Marijuana (Cannabis sativa L.) is the most commonly abused drug in the United States¹, with a similar tendency worldwide². In the last few years its use for recreational and medical purposes has become legal in an increasing number of countries³. Particularly, in the United States marijuana use is illegal under federal law, nevertheless, it has been legalized by individual state laws for non-medical use in 8 states and is allowed for medical use by another 29 states⁴. A primary component of marijuana, Δ⁹-tetrahydrocannabinol (∆⁹-THC)⁵, mediates a psychoactive effect on the nervous system via cannabinoid receptor 1 (CB₁R), for which it has a low binding affinity (K_i = 39.5 nM)⁶. However, in recent years, strains of herbal cannabis with an increased CB₁R potency have appeared: one example is sinsemilla, a female non-pollinated cannabis plant in which the content of ∆⁹-THC has been raised from 3.96% in 1995 to 12.3% in 2014¹. This recent shift in the generation of high-potency cannabis plant material has led to an increasing demand for cannabis-related treatments and emergency department admissions stemming from acute anxiety, psychosis or cognitive impairment^7,8,9.

The advent of synthetic cannabinoids (SCs), a type of new recreational drug commonly known on the market as “Spice” created additional challenges. Since 2008 SCs have led to increased public awareness as the most rapidly growing group of new “legal highs”, with the ability to escape detection by standard cannabinoid screening tests. Once some substances have become regulated, novel analogues have emerged on the market to satisfy demand; and the speed of production is outpacing lawmakers’ attempts to ban them¹⁰. Most SCs act as full agonists at CB₁R, with a much higher binding affinity (JWH-018, K_i = 9 nM; CP 47,497, K_i = 2.2 nM; HU210, Ki = 0.06 nM) compared to ∆⁹-THC^{6, 11}. SCs can not only permeate the blood–brain barrier, but they also accumulate in CB₁R-rich areas of the brain^{12, 13}. Recently, the number of users displaying pathological behaviors after consuming SCs has dramatically increased; symptoms include anxiety, agitation, tachycardia, cardiotoxicity and seizures or status epilepticus^{14, 15}. Deaths after SC use have also been documented: SCs were reported in 2014 to have killed 25 people and sickened more than 700 in northern Russia¹⁶. Controlled studies on humans to examine the action of cannabinoid ligands are difficult, and human data on induction, pharmacokinetics and adverse effects are therefore limited to case studies on users after voluntary drug consumption^{17, 18}. At the same time, rapid analogue development and the growing popularity of SCs impose a strong demand for evaluation of their pharmacology and toxicology to reveal the mechanism of action and to facilitate future development of a drug-specific therapy for intoxication.

Here, we report, for the first time, that acute administration of a natural (∆⁹-THC, 10 mg/kg) or synthetic (JWH-018, 2.5 mg/kg) cannabinoid triggered electrographic seizures in mice, recorded by electroencephalography (EEG) and videography. We further show that cannabinoid-induced seizures were completely prevented by pretreatment of the animals with AM-251 (5 mg/kg), a CB₁R-selective antagonist. Our data imply that even single use of cannabinoids may result in significant adverse neurological and physical effects and negatively affect human health. Finally, based on our results, AM-251 has a strong therapeutic potential for the suppression of toxic symptoms induced by cannabinoid abuse, although the human safety need to be established in controlled clinical trials.

Cannabinoids induce electrographic seizures

We administered ∆⁹-THC (10 mg/kg) or JWH-018 (2.5 mg/kg), intraperitoneally to two independent groups of mice and recorded video, EEG, electromyogram (EMG) and locomotor activity (LMA). Shortly after administration, we observed a significant decrease in LMA and EMG activity that coincided with low-intensity behavioral seizures. Several minutes after cannabinoid administration, the EEG trace resembled non-rapid eye movement sleep (Fig. 1A,B). However, detailed inspection of a magnified view (Fig. 1A,B inset) revealed the presence of electrographic seizures in the form of frequent EEG seizure spikes. The average onset latency of electrographic/behavioral seizures was shorter after JWH-018 (5.4 ± 0.6 min; p < 0.05) compared to ∆⁹-THC administration (11.1 ± 2.5 min) (Fig. 1A,B). Electrographic seizures were apparent for 256 ± 15.3 minutes after ∆⁹-THC; however, after JWH-018 administration they persisted for longer (344 ± 12 min; p < 0.001) and were observed in all tested animals (Fig. 2A). Seizure spike quantification analysis revealed significantly more frequent spikes after JWH-018 (25.1 ± 3.1 spikes/min; p < 0.001) compared to ∆⁹-THC administration (12.3 ± 1.4 spikes/min) (Fig. 2B). Isolated EEG seizure spikes were still observed in both groups 4 h after administration (Suppl. Figure 1A,C); on the next day, the effect was diminished and no electrographic seizures could be detected (Suppl.