The apolar coumarin scoparone (6,7-dimethoxycoumarin) is mainly associated with the consumption of Génépi liquors that are very popular in the Alpine regions of Europe⁶. Distillation recipes of the French apéritif quina and some liquors like the Swedish-made Jeppson’s Malört use Artemisia species containing scoparone in Amaro. We hypothesized that scoparone might contribute to the psychoactive activity of liquors mentioned above and influence the brain and behavioral changes. We found that scoparone is a major coumarin in Artemisia umbelliformis Lam., a plant frequently used to distill mentioned above liquors, with an estimated amount of around 0.1–0.5% of the dried aerial parts.

For the peripheral effects, various mechanisms of action of scoparone have been proposed in vitro, mainly at concentrations in the micromolar concentration range^19–21. However, these concentrations are not achieved in the brain. Thus, despite the wealth of data related to the peripheral anti-inflammatory and liver-protective effects of scoparone, it was hitherto unclear whether this coumarin also affects the brain. Here we show that scoparone exerts potent differential neuropharmacological effects in mice.

We used quantitative LC–MS/MS to measure scoparone in total brain homogenates at different time points after the FST and EPM assays. Because the administration of 5 or 12.5 mg/kg gave similarly low scoparone concentrations in the brain after 1 or 2 h (< 2 ng/g), respectively, we conclude that this coumarin is rapidly metabolized in the periphery, in agreement with previous literature. A fast metabolism of scoparone was described in the periphery after oral administration of herbal drugs²². After 30 min, around 0.1 µg/ml of scoparone was measured that further decreased to 0.01 µg/ml after 2 h. Overall, the bioavailability of coumarins was limited by liver metabolism by CYP1A2 enzymes that demethylate scoparone²³. Scoparone was easily O-demethylated to scopoletin in the liver²⁴, and different metabolites were generated²². Our study did not detect the accumulation of the major metabolites (Fig. 1) in the brain. However, after treatment with borneol and the highest dose of scoparone (12.5 mg/kg), traces of isofraxidin were detected. In the experiment in which borneol was co-administered with scoparone, no modulation of the scoparone-induced memory improvement was observed (Supplementary Fig. S2). Co-administration of 50 mg/kg of borneol, a monoterpene reported previously to specifically increase brain permeability of different drugs, hypothetically via p-glycoprotein inhibition¹⁸, increased scoparone concentrations in the brain by approximately 300%. The effect of borneol enantiomers on the pharmacokinetics of the simple coumarin osthole was previously investigated. Co-administration of these two compounds increases the bioavailability of osthole even by up 270%²⁵. We found a similar acute administration of 12.5 mg/kg of scoparone yielded approximately 3–4 times the brain exposure of scoparone at 5 mg/kg. Our data thus suggest that scoparone could be eliminated from the brain via p-glycoprotein associated mechanisms. The concentration of scoparone (5 and 12.5 mg/kg) in the brain was highest at the earliest time point measured, after 15 min from the injection, and decreased in a time-dependent manner showing first-order kinetics.

Despite the low central bioavailability of scoparone the neuropharmacological profiling in different behavioral paradigms revealed an apparent effect specificity. Scoparone administered acutely (Supplementary Table S2) and subchronically (Supplementary Table S3) did not affect stress and depressive-like behaviors in the FST in both sexes, even though it exerted significant effects on endocannabinoids and related lipids (see below). We employed imipramine as a positive control in this model. Contrary to our expectations, we did not observe sex differences in behavior and acute drug action in the FST.

Upon acute administration of scoparone, in the EPM paradigm we observed anxiogenic-like effects in a bell-shaped curve of changes in the percentage of open arms time and entries (Fig. 2). Interestingly, the anxiogenic-like effect was not observed upon subchronic administration of scoparone (Supplementary Table S4). Unlike the acutely injected mice, these mice were more stressed (receiving repetitive daily injections) as judged by the corticosterone levels between vehicle control groups. Lipidomic analyses showed that subchronic administration of scoparone significantly reduced NAEs, including the endocannabinoid anandamide (AEA) reported to mediate anxiolytic effects in different preclinical models like the EPM²⁶. The reversal of stress-induced anandamide deficiency is a postulated mechanism subserving the therapeutic effects of FAAH inhibition²⁷.

Conversely, acute scoparone administration led to a very strong and unexpected increase of AA, prostaglandins, but also NAEs like anandamide and a concomitant decrease of 2-AG, which points towards a potent effect on lipid remodeling. 2-AG has been implicated in anxiety and is currently being studied in the context of stress-related conditions^13,28. We have previously shown that endocannabinoid reuptake inhibitors lead to anxiolytic effects in the EPM via CB1 receptor activation, possibly mediated via 2-AG²⁹. Thus, we speculate that the anxiogenic effect observed in the acute scoparone treatment could be related to the acute dysregulation of the arachidonate-endocannabinoid axis, i.e., resulting in increased prostaglandin and decreased 2-AG levels (see below). 2-AG deficiency has been implicated in anxiety and 2-AG can directly activate GABA-A receptors³⁰, which mediates anxiolytic effects. Future studies will need to address this modulation in a brain region-specific manner, focussing on the amygdala, which is involved in anxiety-like behaviors in mice.

Dependent on the type of stress stimulus, PGE2 released from neurons, astrocytes, or microglia may trigger multiple neuronal pathways that result in behavioral stress responses including anxiety³¹. The observed effects on endocannabinoids are peculiar and suggest that scoparone significantly modulates the ECS in a behaviour context-dependent manner, a new finding in coumarin pharmacology. The inverse correlation between AA, prostaglandins, and 2-AG in the brain was first described by Nomura et al.³² and generally relates to changes in monoacylglycerol (MAGL) lipase expression. Thus, we measured whether MAGL or related serine hydrolases were a target of scoparone. Data obtained in ABPP experiments show that the serine hydrolases regulate endocannabinoid brain levels, namely MAGL, FAAH, and that ABHD6 were not inhibited by scoparone, and thus are not direct targets of this coumarin. While different natural products have been shown to interact with ECS targets³³. Scoparone does not seem to exert a direct ECS modulation. Therefore, we will aim to elucidate the mode of action of scoparone in future studies.

Since scoparone did not modulate the FST behavior itself, we assume that this lipid remodeling was not causally involved in the FST behavior. Given the association of anandamide with depression-like behaviors in mice, this finding is unexpected because FAAH inhibitors are effective in the FST paradigm³⁴.

The only emerging potentially beneficial effect was related to the PA test paradigm. It showed that scoparone injected both acutely and subchronically improved memory and learning. Because only male mice were tested in the PA assay, we currently do not know the procognitive effects in female mice, which is a limitation of this study. Interestingly, nootropics showing procognitive effects in the PA tests often also show anxiolytic-like effects³⁵, in agreement with our observation in the EPM. To further explore the mechanisms of the procognitive effects of scoparone, we applied two different amnesia models related to scopolamine and LPS treatment, respectively. LPS as bacterial endotoxin causes acute systemic neuroinflammation and amyloidosis in vivo and triggers the production of pro-inflammatory cytokines³⁶. In our study, scoparone did not reverse LPS-mediated cognitive impairment suggesting a lack of central anti-inflammatory effects. Acutely, we observed a significant increase of prostaglandins PGE2 and PGD2 in the FST, indicating at least temporary pro-inflammatory effects. Because timing issues hampered the experimental setup in the PA test for lipidomic analyses, we cannot be sure whether in all behavioral paradigms these lipids were modulated equally, which is a limitation of the study. However, the decrease of NAEs observed after subchronic scoparone treatment in the FST and EPM, if also present in the PA test, could limit the known anti-inflammatory action of these NAEs³⁷.

On the other hand, the scopolamine model of dementia relates to cholinergic processes underlying memory functioning. Scopolamine was also found to cause lipid peroxidation by triggering free radical production, especially malondialdehyde (MDA) in combination with a decrease of antioxidative enzymes in the brain³⁸. In our study, the scopolamine-induced memory impairments were attenuated by scoparone, indicating that scoparone treatment may also alter cholinergic neurotransmission. Biochemical data suggested that scoparone (5 and 12.5 mg/kg) inhibits the increase of AChE, but no BChE was observed upon scopolamine (1 mg/kg) injection. This is in line with our previous studies where the furanocoumarins xanthotoxin and bergapten diminished the level of AChE and increased the antioxidative capacity in the hippocampus and prefrontal cortex in the scopolamine model of dementia in mice^14,38.

The antioxidant capacity was marginally but significantly higher in the scopolamine model after a single scoparone injection (5 and 12.5 mg/kg). Thus, in addition to effects on AChE, scoparone increased the antioxidative capacity in the brain that may, at least in part, underlie the procognitive effects.

The low concentration of 5 ng/g scoparone in brain tissue detected after 0.5 h after the injection of 5 mg/kg (i.p.) of scoparone was high enough to improve memory acquisition compared to the control group points towards a potent mode of action. Nevertheless, the Cmax before the behavioral assay was estimated to be higher than 10 ng/g. Meanwhile, after 0.5 h from the injection of 12.5 mg/kg (i.p.) of scoparone, we observed an average concentration of around 20 ng/g in brain tissue which was four times higher than with the lower dose. Thus, the two times higher dose gave a four times higher concentration in the healthy brain. According to the results from the PA assay, the procognitive effect was higher for the lower dose (5 mg/kg) of scoparone in both acute and subchronic treatment regiments, suggesting the existence of high-affinity molecular targets for scoparone (see below). After acute treatment, the higher dose of scoparone (12.5 mg/kg) caused a twofold lower procognitive effect than the lower dose (5 mg/kg). Thus, it will be interesting to elucidate the precise mode of action underlying the effects on behavior and lipid remodeling within the ECS.

In conclusion, our data demonstrate that scoparone reaches the brain rather inefficiently. Since scoparone may be effluxed via p-glycoprotein and eliminated efficiently through fast peripheral metabolism, its overall brain exposure might be restricted. However, already low nanomolar concentrations of scoparone in the brain mediated significant precognitive and anxiogenic-like effects in mice, which is a remarkable finding. The potent effects of scoparone on lipid remodeling in the ECS may, at least in part, account for the differential behavioral anxiogenic effects between the acute and subchronic treatments, the latter treatment may have a significant impact on stress. There are several possible limitations of this study that relate to the complexity of the behavioral experiments, the lipidomic analyses which were only performed in the behavioral experiments, and the preliminary nature of the pharmacokinetic data that require confirmation in fully dialyzed mice to determine the blood–brain ratio. Despite the preliminary nature of our experiments, the findings could have translational implications. Given its apparent potency, scoparone may contribute to the alleged psychoactive effects of liquors containing Artemisia species containing high amounts of this coumarin as shown for A. umbelliformis. In future studies, we aim to address the mode of action of scoparone and related coumarins on lipid remodeling and its overall effect on the central ECS.