A Preclinical Systematic Review of Ginsenoside-Rg1 in Experimental Parkinson's Disease

To date, no drug has been proven to be neuroprotective or disease-modifying for Parkinson's disease (PD) in clinical trials. Here, we aimed to assess preclinical evidence of Ginsenosides-Rg1 (G-Rg1), a potential neuroprotectant, for experimental PD and its possible mechanisms. Eligible studies were identified by searching six electronic databases from their inception to August 2016. Twenty-five eligible studies involving 516 animals were identified. The quality score of these studies ranged from 3 to 7. Compared with the control group, two out of the 12 studies of MPTP-induced PD showed significant effects of G-Rg1 for improving the rotarod test (P < 0.01), two studies for improving the swim-score values (P < 0.01), six studies for improving the level of TH protein expression (P < 0.01), and two studies for increasing the expression of TH mRNA in the substantia nigra of mice (P < 0.01). The studies reported that G-Rg1 exerted potential neuroprotective effects on PD model through different mechanisms as antineuroinflammatory activities (n = 10), antioxidant stress (n = 3), and antiapoptosis (n = 11). In conclusion, G-Rg1 exerted potential neuroprotective functions against PD largely by antineuroinflammatory, antioxidative, and antiapoptotic effects. G-Rg1 as a promising neuroprotectant for PD needs further confirmation by clinical trials.


Introduction
Parkinson's disease (PD) is the second most frequent neurodegenerative disease after Alzheimer's disease characterized by the loss of dopamine-containing cells in the substantia nigra (SN) [1]. The clinical symptoms of PD are a wealth of motor symptoms and nonmotor symptoms. The treatment is divided into two directions: symptomatic therapy for motor symptoms and modifying the underlying disease process through neuronal protection or restoration. However, current treatments of PD are mainly symptomatic therapies and no treatment has yet been proven to be truly neuroprotective [2]. Dopamine replacement therapy (L-DOPA and dopamine agonists) is still the most effective symptomatic treatment of PD, but this treatment frequently induces therapy-related motor complications such as dyskinesia, choreoathetosis, and fluctuations in motor function [3]. Thus, a number of PD patients resort to various kinds of complementary or alternative medicine (CAM) to improve their motor and/or nonmotor symptoms [4]. Traditional Chinese medicine (TCM), as one of the most important parts in CAM, has played a vital role in the medical care of PD patients for thousands of years [5]. Ginseng, the root of Panax species (C.A. Meyer Araliaceae), is a well-known traditional Chinese herbal medicine that has been used for various kinds of diseases in China, Japan, and Korea for thousands of years and is still a popularly and worldwide used natural medicine in modern time [6]. The major pharmacologically active ingredients of ginseng are Ginsenosides and they are responsible for most of the activities of ginseng [7]. Ginsenosides are divided into two categories as follows: (1) the protopanaxadiol (PPD) type: Rb1, Rb2, Rb3, Rc, Rd, and Rg3; (2) the protopanaxatriol (PPT) type: Rg1, Re, Rf, and Rg2 [8]. It has been reported that Ginsenosides-Rg1 (G-Rg1) might have neuroprotective 2 Oxidative Medicine and Cellular Longevity effects and little toxicity both in vitro and in vivo [9]. It also has beneficial effects on many neurological conditions, including the progressive neurodegenerative diseases such as PD [10]. The mechanisms of the neuroprotective effect of G-Rg1 include potentiating nerve growth factors, increasing anti-inflammation, antioxidation, and antiapoptosis, inhibiting excite toxicity and Ca 2+ overinflux into neurons, maintaining cellular ATP levels, and preserving structural integrity of neurons [11]. However, no systematic review has been conducted to assess the effect of G-Rg1 on experimental PD models to date. Systematic review of all available evidence from animal experiments before clinical trials can provide us adequate interpretation of the limitations and potential of a novel treatment strategy [12]. Therefore, in the present study, we conducted a systematic review of all available animal studies to evaluate the preclinical evidence of G-Rg1 for experimental PD.

Search Strategy. Two trained researchers independently
searched studies on the effects of G-Rg1 on PD from their inception to August 2016 in the following databases: PubMed, the Cochrane Database, Excerpta Medica (EMBASE), Chinese National Knowledge Infrastructure (CNKI), Wanfang database, and VIP Information Database. The following search terms were used: (Ginseng OR Ginsenoside OR Ginsenoside-Rg1 OR G-Rg1 OR Ginseng saponin) AND (Parkinson disease OR Parkinson's disease OR PD) in both English and Chinese.

Inclusion Criteria.
Studies that were included met all of the following criteria: (1) all studies should test the effect of G-Rg1 on animal models of PD, regardless of language, blinding, or publication status; (2) in the treatment group, any intervention that used G-Rg1 for PD should be included irrespective of the frequency, dose, the method of injection, and intensity; (3) in the control group, animals were treated with normal saline or nothing.

Exclusion Criteria.
Prespecified exclusion criteria were the following: (1) not reporting the efficacy of G-Rg1 on animal experiment of PD; (2) another neuroprotective agent being administered in the treatment group in addition to G-Rg1; (3) no control group; (4) reviews, case reports, abstracts, letters, comments, study protocol, editorials, and clinical guidelines; (5) duplicate publication.

Outcome Measurements.
The primary outcome of the interest was the behavioral assessments, including rotarod test, pole test, wire suspension test, and the values of swimscore. Secondary outcomes were the number of Tyrosine Hydroxylase-(TH-) positive dopamine neurons in the substantia nigra pars compacta (SNpc), levels of TH protein expression in the SNpc, and the mechanisms of G-Rg1.

Data Extraction.
Two investigators independently extracted information from each study, including (1) the first author's name and publication year; (2) individual data obtained for experimental animals including species, sex, number, weight, and anesthetic used; (3) experimental model; (4) information on treatment group including route of administration, dosage, and time for treatment; (5) data of control groups extracted as well as route of administration, dosage, and time of administration; (6) intergroup difference of each outcome measure; (7) outcome measures, including the behavioral exhibition of PD animal models, the number of TH neurons, and possible mechanisms of neuroprotective effects of G-Rg1 against PD. If outcomes were presented at different time points, data from the last time point were extracted. If the outcome data for metaanalysis were only expressed graphically or missing, we made attempt to contact authors for further information. When a response was not received, we used digital ruler software to measure the data from the graphs. We extracted data of mean value and standard deviation for each comparison from every study. Any disagreements were resolved through consultation with a corresponding author (Guo-qing Zheng).

Quality Assessment.
The methodological quality of the included studies was assessed based on a nine-item modified scale from the Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES) [13]. The modified CAMARADES includes the following criteria: (1) peer reviewed publication; (2) control of temperature; (3) random allocation to groups; (4) blinded assessment of behavioral outcome; (5) use of anesthetic without significant intrinsic neuroprotective activity; (6) calculation of the sample size necessary to achieve sufficient power; (7) appropriate animal model which uses animals without relevant comorbidities (aged, diabetic, or hypertensive); (8) compliance with animal welfare regulations; (9) statement of potential conflict of interests. For quality assessment score, the interquartile range of score across studies was reported.

Statistical Analysis.
We conducted statistical analysis using Cochrane's Review Manager (version 5.3) software. Data extracted from each study were considered as continuous data. WMD (weighted mean difference) is a standard statistic that measures the absolute difference between the mean values in two groups. Meanwhile, standardized mean difference (SMD) is also used as a summary statistic in metaanalysis when all the studies assess the same outcome but measure it in a variety of ways [14]. Heterogeneity among studies was estimated using Cochran's test (reported with 2 value and value) and 2 statistic. 2 values of 75, 50, and 25% correspond to high, medium, and low levels of heterogeneity, respectively; 2 values less than 50% indicated an acceptable degree of heterogeneity between studies [15]. Probability values of 0.05 were considered significant. Sensitivity analyses omitting each study at a time from the original analysis were conducted to verify our main results to be robust.

Risk of Bias in Included Studies.
According to the nineitem modified CAMARADES checklist, the mean quality score of the 25 included studies was 5.12 (interquartile range: 4.75-6.0), with scores ranging from 3 to 7 (Table 2), of which one study [21] Figure 2). Meta-analysis of 2 studies [22,33] showed that the G-Rg1 group significantly improved the swim-score values compared with the MPTPinduced PD group ( = 48; WMD: 8.56; 95% CI: 7.61 to 9.52; < 0.00001; heterogeneity: 2 = 0.13; df = 1; = 0.72; 2 = 0%) ( Figure 3). Four studies [16,17,23,39] indicated that mice treated with G-Rg1 spent less time descending the pole compared with mice treated with MPTP ( < 0.01 or < 0.05 at different time point). There are two time durations that should be recorded in the pole test: one is the time it took the mouse to turn completely downward (T-turn) and the other one is the time it took the mouse to descend to the floor (T-total). But only one study recorded the two time durations; other studies did not clearly record them in detail. Meanwhile, the climbing pole time of mice in each study was conducted at different days. Owing to the above reasons, meta-analysis for this pole test could not be performed. The other 6 studies [24,26,28,29,32,34] described that G-Rg1 group significantly improved the motor symptoms of PD induced by MPTP in mice, including the symptoms of thrilling, piloerection, raising tail, activity decrease, postural bradykinesia, and staggering gait but also failed to make a meta-analysis because they were just descriptive studies without any data. For the 3 studies on the motor dysfunction of 6-OHDA induced PD model, 2 studies [35][36][37] indicated that G-Rg1 group showed significant improvement in the rotational behavior in 6-OHDA-lesioned rats compared with      Note: (1) peer reviewed publication; (2) control of temperature; (3) random allocation to groups; (4) blinded assessment of behavioral outcome; (5) use of anesthetic without significant intrinsic neuroprotective activity; (6) calculation of the sample size necessary to achieve sufficient power; (7) appropriate animal model (aged, diabetic, or hypertensive); (8) compliance with animal welfare regulations; (9) statement of potential conflict of interests.
control group; one study [17] reported that, in comparison with G-Rg1, mice treated with MPTP spent much more time reaching the platform during the wire suspension test ( < 0.05).
Antioxidant Stress. Three studies including two MPTPinduced PD [19,30,36] and one 6-OHDA induced PD [36] studies reported the antioxidant stress effect of G-Rg1 on PD models. Two studies [19,36] detected that G-Rg1 significantly increased glutathione (GSH) level and decreased total superoxide dismutase (T-SOD) activity and lactate dehydrogenase (LDH) levels in the SN compared with the control group ( < 0.01). The remaining study [30] showed significant effects of G-Rg1 for reducing the numbers of ironstaining cells compared with the control group ( < 0.01).

Summary of Evidence.
Twenty-five studies with 516 animals were identified. This study found that G-Rg1 could improve the neurobehavioral abnormality and exert potential neuroprotective effects on PD model through different mechanisms such as antineuroinflammation, antioxidant stress, and antiapoptosis. However, we should treat the preclinical evidences cautiously because the methodological flaws undermine the validity of outcomes.

Methodological Considerations.
This systematic review has a number of weaknesses. Firstly, animal studies with neutral or negative results may be more likely to remain unpublished and will be missed. Therefore, the effect size may be overstated. Secondly, our search strategy includes only Chinese or English databases, which may cause a certain degree of selective bias [41]. Thirdly, previous meta-analyses have suggested that animal studies that are less rigorously designed may overestimate treatment effects [42]. In the present study, all the studies failed to mention the blinded assessment of behavioral outcome. It may lead to performance bias and detection bias [43]. Sufficient size is essential to determine the efficacy of a new therapy or drug [44]. No study reported the calculation of the sample size that was necessary to achieve sufficient power, which indicated the lack of statistical power to ensure suitable estimation of the therapeutic effect [45]. Finally, the results from individual studies were inconsistent, and most of the studies used the graph rather than original data to present the outcomes. Therefore, we could not synthesize these data into the quantity.  [19,20,24,26,30,31]. Therefore, G-Rg1 exerts beneficial effects on multiple aspects of the pathophysiology in PD.

Implications.
It is well known that animal experiments have contributed to our understanding of mechanisms of diseases, but the translation of preclinical experiment which results in a prediction of the effectiveness of treatment strategies in clinical trials is still challenging [46,47]. Previous studies [13] suggested that the quality of the research design is an important factor affecting the outcome. The main causes for the failure of translation of animal studies to human clinical trials include inadequate animal data and overoptimistic conclusions about efficacy drawn from methodologically flawed animal studies. Thus, it is essential to improve the methodological standards in the design, execution, and reporting of preclinical PD studies in the future.
Quantitative and statistical analysis of Ginsenosides in plasma indicates that PD type exhibits higher concentration and longer half-life than PT type [48]. Due to the low membrane permeability, active biliary excretion, and biotransformation, the oral bioavailability of G-Rg1 is very low [49][50][51]. After an oral administration of G-Rg1, the experiment in rats indicated that the area under the curve of G-Rg1 is 28.93 g⋅h⋅L −1 and the mean value of half-life is 15.26 hours. The peak concentration is 7.15 g⋅L −1 , while max is 2.19 hours. In clinic, the use of ginseng for the suggestive symptoms of PD could date back to 1623-1670 AD recorded in Yizong Jiren Bian (Compiled Texts on Self Duty of Medicine) by Gao Gufeng who discussed the pathogenesis of tremor syndrome in the chapter shiver, shake, tremble: "Pathogenesis is mainly due to deficiency of Qi and Blood. The bones and muscles could not get enough nourishment, causing tremble that could not be controlled." Ginseng Tonic Decoction should be used for treatment to invigorate Qi and Blood [5]. In fact, ginseng was one of most commonly used herbs for tremor syndrome from the Han Dynasty to the end of the Qing Dynasty (206 BC-1911 AD) in China by using the frequency statistics according to 232 prescriptions involving 193 herbs and 2529 total frequency of herbs [5]. In modern time, several clinical studies have been conducted to assess the efficacy and safety of ginseng prescription for PD, and the results indicated that ginseng prescriptions could significantly ameliorate the motor symptoms and improve the quality of life [52,53]. However, no clinical study of G-Rg1 for PD has been yet conducted. In the present study, the findings indicated that G-Rg1 exerted potential neuroprotective functions against PD and its mechanisms are involved with on multiple aspects of the pathophysiology in multiple PD models. Thus, G-Rg1 may be a promising candidate neuroprotectant from bench to bedside. In addition, high-quality randomized controlled trials (RCTs) and a systematic review of those RCTs are commonly regarded the highest level of evidence in judging the treatment efficacy and safety of interventions [54]. Given the huge gap between the animal studies and the clinical trials, seeking and developing innovative neuroprotectants and further rigorous RCTs are urgently needed. In the present study, the findings indicated that G-Rg1 exerted potential neuroprotective functions against PD and its mechanisms are involved with multiple aspects of the pathophysiology of PD. Thus, G-Rg1 may be a promising candidate neuroprotectant from bench to bedside.

Conclusion
G-Rg1 exerted potential neuroprotective functions against PD despite of the methodological flaws. In addition, we identified an important area, which is worthy of further study. G-Rg1 as a promising clinical candidate neuroprotectant for PD needs to be further confirmed by clinical trials.