Melissa officinalis essential oil modulates oxidative balance, cholinergic activity, and cognitive performance in a scopolamine-induced zebrafish model: implications for neuroprotective strategies in cognitive disorders

Abstract

Melissa officinalis essential oil (MEO) possesses documented neuroprotective, antioxidant, and anxiolytic properties. This study investigated the effects of MEO (150 and 300 µL/L) in a scopolamine (SCO, 100 µM)-induced zebrafish model of cognitive impairment. Cognitive performance was assessed using the Y-maze and novel object recognition (NOR) tests for spatial and recognition memory. At the same time, anxiety-like behaviors were evaluated via the novel tank diving test (NTT) and the novel approach test (NAT). MEO was administered daily for 21 days, whereas SCO and the reference drug galantamine (GAL, 1 mg/L) were administered acutely before behavioral testing. MEO significantly ameliorated SCO-induced cognitive deficits, reduced anxiety-like behaviors, inhibited brain acetylcholinesterase (AChE) activity, and mitigated oxidative stress markers. In silico ADMET predictions for MEO’s major constituents (citral, β-caryophyllene, limonene, and α-pinene) indicated high gastrointestinal absorption, blood-brain barrier permeability, and favorable safety profiles, with no predicted mutagenic or hepatotoxic effects. Collectively, these findings support the neuroprotective and anxiolytic potential of MEO in a zebrafish model of cognitive dysfunction and suggest its promise as a candidate for further investigation in neurodegenerative disease research.

Introduction

Alzheimer’s disease (AD) is a complex, progressive neurodegenerative disorder with a multifactorial etiology involving both genetic and environmental components. The main pathological hallmarks of AD include extracellular amyloid plaques, aggregates of β-amyloid (Aβ) peptides that disrupt neuronal communication, and intracellular neurofibrillary tangles formed by hyperphosphorylated tau protein, which impair cytoskeletal integrity and axonal transport^1,2,3. In addition to these classical features, mitochondrial dysfunction has emerged as a key contributor to AD pathogenesis. Aβ has been shown to impair mitochondrial enzymes, particularly cytochrome c oxidase, leading to disruptions in the electron transport chain, altered mitochondrial membrane potential, and impaired oxygen utilization. Elevated levels of Aβ40, along with evidence of mitochondrial damage, have been correlated with neurodegeneration⁴. Moreover, decreased expression of mitochondrial biogenesis regulators such as PGC-1α, NRF1, NRF2α/β, and TFAM suggests a pivotal role of mitochondrial impairment, oxidative stress, and apoptosis in disease progression⁵. Clinically, AD manifests as a gradual decline in multiple cognitive domains. Early symptoms include short-term memory loss, followed by deterioration in long-term and procedural memory, language comprehension, learning ability, and executive functions. Disorientation in time and space, impaired judgment, and reduced problem-solving capacity are commonly observed. Neuropsychiatric symptoms such as anxiety, apathy, irritability, depression, agitation, and hallucinations are also prevalent in advanced stages^6,7.

The global use of herbal remedies, either alone or as adjuvants to conventional therapies, has gained considerable attention due to their perceived safety and efficacy. Numerous preclinical and clinical studies have demonstrated the cognitive-enhancing effects of plant-based therapies^8,9. Natural products and their derivatives constitute over 50% of pharmaceuticals currently used in clinical practice, with approximately 25% derived directly from higher plants. The wide range of bioactive compounds in medicinal plants, coupled with their biocompatibility and relatively low toxicity, underscores their potential in the prevention and treatment of various disorders, including neurodegenerative diseases¹⁰.

Melissa officinalis (lemon balm) has been used medicinally for over two millennia. First recommended by Dioscorides for treating animal bites and digestive issues, it was later employed in traditional medicine for pain relief, bleeding, and melancholia. Historically praised for its memory-enhancing and sedative effects, M. officinalis remains a staple in ethnopharmacology across cultures¹¹. The plant’s pharmacological value is attributed to its diverse phytochemical profile, which includes essential oils (e.g., neral, geranial, citronellal), phenolic acids (e.g., rosmarinic acid), flavonoids (e.g., luteolin), and triterpenoids such as ursolic and oleanolic acids¹². These constituents contribute to a wide range of biological activities, including antiproliferative¹³, antitumor¹⁴, antioxidant and antiangiogenic¹⁰, cardioprotective¹⁵, antinociceptive and antihyperglycemic¹⁶, antidepressant¹⁷, anxiolytic and neuroprotective¹⁸, antiviral¹⁹, antidiabetic²⁰, antispasmodic²¹, antifungal²², and antibacterial²³ effects.

It is important to note that scopolamine (SCO) does not induce frank neurodegeneration but rather produces reversible cholinergic hypofunction by antagonizing muscarinic receptors. This transient blockade impairs learning and memory and leads to downstream biochemical alterations, such as oxidative stress, increased lipid peroxidation, and changes in antioxidant enzyme activity, which are consistently observed in both mammalian and zebrafish models of SCO-induced amnesia^24,25,26. These neurochemical disturbances serve as valuable correlations of cholinergic dysfunction and cognitive decline, even in the absence of structural neuronal loss, thereby supporting the relevance of the SCO model for screening potential neuroprotective interventions.

Beyond memory impairment, SCO also alters affective behavior. In rodents, acute or repeated SCO typically produces anxiogenic-like effects in standard assays (elevated plus maze, open field), consistent with cholinergic disruption and stress-axis activation²⁷. By contrast, zebrafish findings are more nuanced: one study reported anxiolytic-like effects of SCO in the novel tank diving test (NTT), aligning with certain human observations and highlighting a species-, dose-, and context-dependent profile for SCO’s effects on anxiety²⁸. These discrepancies underscore that, while SCO reliably models transient cholinergic hypofunction relevant to cognitive deficits, its effects on anxiety require careful interpretation and explicit behavioral validation in each paradigm²⁹.

The present study aimed to investigate the neuroprotective potential of M. officinalis essential oil (MEO) in a zebrafish model of cognitive impairment induced by SCO, a muscarinic receptor antagonist commonly used to mimic the cholinergic deficits observed in dementia. By evaluating behavioral and neurochemical endpoints, this study sought to elucidate the mechanisms underlying MEO’s protective effects. The results contribute to the growing body of evidence supporting the therapeutic use of essential oils, particularly MEO, as natural agents in the prevention or management of dementia-related cognitive decline.

Results

Chemical composition of the MEO

Gas chromatography-mass spectrometry (GC-MS) analysis of MEO revealed the presence of several volatile constituents. A total of ten major compounds were identified, and their retention times, molecular formulas, and relative percentages are listed in Table 1.

Table 1 Chemical composition of MEO as determined by GC-MS.

The compositional profile is dominated by two isomeric monoterpene aldehydes, citral components geranial (35.2%) and neral (25.4%), which together account for 60.6% of the oil. These compounds are largely responsible for the characteristic lemon-like fragrance of MEO. Other significant constituents included the sesquiterpenes caryophyllene (8.7%) and germacrene D (5.6%), as well as monoterpenes such as β-pinene (4.1%), α-pinene (3.5%), and limonene (2.9%). Compounds present in lower concentrations included β-caryophyllene (2.3%), α-humulene (1.9%), and eucalyptol (1.5%). Overall, the essential oil composition is rich in both monoterpenes and sesquiterpenes, suggesting a notable potential for bioactivity, including antioxidant and anti-inflammatory effects. The high citral content is consistent with previous reports of MEOs, particularly those sourced from Elazığ, Turkey. While this study focused on the volatile fraction of MEO, previous investigations have reported that phenolic and flavonoid constituents also contribute to its biological activity. Noguchi-Shinohara et al.³⁰ identified rosmarinic acid as a major active compound, particularly with antiviral and antioxidant properties. In addition, Kara et al.³¹ described the presence of phenolic acids such as caffeic, ferulic, and gallic acids, as well as flavonoids including luteolin, quercetin, apigenin, and hesperidin in standardized M. officinalis extracts.

Analysis of biomedical signaling pathways and Pharmacokinetic characterization of MEO

The physicochemical properties and oral bioavailability of the major compounds identified in MEO were evaluated using the SwissADME web platform³². This tool provides an intuitive visualization of “drug-likeness” through the bioavailability radar, which integrates six key parameters: lipophilicity, molecular size, polarity, solubility, flexibility, and saturation level³³. As shown in Fig. 1, all evaluated compounds exhibited values falling within the optimal physicochemical space for oral absorption, as indicated by their complete inclusion within the pink area of the radar plot. This alignment with the defined drug-likeness criteria suggests that the compounds possess favorable pharmacokinetic profiles, supporting their potential utility as orally bioavailable bioactive agents.

Fig. 1

Bioavailability radar plots for the evaluated compounds. The radar plots delineate the optimal physicochemical space required for oral bioavailability, represented by the pink area. Six key parameters are integrated: lipophilicity (logP): optimal range between − 0.7 and + 5.0; size (molecular weight): Ideal range from 150 to 500 g/mol; polarity (topological polar surface area, TPSA): between 20 and 130 Å; solubility (logS): should be greater than − 6; saturation (fraction of sp³-hybridized carbon atoms): greater than 0.25; flexibility (number of rotatable bonds): no more than 9. Each compound’s physicochemical profile was evaluated according to these criteria using the SwissADME platform. All tested molecules fell within the defined optimal zones, suggesting good oral bioavailability potential. The analyzed compounds included: (A) Scopolamine (SCO); (B) Galantamine (GAL); (C) Citral (Geranial); (D) Citral (Neral); (E) Caryophyllene; (F) Germacrene D; (G) β-Pinene; (H) α-Pinene; (I) Limonene; (J) β-Caryophyllene; (K) α-Humulene; and (L) Eucalyptol.

The drug-likeness properties and medicinal chemistry parameters estimated for the selected compounds are presented in Table 2. According to the analysis performed using the SwissADME platforms³² and ADMETlab 3.0³⁴. The reference compounds SCO and GAL met al.l the filters used in evaluating their pharmacological potential (Lipinski, Ghose, Veber, Egan, Muegge, Pfizer, GSK, and Golden Triangle), with no deviations. This profile gives them a high potential for development as therapeutic agents due to the balance between their physicochemical and pharmacokinetic properties. In contrast, most of the analyzed volatile compounds (including citral, geranial and neral, caryophyllene, germacrene D, limonene, pinene, and eucalyptol) exhibited partial compliance with the applied filters. Although the analyzed compounds fully complied with the Veber and Egan rules, which are associated with favorable intestinal permeability and oral absorption, several deviations were observed in the Ghose and Muegge filters. These deviations were primarily attributed to the low molecular mass (< 160 Da or < 200 Da) and the reduced number of heteroatoms. Such limitations may potentially diminish the affinity and specificity for binding to pharmacological targets.

Table 2 Predicted Pharmacokinetic properties and medicinal chemistry characteristics of the analyzed compounds, indicating their compatibility with a drug-like profile.

Several of the analyzed compounds, specifically β-pinene, α-pinene, caryophyllene, and germacrene D, displayed MLOGP values exceeding 4.15, indicating elevated lipophilicity. This high degree of hydrophobicity may negatively impact solubility in aqueous biological environments, resulting in suboptimal biodistribution. Evaluation of medicinal chemistry filters revealed that only SCO, GAL, and eucalyptol fully satisfied both the Pfizer and GlaxoSmithKline (GSK) filters, which are used to predict the risk of systemic toxicity and poor absorption. This suggests a comparatively higher pharmacological safety profile for these three molecules. Furthermore, assessment using the Golden Triangle rule, which integrates key physicochemical parameters such as logP and topological polar surface area (TPSA), identified six compounds: SCO, GAL, caryophyllene, germacrene D, β-caryophyllene, and α-humulene as falling within the favorable range for oral drug-likeness. Notably, despite varying performance across individual drug-likeness filters, all compounds demonstrated an estimated oral bioavailability score of 55%, indicating a moderate potential for oral absorption. These findings were consistent across both the SwissADME and ADMET lab 3.0 prediction platforms.

Pharmacokinetic and Pharmacological profiling of bioactive Melissa officinalis compounds

As summarized in Table 3, the in silico pharmacokinetic evaluation of the selected compounds was conducted using the pkCSM platform³⁵. The results revealed that all investigated molecules demonstrated high predicted human intestinal absorption (HIA) values, exceeding the 30% threshold typically considered favorable for oral bioavailability. Among the reference drugs, GAL exhibited a particularly high absorption rate (94.99%), while SCO showed a slightly lower but still favorable value (72.63%). Notably, the phytocompounds derived from M. officinalis, including citral (geranial and neral), eucalyptol, α- and β-pinene, and both β- and α-caryophyllene, exhibited excellent absorption rates, all surpassing 94%. Eucalyptol demonstrated the highest value (96.51%), indicating strong theoretical oral bioavailability. Regarding skin permeability, estimated via log Kp values, significant differences were observed among the compounds. SCO (− 4.10) and GAL (− 3.75) presented low values, suggesting limited capacity for transdermal delivery. In contrast, monoterpenes and sesquiterpenes such as germacrene D (− 1.43) and β-caryophyllene (− 1.58) had values approaching the threshold for high skin permeability (around − 2.5), indicating a greater potential for topical or transdermal administration. In terms of P-glycoprotein (P-gp) interactions, SCO was identified as a P-gp substrate, implying that its central nervous system (CNS) penetration might be restricted by active efflux mechanisms. Most natural compounds, including GAL and both citral isomers, were not predicted to be substrates. However, β-caryophyllene, α-humulene, limonene, and eucalyptol were flagged as P-gp substrates, which could potentially influence their tissue distribution and CNS access. Importantly, none of the analyzed molecules were identified as inhibitors of P-gp I or II isoforms, suggesting a low risk for transporter-mediated pharmacokinetic interactions. Overall, the data indicate that the majority of these compounds possess favorable oral absorption profiles. Moreover, compounds such as germacrene D (log Kp = − 1.43), β-caryophyllene (− 1.58), α-pinene (− 1.83), and limonene (− 1.72) show promise for transdermal delivery, expanding their potential use in systemic and neurological therapeutic applications—including alternative routes of administration that could facilitate CNS targeting, bypassing the blood-brain barrier (BBB).

Table 3 In Silico ADMET prediction results for reference drugs and major Melissa officinalis compounds, including estimates of human intestinal absorption, skin permeability, and P-glycoprotein interaction profiles.

The predicted distribution parameters revealed considerable variability in the apparent volume of distribution (VDss) among the analyzed compounds. Most compounds exhibited log VDss values above the 0.45 log L/kg threshold, indicative of a broad distribution beyond the vascular compartment and into peripheral tissues. The highest log VDss values were observed for GAL (0.89), β-pinene (0.685), α-pinene (0.667), and β-caryophyllene (0.652), suggesting an enhanced potential for extravascular tissue penetration. The fraction unbound (fu) in plasma, reflecting the proportion of drug not bound to plasma proteins and thus freely available for pharmacological activity, varied between 0.261 for germacrene D and 0.553 for eucalyptol. Compounds such as eucalyptol, α-pinene, and limonene exhibited relatively high free fractions, which may enhance their systemic bioavailability and interaction with molecular targets. Predicted BBB permeability was favorable for the majority of terpenic compounds, as reflected by log BB values greater than 0.3, commonly considered a threshold for efficient BBB penetration. Notably high values were recorded for β-pinene (0.818), α-pinene (0.791), germacrene D (0.723), limonene (0.732), and β-caryophyllene (0.733), suggesting strong potential for CNS accessibility and positioning these compounds as promising candidates for neurological therapeutic applications.

As illustrated in Fig. 2, the boiled-egg model was employed to visualize the predicted absorption and brain penetration profiles of the investigated compounds. This model integrates lipophilicity and polarity to estimate the likelihood of passive gastrointestinal absorption and BBB permeability. Among the analyzed compounds, SCO (A) was in the white area, indicating a high probability of passive intestinal absorption, and is marked with a red dot, denoting that it is not a P-glycoprotein (P-gp) substrate, which favors sustained systemic presence. In contrast, GAL (B) falls within the yellow-green area, signifying good BBB penetration, but is identified as a P-gp substrate (blue dot), suggesting it may be subject to active efflux and rapid elimination from the CNS. Several phytochemicals from MEO, including citral (geranial) (C) and citral (neral) (D), β-pinene (G), α-pinene (H), limonene (I), and eucalyptol (L), are positioned in the yellow region, indicative of favorable BBB permeability, and are marked with red dots, suggesting they are not P-gp substrates. This profile implies a potential for longer retention within the brain, supporting their relevance in neurotherapeutic strategies. On the other hand, caryophyllene (E), germacrene D (F), β-caryophyllene (J), and α-humulene (K) were located outside the optimal absorption and brain penetration zones (gray area). These compounds show moderate predicted intestinal absorption and lower BBB permeability, potentially limiting their CNS availability via conventional administration routes.

Fig. 2

Boiled-Egg predictive model illustrating passive gastrointestinal absorption and blood-brain barrier (BBB) permeability for the reference compounds (A) scopolamine (SCO) and (B) galantamine (GAL), as well as for selected bioactive constituents of MEO: (C) citral (geranial), (D) citral (neral), (E) caryophyllene, (F) germacrene D, (G) β-pinene, (H) α-pinene, (I) limonene, (J) β-caryophyllene, (K) α-humulene, and (L) eucalyptol. The model is based on molecular positioning in a WLOGP vs. TPSA graph and enables intuitive classification: the white area indicates a high probability of passive gastrointestinal absorption; the yellow area denotes a high probability of BBB permeability. These regions are not mutually exclusive. Each compound is also marked based on its interaction with P-glycoprotein (P-gp): blue dots (PGP+) indicate compounds likely to be actively effluxed by P-gp, while red dots (PGP–) indicate compounds not subject to P-gp efflux, suggesting longer retention and improved CNS availability.

The predicted CNS permeability (log PS) values for all analyzed compounds ranged between − 1.85 and − 2.97. Notably, citral (geranial and neral), β-pinene, and germacrene D surpassed the − 2.0 threshold, suggesting a higher ability to penetrate the central nervous system. In contrast, scopolamine (− 3.031) and galantamine (− 2.511) exhibited lower predicted permeability despite their clinical use in neurological disorders. These findings imply that certain volatile compounds may serve as effective alternatives for CNS-targeted delivery due to their superior passive permeability (Table 3). In terms of metabolic profiles, only SCO and GAL were identified as substrates for the cytochrome P450 enzyme CYP3A4, indicating their hepatic metabolism may be significantly influenced by co-administered CYP3A4 inhibitors or inducers. Conversely, none of the MEO constituents, including citral, caryophyllene, limonene, and eucalyptol, were identified as CYP3A4 substrates, suggesting greater metabolic stability and lower interindividual variability. Among all compounds, only β-caryophyllene was predicted to inhibit CYP1A2, which could affect the metabolism of other xenobiotics, although the clinical relevance of this interaction warrants further investigation. Regarding systemic elimination, the compounds demonstrated a broad range of predicted total clearance (log mL/min/kg). Germacrene D (1.420), α-humulene (1.282), and caryophyllene (1.088) exhibited the highest clearance values, suggesting rapid elimination from the body. In contrast, β-pinene (0.030), α-pinene (0.043), and limonene (0.213) showed much lower clearance rates, indicating longer systemic retention and potentially prolonged pharmacological action. Among all the compounds, only GAL was identified as a substrate for the renal organic cation transporter OCT2, suggesting active renal excretion; this was not the case for any of the phytocompounds, which appear to rely more on hepatic metabolism or passive diffusion. The toxicological profile was evaluated using ADMETlab 3.0³⁴ and pkCSM³⁵, included assessments of cardiotoxicity, hepatotoxicity, genotoxicity, neurotoxicity, nephrotoxicity, cytotoxicity, and hematotoxicity.

Cardiotoxicity, as evaluated by hERG channel inhibition, showed minimal concern. All compounds scored below the 0.1 alert threshold: SCO (0.084) and citral (neral) (0.027) were well within safe limits. Caryophyllene had a slightly elevated score (0.065) but remained under the acceptable range, suggesting low risk of arrhythmogenic potential.

For hepatotoxicity, SCO (0.429), GAL (0.461), and caryophyllene (0.394) all exhibited moderate Drug-Induced Liver Injury (DILI) risk, which may require attention in long-term use, although short-term administration appears safe.

Regarding genotoxicity, SCO and GAL showed low but notable risk values (0.609 and 0.739, respectively), suggesting minimal potential for DNA damage.

Carcinogenicity scores were low, with SCO (0.021) and GAL (0.428), indicating a very low risk of cancer induction. Neurotoxicity was more significant for the clinically used compounds: SCO (0.883) and GAL (0.744), indicating a notable risk of CNS toxicity that should be carefully monitored in therapeutic contexts.

Nephrotoxicity scores were low overall, with SCO (0.409) and caryophyllene (0.023) both indicating low renal toxicity risk. Cytotoxicity, evaluated via cell culture assays, was minimal for most compounds: SCO (0.304) and GAL (0.069), confirming low impact on cellular viability. Hematotoxicity risk was also low, with SCO (0.412) and GAL (0.505), suggesting limited adverse effects on the hematologic system.

Immunotoxicity, evaluated in the RPMI-8226 model, showed low risk across compounds: SCO (0.053) and GAL (0.097), indicating minimal immunosuppressive or immunotoxic potential. Finally, bioaccumulation potential, assessed via the bioconcentration factor (BCF), was low for all compounds, including SCO (0.469) and GAL (0.486), suggesting a low risk of long-term accumulation and toxicity.

Computational predictions regarding toxicity and bioactivity

The molecular interaction potential and bioactivity profiles of the selected compounds were assessed using the ProTox-II platform³⁶. This tool provides predictive scores based on the likelihood of interactions with various biological targets and signaling pathways associated with neurotoxicity, metabolic regulation, and cellular stress responses³⁷.

The evaluated compounds, including SCO, GAL, citral (geranial and neral), caryophyllene, germacrene D, β-pinene, α-pinene, limonene, β-caryophyllene, α-humulene, and eucalyptol, were analyzed in the context of multiple Molecular Initiating Events (MIEs), as well as their involvement in cellular defense and stress response pathways.

These predictive interactions, detailed in Table 4, provide insights into the possible mechanisms through which the compounds may exert therapeutic or adverse effects, particularly within the central nervous system and in systemic metabolic pathways.

Table 4 Predicted molecular bioactivity profiles of the analyzed volatile compounds, highlighting potential interactions with key cellular targets involved in oxidative stress, inflammation, and neurotransmission pathways.

According to Table 4, only three compound-target pairs were predicted to fall within active clusters, indicating a higher confidence in potential biological interaction. Specifically, citral (both geranial and neral isomers) demonstrated a predicted interaction with acetylcholinesterase (AChE), suggesting a possible inhibitory effect that may underlie its proposed neuroprotective role. In addition, SCO was correctly associated with GABA-A receptor (GABAR) modulation, a well-established pharmacological target, thereby confirming the predictive validity of the computational model. In contrast, the remaining compound, target predictions, despite some showing elevated probability scores, were categorized into inactive clusters. This classification implies a low likelihood of verified biological activity, particularly in cases where the molecular structures did not share sufficient similarity with known active ligands from the model’s training data. These findings emphasize that high theoretical interaction scores alone are not definitive indicators of functional activity but rather represent hypothetical affinities that require experimental validation. Consequently, the predicted neuroprotective, anti-inflammatory, or antioxidant effects of the volatile compounds, especially those not placed in active clusters, should be further investigated through in vitro and in vivo studies to confirm their pharmacological relevance and therapeutic potential.

Impact of MEO on the response of amnesic zebrafish in the novel tank diving test (NTT)

Following chronic exposure to MEO at doses of 150 and 300 µL/L for 7 days, zebrafish displayed no signs of behavioral abnormalities, toxicity, or mortality, confirming the safety profile of the tested concentrations. The effects of SCO (100 µM) and MEO (150 and 300 µL/L) on anxiety-like behavior were evaluated using the NTT, as illustrated in Fig. 3. Representative locomotor tracking patterns clearly show altered swimming behavior between the top and bottom zones of the tank across treatment groups (Fig. 3A). Statistical analysis by one-way ANOVA revealed significant treatment effects for multiple behavioral parameters: latency to enter the top zone (F(4, 45) = 8.348, p < 0.0001) (Fig. 3B), time spent in the top/bottom zones (F(4, 45) = 21.30, p < 0.0001) (Fig. 3C), distance traveled in the top/bottom zones (F(4, 45) = 6.482, p < 0.0001) (Fig. 3D), swimming velocity (F(4, 45) = 11.37, p < 0.0001) (Fig. 3F), freezing duration (F(4, 45) = 7.196, p < 0.0001) (Fig. 3G). However, no significant differences were observed among the groups in terms of total distance traveled (F(4, 45) = 1.149, p = 0.3459) (Fig. 3E), suggesting that overall locomotor capacity was unaffected by treatment. These results indicate that MEO administration modulates specific anxiety-like behaviors in zebrafish without impairing general locomotion, supporting its anxiolytic potential.

Fig. 3

Effects of Melissa officinalis essential oil (MEO, 150 and 300 µL/L) on anxiety-like behavior in zebrafish, as assessed by the Novel Tank Test (NTT). (A) Representative swim trace patterns illustrating vertical exploration; (B) Latency to enter the top zone (s); (C) Time spent in the top-to-bottom zone ratio (s); (D) Distance traveled in the top-to-bottom zone ratio (m); (E) Total distance traveled (m); (F) Swimming velocity (m/s); (G) Freezing duration (s). Data are presented as means ± SEM (n = 10 per group). Statistical significance was determined using one-way ANOVA followed by Tukey’s post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Galantamine (GAL, 1 mg/L) was used as a positive control for anxiolytic activity.

Tukey’s post hoc analysis revealed that SCO treatment significantly impaired anxiety-related behaviors in zebrafish compared to the control group. SCO exposure resulted in a prolonged latency to enter the top zone (p < 0.05; Fig. 3B), a reduction in time spent in the top/bottom ratio (p < 0.001; Fig. 3C), and a decrease in distance traveled in the top zone (p < 0.05; Fig. 3D). Additionally, SCO significantly increased swimming velocity (p < 0.0001; Fig. 3F) and freezing duration (p < 0.01; Fig. 3G), while it had no significant effect on the total distance traveled (p = 0.3459; Fig. 3E). These results confirm SCO’s anxiogenic effects, consistent with its known neuropharmacological profile. Treatment with GAL (1 mg/L), a standard drug for mild to moderate dementia³⁸, significantly attenuated the anxiogenic effects of SCO. GAL treatment reduced latency to the top zone (p < 0.001; Fig. 3B) and increased both time and distance spent in the top zone (p < 0.0001 and p < 0.01, respectively; Figs. 3C, D). It also normalized swimming velocity (p < 0.0001; Fig. 3F) and reduced freezing behavior (p < 0.05; Fig. 3G), compared to SCO alone, suggesting effective anxiolytic modulation. Importantly, the co-administration of MEO with SCO also significantly reversed anxiety-like behaviors, with more pronounced effects observed at the 300 µL/L dose. MEO treatment reduced latency to enter the top zone (p < 0.05; Fig. 3B), increased time spent in the top zone (p < 0.0001; Fig. 3C) and enhanced vertical exploration by increasing the distance traveled in the top zone (p < 0.01 for 150 µL/L; p < 0.001 for 300 µL/L; Fig. 3D). MEO also significantly normalized swimming speed (p < 0.0001 and p < 0.001 for 150 and 300 µL/L, respectively; Fig. 3F) and reduced freezing duration (p < 0.01 and p < 0.001; Fig. 3G). Similar to GAL, MEO did not affect total distance traveled, indicating that the observed effects were specific to anxiety-related behavior, not general locomotion. These findings support the anxiolytic potential of MEO in a zebrafish model of SCO-induced anxiety and are consistent with previous clinical evidence. A clinical study demonstrated that administration of 1.5 g/day of powdered M. officinalis leaves for seven days led to a 49% reduction in anxiety levels and 54% improvement in sleep quality in patients undergoing coronary artery bypass surgery³⁹. Additionally, a systematic review and meta-analysis confirmed that M. officinalis significantly alleviates symptoms of anxiety and depression with a favorable safety profile, although high inter-study variability warrants further investigation⁴⁰.

Impact of MEO on the response of amnesic zebrafish in the novel approach test (NAT)

Behavioral assays in fish models, particularly zebrafish, are widely used to assess anxiety-like behavior, with the NAT being one of the most reliable tools for evaluating locomotion and anxiety in individually tested fish⁴¹. The NAT measures parameters such as time spent in the central (inner) zone and peripheral (outer) zone, which are indicative of anxiety-related exploratory behavior. The effects of SCO (100 µM) and MEO (150 and 300 µL/L) on anxiety-like behavior were assessed using the NAT, and results are shown in Fig. 4. Representative swim trace patterns illustrate treatment-induced differences in zone preference and exploration (Fig. 4A). One-way ANOVA revealed that treatment significantly affected multiple behavioral parameters: time spent in the outer zone (F(4, 45) = 7.426, p = 0.0001) (Fig. 4B), time spent in the inner zone (F(4, 45) = 15.53, p < 0.0001) (Fig. 4C), distance traveled (F(4, 45) = 26.11, p < 0.0001) (Fig. 4D), latency to enter the inner zone (F(4, 45) = 10.05, p < 0.0001) (Fig. 4E). Swim trace analysis (Fig. 4A) showed that SCO-treated zebrafish exhibited reduced exploration of the inner zone, a behavioral marker of heightened anxiety. In contrast, zebrafish treated with GAL (1 mg/L) or MEO (150 and 300 µL/L) display enhanced exploration of the inner zone, suggesting an anxiolytic effect. These effects were especially evident in SCO-pretreated animals co-treated with MEO, indicating that chronic MEO administration may reverse SCO-induced anxiety-like behavior.

Fig. 4

Effects of Melissa officinalis essential oil (MEO; 150 and 300 µL/L) on anxiety-like and novelty-induced behaviors in the NAT in zebrafish. Galantamine (GAL; 1 mg/L) was used as a positive control. (A) Representative swim trace plots illustrating exploratory behavior; (B) Time spent in the outer zone (s); (C) Time spent in the inner zone (s); (D) Total distance traveled (m); (E) Latency to enter the inner zone (s). Data are presented as means ± SEM (n = 10 per group). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Tukey’s post hoc analysis demonstrated that SCO significantly exacerbated anxiety-like behaviors in zebrafish during the NAT. Specifically, SCO reduced the time spent in the inner zone (p < 0.0001; Fig. 4C), with a corresponding increase in time spent in the outer zone (p < 0.001; Fig. 4B), and decreased latency to approach the central object (p < 0.01; Fig. 4E). These behavioral changes reflect an anxiogenic response. Notably, SCO did not significantly alter total locomotor activity, as shown by the absence of differences in total distance traveled (Fig. 4D), suggesting that the observed effects were specific to anxiety and not due to impaired mobility. In contrast, treatment with MEO, particularly at 300 µL/L, reversed the SCO-induced behavioral alterations. MEO decreased the time spent in the outer zone (p < 0.05; Fig. 4B), while increasing time spent in the inner zone (p < 0.01; Fig. 4C), indicating a reduction in anxiety-like behavior. Additionally, MEO shortened the latency to initiate exploration of the novel object, with significance observed at both doses (p < 0.01 for 150 µL/L and p < 0.0001 for 300 µL/L; Fig. 4E). However, MEO treatment also exhibited a hypolocomotor effect, as reflected by a reduction in total distance traveled at both concentrations (Fig. 4D), suggesting a possible sedative component to its anxiolytic action. These results are consistent with previous studies in mammalian models. For example, Stojanović et al.⁴² reported significant anxiolytic effects of MEO in mice across several behavioral tests, although its main constituent, citronellal, showed mixed effects, including both anxiolytic and locomotor-inhibiting properties. Similarly, Ghazizadeh et al.¹⁷ demonstrated that a hydroalcoholic extract of M. officinalis significantly alleviated anxiety- and depression-like behaviors in a mouse model of restraint stress by reducing oxidative stress and enhancing antioxidant and anti-apoptotic markers in the prefrontal cortex and hippocampus. These findings collectively support the potential of M. officinalis as a natural anxiolytic agent, with effects that may involve both behavioral modulation and neuroprotective biochemical pathways.

Impact of MEO on the response of amnesic zebrafish in the Y-maze test

Deficits in executive functions, including working memory and cognitive flexibility, are hallmark features of numerous neurodegenerative and psychiatric disorders. Addressing these impairments is critical for the development of more effective treatments⁴³. The use of robust, translationally relevant animal models with validated behavioral paradigms is essential for accurately modeling cognitive deficits observed at various stages of disease progression. Zebrafish have emerged as a valuable model in this context due to their rich behavioral repertoire and increasing standardization of cognitive tests⁴⁴. Among the most employed apparatuses for assessing spatial memory and cognition in zebrafish and rodents are maze-based tasks such as the plus maze, T-maze, and Y-maze^45,46. The Y-maze test is widely used to evaluate working memory, novelty-seeking behavior, and spontaneous alternation performance. In the present study, SCO-treated zebrafish showed reduced exploration of the novel arm of the Y-maze compared to the control group, as illustrated by representative trace plots (Fig. 5A). In contrast, zebrafish co-treated with either GAL (1 mg/L) or MEO (150 and 300 µL/L) demonstrated increased exploratory behavior in the novel arm, suggesting a rescue of SCO-induced cognitive deficits. Statistical analysis using one-way ANOVA revealed the following treatment effects: no significant differences were observed for the number of arm entries (F(4, 45) = 1.717, p = 0.1629) (Fig. 5B), indicating that general locomotor activity was not substantially affected. However, significant treatment effects were found for: distance traveled (F(4, 45) = 7.419, p = 0.0001) (Fig. 5C), turn angle (F(4, 45) = 6.939, p = 0.0002) (Fig. 5D), number of line crossings (F(4, 45) = 3.018, p = 0.0274) (Fig. 5E), spontaneous alternation percentage (F(4, 45) = 7.421, p = 0.0001) (Fig. 5F), time spent in the novel arm (F(4, 45) = 13.29, p < 0.0001) (Fig. 5G). These results indicate that while total movement between arms remained stable, SCO impaired cognitive performance, particularly spatial memory and novelty-seeking behavior. Conversely, both GAL and MEO treatments significantly improved cognitive parameters, suggesting restoration of memory function and increased exploratory drive.

Fig. 5

Effects of Melissa officinalis essential oil administration (MEO, 150 and 300 µL/L) on spatial memory and novelty-seeking behavior in the Y-maze test. Galantamine (GAL, 1 mg/L) was used as a positive control. (A) Representative trajectories of zebrafish in the Y-maze; (B) number of arm entries; (C) distance traveled (m); (D) turn angle (°); (E) number of line crossings; (F) spontaneous alternation percentage (%), and (G) time spent in the novel arm (s). For the Y-maze arms: A- start arm, B – the other arm, and C – novel arm. Data are presented as mean ± SEM (n = 10 per group). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test: *p < 0.05, **p < 0.01, and ****p < 0.0001.

Post hoc analysis using Tukey’s test revealed that acute administration of SCO significantly impaired spatial memory in zebrafish, as evidenced by a reduction in spontaneous alternation (%) (p < 0.01; Fig. 5F) and time spent in the novel arm (p < 0.0001; Fig. 5G). Except for a modest increase in turn angle (p < 0.05; Fig. 5D), SCO treatment did not significantly alter other parameters of locomotor activity compared to control fish.

In contrast, chronic administration of MEO at concentrations of 150 and 300 µL/L followed by acute SCO exposure (100 µM) significantly ameliorated memory deficits. MEO treatment improved spontaneous alternation (p < 0.01 at 150 µL/L; p < 0.0001 at 300 µL/L; Fig. 5F) and time spent in the novel arm (p < 0.05 at 150 µL/L; p < 0.0001 at 300 µL/L; Fig. 5G). Galantamine (GAL, 1 mg/L), used as a reference drug, showed a significant effect only on spontaneous alternation (p < 0.05; Fig. 5F). Moreover, MEO exerted additional effects on locomotor parameters, including increased distance traveled (p < 0.01 at 300 µL/L; Fig. 5C) and elevated turn angle (p < 0.01 at 150 µL/L; p < 0.0001 at 300 µL/L; Fig. 5D), suggesting a broader neuromodulatory role.

Previous studies corroborate these findings. Soodi et al.⁴⁷ demonstrated that M. officinalis extracts significantly improved learning and memory in both naïve and SCO-treated rats, particularly at a dose of 200 mg/kg, likely through cholinergic mechanisms. Similarly, Naser et al.⁴⁸ reported cognitive enhancement in diabetic rats following M. officinalis treatment, associated with increased hippocampal expression of brain-derived neurotrophic factor (BDNF) and nitric oxide synthase, along with improved Y-maze performance. Furthermore, Kennedy et al.⁴⁹ found that a high dose (1600 mg) of M. officinalis enhanced memory performance and induced calmness in healthy human subjects, although lower doses resulted in decreased cognitive function, highlighting dose- and preparation-dependent effects.

Impact of MEO on the response of amnesic zebrafish in NOR

Recognition memory in various model organisms is commonly assessed using the NOR test, also referred to as the novel object preference test. In this paradigm, subjects are initially exposed to two identical objects. After a retention interval, they are reintroduced to the testing arena containing one familiar object (FO) and one novel object (NO). A preference for the NO reflected by increased exploration time indicates intact recognition memory⁵⁰. The conventional tracking data of zebrafish locomotion (Fig. 6A) illustrates clear differences in exploration patterns between FO and NO during the NOR test. Notably, zebrafish acutely exposed to SCO (100 µM) exhibited a marked preference for the FO, indicative of impaired recognition memory. One-way ANOVA revealed a significant effect of treatment on NO preference, as reflected by the exploration percentage (F(4, 45) = 5.660, p = 0.0009; Fig. 6B).

Fig. 6

Effects of Melissa officinalis essential oil administration (MEO, 150 and 300 µL/L) on recognition memory in the novel object recognition (NOR) test. Galantamine (GAL, 1 mg/L) was used as a positive control. (A) Representative tracking plots of zebrafish exploratory behavior in the NOR arena; (B) Novel object (NO) preference, expressed as percentage of time spent exploring the novel object. FO – familiar object; NO – Novel object. Data are presented as mean ± SEM (n = 10 per group). Statistical significance was assessed by one-way ANOVA followed by Tukey’s post hoc test: *p < 0.05, and ***p < 0.001.

Tukey’s post hoc analysis indicated that SCO significantly impaired recognition memory in zebrafish, as evidenced by a decreased preference for the NO (p < 0.001; Fig. 6B). In contrast, chronic administration of MEO at both tested concentrations significantly improved NO preference (p < 0.05), reflecting enhanced recognition memory. A comparable effect was observed in the GAL-treated group (p < 0.05; Fig. 6B). These findings are supported by previous research. Ozarowski et al.⁵¹ demonstrated that subchronic administration of a 50% ethanol extract of M. officinalis (200 mg/kg), alongside huperzine A (0.5 mg/kg), improved long-term memory in SCO-induced memory-impaired rats. The memory-enhancing effects of M. officinalis were associated with significant reductions in cortical and hippocampal mRNA expression of AChE and beta-secretase (BACE-1), although the extract had minimal influence on AChE enzymatic activity.

Impact of MEO on the response of amnesic zebrafish on brain ache activity

Current research suggests that enhancing cholinergic neurotransmission by inhibiting AChE activity remains one of the most effective therapeutic strategies for AD, despite the complex and not yet fully elucidated pathophysiology of the disorder. AChE plays a critical role in the central nervous system by hydrolyzing ACh, thus regulating cholinergic synaptic transmission. Clinically approved AChE inhibitors such as rivastigmine, donepezil, and galantamine are widely employed to manage cognitive symptoms in AD. Consequently, targeting AChE continues to be a promising approach in the development of novel therapeutic agents for the treatment of AD⁵². In our study, zebrafish exposed to SCO (100 µM) exhibited a significant elevation in brain AChE activity compared to control fish (p < 0.0001; Fig. 7A), indicating cholinergic dysfunction. Conversely, chronic treatment with MEO significantly attenuated this SCO-induced increase in AChE activity at both tested concentrations (p < 0.0001 vs. SCO; Fig. 7A). Similarly, GAL (1 mg/L) exerted a robust anti-AChE effect, significantly reducing AChE activity compared to the SCO group (p < 0.0001; Fig. 7A). These findings suggest that MEO may exert cognitive-enhancing effects, at least in part, through inhibition of AChE activity, thereby restoring cholinergic balance in the scopolamine-induced amnesic zebrafish model.

Fig. 7

Effects of Melissa officinalis essential oil administration (MEO, 150 and 300 µL/L) on enzymatic antioxidant defense, oxidative stress markers, and cholinergic function in the zebrafish brain. Galantamine (GAL; 1 mg/L) was used as a positive control. (A) Acetylcholinesterase (AChE) activity; (B) Superoxide dismutase (SOD) activity; (C) Catalase (CAT) activity; (D) Glutathione peroxidase (GPX) activity; (E) Total glutathione (GSH) content; (F) Protein carbonyl content; (G) Malondialdehyde (MDA) levels. Data are presented as mean ± SEM (n = 3 per group). Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test: **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Impact of MEO on the response of amnesic zebrafish to brain oxidative stress

Oxidative stress plays a pivotal role in the onset and progression of aging and age-related neurodegenerative diseases, including AD. It results from a redox imbalance characterized by excessive production of reactive oxygen species (ROS), impaired antioxidant defense mechanisms, and dysregulated cellular redox signaling. Accumulating evidence suggests that elevated ROS levels in the brain contribute significantly to the pathogenesis of late-onset neurodegenerative disorders, particularly those involving neuroinflammation, such as AD. Although the precise molecular cascade linking ROS to AD remains incompletely understood, oxidative stress is known to damage key cellular components, including proteins, lipids, and DNA. Given the brain’s high oxygen consumption and abundance of polyunsaturated fatty acids, it is especially vulnerable to oxidative damage. These changes disrupt neurotransmitter regulation and intracellular signaling, ultimately leading to neuronal and glial cell dysfunction and triggering apoptosis. Furthermore, increased ROS levels have been strongly correlated with tau hyperphosphorylation, tau aggregation, and cytoskeletal abnormalities, factors that promote the formation of beta-amyloid (Aβ) plaques and contribute to AD pathology^53,54,55. In the present study, acute exposure of zebrafish to SCO (100 µM) for 30 min significantly reduced the brain-specific activities of key antioxidant enzymes, including SOD (p < 0.001; Fig. 7B), CAT (p < 0.001; Fig. 7C), and GPX (p < 0.0001; Fig. 7D), compared to control fish. SCO exposure also markedly decreased total GSH levels (p < 0.0001; Fig. 7E) and significantly increased markers of oxidative damage, namely MDA (p < 0.0001; Fig. 7G) and carbonylated proteins (p < 0.0001; Fig. 7F). These changes indicate the induction of oxidative stress and the associated biomolecular damage in the zebrafish brain. Conversely, chronic treatment with MEO significantly counteracted these effects in a dose-dependent manner. MEO administration restored the activities of SOD (p < 0.01 at 150 µL/L; p < 0.001 at 300 µL/L; Fig. 7B), CAT (p < 0.01 at 150 µL/L; p < 0.001 at 300 µL/L; Fig. 7C), and GPX (p < 0.0001 at both concentrations; Fig. 7D) compared to the SCO group. Furthermore, MEO significantly reduced levels of MDA (p < 0.01 at 150 µL/L; p < 0.0001 at 300 µL/L; Fig. 7G) and carbonylated proteins (p < 0.0001 at both doses; Fig. 7F). These findings suggest that MEO exerts robust antioxidant effects in a zebrafish model of SCO-induced cognitive dysfunction by enhancing endogenous antioxidant enzyme activity and attenuating oxidative damage to lipids and proteins.

Correlational studies of behavioral and biochemical parameters

To investigate the interrelationships between behavioral performance, enzymatic antioxidant activity, and oxidative stress, Pearson’s correlation coefficient (r) was calculated. The variables analyzed included behavioral parameters such as latency to swim vertically toward the top zone in NTT, time spent in the inner zone in the NAT, time spent in the novel arm of the Y-maze, and NO preference (%) in the NOR test. Biochemical markers included the specific activities of AChE, SOD, CAT, and GPX, as well as levels of carbonylated proteins and MDA in zebrafish brain tissue. The results revealed several significant correlations between behavioral performance and MDA levels. A moderate positive correlation was observed between latency to enter the top zone in the NTT and MDA concentration (r = 0.6723, p < 0.01; Fig. 8A), suggesting that increased oxidative stress is associated with heightened anxiety-like behavior. In contrast, negative correlations were identified between MDA levels and time spent in the inner zone during NAT (r = − 0.6092, p < 0.05; Fig. 8B), time spent in the novel arm of the Y-maze (r = − 0.5961, p < 0.05; Fig. 8C), and novel object preference in the NOR test (r = − 0.5371, p < 0.01; Fig. 8D), indicating that higher oxidative stress is associated with impairments in exploration and recognition memory. Additionally, biochemical parameters showed strong associations with MDA levels. A significant positive correlation was observed between AChE activity and MDA levels (r = 0.7239, p < 0.01; Fig. 8E), while inverse correlations were found between MDA levels and the activities of SOD (r = − 0.7847, p < 0.001; Fig. 8F), CAT (r = − 0.9253, p < 0.0001; Fig. 8G), and the levels of carbonylated proteins (r = − 0.8648, p < 0.0001; Fig. 8H). These findings highlight a strong link between oxidative stress and both behavioral deficits and neurochemical imbalances in the SCO-induced zebrafish model of cognitive dysfunction.

Fig. 8

Pearson correlation analysis between behavioral and biochemical parameters in zebrafish. Data represent the correlation between MDA levels and: (A) Latency to enter the top zone in the NTT (n = 3, r = − 0.6723, p < 0.01); (B) Time spent in the inner zone during the NAT (n = 3, r = − 0.6092, p < 0.05); (C) Time spent in the novel arm of the Y-maze (n = 3, r = − 0.5961, p < 0.05); (D) Novel object preference (%) in the NOR test (n = 3, r = − 0.8266, p < 0.0001); (E) Acetylcholinesterase (AChE) activity (n = 3, r = 0.7239, p < 0.01); (F) Superoxide dismutase (SOD) activity (n = 3, r = − 0.7847, p < 0.001); (G) Catalase (CAT) activity (n = 3, r = − 0.9253, p < 0.0001); (H) Carbonylated protein content (n = 3, r = 0.8648, p < 0.0001).

These correlations suggest a complex interplay between behavioral outcomes, enzymatic antioxidant defenses, and lipid peroxidation, underscoring key pathophysiological mechanisms involved in neurodegeneration and therapeutic response. Given its ability to mitigate oxidative stress and improve behavioral performance, MEO emerges as a promising candidate for alleviating symptoms associated with amnesia and anxiety.

Discussion

This study demonstrates that MEO exerts significant neuroprotective and anxiolytic effects in a SCO-induced zebrafish model of cognitive dysfunction. These outcomes are consistent across behavioral assays and biochemical markers, suggesting a strong therapeutic potential for MEO as a natural alternative to conventional cholinesterase inhibitors such as GAL. Importantly, our protocol evaluated MEO as a pretreatment in the context of an acute SCO challenge; thus, the observed effects should be interpreted as a reversal of SCO-induced changes, not baseline actions of MEO in naïve fish.

Our findings should be interpreted within the context of the SCO model. SCO primarily impairs cognition by reversibly blocking muscarinic cholinergic transmission, and while it does not produce overt neuronal death or synapse loss, it reliably induces oxidative stress, cholinergic imbalance, and behavioral impairments that mimic early features of AD^24,56,57. The biochemical changes observed here, including increased AChE activity, decreased antioxidant enzyme defenses, and elevated lipid peroxidation, are therefore best understood as secondary consequences of cholinergic dysfunction and redox imbalance, rather than direct indicators of neurodegeneration. This distinction underscores the utility of the SCO model as a tool for probing the mechanisms of memory impairment and testing candidate neuroprotective agents, while also highlighting its limitations in recapitulating the full neurodegenerative cascade of AD.

The essential oil’s efficacy, confirmed through Y-maze, NOR, and NTT tests, appears to be mediated by multiple synergistic mechanisms, including inhibition of AChE, restoration of antioxidant enzyme activity, and attenuation of oxidative damage.

The zebrafish model used in this study effectively mimics cholinergic deficits and behavioral impairments associated with neurodegenerative diseases, particularly AD. The observed SCO-induced cognitive impairments parallel previous findings, and the MEO-mediated reversal of these deficits supports earlier reports of its memory-enhancing properties⁵⁸. Notably, MEO restored performance in spontaneous alternation and object recognition tasks, behaviors that are directly linked to hippocampal and prefrontal cortex function, reinforcing the central role of the cholinergic system in these processes.

Interpreting the anxiolytic-like action of MEO requires situating it within the mixed literature on SCO and anxiety. In rodents, SCO generally increases anxiety-like behavior, whereas zebrafish studies have reported both anxiogenic and anxiolytic-like outcomes depending on assay parameters^28,59. Our NTT/NAT data therefore reinforce the need to read SCO’s affective effects as model- and context-dependent, while showing that MEO consistently reverses SCO-evoked anxiety-like readouts in zebrafish.

Notably, independent evidence supports an anxiolytic profile for M. officinalis in people: randomized and controlled clinical studies (oral preparations or inhaled essential oil) report reductions in state anxiety in cardiac surgery, hemodialysis, and metabolic cohorts, and broader reviews synthesize anxiolytic signals with good tolerability^60,61,62. Together with preclinical summaries pointing to terpenoid constituents (e.g., citronellal/citral) as candidate actives, these data provide external validity for the anxiolytic-like effects we observed⁶³. Oxidative stress is a hallmark of AD, contributing to neuronal damage and dysfunction through lipid peroxidation, protein oxidation, and mitochondrial damage⁶⁴. In this study, SCO administration significantly increased MDA and protein carbonyls, while reducing antioxidant enzyme activities such as SOD, CAT, and GPX. These changes were effectively reversed by MEO, indicating potent antioxidant activity. The oil’s major constituents - citral, β-caryophyllene, and limonene have well-documented free radical-scavenging and cytoprotective properties, and their synergistic activity likely underpins the overall efficacy of the oil^65,66.

Mechanistically, our biochemical results (decrease of AChE activity; increase of SOD/CAT/GPX activities; decrease of MDA/carbonyls) suggest that restoring cholinergic system activity and mitigating oxidative stress are plausible pathways for MEO’s anxiolytic-like action. This is consistent with reports that NTT metrics capture anxiety rather than nonspecific hypo-/hyperlocomotion when appropriately controlled, and with evidence linking redox imbalance to heightened anxiety phenotypes²⁹. While SCO does not recapitulate neurodegeneration per se, its cholinergic blockade reliably unmasks anxiety- and memory-related deficits that can be rescued by agents with cholinergic and antioxidant activity, as observed here for MEO.

Furthermore, in silico ADMET predictions support the pharmacokinetic viability of MEO constituents. Compounds such as citral, eucalyptol, and α-pinene exhibited good BBB permeability and favorable safety profiles. The lack of mutagenicity, hepatotoxicity, or neurotoxicity in computational models reinforces their translational potential, especially for chronic use. These findings align with current efforts in drug discovery, where natural compounds with multi-target actions are being prioritized for complex diseases like AD⁶⁷.

The behavioral effects observed, improved exploration, reduced latency, and normalized anxiety metrics, also suggest that MEO modulates stress-responsive neural circuits. Prior studies have linked MEO to activation of the GABAergic system and reduced corticosterone levels in rodents⁶⁸, mechanisms that could explain the anxiolytic effects seen here. Interestingly, the correlation analysis in our study demonstrated strong associations between behavioral outcomes and oxidative stress markers, supporting the hypothesis that oxidative stress underlies anxiety-like and memory-deficit phenotypes.

Study limitations

Despite the promising findings, several limitations should be acknowledged. First, the study employed only a single animal model (zebrafish), which, although well-suited for behavioral and neurochemical screening, does not fully replicate the complexity of mammalian neurodegeneration. Second, the SCO-induced model reflects an acute and reversible cholinergic deficit rather than the chronic, multifactorial pathology characteristic of AD. Third, while in silico analyses provided useful insights into pharmacokinetic and toxicological profiles of major constituents, these predictions require validation through in vivo pharmacokinetic and safety studies. Fourth, we did not include an MEO-alone cohort; therefore, attribution should be limited to MEO antagonism of SCO-evoked behavioral and biochemical alterations, and future confirmatory work will include MEO-only groups to characterize baseline effects. Finally, sex differences were not examined, and the relatively small sample size may restrict the generalizability of behavioral findings. Future studies should aim to address these limitations by incorporating additional models, longer treatment durations, and mechanistic investigations at the molecular level. Because SCO’s effects on anxiety vary across species and paradigms, future work should include dose-response and time-course designs and benchmark anxiolytics to further anchor MEO’s behavioral profile²⁹.

Taken together, the results of this study affirm the therapeutic relevance of MEO in managing both cognitive and emotional disturbances. While our study is limited to an aquatic model and the essential oil’s volatile fraction, the data lay a foundation for future research in mammalian systems and clinical contexts. Investigating downstream molecular targets such as BDNF, Nrf2, and synaptic proteins will further clarify MEO’s mechanism of action. Additionally, the exploration of non-volatile bioactives in M. officinalis, including rosmarinic acid and flavonoids, may yield further insights into the plant’s full therapeutic potential.

In conclusion, MEO represents a promising phytotherapeutic agent with multi-targeted effects against cholinergic dysfunction and oxidative stress. Its behavioral, biochemical, and computational profiles support its candidacy for further development in the treatment of AD and related neurodegenerative conditions.

Methods

Ethics statement and animal husbandry

Fifty adult wild-type short-fin zebrafish (Danio rerio), aged 5–7 months with an approximately equal sex ratio, were sourced from the European Zebrafish Resource Center at the Institute of Toxicology and Genetics in Germany. Upon arrival, fish were acclimatized for a minimum of 14 days in a dedicated zebrafish facility at the Faculty of Biology, Alexandru Ioan Cuza University of Iași, Romania, following established husbandry guidelines⁶⁹. Fish were housed in groups of 10 in 10 L glass tanks with dechlorinated tap water, maintained in a recirculating system. Water parameters were monitored twice daily to ensure optimal conditions: pH 7.0-7.5, dissolved oxygen 8 ± 1 mg/L, conductivity 1500–1600 µS/cm, and ammonia/nitrite levels < 0.001 mg/L. The lighting regimen followed a 14-h light/10-h dark cycle. Fish were fed a commercial flake diet (Norwin Norvital, Norwin, Gadstrup, Denmark) twice daily (06:00 and 17:00) ad libitum for 10 min, with the air pump paused during feeding to improve consumption. All procedures were conducted in compliance with Directive 2010/63/EU of the European Parliament and the Council on the protection of animals used for scientific purposes. The experimental protocol was reviewed and approved by the Animal Research Ethics Committee of the Faculty of Biology, Alexandru Ioan Cuza University of Iași (approval number: BIO-UAIC-02, dated June 30, 2020). Every effort was made to minimize animal suffering, following the principles of the 3Rs: replacement, reduction, and refinement.

Experimental design and treatments

Zebrafish were randomly assigned to five experimental groups (n = 10 fish per group): (I) Control group – maintained under standard housing conditions without any pharmacological intervention; (II) Scopolamine group (SCO, 100 µM) – received an acute exposure to SCO hydrobromide (Sigma-Aldrich, Darmstadt, Germany) at a concentration of 100 µM, administered 30 min prior to behavioral testing (refer to Fig. 9 for administration details). SCO served as a negative control due to its well-established anticholinergic effects, which disrupt central cholinergic neurotransmission and induce cognitive impairments⁷⁰; (III) Galantamine group (GAL, 1 mg/L) – exposed to GAL hydrobromide (Sigma-Aldrich, Darmstadt, Germany) at 1 mg/L for 30 min, followed by acute SCO treatment (100 µM) administered 30 min later. GAL, a clinically approved AChE inhibitor, was used as a positive control to enhance synaptic acetylcholine (ACh) levels and improve cognitive performance⁷¹; (IV) MEO group – 150 µL/L – chronically exposed to MEO at a concentration of 150 µL/L via water immersion for 21 consecutive days, with routine water changes to maintain consistent exposure, and (V) MEO group – 300 µL/L – received the same chronic treatment protocol as above, using a higher concentration of 300 µL/L MEO. All exposures were conducted under identical environmental conditions, and treatment schedules are illustrated in Fig. 9. For groups IV and V, fish were chronically preexposed to MEO and then received an acute SCO challenge (100 µM, 30 min) immediately prior to each behavioral test to assess antagonism of SCO-evoked alterations. An MEO-alone cohort was not included in this study design. The design aimed to compare the cognitive and behavioral effects of MEO with those of GAL in a SCO-induced zebrafish model of cognitive impairment. The data analysis and presentation, as well as this study’s experimental design, all followed the ARRIVE guidelines⁷² for planning and organizing animal testing and research, respectively.

Fig. 9

Schematic representation of the experimental design of the study.

Behavioral testing

Zebrafish behavior was recorded using a Logitech C922 Pro HD Stream webcam (Logitech, Lausanne, Switzerland). Video recordings were analyzed offline using ANY-maze software version 7.48 (Stoelting Co., Wood Dale, IL, USA).

Novel tank diving test (NTT)

The NTT was used to evaluate anxiety-like behavior in zebrafish, following the protocol described by Dumitru et al.⁷³. This assay exploits the species’ natural tendency to initially dwell near the bottom of a novel environment, a behavior that diminishes as anxiety levels decrease. In anxiogenic states, zebrafish exhibit prolonged bottom-dwelling, increased immobility, and reduced exploration of the upper zone⁷⁴. To standardize the analysis, the tank was virtually divided into top and bottom zones. Zebrafish were tested individually in a 6-min session, during which exploratory activity was tracked. Testing was performed sequentially across treatment groups, with a 2-day interval between behavioral tasks to minimize stress-induced carryover effects. All tests were conducted between 08:00 and 17:00 under consistent lighting and environmental conditions. The following behavioral parameters were quantified using ANY-maze: latency to enter the top zone (s); time spent in top vs. bottom zones (s); distance traveled in top vs. bottom zones (m); total distance traveled (m); average swimming velocity (m/s); freezing duration (s). Galantamine (GAL, 1 mg/L) served as the reference anxiolytic compound for the positive control group.

Novel approach test (NAT)

Locomotion and anxiety-like behavior were assessed using a NAT test adapted from Hamilton et al.²⁸. Fish were tested individually in an opaque white cylindrical tank (diameter, 34 cm; wall height, 15 cm) filled with 6 cm of system water maintained at 25–28 °C using a sub-tank heating pad (Seedling Heat Mat, Hydrofarm). Water was replaced between trials. Testing was conducted between 08:00 and 16:00. Each zebrafish underwent one 5-min trial. Animals were gently released facing a novel object (a 5 cm multicolored Lego figure) affixed at the center of the tank. A top-mounted video camera (Logitech, Lausanne, Switzerland) recorded behavior for subsequent analysis in ANY-maze. The virtual arena comprised two analysis zones: (i) a central (inner) zone defined as a 10 cm-diameter circle centered on the object, and (ii) a peripheral (thigmotaxis) zone defined as a 4.5 cm-wide annulus adjacent to the tank wall. The following dependent variables were extracted: time in inner zone (s) (exploration/approach), time in outer zone (s) (thigmotaxis/anxiety-like behavior), total distance travelled (m) (locomotor activity), and latency to enter the inner zone (s) (initiation of approach).

Y-maze test

Spatial working memory and novelty-seeking behavior were assessed using the Y-maze test, adapted from Cognato et al.⁷⁵. The apparatus consisted of a transparent Y-shaped glass tank with three arms (25 cm length × 8 cm width × 15 cm height). To aid spatial orientation, each arm was marked with distinct geometric visual cues (squares, triangles, or circles) affixed to the outer surfaces. All remaining exterior surfaces were covered with black adhesive foil to minimize external visual stimuli. The maze was filled with 3 L of home tank water for each test. Prior to behavioral evaluation, zebrafish were acclimated to the testing room in a separate holding tank for 5 min. The arms of the maze were randomly designated as: start arm – initial placement site; novel arm – closed during the first trial, opened during the second and permanent arm – open during both trials. The central junction of the maze served as a neutral zone and was excluded from the analysis⁷⁵. The Y-maze task comprised two sequential trials separated by a 1-h inter-trial interval. Trial 1 (familiarization phase, 5 min): fish were allowed to explore the start and permanent arms, while the novel arm remained closed. Trial 2 (test phase, 5 min): all three arms were accessible, enabling assessment of spatial recognition and novelty preference. The following parameters were quantified using ANY-maze software: Time spent in the novel arm (s) – an index of spatial novelty preference; spontaneous alternation (%) – calculated as successive entries into all three arms without repetition; number of line crossings – a proxy for spatial orientation; number of arm entries; total distance traveled (m), and turn angle (°) – to assess movement complexity.

Novel object recognition test (NOR)

Recognition memory was evaluated using the Novel Object Recognition (NOR) test, a validated paradigm for assessing cognitive performance in zebrafish⁷⁶. The test was conducted in a novel glass tank (30 × 30 × 30 cm) filled with 6 cm of clean, dechlorinated water. The NOR procedure consisted of three phases: (1) Habituation phase (days 1–3): each fish was individually placed in the novel tank for 5 min daily over three consecutive days, with no objects present, to minimize novelty-induced anxiety and ensure environmental familiarization; (2) training phase (day 4): fish were exposed to two identical objects placed in opposite corners of the tank for 10 min. The objects were constructed from inert plastic and matched for size, shape, and color to control for innate preference. (3) testing phase (1 h post-training): after a 1-h retention interval, one of the familiar objects (FO) was randomly replaced with a novel object (NO) of similar dimensions. Each fish was reintroduced to the tank for 10 min, and interactions with both objects were recorded. Object exploration was defined as active investigation within a proximity (within 2 cm) oriented towards the object. Passive swimming or time spent in adjacent zones without active orientation was excluded from the analysis. The primary outcome was the novel object preference (%), calculated as: Preference (%) = (time exploring NO /time exploring FO + NO) x 100. A preference score significantly above 50% was interpreted as an intact recognition memory, indicating successful discrimination between familiar and novel stimuli. All tests were conducted between 08:00 and 17:00 under consistent environmental conditions. Objects and tanks were thoroughly cleaned with 70% ethanol and rinsed with water between trials to eliminate olfactory cues.

Preparation of enzyme homogenates and biochemical analysis

Immediately following behavioral testing, zebrafish were euthanized via rapid cooling (ice-water immersion), a method recognized for minimizing post-mortem biochemical alterations and preserving enzymatic activity for subsequent analyses⁷⁷. Pooled brain tissue from two individuals was combined to form a single independent biological sample (n = 5 per group). Brain tissues were homogenized using a Mikro-Dismembrator U mill (Sartorius, New York, NY, USA) fitted with 3 mm magnetic balls (Sartorius Stedim Biotech GmbH, Goettingen, Germany). Homogenization was performed at 1000 rpm for 1 min in a 1:10 (w/v) ratio using ice-cold 0.1 M potassium phosphate buffer (pH 7.4) containing 1.15% KCl. The resulting homogenates were centrifuged at 14,000 rpm for 15 min at 4 °C. Supernatants were carefully collected and stored on ice for immediate biochemical assays. The following parameters were measured: total protein content – determined using the Bradford method⁷⁸, antioxidant enzyme activities – superoxide dismutase (SOD) activity⁷⁹, catalase (CAT) activity⁸⁰, glutathione peroxidase (GPX) activity⁸¹, cholinergic marker – acetylcholinesterase (AChE) activity⁸², non-enzymatic antioxidant – reduced glutathione (GSH) content⁸³, oxidative stress markers – malondialdehyde (MDA) levels, as a measure of lipid peroxidation⁸⁴, and protein carbonyl content, as an index of protein oxidation⁸⁵. All assays were performed in triplicate using spectrophotometric or colorimetric methods according to standardized protocols and expressed relative to protein concentration (U/mg protein or nmol/mg protein, as appropriate).

Data analysis

All data are expressed as mean ± standard error of the mean (SEM). Statistical analyses were performed using one-way analysis of variance (ANOVA) to evaluate differences among treatment groups, with treatment considered as the fixed factor. Post hoc comparisons were conducted using Tukey’s multiple comparison test. Statistical significance was set at p < 0.05. Pearson’s correlation coefficient (r) was calculated to assess the strength and direction of associations between behavioral outcomes, antioxidant enzyme activities, acetylcholinesterase activity, and lipid peroxidation levels (MDA). All analyses were conducted using GraphPad Prism version 9.5 (GraphPad Software, Inc., San Diego, CA, USA).