The Neuroprotective Effects of Cannabis-Derived Phytocannabinoids and Resveratrol in Parkinson’s Disease: A Systematic Review of Pre-clinical Studies

Currently, there are no pharmacological treatments able to reverse nigral degeneration in Parkinson’s disease (PD), hence the unmet need for the provision of neuroprotective agents. Cannabis-derived phytocannabinoids (CDCs) and resveratrol (RSV) may be useful neuroprotective agents for PD due to their anti-oxidative and anti-inflammatory properties. To evaluate this, we undertook a systematic review of the scientific literature to assess the neuroprotective effects of CDCs and RSV treatments in pre-clinical in vivo animal models of PD. The literature databases MEDLINE, EMBASE, PsychINFO, PubMed and Web of Science core collection were systematically searched to cover relevant studies. A total of 1034 publications were analyzed, of which 18 met the eligibility criteria for this review. Collectively, the majority of neurotoxin-induced PD rodent studies demonstrated that treatment with CDCs or RSV produced a significant improvement in motor function and mitigated the loss of dopaminergic neurons. Biochemical analysis of rodent brain tissue suggested that neuroprotection was mediated by anti-oxidative, anti-inflammatory, and anti-apoptotic mechanisms. This review highlights the neuroprotective potential of CDCs and RSV for in vivo models of PD, and therefore suggests their potential translation to human clinical trials to either ameliorate PD progression and/or be implemented as a prophylactic means to reduce the risk of development of PD.


Introduction
Parkinson's Disease (PD) is a neurodegenerative motor disorder that primarily affects the elderly. PD is progressive, and patients typically display a clinical triad of motor symptoms that are postural rigidity, bradykinesia and resting tremor [1]. Approximately 1% of the population over the age of 70 and an estimated 6.2 million individuals worldwide are affected by PD, and this is expected to increase yearly in line with a burgeoning geriatric population [2][3][4]. PD is characterized histopathologically by the loss of dopaminergic neurons within the substantia nigra pars compacta (SNpc) and accumulation of protein aggregates including α-synuclein within Lewy-bodies (LBs) [5][6][7][8]. The oligomeric and aggregated forms of α-synuclein can be neurotoxic and can promote loss of dopaminergic neurons [5][6][7][8]. PD is primarily an idiopathic disease, for which age is the major risk factor [9]. However, other risk factors associated with lifestyle and environmental exposures, such as alcohol intake and pesticide exposures have also been proposed; although a causal relationship between these and disease pathogenesis has yet to be clearly established [10,11]. Genetic vulnerability to PD has been observed in a minority of PD cases (10-15%) via rare familial mutants, including those in α-synuclein that trigger early onset Parkinsonian phenotypes, and other inheritable gene mutations that may contribute to cellular damage, oxidative stress, and inflammation [10,12].

Eligibility Criteria
All search results (n = 1034) were exported into Endnote for removal of duplicate publications and text analysis with respect to predefined eligibility criteria. Included articles were original, full-text publications investigating the direct neuroprotective effect of CDCs or RSV in PD animal models in vivo, with no restriction on age, dosing, length of study or outcome measures. Studies were excluded if the experiment was performed on non-rodent animals or artificial CDCs or RSV derivatives were used.

Data Acquisition and Analysis
A total of 18 eligible publications were included for review, from which the following variables were extracted to an Excel data spreadsheet: author(s), year of publication, the aim of the study, population, intervention, dosing, length of study, outcome measures and results. For methodological quality assessment, the Systematic Review Centre for Laboratory animal Experimentation (SYRCLE) risk of bias tool was considered (Supplementary Table S1), which has been adapted in accordance with methodology used in animal studies [31].

Results
The preliminary data search generated 1021 results, reduced to 898 following the removal of duplicates. The addition of 13 hand-searched studies from references and bibliographies of related publications, resulted in a total of 911 papers, which were screened with respect to their titles and abstracts. Of these 911 papers, 836 did not meet the predefined inclusion criteria and were excluded for the following reasons: irrelevant, performed on non-rodent models, in vitro studies, focused on other neurodegenerative diseases, full-text inaccessibility and lack of specificity. Full-text articles were read in full for 75 studies, of which 57 were removed on the grounds of investigating synthetic cannabinoids, cannabinoid receptor agonists or resveratrol-related compounds, and therefore did not meet the eligibility criteria. The remaining 18 studies fulfilled the eligibility criteria and were included in the final analysis of this review (Figure 1). The majority of studies investigated RSV (n=12) and the remaining studies CDCs (n=6); specifically, tetrahydrocannabinol (THC) (n=2), cannabidiol (CBD) (n=1), tetrahydrocannabivarin (THCV) (n=1) and βcaryophyllene (BCP) (n=2).

Figure 1.
Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow chart detailing the stages of study retrieval and selection [30].

Study Characteristics
A total of six strains from two different rodent species were assessed: Wistar rats (n=252), Sprague Dawley rats (n=183), C57BL/6 mice (n=156), Balb/C mice (n=60), Swiss Albino mice (n=42) and A53T-α synuclein mice (n=40). Two studies did not report the strain of mice investigated (n=39). An average of 43 rodents was used per study, with a median of 41 and a range of 24-60. All rodents were male. The weight of rats and mice ranged from 180-350g and 20-35g, respectively.

Behavioural Outcomes
There were 14 studies that incorporated an array of behavioural outcome measures to observe cognitive and motor changes in rodent PD models and assessed whether these impairments were mitigated by the administration of CDCs or RSV. The results of these studies are listed in Table 3.

Open Field Testing and Movement
A single study evaluated the neuroprotective effect of CBD in a risperidone-induced PD rat model using an open field circular arena and concluded that the locomotor activity of the risperidone group was significantly lower than the control group [41]. CBD treatment ameliorated risperidone-induced memory deficits but not locomotor activity, although oral movements as vacuous chewing were also reduced after treatment with CBD [41].
Motor activity was also examined using a computer-aided actimeter (CAA) and reported a significant increase in distance travelled and mean velovity in THCV-treated rodents when compared to the PD model [33].
PD model rodents displayed reduced velocity, rearing and distance travelled and these impairments were significantly improved by RSV treatment [37,42,44,46]. By contrast, a single study reported increased velocity in a PD model (A53T α-synuclein mice), but that was returned to control levels after treatment with RSV [43].
Rotational (circling) behaviour was examined in three studies and demonstrated the benefit of RSV treatment to reduce apomorphine-induced circling behaviour [34][35][36].

Rotarod and Grasp Strength
A rotarod test was used to assess grip strength and balance. Measurements were based on the time rodents remained on a rotating metal rod before falling off. Across six studies that used this method, all investigated the effects of RSV, but with slightly different experimental methods. The size of the rods ranged from 0.75-6 cm in diameter and maximum rotational speed varied from 12-20 rotations per minute. All studies showed reduced retention time on the rotarod in PD model groups relative to control groups [35,37,40,42,46,48]. One study also assessed grasp strength (g), measured by having rodents hold onto a horizontal bar over six trials. In comparison with the control group, the PD model group displayed increased grasp strength, indicating muscle rigidity, and this grasp strength was reduced by RSV administration [48].

Pole and Beam test
A pole test was used to assess bradykinesia and was performed by placing the rodent at the top of a pole with its head facing upwards. The time that was taken for a rodent to turn around and descend the pole was recorded. Of the three studies that used this measure, the height of the pole varied from 50-55 cm. An average of at least 3 trials was recorded per study. The beam test is similar, with rodents placed at one end of a narrow beam, and the time taken to reach the other end measured. Rodents within the PD model groups spent a significantly increased time in the beam and pole tests compared to controls, whereas rodents that received BCP both orally and i.p., or RSV, displayed significantly decreased beam and pole test times [43,44,46,49].

Gait Assessment and Stepping Test
Gait assessment was used to monitor a change in stride length by measuring the distances between forepaw prints in rodents. A shortened stride length is observed for rodent models of PD [50]. Provision of oral or i.p. BCP or administered RSV countered this reduced stride length [44,49]. A stepping test was also performed in a single study, and the number of adjusting steps taken when forced to walk on one forepaw was recorded. There was a reduction of adjustment steps in the PD model group, which was significantly rectified by RSV administration [35].

Catalepsy
Catalepsy, a decrease in movement and inability to correct abnormal posture, was assessed by both the bar test and the grid test across four studies. The bar test involved placing a rodent's hindlimbs on a bench and their forepaws on an elevated horizontal bar, and quantification of the time that they remain in this position. Catalepsy was attenuated in CBD and THC administered groups, resulting in a significantly decreased time spent in the bar test [38,41]. Two studies employed a grid test in which rodents were hung from a vertical grid approximately 0.5m high. One study measured the time taken for the rodent to initiate corrective movement [40], whilst the other measured the time taken to fall from the grid [37]. Both studies showed significantly increased catalepsy in PD model groups, and this was significantly rectified after receiving RSV [37,40].

Apomorphine-induced circling behaviour
Apomorphine-induced circling behaviour was assessed in three studies. Apomorphine, a non-selective DA receptor agonist influences rotational behaviour in rodents [51]. Apomorphine was administered subcutaneously to rats and the net number of contralateral rotations was measured over a time course. All studies demonstrated increased rotational behaviour in the PD model group of rats, and that RSV significantly decreased the number of rotations [34][35][36]. Abbreviations: BCP, β-caryophyllene; CAA, computer-aided actimeter; ND, Not determined; RSV, resveratrol; THC, ∆ 9 -tetrahydrocannabinol.

Biochemical and Immunohistochemical Outcomes
Thirteen studies assessed the neuroprotective effect of CDCs or RSV via an assessment of the levels of dopaminergic neurons and dopamine in the striatum of rodents, and related metabolites such as 3,4-dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA) [32][33][34]36,37,39,40,[43][44][45][46][47]49]. The results of these studies have been detailed in Table 4. The striatal concentration of tyrosine-hydroxylase (TH) markers were quantified in nine studies [32,33,36,37,[43][44][45]47,49]. One study showed a significant 5-fold decrease in striatal dopamine levels in a 6-OHDA-induced PD model, that was partially restored by THCV administration, although this did not reach significance [33]. The same study showed a significant decrease in SNpc dopaminergic neurons in LPS-lesioned mice, which was significantly restored via administration of THCV or cannabidiol (CBD)-derived drug (HU-308) in interventional groups [33]. There was a substantial decrease in striatal dopamine and DOPAC in 6-OHDA treated mice, in addition to decreased THimmunostaining and TH mRNA when compared to control groups, and these decreases were significantly improved by administration of THC in interventional groups [32]. TH immunoreactivity levels also showed a significant decrease in two studies using neurotoxin-PD-induced Wistar rats and C57BL/6J mice, which was significantly restored by BCP administration [39,49].
The results of these studies are shown in Table 5 elevated free radical levels, and these were increased in the PD model groups for three studies investigating the effects of RSV. These increases were significantly attenuated by RSV administration [36,43,48].
Endogenous anti-oxidative agents including reduced glutathione (GSH), and the enzymes, superoxide dismutase (SOD), catalase (CAT), glutathione reductase (GR), glutathione peroxidase (GPx), xanthine oxidase (XO), as well as the citric acid cycle enzymes aconitase, citrate synthase (CS), and succinate dehydrogenase (SDH) were investigated as markers of oxidative stress across seven studies that investigated the effects of BCP, THC and RSV [35,[38][39][40]42,43,46]. These studies showed decreased antioxidant capacity in PD model groups, which could be significantly mitigated by BCP, TCH or RSV administration [35,[38][39][40]42,43,46], except for the study of Anandhan et al. (2010) [46] that reported elevated SOD and CAT activities in their PD model. Total antioxidant capacity (T-AOC) was also increased in response to RSV treatment in the intervention group [36]. One study also monitored mitochondrial complex-I (MC-1) activity, which was decreased in the PD model group but significantly increased by RSV treatment [42]. Antioxidant defence was driven by increased Nrf-2 DNA binding activity in the RSV treated group relative to the PD model group [40].

Table 5. Summary of biochemical and immunohistochemical analysis for oxidative stress markers inPD model groups and interventions.
Abbreviations

The effectiveness of CDCs or RSV to combat neuroinflammation
Seven studies investigated the effects of RSV and CDCs on neuroinflammation in rodent brain tissue of the striatum and SNpc [33,34,39,40,43,45,49]. The results of these studies are summarized in Table 6. Five studies showed increased markers of microglia and astrocytes activation via quantification of glial fibrillary acidic protein (GFAP) and ionized calcium-binding adaptor molecule 1 (Iba-1) protein or mRNA levels and these were significantly reduced via administration of THCV, BCP, or RSV [33,39,43,45,49]. Inflammatory protein markers and their complementary mRNA levels were significantly increased in the PD model groups and this was significantly countered with BCP or RSV treatment [34,39,40,43,45,49]. The suppressor of cytokine signalling protein 1 (SOCS-1) was detected in α-synuclein transgenic mice and was signficantly upregulated by RSV treatment [45].

RSV anti-apoptotic effects
Four studies assessed the anti-apoptotic effects of RSV, there were no studies undertaken for the CDCs [36,37,40,44]. A summary of the RSV studies is included in Table 7. One study observed upregulation of apoptotic mediators in the PD model group that was decreased by RSV administration, with reduced neuronal apoptosis [36]. Procaspase and activated caspase-3 as key inducers of neuronal apoptosis were assessed by three studies, and all displayed increased caspase levels in PD model groups relative to controls [37,40,44]. Caspase levels were significantly decreased in groups receiving RSV treatment [37,40,44]. Bcl-2-associated X protein (Bax) and other pro-apoptotic regulators from the B-cell lymphoma 2 (Bcl-2) family were upregulated in PD model groups and were significantly reduced by RSV administration [37]. Increased p62 in nuclear factor kappa beta (NF-κβ) induced autophagy and increased acetylated microtubule-associated protein 1A/1B-light chain 3 (LC3-II) were countered with RSV treatment [44]. Increased C/EBP homologous protein (CHOP) and glucose regulated protein (GRP78), both apoptotic markers of endoplasmic reticulum (ER) related oxidative stress, were significantly reduced by RSV treatment [40].  Abbreviations: Bax, Bcl-2-associated X protein; Bcl-2, B-cell lymphoma 2; CHOP, C/EBP homologous protein; GRP78, glucose regulated protein 78; LC3-II, microtubule-associated protein 1A/1B-light chain 3; RSV, resveratrol.

Discussion
This review considered the neuroprotective effects of CDCs and RSV for a range of neurotoxin-induced rodent PD models. Since degeneration of dopaminergic neurons (in the SNpc) is the key pathological hallmark of PD, quantitative analysis of dopamine levels and dopaminergic neurons was utilized as the main indicator for neuroprotection. The in vivo rodent studies supported the hypothesis that these agents were neuroprotective against PD, and resulted in increased dopamine and dopaminergic neuron levels in response to CDCs or RSV treatment, consistent with a recent meta-analysis [52]. Collectively, the molecular mechanisms associated with neuroprotection reflected anti-oxidative, anti-inflammatory and anti-apoptotic capabilities (Figure 2).

Behavioural improvements indicative of neuroprotection
The reduction in motor and cognitive functions in PD mice models were attributed to neurotoxin-induced dopaminergic neuron loss within the SNpc, resulting in a dysfunctional striatal pathway and overstimulation of GABAergic neurons innervating the thalamus [53]. Reduced interconnectivity between the cerebral cortex and basal ganglia results in an impairment of motor functions [54], and this was evidenced via reduced performance in rodent behavioural tests. Treatment with CDCs or RSV improved motor performance and reduced PD symptoms associated with bradykinesia, rigidity and postural control. Animals treated with neuroprotective agents displayed reduced catalepsy, reduction in abnormal behaviours and an overall improvement in movement, strength, speed or balance [33][34][35][36][37][38][40][41][42][43][44]46,48,49].

Anti-inflammatory effects of neuroprotective agents
Neuroinflammation is likely to have a critical role in neurodegenerative diseases including PD [57]. Treatment with CDCs and RSV proved effective at reducing neuroinflammation in rodent PD models, as quantified by decreased levels of inflammatory protein markers and/or mRNA levels, as well as markers of astrocyte activation and microglia [33,34,39,40,43,45,49]. Collectively, beneficial anti-oxidative and anti-inflammatory effects of polyphenols including RSV, may in part relate to activity as free radical scavengers [58].

Anti-apoptotic effects of RSV
RSV displayed anti-apoptotic effects against nigral degeneration, with increased apoptotic markers Bcl-2 and pro-caspase 3 observed that paralleled decreased Bax and activated caspases [37,40,44]. The phosphoinositide 3-kinase (PI3K)/ protein kinase B (Akt) signalling pathway was upregulated by RSV to reduce dopaminergic neuron injury [37]. Akt is involved in homeostatic regulation and is recruited to cell plasma membranes in response to cell stress, where it is phosphorylated by PDK1 at serine-437 and threonine-308 [59]. Akt activation can reduce Bax and activated caspase-3 levels to inhibit apoptosis. RSV treatment resulted in the upregulation of proteins involved in this pathway, and induced an increased p-Akt (ser437), PI3K-110α and PDK-1 levels, thus inhibiting neuronal apoptosis in PD rodents [37]. Consistent with a role for Akt, reduced p-Akt was detected in dopaminergic neurons from the SN in PD patients after analysis of post-mortem brain tissue [60].
Endoplasmic reticulum (ER) stress occurs in response to an imbalance in ER homeostasis as a result of a prolonged accumulation of misfolded or damaged proteins such as α-synuclein [63]. ER stress activates the unfolded protein response (UPR), and the apoptotic division of the UPR pathway contributes to the loss of striatal dopaminergic neurons in PD [63,64]. RSV treatment was able to reduce ER stress through downregulation of the glucose-regulated protein 78 (GRP78) and CCAAT/enhancer binding protein homologous protein (CHOP) [40,65]. GRP78 forms a complex with misfolded proteins, in turn initiating the UPR pathway. Overexpression of CHOP in ER stress stimulates the activation of the pro-apoptotic Bax protein facilitating activation of caspases. This may be one of the potential mechanisms influencing the beneficial descrease in apopotic nigral cells following RSV treatment in the rodent studies [36,37,40,44].

Pro-dopaminergic properties of CDC's may involve cannabinoid receptors
THC is a major cannabinoid constituent of the Cannabis sativa plant and interacts with the G-protein coupled cannabinoid receptor 1 (CB1-R) and has a weak affinity for the cannabinoid receptor 2 (CB2-R) [66]. It might be expected that the attenuation of DA neuron loss in rodents would arise via stimulation of CB1 receptors in the CNS; however, a similar neuroprotective response was demonstrated using CBD [32], yet CBD has a low affinity for CB1 receptors. This suggests that the increase in TH-positive neurons and reduction in dopaminergic neuron loss can be mediated by a CB1-independent mechanism [32]. This was further supported by a demonstration that the minor phytocannabinoid THCV, which at low dose is a CB1R antagonist and a CB2R agonist, proved effective at countering a reduction in dopaminergic neuron levels with neuroprotective effects not attributed to CB1 binding, as a similar result was observed using a CBD derived drug [33]. Furthermore, the pro-dopaminergic action of THCV as a CB2 receptor agonist proved viable in an LPS-mediated inflammatory model of PD and demonstrated its anti-inflammatory potential for treatment in PD [33]. The stimulation of the CB1R by THC offers symptomatic relief against resting tremor in PD, but there are concerns regarding its ability to worsen hypokinesia and exhibit unwanted psychotropic effects [67]. In contrast, CBD is effective at attenuating dopaminergic neuron loss in a PD model [32] and may not induce the symptomatic complications associated with THC.
Furthermore, the potentially beneficial immunomodulatory actions of cannabinoids such as CBD are assocaited with agonism to CB2 receptors [68].

Clinical Trials using cannabinoids and RSV
Although there is a growing body of evidence from preclinical animal studies that support the neuroprotective effects of CDCs and RSV, there remains a lack of human clinical trials that consider the effects of BCP, THC, THCV or RSV. There are, by contrast, several studies that have investigated the effects of CBD in PD and RSV in Alzheimer's disease (AD), with study characteristics summarised in Table 8. An open label dose-escalation study investigating the efficacy of CBD in individuals with PD reported that CBD administration improved motor movement and sleep quality, evidenced via increased Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) scores [69]. However, this small-scale trial provided evidence of adverse effects of CBD that included diarrhoea and hepatoxicity [69]. An exploratory double-blind trial investigating CBD reported no significant changes in movement scores, although there was an improvement in the Parkinson's Disease Questionnaire (PDQ) scores and overall emotional well-being [70]. A randomised, double-blind, placebo-controlled trial of RSV for patients with mild-to-moderate AD reported a decrease of some CSF immunomodulatory effects but was also without significant improvements in cognition [71,72]. A separate randomised, double-blind, placebo-controlled trial of RSV with glucose and malate displayed reduced neuropsychiatric deterioration in treatment groups; however, this improvement too remained insignificant [73]. Collectively, these studies have proved useful in evaluating safety of these compounds in patients with neurodegenerative diseases. However, evidence supporting the use of CBD and RSV in clinical treatment remains insufficient, and further human trials assessing the benefits of CDCs and RSV specifically for PD patients is required.

Bioavailability of CDCs and RSV
The efficacy of CDCs and RSV has been demonstrated in rodent studies and data from clinical trials has supported the potential safety of CBD and RSV in humans. Nonetheless, compounds such as RSV have limitations associated with relatively rapid metabolism and low bioavailability, although this may be combatted by acute dose escalation and/or more chronic dosing regimens, or utilization of derivatives of RSV [74][75][76][77][78].
Details regarding the pharmacokinetics and bioavailability of CBD are limited, with a report of 31% after smoking [79,80], however, when consumed orally, the bioavailability may only be as low as 6% [79,81]. In accordance with CBD's lipophilic nature, it is often prescribed as oromucousal sprays or gel-encapsulations which utilize its oil/alcohol-soluble properties [79]. Moreover, the lipophilicity of CBD can be intentionally exploited when it is consumed, given that it can dissolve in high-fat nutrients to form micelle-structures favourable for gastrointestinal tract transport, thus increasing solubility and bioavailability [81]. However, whether these modifications result in increased bioavailability remains largely undetermined in humans, and also of concern is whether increased bioavailability impacts upon side effects, as short-term use of medicinal cannabinoids has a risk of non-serious adverse events and long-term use is as yet poorly characterized [82].

Study Limitations
A general limitation of the PD models described is that the neurotoxins used to induce PD phenotypes are fast-acting and cause a rapid depletion of dopaminergic neurons, typically within a period of several days, whereas in humans disease progression is much slower, and may take two decades for dopaminergic neuron depletion and establishment of clinical sequelae. Hence, although the described models have proved useful in establishing pharmacological effects of CDCs and RSV to combat PD pathology, the acute action of the neurotoxins will not replicate a long-term disease course and the associated symptomology observed in humans.
Therefore, one cannot predict the extent of the protective effects of CDCs or RSV in the early-to-moderate stages of PD. Additionally, individual variability with respect to patient age, co-morbidities, and genetic influences on disease progression cannot be anticipated using rodent models. An extension of the length of studies may prove useful to better evaluate behavioural changes in motor and cognitive function in rodents, as well as increasing the validity of outcome measures.
There is also a risk of experimental and publication bias in these pre-clinical animal trials. For example, a lack of details concerning blinding in allocation and outcome assessments. Furthermore, all studies cited in this review only performed experiments with male rodents, and there may well be sex-specific differences in PD pathology. Lastly, although the majority of studies employed oral and intragastric administration,particularly for RSV, some of the studies used subcutaenous, intraperitoneal, or intravenous administration routes, ones less likely to be adopted routinely by humans, and that will also influence compound bioavailability and pharmacokinetic data.

Conclusion
To our knowledge, this is the first systematic review that has directly considered the effects of both selective CDCs and RSV in the neuroprotective treatment of PD. Collectively, in vivo rodent studies have demonstrated that these natural compounds are efficacious in their neuroprotection of PD and produced symptomatic benefits. However, there remains a need to expand these studies to a more chronic model and for additional studies that consider the benefits of formulations and derivatives with improved bioavailability, but ultimately, further human clinical trials are required with patients with early-stage or early-onset PD.