:GFP mice (Herman et al., 2013) and chronic ethanol vapor increased excitability and decreased inhibition in CeAm CRF1+ neurons (Herman et al., 2016). Psilocin increased activation of the CeAl which receives input from brain regions with dense 5HT2A expression including basolateral amygdala and dorsal raphe (Bombardi and Di Giovanni, 2013). The CeAl projects to the CeAm mainly through GABAergic signaling (Melkumyan and Silberman, 2022) which could dampen activity of CeA CRF1 neurons. The subregion- and population-specific changes in CeA activity following psilocin suggest the potential engagement of CeA microcircuitry with increases in activation reducing the proportion of CRF1 neuron activation contributing to reductions in drinking (Fig. 6).
Working diagram illustrating potential microcircuitry underlying effects of psilocin on CeA non-CRF1 and CRF1 neuronal activation following CIE-2BC and AIR-2BC drinking paradigms. Psilocin increases CeA and CeAl activation which are likely to be GABAergic non-CRF1 neurons. As there are inhibitory projections from CeAl to CeAm, these non-CRF1 neurons may form inhibitory synapses with neurons in the CeAm, including CRF1 neurons, suppressing the proportion of these neurons in the overall active population. These changes may contribute to the acute reduction in alcohol drinking observed after psilocin treatment in CIE-2BC and AIR-2BC exposed mice.
Stress can contribute to AUD by promoting alcohol seeking through active coping, which could be a factor in clinical effects of psychedelics. The hypothalamic-pituitary-adrenal (HPA) axis promotes release of stress hormones like cortisol in humans (corticosterone in rodents) and can become dysregulated with chronic alcohol use (Stephens and Wand, 2012). 5HT2A agonists, such as psilocybin, have been shown to increase corticosterone in mice (Jones et al., 2023; Farinha-Ferreira et al., 2025). The paraventricular nucleus of the hypothalamus (PVN) is activated by psilocin, likely contributing to increased corticosterone levels; however, corticosterone levels following acute restraint stress were not different from vehicle controls 1 week after psilocin administration (Effinger et al., 2024). Corticosterone was measured after psilocin injection at 24 and 72 h into withdrawal to investigate changes in hormonal stress response. Psilocin increased corticosterone at 24 h but not 72 h withdrawal, which could be due to increased sensitization at 24 h or to desensitization of PVN 5HT2A receptors at 72 h, as has been observed with chronic stress (Lee et al., 2009; Jones et al., 2023). Following chronic ethanol exposure, corticosterone, CeA CRF1 activation, and ethanol intake were not different between vehicle and psilocin groups 1 week after injection, suggesting drug effects were not persistent. While both HPA and extrahypothalamic limbic CRF1 circuitry are engaged by stress and alcohol (Stephens and Wand, 2012), the effects of psilocin on corticosterone are likely independent of changes in CeA activation, reflecting the different roles of the endocrine and CeA CRF1 systems.
Although CRF1 antagonists reduce drinking (Agoglia et al., 2022) and anxiety-like behavior (Valdez et al., 2003) in preclinical studies, clinical studies have not replicated these effects (Kwako et al., 2015; Schwandt et al., 2016). It is possible that systemically blocking only CRF1 receptor activity in humans is not sufficient to reduce drinking because there are other cell types and receptors involved in AUD pathology. Our study provides insight into the mechanisms related to the acute, active drug effects of psilocin on alcohol drinking. During the active drug period, psilocin increased CeA activation and decreased the proportion of CRF1 neurons within that activated population which may contribute to the acute reduction in drinking immediately following treatment. This single psilocin dose and change in activation may be transient and insufficient to produce long-lasting effects on drinking as observed in clinical studies; however, it may contribute to a sustained reduction in drinking with multiple doses. Psilocin may also have had more prolonged effects if ethanol access was restricted before reinstatement. Another consideration is that psilocin may induce anxiety-like behavior or alter taste perception or the hedonic value of ethanol. Previous studies report that psilocybin does not alter taste perception as measured by quinine preference (Alper et al., 2023) and does not alter sucrose preference in corticosterone-exposed mice (Zhao et al., 2024). Locomotor testing found no differences in activity or center time, suggesting psilocin did not impact locomotion or increase anxiety-like behavior and that the mice did not drink less due to psilocin-induced sedation or hyperactivity. Psychedelics have been shown to reduce the motivation for food reinforcers, and since alcohol is caloric, it is possible this may play a role in the acute reduction in drinking observed following psilocin injection. Future studies could examine the effects of psilocin on motivation to feed using a sucrose consumption test or novelty-suppressed feeding test to understand any influence this may have on alcohol drinking.
In humans, the risk of developing AUD or comorbid psychiatric conditions can be influenced by sex differences in the motivation to drink and the emergence of negative affect during withdrawal. We used female mice as they drink more than males (Salazar and Centanni, 2024) and thus provide a model for investigating psilocin effects related to severity of exposure. Additionally, there are sex differences in CRF/CRF1 signaling in response to alcohol and stress as well as CRF regulation of stress-related behaviors (Agoglia et al., 2020a,b). These sex differences can contribute to differences in AUD pathology that should be considered when exploring therapeutic targets and outcomes. Further, there are sex differences in cellular and behavioral effects of psilocin (Effinger et al., 2023), and more research is needed to examine potential sex-specific effects of psilocin on drinking. As withdrawal/negative affect plays a crucial role in alcohol seeking and consumption, future studies should consider the effects of psilocin on withdrawal behavior (e.g., somatic signs, anxiety-like behavior) and additional withdrawal timepoints. Future studies using direct measures of activity to investigate psilocin effects on CeA microcircuits or other AUD/stress-relevant circuits should be performed to gain a more holistic understanding of the neurocircuitry involved in drug and behavioral effects. Lastly, future studies should consider the effects of psilocin in other models (e.g., chronic stress, stress-induced relapse, comorbid conditions) to determine the effects of psilocin in other disease states.
Overall, this study demonstrates that psilocin engages CeA circuitry and reduces the proportion of CRF1 neurons within the activated CeA population in ethanol-naive and ethanol-exposed female mice. Psilocybin may alleviate symptoms of AUD (e.g., negative withdrawal state, alcohol seeking/consumption), in part through decreases in CeA CRF1 activation. These findings extend our understanding of the potential mechanisms underlying the effects of psilocin and may inform future therapeutic strategies for AUD.
Footnotes
This work was supported by the National Institute on Alcohol Abuse and Alcoholism (R01AA026858, P60AA011605; M.A.H.) and by the Brain and Behavior Research Foundation NARSAD Young Investigator Award (M.A.H.).
The authors declare no competing financial interests.
This paper contains supplemental material available at: https://doi.org/10.1523/JNEUROSCI.0652-25.2025
- Correspondence should be addressed to Melissa A. Herman at melissa_herman{at}unc.edu.
