Experimental Drug Blocks THC’s Psychosis-Like Effects in Mice

Scientists have long known that tetrahydrocannabinol, better known as THC, can produce dramatically different effects depending on the dose, the person, and the circumstances in which it is consumed. For some people, intoxication may involve relaxation or altered sensory perception. For others, it can include paranoia, disorganized thinking, or temporary psychotic symptoms.

What researchers have understood less clearly is how THC produces these psychiatric effects at the level of individual brain cells.

A new study¹ published in Biological Psychiatry Global Open Science provides a potential piece of that puzzle. Working with male mice and cultured cells, researchers identified a signaling pathway connecting THC activation of cannabinoid receptors to a behavioral response used in preclinical psychosis research. They also found that an experimental drug called AEF0117 could interrupt this pathway and prevent the behavioral effect.

The findings do not demonstrate that the drug can prevent cannabis-induced psychosis in people. However, they reveal a potentially useful therapeutic target and illustrate how researchers may be able to block specific harmful effects of THC without disabling the entire endocannabinoid system.

How THC Can Produce Opposing Effects

THC primarily produces its intoxicating effects by activating cannabinoid type 1 receptors, commonly called CB1 receptors. These receptors are abundant throughout the brain and help regulate processes including movement, memory, appetite, reward, stress, and sensory perception.

Activating CB1 does not always produce one predictable response. THC has what researchers describe as biphasic effects, meaning that lower and higher doses can sometimes produce opposite outcomes. In rodents, higher THC doses commonly reduce movement, while sufficiently low doses can increase it.

This increase in movement is known as hyperlocomotion. Researchers use hyperlocomotion induced by psychoactive drugs as an experimental proxy for certain positive psychosis-like effects, particularly during early screening of potential antipsychotic treatments. It is not equivalent to diagnosing psychosis in a human. Instead, it provides a measurable behavioral response that can be connected to particular receptors, cells, and signaling pathways.

This distinction matters because genetic pathways linking cannabis use disorder to psychosis suggest that vulnerability is influenced by multiple interacting biological systems. No single mouse behavior can reproduce the perceptual, emotional, and cognitive complexity of human psychosis.

Researchers Traced the Signal to Specific Brain Cells

The researchers administered a low dose of THC, 0.3 milligrams per kilogram, to adult male mice. They then combined behavioral testing with genetic mouse models, cultured-cell experiments, molecular assays, and targeted viral techniques.

The resulting evidence pointed to a particular intracellular chain involving CB1 receptors, beta-arrestin 1, Akt, and glycogen synthase kinase 3 beta, or GSK3β. When THC activated CB1 receptors under the tested conditions, beta-arrestin 1 helped transmit the signal inside the cell. Akt was subsequently activated, while GSK3β activity was inhibited through phosphorylation.

The researchers tested this relationship from several directions. THC did not produce the same hyperlocomotion in mice lacking beta-arrestin 1. A compound that inhibited an upstream component of the pathway also prevented both the molecular changes and increased movement. Finally, genetically preventing the relevant modification of GSK3β in selected neurons stopped the behavioral response.

Together, these experiments suggest that the pathway was not simply associated with hyperlocomotion. Its operation was required for the response to occur.

Why D2 Striatopallidal Neurons Matter

The team also investigated where the relevant CB1 receptors were located. Their experiments pointed to plasma-membrane CB1 receptors on dopamine D2-positive medium spiny neurons in the striatopallidal pathway.

This pathway is part of the basal ganglia circuitry involved in movement, motivation, and action selection. It operates alongside another major pathway containing neurons that express dopamine D1 receptors. Although these systems interact, they can have different effects on movement and behavior.

Removing CB1 receptors from D2-associated neurons prevented THC-induced hyperlocomotion. Removing them from D1-associated neurons did not eliminate the effect, although the researchers acknowledged that CB1 reduction in that model was incomplete. The study also ruled out a primary role for mitochondrial CB1 receptors in this particular response.

Selective CB1 Modulation Offers a Different Strategy

Identifying a pathway is only part of the story. The researchers also tested whether it could be interrupted without broadly suppressing normal movement.

They examined pregnenolone, a naturally occurring neurosteroid, and AEF0117, a synthetic derivative designed to be orally available and remain active longer. Both are described as signaling-specific inhibitors of CB1 receptors. Rather than competing directly with THC at the receptor’s primary binding site, these compounds are intended to selectively alter particular signals generated after CB1 activation.

This approach differs from conventional CB1 antagonism. Earlier drugs that broadly blocked CB1 receptors could interfere with appetite, mood, reward, and other normal functions controlled by the endocannabinoid system. Selective signaling inhibition aims to preserve more of the receptor’s ordinary activity while restraining pathways associated with harmful cannabinoid effects.

The distinction also helps explain why natural cannabis and powerful synthetic CB1 agonists can have very different safety profiles. As discussed in research examining why synthetic cannabinoids are not the same as cannabis, some synthetic compounds activate CB1 receptors more strongly and completely than THC. Receptor signaling is therefore shaped by more than whether a substance merely binds to CB1.

AEF0117 Blocked the THC Response in Mice

AEF0117 prevented the low-dose THC-induced increase in movement and the associated GSK3β phosphorylation across the dorsal striatum, nucleus accumbens, and cerebellum. Importantly, it did not reduce baseline movement when administered without THC.

That result supports the idea that AEF0117 was selectively modifying a THC-activated mechanism rather than simply sedating the animals. It also provides a possible biological explanation for findings from an earlier human study of AEF0117, in which the compound reduced some positive subjective effects of cannabis without producing obvious withdrawal or broadly disrupting cannabis-associated experiences.

However, the clinical picture remains unsettled. A subsequent Phase 2b study in people seeking treatment for cannabis use disorder did not meet its primary endpoint or key secondary abstinence endpoints, according to an AEF0117 clinical trial update from Indivior. Later analyses reported signals in certain subgroups and quantitative measures, but those findings do not replace the failed primary outcome.

The present mouse study explores a different potential application: acute psychosis-like effects rather than reducing cannabis consumption. A drug can perform differently across indications because addiction, intoxication, and psychosis involve overlapping but nonidentical biological processes.

What the Research Adds Beyond Cannabis Risk Statistics

Public discussions often reduce the cannabis and psychosis relationship to a debate over whether cannabis causes schizophrenia. Biology is more complicated. Cannabis exposure may interact with genetic susceptibility, age, THC potency, frequency of use, psychiatric history, and other environmental factors.

The new study does not resolve those population-level questions. Instead, it contributes something different by tracing one acute THC response from receptor activation to a particular intracellular signal and neuronal population.

This mechanistic perspective could eventually support several advances:

• More targeted treatments for harmful acute THC reactions
• Better screening models for cannabinoid-related therapeutics
• Biological markers that distinguish beneficial and harmful CB1 signaling
• Safer drug designs that modulate rather than eliminate receptor activity

The broader opportunity is precision pharmacology. CB1 receptors perform too many useful functions to treat all their signaling as interchangeable. If researchers can separate pathways associated with therapeutic effects from those involved in intoxication, dependence, or psychiatric harm, future medicines may become more selective.

Why the Findings Remain Preliminary

Several limitations prevent immediate clinical conclusions. The research used an acute THC exposure rather than prolonged or repeated consumption. Chronic exposure can produce tolerance and changes in receptor availability that may substantially alter the pathway.

All animal experiments involved male mice. The authors noted that women may have different vulnerability to cannabis-associated psychosis, making female cohorts essential in future studies. Hyperlocomotion also captures only a narrow behavioral dimension and cannot represent hallucinations, delusions, or disorganized thought.

Commercial involvement requires transparency as well. Aelis Farma supported part of the research, several authors are shareholders or hold stock options, and some researchers are inventors on patents covering AEF0117 and its use for cannabinoid-related disorders. These relationships do not invalidate the experiments, but independent replication would strengthen confidence in the findings.

A More Precise View of THC’s Effects

The study advances cannabis science by showing that a behavioral effect of THC may depend on a specific receptor location, cell type, and intracellular pathway. THC-induced hyperlocomotion required membrane CB1 receptors on D2 striatopallidal neurons and signaling through beta-arrestin 1, Akt, and GSK3β.

AEF0117 interrupted that chain without broadly suppressing movement, illustrating why selective CB1 modulation remains scientifically compelling despite mixed clinical results in cannabis use disorder.

The most important takeaway is not that researchers have discovered a treatment for cannabis-induced psychosis. They have identified a biological route that may help explain one component of THC’s psychotomimetic effects and could be investigated as a therapeutic target. Translating that discovery will require chronic-exposure models, female animals, independent replication, and carefully designed human trials.

References:

^{1. Tomaselli, G., Bellocchio, L., Raux, P.-L., Roullot-Lacarrière, V., Busquets-Garcia, A., Lalanne, V., Mariani, Y., Cathala, A., Zanese, M., Cannich, A., Grel, A., Gisquet, D., Mondesir, M., Metna-Laurent, M., Drutel, G., Marsicano, G., Piazza, P. V., Revest, J.-M., & Vallée, M. (2026). Psychotomimetic cannabinoid hyperlocomotion involves GSK3β signaling at striatopallidal neurons in male mice. Biological Psychiatry Global Open Science, 100816. https://doi.org/10.1016/j.bpsgos.2026.100816}