All cannabinoids used in this study were purchased as 1 mg/ml standards in methanol (Sigma, Ontario, Canada): Cannabidiol (CBD), Cannabichromene (CBC), Cannabidivarin (CBDV), Cannabigerol (CBG), and Cannabinol (CBN). The α-pinene (98% purity), trans-nerolidol (>85% purity), and Methanol (99.9% purity) were also purchased from Sigma (Sigma, Ontario, Canada). D-Limonene (96.9% purity) was purchased from MPBIO (MP Biomedicals LLC, Ohio, USA), Linalool (>96% purity) was purchased from TCI (TCI, Oregon, USA) and Phytol was from Agilent Technologies (Agilent Technologies, Inc., Rhode Island, USA). All terpenes were diluted in methanol. 6-hydroxydopamine (OHDA) (Sigma, Ontario, Canada) was diluted in saline buffer (0.9% NaCl) supplemented with 0.02% Ascorbic acid. Mixtures produced for cell line experiments used equimolar components as follows: MIX-1 = minor cannabinoids (CBC, CBG and CBDV) + Terpenes (Linalool, α-pinene, limonene, t-nerolidol and phytol), MIX-2 = terpenes only, and MIX-3 = Minor cannabinoids only. Individual major cannabinoids, CBD and CBN, were added at the same equimolar amount. Please refer to legends in Figures 3, 6 for specifics regarding mixture compositions.

Cell lines used were from ATCC (Manassas, Virginia, United States). Cath.a cells (sp. = mouse, cat# CRL-11179), a CNS catecholaminergic cell line, were cultured according to ATCC instructions and were induced to CAD differentiated status by serum deprivation (0.5% FBS culture for 36 h) prior to experiments as described (Qi et al., 1997). PC12 cells (sp. = rat, ATCC #CRL-1721) were cultured in RPMI 10% FBS. PC12 differentiation used Minimal Essential Medium containing 1% HS and 0.5% FBS, then the cells were treated with 100 ng/ml NGF, 100 ng/ml basic fibroblast growth factor (bFGF), and serum-starved media containing 2 mg/ml BSA for 2 days (Jeon et al., 2010). Schematic representations of exposure paradigms available in Supplementary Figure S2.

Neuroprotective effects were assessed based on the ability of both individual compounds and mixtures of compounds to protect against neuronal cell death induced by 1-methyl-4-phenylpyridinium (MPP⁺) in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-based selective cytotoxicity assay (Arshad et al., 2014). MPP⁺ is an active metabolite of MPTP that is known to cause human Parkinsonism after injection (Langston and Palfreman, 2014). As in Arshad et al. (2014), MPTP/MPP⁺ assays were performed in vitro on Cath. a cells by applying each compound or mixture of compounds to the cell cultures 18 h after application of MPP⁺ (Arshad et al., 2014). Cell viability assessments were performed using a standard MTT cell viability assay (MTT Cell Proliferation Assay Kit, Cayman Chemicals, Ann Arbor, Michigan, Item No. 10009365). Cell viability was assessed 24 h after exposure to MPP⁺, which is 6 h after exposure to the tested compound or compound mixture. Percent protection was normalized to MPTP control alone (100% cell death). To establish an effective dose range across the compounds in this assay system, the neuroprotective effects of each individual compound were tested at 5 different concentrations (Supplementary Figure S1). Based on these results the individual and equimolar mixtures were tested in the cell assays at 10 µM each (Figure 1). Equimolar mixtures contained 10 µM of each compound. In all cases vehicle controls contained methanol at equal concentrations to those found in test compounds/mixtures, ≤ 5%.

In parallel with the neuroprotection assays, cannabinoid and terpene compounds were tested alone and in mixtures to determine their effects on dopamine release from differentiated PC12 cells (Greene and Tischler, 1976). PC12 cells were differentiated as described (Jeon et al., 2010; Hu et al., 2018). Supernatant samples were collected from 3 replicate wells 30 min after application of PMA/Ionomycin (positive control) or the indicated compounds, and dopamine was measured in the medium using the Dopamine ELISA Kit #KA1887 from Abnova (Abnova, California, US) according to manufacturer’s instructions. For dopamine secretion, dopamine release was normalized to PMA/ionomycin control (0% baseline).

The fish used in this study were wild-type AB/Tubingen hybrids. Age-matched embryos were reared in Pentair Aquatic Ecosystem (Apopka, Florida, USA) nursery baskets (200 embryos per basket) on a ZebTec Recirculation Water Treatment System (Tecniplast, Buguggiate, Varese, Italy) at 28.5 ± 0.5°C, on a 14-h day–10-h night light cycle. All adult zebrafish husbandry and breeding was in accordance with the Canadian Council of Animal Care guidelines.

All compounds were diluted in 100% methanol (MeOH) and experiments were performed in a HEPES buffered E3 (HE3) medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2-2H2O, 0.33 mM MgSO4-7H2O, 10 mM HEPES, pH 7.2). Individual 120 h post fertilization (hpf) larval zebrafish were transferred to a 48-well microtiter plate in 500 μL of HE3 media. Larvae were acclimated for at least 1 hour in a lighted 28.5°C incubator (photon flux: 3–5 μmol s−1 m²) prior to experimentation and larval behavior was analyzed using DanioVision larval tracking systems with EthoVisionXT14/15 software (Noldus Information Technology Inc., Virginia, USA). Distance traveled was measured using dynamic subtraction at 28.5°C over 120 min with the first 90 min under lighted conditions (15 µmol m-2 s-1), followed by alternating 5-min dark/light cycles. Each larvae represents an independent measurement. Any larvae that were dead or displayed phenotypic abnormalities were removed from analysis. 12 larvae were used in each experimental condition, and at least 2 replicates of each concentration were performed.

Larvae were exposed to varying concentrations of 6-hydroxydopamine (OHDA) from 48 to 120 h post fertilization (hpf). 15 dechorionated larvae were transferred in 5 ml of HE3 to each well of a six-well plate. The 5 ml exposure media was replaced daily, and ascorbic acid/saline buffer (used to resuspend OHDA) was used as a vehicle control. Larvae were then loaded into 48-well microtiter plate in 500 μL of HE3 media as described above. Schematic representations of exposure paradigms available in Supplementary Figure S2.

In addition to distance travelled, activity was measured as a percentage change in pixel density during data acquisition. The integrated visualization feature in EthoVision software (Noldus Information Technology Inc., Virginia, USA) was used to detect larval activity during three distinct activity types: high (greater than 0.5% pixel change per sampling), moderate (between 0.03 and 0.5%), and inactive (less than 0.03%) states. The frequency with which larvae switched between activity states and the cumulative duration of time spent in each activity state was then measured. Metrics were captured in 1 min bins, and the average over 90 min was used to calculate each metric in each activity state. For frequency calculations, the number of times larvae switched between activity states (high, moderate, and low) was calculated (termed Total Frequency). The cumulative amount of time spent in the high and moderate activity states combined (termed Cumulative Duration) was averaged over 90 min and measured as a percentage.

Acute behavioral assays were performed on the cannabinoids selected for the study as described above except larvae were loaded into the well plates with 450 µL HE3. Immediately prior to recording 50 μL of 10 × test compound solution was added to each well. For the OHDA challenge experiments 150 µM OHDA was used for all challenge experiments. During media replacements at 72 and 96 hpf, 25 µL of 200X stock solutions of the respective cannabinoid or E-MEM was added to the exposure media. At 120 hpf larvae were washed with HE3 and transferred in 500 μL to 48-well plates. Behavioral analysis was then performed as described above. All experiments were performed at least in duplicate. The E-MEM selected for further study (above) were combined in non-equimolar ratios (defined cannabinoid-ratios DCR-MEM) and subjected to both the total distance and activity analysis metrics, as described above. This procedure was done in 3 steps, reducing a single component by 50% (a 1:2 ratio relative to their original equimolar concentrations) or 90% (a 1:10 ratio relative to their original equimolar concentrations), reducing 2 components by 50% (1:2) or 90% (1:10), or a defined cannabinoid-ratio (DCR)-MEM reducing 2 components by different cannabinoid-ratios (1:2 or 1:10 relative to their original equimolar concentrations).

The mean ± the standard deviations were calculated for all samples in the cell assays. A two-tailed student’s t-test was used to evaluate the statistical differences between sample types. Calculation of statistical significance for total distance traveled over 90 min was performed by one-way ANOVA using a Dunnett’s multiple comparison test using GraphPad Prism 7.04 software (La Jolla, California, United States). Comparisons of either OHDA + Drug vs. OHDA, or Drug vs. MeOH were performed as an unpaired Student’s t-test. An asterisk is used to represent a p-value less than 0.05 or lower, unless otherwise defined.

At 10 μM, all of the individual cannabinoids except CBC displayed some neuroprotective effects while none of the terpenes were able to prevent MPTP/MPP⁺ induced apoptosis (Figure 1). A mixture of the minor cannabinoids and the terpenes (MIX-1) showed a substantial neuroprotective effect with an increase in cell survival of 37 ± 3.8%. Mixtures of either the terpenes alone (MIX-2) or the minor cannabinoids (MIX-3) were then created to assess their contribution to the overall activity of MIX-1. MIX-2 (terpenes) showed a limited overall protection (4 ± 1.1%) while MIX-3 (minor cannabinoids) demonstrated similar activity to MIX-1 (25 ± 0.5%). The effects of the major cannabinoids were then assessed by adding each individually to the mixtures. CBD increased cell survival in all three mixtures (62 ± 3.1% vs. 37 ± 3.8% (MIX1), 16 ± 3% vs. 4 ± 1.1% (MIX-2) and 31 ± 2.2% vs. 25 ± 0.5% (MIX-3). The effects of the second major cannabinoid (CBN) were similar but less pronounced than those seen for CBD. Importantly, in all cases the mixtures were more neuroprotective than would have been estimated from the sum (hashed shading area of bars) of their individual effects.

In parallel with the MPTP testing described in the previous section, we tested the effects of the same individual compounds and mixtures on dopamine release from PC12 cells (Figure 2). The major cannabinoids, CBD (6.0 ± 0.3% DOPA release) and CBN (8.1 ± 0.4% DOPA release), were the only individual compounds tested that produced a statistically significant (p < 0.05) increase in dopamine release. The relative performance of the mixtures in the dopamine assay was similar to the trend observed in the MPTP/MPP⁺ assay. MIX-1 led to the largest increase in DOPA release of 31 ± 1.6%, followed by MIX-3 (21 ± 1.1%), while the DOPA release for MIX-2 was not significant. Again, we also evaluated the effectiveness of adding the major cannabinoids (CBD or CBN) to each MIX relative to the effectiveness of each MIX alone. When CBD was added to MIX-1, the effectiveness was increased leading to a DOPA release of 46 ± 2.3% for MIX-1 + CBD, while adding CBN to MIX-1 did not significantly increase the DOPA release. The addition of CBD or CBN to MIX-2 were able to produce modest but significant increases in dopamine secretion up to 7.0 ± 0.4% DOPA release for MIX-2 + CBD and 6.0 ± 0.3% DOPA release for MIX-2 + CBN. Similarly, the addition of CBD to MIX-3 produced an increase in DOPA release relative to MIX-3 alone (MIX-3 + CBD produced 18 ± 0.9% DOPA release), while the addition of CBN to MIX-3 produced a significant reduction in the DOPA release compared to MIX-3 alone (MIX-3 + CBD produced 14 ± 0.7% DOPA release). As in the neuroprotection assays, mixtures of the components were able to elicit significantly more robust responses than the individual components.

Taken together, these results demonstrate that the effects of the mixtures cannot be attributed to a single ingredient. On the contrary, it suggests that interactions between the components in the mixtures are critical for the maximal efficacy of the mixture. In addition, it appears from the cell assay data that the terpene components of the mixtures have a minimal contribution on their efficacy and that the cannabinoid components are sufficient to use as a potential therapeutic.

In order to test whether the cannabinoids and cannabinoid-based mixtures identified by the cell screening assays would potentially alleviate symptomatic effects in an animal model of Parkinson’s disease, we applied a previously developed zebrafish larval model of dopamine cell loss caused by exposure to 6-hydroxydopamine (OHDA) (Feng et al., 2014; Cronin and Grealy, 2017; Benvenutti et al., 2018) and refined it based on a determination of OHDA dose and time conditions. We found that larval exposure to OHDA produced a concentration-dependent decrease in the baseline activity during an initial 90-min period in the light (Figures 3A,B). It also led to a non-significant increase in the maximum response to the light/dark transition (startle response) at 225 µM. Based on these findings, we selected 150 µM OHDA as our testing model as it produced a decrease in activity during the 90-min baseline period, which can be considered a model of bradykinesia, while not impacting the startle response which suggests a minimal effect on general locomotor function (Figure 3). Visual assessment of the OHDA treated larvae appeared to show a more complex pattern of behavior than could be assessed by a simple measure of distance travelled. We observed that OHDA treated larvae, when at rest, displayed a small, periodic side-to-side movement with no velocity that may represent a ‘resting tremor’. This behavior had not been previously defined and further highlights the significance of the OHDA exposure to act as a model of PD. Thus, in addition to analyzing distance travelled by the larvae as described above, larval activity was also analyzed using a % pixel-change based assessment (Figures 3C,D). The activity was divided into three types: high, moderate, and inactive states. The high and moderate activities reflect burst swimming (escape behavior) and slower speed foraging swimming respectively (Budick and O'Malley, 2000). The inactive state was set between 0 and 0.03% change as a way to quantify the ‘resting tremor’ as a unique phenotype in the OHDA treated zebrafish larvae. OHDA treatment caused a reduction in the frequency of all states (Figure 3C). Typically, OHDA treatment resulted in a 50–60% reduction in activity state transitions for all three states (Figure 3C). As shown in Figure 3D, all larvae spent the majority of their time in the inactive state, with untreated larvae spending ∼80% of their activity in the inactive state, ∼15% in the moderately active state and ∼5% in the highly active state (Figure 3D). The OHDA treated larvae spent a greater fraction of their time in the inactive state (∼90%) with a concomitant decrease in moderate and high activities (5 and 1% respectively).

Initial experiments were conducted to determine the effective concentration ranges of each of the five pure, individual cannabinoids. In general, the cannabinoids tested acutely showed a similar effect on baseline larval behavior to that previously profiled for CBD (Achenbach et al., 2018). As the concentration was increased, the normal response to a light to dark transition was abolished. The effective concentration range was considered the concentrations between a no observable effect level (NOE) and a level that had a minimal statistically significant effect on behavior (Table 1). At the concentrations of the cannabinoids tested there was a slight opposition to the 150 µM OHDA induced hypoactivity, however, the effects were not significant.

In order to assess possible potentiating effects between the cannabinoids, three component, equimolar minimum essential mixtures (E-MEM) of the 5 cannabinoids were prepared (Figure 4A). Three of these E-MEMs; A, C and G showed a significant opposition to the OHDA induced hypoactivity as measured by the total distance travelled (Figure 4C) while not displaying any effect on carrier control larvae (Figure 4B).

The equimolar MEMs (A, C and G) were further analyzed using the refined activity metrics. At 500 nM all three E-MEMs except E-MEM-C, significantly affected Frequency, Cumulative Duration (CD) and Total distance as measured using refined activity metrics vs OHDA (Figures 5B,D,F). 250 nM dilutions of the E-MEMs abolished these changes, except with relation to Frequency and CD in 250 nM E-MEM A (Figure 5B). No recovery was apparent for the 100 nM E-MEM experiments for Frequency or CD metrics. 500 nM MEM G increased the activity of MeOH treated control larvae as measured by Frequency and CD, (Figures 5A,C), but not total distance (Figure 5E).

Based on the results from the equimolar mixture experiments, a comprehensive series of experiments were performed where one or two components of each mixture were reduced by 50% (1:2 ratio relative to the equimolar MEM) and/or 90% (1:10 ratio relative to the equimolar MEM) to produce novel molar ratios for further efficacy studies. The defined cannabinoid-ratio minimum essential mixtures (DCR-MEMs) that produced the most significant opposition to the PD-like effects of OHDA will be described herein (Figure 6) and a comprehensive summary of all of the results is included in Table 2. Optimization of all three E-MEMs led to a substantial opposition to PD-like, OHDA mediated changes in Frequency and Cumulative Duration, while having no effect on methanol treated controls (Figure 6). For E-MEM A, the 250 nM equimolar mixture performed as well as the optimized version (a 50% decrease in CBDV from the original 500 nM dilution). None of the MEM C optimized ratios were able to improve upon the 500 nM original dilution without causing increased activity against methanol controls. Several optimized ratios were developed for MEM G that improved upon the original equimolar response to OHDA, while not affecting methanol treated larvae. Specifically, ratio 5 displayed the greatest response in both the Frequency and Cumulative Duration metrics (Figure 6).

Anecdotal and Patient Reported Outcomes (PRO) suggest that c Cannabis can alleviate the symptoms of several neurological disorders, including PD. Analysis of clinical meta-data offers mixed evidence in support of medical cannabis. While well tolerated, medical cannabis has been shown to have limited benefit in improving dyskinesia and motor function (Thanabalasingam et al., 2021), although patients did report improvements in sleep quality and quality of life (Urbi et al., 2022). Conversely, cannabis-derived phytocannabinoids have shown a clear neuroprotective effect in rodent models (Prakash and Carter, 2021). While these meta-analyses reveal that medicinal cannabis has therapeutic potential, in order to become a recommended intervention in the treatment of PD, more studies will be required.

However, the use of cannabis as a therapeutic is hampered by a number of factors, including the presence of psychoactive compounds, potential negative chemical interactions, and non-standardized drug delivery methods (inhalation, edibles, dermal). Because there are hundreds of ingredients in cannabis extracts that may have therapeutic potential along with an extremely large number of possible combinations of these constituents, it may be impossible to systematically test every combination in order to identify the optimal mixture of purified components that make cannabis derived compounds good potential therapeutics. Therefore, one approach to identify the essential elements in cannabis is to logically and experimentally reduce the complexity of these mixtures while retaining as much of the original bioactivity as possible. More specifically, one may increase the probability of a positive therapeutic outcome by identifying the top performing mixture at each level of reduced complexity without exhaustively (and blindly) attempting to test every possible variation.

In this paper, we demonstrated the utility of a multitiered approach to identify minimum essential mixtures that are pharmacologically active in cell and animal models of PD. The research presented in this manuscript is predicated on a previously identified patented mixture of cannabis derived compounds (U.S. Patent Number 10,653,640). In our previous work, an in-silico database containing the relative percentages (% wt/wt) of the components found in extracts from different varieties of the cannabis plant (Reimann-Philipp et al., 2020) was used to identify a pool of cannabis components with therapeutic potential. The efficacy of >1,000 complex mixtures derived from this original pool of cannabis components were tested using dopaminergic cell models of PD (U.S. Patent Number 10,653,640). Based on this information, the most effective preliminary mixtures were reduced to complex mixtures derived from 10 cannabis-based compounds containing both cannabinoids and terpenes for further testing. We used two cell models to initially reduce the complexity of this patented mixture and to define the most efficacious minimum essential mixtures in the animal model. The MPTP/MPP⁺ assay used in this study models several aspects of PD pathology, including mitochondrial dysfunction and calcium dysregulation, neuronal cytotoxicity resulting from the calcium dysregulation in the mitochondria and the concomitant increase in the production of reactive oxygen and nitrogen species (Cassarino et al., 1997). For this assay we chose Cath.a cells, which are murine neurons from the Locus Coeruleus (LC), one of the earliest sites of PD neurodegeneration (Paredes-Rodriguez et al., 2020). Additionally, because cannabinoids reportedly modulate DA secretion (More and Choi, 2015), we also used dopamine-release assays to measure the ability of these compounds to supplement low dopamine production levels through increased DA secretion. PC12 cells are dopaminergic neuroendocrine cells that have been extensively characterized and are highly amenable to pharmaceutical manipulation (Zhang G. et al., 2019). The cell-based data demonstrated clearly that the cannabinoids provided both neuroprotective and dopamine secretion abilities not seen in the terpene mixtures alone (Figures 1, 2). Because PD has a complex pathology, no single cell line can recapitulate all of the mechanistic details. Thus future work will endeavor to continue to validate the cell line observations in additional animal models, potentially using ex vivo cultures of dopaminergic neurons. Importantly, in both of these assays, the cannabinoids demonstrated positive interactions that may be explained by the entourage effect. While more data and modeling are required to determine whether the activity of these combinations represent true synergy (Lederer et al., 2019), the maximal effects seen in the mixtures do represent more than the expected sum of their individual effects. Based on these findings, the terpenes were eliminated from the pool of compounds prior to testing the candidate Minimum Essential Mixtures (MEMs) in the zebrafish model of PD.

A larval zebrafish model was selected to assess the ability of the cannabinoid mixtures to alleviate the movement disorders associated with Parkinson’s disease. When larval zebrafish are exposed to 6-hydroxydopamine (OHDA), there is a dose-dependent effect that leads to the inactivation and eventual death of DA-producing cells of the substantia nigra (Zhang W. et al., 2019). DA-producing neuronal cell loss and associated DA depletion in the striatum are correlated with altered motor behavior and changes in the movement patterns of zebrafish (Feng et al., 2014; Cronin and Grealy, 2017; Benvenutti et al., 2018). Zebrafish larval locomotion is often divided into burst swimming (high velocity) and slow swimming (lower velocity) states (Budick and O'Malley, 2000). Not surprisingly, these disparate behaviors have been mapped to multiple different regions of the brain that are enervated substantially by dopaminergic neurons (Drapeau et al., 2002; McLean and Fetcho, 2004; Severi et al., 2014).

In the current study, we observed a third activity state, characterized by a zero-velocity, tremor-like movement. Attempts to quantify this activity led us to develop advanced analytics of larval behavior, and further evaluation of the etiology of this phenotype is ongoing. The use of the zebrafish OHDA model allowed for a reduction in the complexity and a refinement of the ratios of the original cannabinoid mixtures that demonstrated efficacy in the cell-based assays. MEMs containing CBD and CBDV or CBD and CBC demonstrated the greatest therapeutic potential. These MEMs can now be tested in additional, higher-cost, preclinical model systems.

While cannabinoids have been suggested as potential agents for treating a spectrum of neurological disorders, including PD, the mechanism by which they exert these actions in unknown. A priori, cannabinoids would appear to be promising drugs for targeting the mechanistic pathways that underlie PD with respect to their effects on movement and potentially modifying disease progression through neuroprotective actions. The minor cannabinoids are ligands for a number of receptors, including CB1, CB2, peroxisome proliferator-activated receptors (PPARs), serotonin 5-HT1a receptors, TRPV1 and others (Walsh et al., 2021). These receptors likely play roles in a number of the known multifactorial etiologies of PD, including mitochondrial Ca2+ homeostasis, intracellular Ca2+ signaling, reactive oxygen species, and neuro-inflammatory pathways (More and Choi, 2015). Increasing evidence supports a modulatory role for the ECB system in movement and movement disorders through bi-phasic modulation of dopaminergic, glutamatergic, and GABAergic receptors within ECB retrograde-signaling systems (Catlow and Sanchez-Ramos, 2015; More and Choi, 2015). The human endocannabinoid (ECB) system is therefore an attractive target for the development of novel treatments for neurologic disorders.

Preclinical studies using exogenous cannabinoids have shown their ability to act as neuroprotectants for dopamine (DA)-producing neurons, to reduce oxidative stress and neuroinflammation, and to provide relief from the motor symptoms of PD (More and Choi, 2015; Bandres-Ciga et al., 2020). Loss of dopaminergic neurons in the substantia nigra pars compacta (SNpc) region and their projecting fibers in the striatum are one of the core pathological features of Parkinson’s disease (Gonzalez-Rodriguez et al., 2020). A significant body of literature also demonstrates that cannabinoids can individually restore mitochondrial Ca2+ homeostasis, intracellular calcium signaling, and protect against cytotoxic oxidative stress that would otherwise harm DA-producing neurons (Ryan et al., 2009; Turner et al., 2017; Peres et al., 2018). PD patients in PRO studies reported motor symptom relief from cannabis use; where they reported improvements/decreases in bradykinesia, muscle rigidity, tremor, and dyskinesia in decreasing order of significance (More and Choi, 2015). The data in this study support the idea that ECB-targeting will modify Parkinsonian movement disorders and provides a MEM for further experimentation.

In this work, cell models of PD allowed for the screening of complex mixtures of cannabis-derived ingredients and identified the cannabinoids as those being responsible for the neuroprotective (MPTP/MPP⁺ assays) and dopamine secretory effects. Using the OHDA zebrafish model of Parkinsonian movement disorders, three MEMs containing equal parts of three cannabinoids each were identified that could significantly relieve the OHDA-related motor symptoms. An additional 63 variations using different ratios of the 3 original MEM were also tested, and 22 out of 63 of the defined cannabinoid ratio variations also significantly improved OHDA-related symptoms in zebrafish. Five of these 22 MEMs outperformed the original mixtures and will be further tested in more advanced animal models to develop new therapeutic options for Parkinson’s patients.