Ginsenoside Rk1 bioactivity: a systematic review

Ginsenoside Rk1 (G-Rk1) is a unique component created by processing the ginseng plant (mainly Sung Ginseng (SG)) at high temperatures. The aim of our study was to systematically review the pharmacological effects of G-Rk1. We utilized and manually searched eight databases to select in vivo and in vitro original studies that provided information about biological, pharmaceutical effects of G-Rk1 and were published up to July 2017 with no restriction on language or study design. Out of the 156 papers identified, we retrieved 28 eligible papers in the first skimming phase of research. Several articles largely described the G-Rk1 anti-cancer activity investigating “cell viability”, “cell proliferation inhibition”, “apoptotic activity”, and “effects of G-Rk1 on G1 phase and autophagy in tumor cells” either alone or in combination with G-Rg5. Others proved that it has antiplatelet aggregation activities, anti-inflammatory effects, anti-insulin resistance, nephroprotective effect, antimicrobial effect, cognitive function enhancement, lipid accumulation reduction and prevents osteoporosis. In conclusion, G-Rk1 has a significant anti-tumor effect on liver cancer, melanoma, lung cancer, cervical cancer, colon cancer, pancreatic cancer, gastric cancer, and breast adenocarcinoma against in vitro cell lines. In vivo experiments are further warranted to confirm these effects.


Eligibility criteria
We selected only original studies published up to July 2017 that provided information about the biological and pharmaceutical effects of G-Rk1. We included articles with G-Rk1 biological effects on human and animals either in vivo or in vitro with no restriction regarding publication language, publication date, or study design. We excluded three main types of studies which are: (1) Studies with unreliable extracted data or overlapping data set; (2) studies with only abstract available or no full-text available; (3) books, reviews, meta-analysis studies, conference papers, and thesis. Any disagreement was discussed carefully among three reviewers to get a final decision.

Information sources and search strategies
We conducted electronic searches using eight databases which include: PubMed, Scopus, ISI Web of Science, Google Scholar, SIGLE (System for Information on Grey Literature in Europe), Virtual Health Library (VHL), World Health Organization Global Health Library (GHL), and POPLINE. A Manual search using reference lists of studies was performed to find more relevant studies. The search strategy was performed by (AE, NXT, SS, YSS, EBO, MTE, ARK) and more information on search strategy was provided in Table S1.

Study selection
We selected articles in two phases: (1) Title and abstract screening of all searched articles; (2) full-text screening. The articles which were not in agreement with our inclusion and exclusion criteria were excluded. Three independent reviewers completed these two selecting phases. When disagreement occurred, a consensus decision was made following a discussion with supervisor (NTH).

Data collection process and data items
We prepared our primary extraction form, extracted three papers with it one by one, modified our form after each paper extraction and finally developed the extraction sheet that we used in the remaining articles. Three independent reviewers extracted the data from each paper. A discussion among the three reviewers was held to reach a consensus whenever there was a disagreement in any information retrieved. If three reviewers could not come to an agreement, the supervisor (NTH) was consulted.
The extracted data items included the last name of the first author, year of publication, year of subject recruitment, journal name, study design, country and city of origin of cell lines, the name of the plant, and method of extraction of our targeted material (G-Rk1). If the study included animals, we extracted their species, sex, age, and weight. If it had been done in vivo, we extracted the name of the cell line, its origin, the main medium used in terms of either primary (isolated by authors) or commercial cell lines. Also, we extracted the name of the measured parameter, an assay for its measurement, time effect, administration time, active substance name, its concentration, mean, standard deviation, standard error, a P value of results and the statistical test. When the data was presented as graphs, we used Web blot digitizer software, and the average of the results from three reviewers was calculated to obtain one result.

Risk of bias in individual studies
Two independent reviewers assessed all of the selected studies according to the GRADE method (Guyatt et al., 2011) to judge the quality of evidence, and any disagreement was resolved by discussion between them. Items such as limitation, inconsistency, indirectness, imprecision, publication bias, and moderate or large effect size were to be scored as ''1'' if there is no serious limitation or ''0'' if there is a serious limitation that has been defined according to GRADE criteria. Then the overall quality was to be scored as ''high'', ''moderate'', ''low'', or ''very low'' quality, according to their analysis of each study. The supervisor (NTH) was consulted when a disagreement occurs.

Summary measures
Inhibition of cell proliferation, apoptosis, and regulation of protein expression were the main evaluated outcomes.

Study selection
We identified 317 citations using the search strategy. From these, we included 156 articles after removing the duplicates. After that, we examined the title, abstract and further excluded 107 articles. We retrieved and evaluated the full-text of the remaining 49 articles, of which 25 articles were excluded, leaving 24 articles that were eligible, in addition to four articles that were retrieved from manually searching the included references. A flowchart described in details the process of identification, inclusion, and exclusion of articles was presented in Fig. 2.
One study (Ju et al., 2012) was downgraded as it has statistical typing mistake of one of its values (Table S2).
Liver cancer. Toh et al. (2011) evaluated the inhibitory effects of G-Rk1 (0.25 µg/ml) on cell growth of liver cancer cell lines (human hepatocellular carcinoma cells (HepG2), SNU449, and SNU182). A significant reduction of cell viability caused by G-Rk1 at 0.25 mg/ml was recorded (p < 0.001). The inhibition concentration (IC 50 ) value of G-Rk1 for inhibiting growth in the SNU449 cell line for 48 h was evaluated 0.08 mg/ml (100 µM) by using the WST-1 assay. These results indicated that G-Rk1 is one of the most anti-proliferative ginsenosides of raw and steamed P. notoginseng. Similarly, Quan et al. (2015) revealed that the HepG2 cell viability was reduced to 23% and 15% compared to the vehicle control when treated with G-Rk1 at 40 µM and 80 µM for 24 h, respectively (Quan et al., 2015).  evaluated the effect of G-Rk1 on cell viability of HepG2 cells after 24 h incubation in concentrations of 50, 75, 100 µM in the presence of 0.1 µM taxol which was used as a positive control. Compared with the vehicle control, G-Rk1 (at a dose of 100 µM) inhibited HepG2 cell proliferation by about 40%. When HepG2 cells were exposed to various concentrations of G-RK1 for 24 h (from 50 to 100 µM), the inhibitory effect on growth rate raised significantly, from 8 to 37.5%, in a dose-dependent manner. In addition, the cell viability was also tested when bafilomycin A1 was added to G-Rk1 (100 µM) and then, three independent experiments showed that this co-treatment enhanced HepG2 cell death more than the cells that were treated with 100 µM of G-Rk1 alone. In this experiment, to verify the effects of this combination and exclude cytotoxicity of bafilomycin A1, cytotoxicity was measured after 24 h and no cytotoxicity was detected.
In the study of  they assessed the effects of G-Rk1 on cell viability of HepG2 cells. The concentrations of G-Rk1 ranging from 12.5 to 100 µM with 0.5   (v/v) dimethyl sulphoxide added as control and incubated the cells for 48 h were used in this study. At 75 and 100 µM of G-Rk1, the effect of G-Rk1 induced cell death was maximized to 55% and 95% cell death respectively. In addition, the results revealed that the treatment of HepG2 cells with 100 µM G-Rk1, the fraction of early apoptotic cells increased from 0.46 to 16.23% and the underlying mechanism by which G-Rk1 induces the mitochondria-independent apoptosis can be through the activation of caspase-8, the signaling cascade of the one not associated with Fas-associated death domain expression.
To increase their cytotoxicity against Sk-Hep-1 hepatoma cancer cells, Park et al. (2002) used steamed ginseng which was separated by HPLC and tested with MTT assay to produce many active ginsenosides including G-Rk1. In this study, they found that the isolated G-Rk1 was associated with an inhibitory effect on cell viability in Sk-Hep1 cells. The growth inhibition concentration of G-Rk1 was 13 µM.
Lung cancer. G-Rk1 was evaluated in human lung cancer A549, and cell viability (% to control) was assessed using MTT assay. At the concentration of 50 µM, there was a statistically significant difference between cisplatin treated cell lines and Rk1 treated cell lines. However, G-Rk1 showed approximately two times higher anticancer activity than Rg5 when treated at 100 µM. After 24 h treatment, the IC50 values of G-Rk1 and cisplatin were 70, and 50 µM, respectively. Several proteins were found to be related to the apoptotic effect of G-Rk1 such as calmodulin-like protein, purine nucleoside phosphorylase, adaptor molecular crk, and transaldolase enzyme were increased while biliverdin reductase, aldehyde dehydrogenase, dihydropteridine reductase, and transactive response DNA binding protein-43 were decreased (Kwak & Pyo, 2016). In another study, A549 cell viability was reduced to 47% and 3.6% compared to the vehicle control when treated with G-Rk1 at 40 µM and 80 µM for 24 h, respectively (Quan et al., 2015). The results showed that Kp 7-6 treatment alone did not induce cell death or cell proliferation. Therefore, they concluded that Kp 7-6 has no effect on cell viability when used alone. However, when the cells were treated with Kp 7-6 followed by G-Rk1 (100 µM) treatment, the effect of G-Rk1 was reduced by 32 % compared to the control (no treatment of Kp 7-6). Moreover, they also assessed the induction of apoptosis by G-Rk1 in SK-MEL-2-Human Melanoma and their findings showed that when the concentration of G-Rk1 increased, the number of apoptotic cells also increased. More importantly, the cell lines responded in a dose-dependent manner.
Other types of cancer. Kim et al. (2013b) evaluated the effect of the combination of G-Rg5/G-Rk1 on cell viability of gastric cancer cells. After treatment with this combination at different concentrations (12.5, 25, 50 and 100 µM) for 24 h, the results showed an inhibitory effect on cell viability and proliferation of these cells in a dose-dependent manner (99, 93.5, 37.5, 3 %) respectively. In another study, cell viability was assessed using different cancer cell lines including human colon carcinoma (HCT-116), human cervical carcinoma (Hela), human breast adenocarcinoma (MCF-7), and human pancreatic cancer (PANC-1). When they were treated with 80 µM of G-Rk1, cell viability was reduced by 5.4%, 11%, 8.6%, and 9.9%, respectively (Quan et al., 2015).

Antiplatelet aggregation activity
Two studies evaluated the anti-aggregation effects of G-Rk1 both in vivo and in vitro (Ju et al., 2012;Lee et al., 2009) respectively. Ju et al. (2012 compared the antiplatelet aggregation activity of G-Rk1 and acetylsalicylic acid (ASA). The results indicated that G-Rk1 exhibits a stronger antiplatelet aggregation activity than ASA in which the action of G-Rk1 in platelets might be related to arachidonic acid (AA) metabolism. In addition, the alteration of (S) hydroxyl eicosatetraenoic acids and thromboxane B2 levels were determined using an immunoassay kit and UPLC/Q-TOF MS system, respectively. The 12-hydroxyleicosatetraenoic acid level was remarkably decreased in the G-Rk1 group but increased in the ASA-treated group. The thromboxane B2 level in the washed platelets decreased significantly by 66% when treated with 100 µM ASA and 77% when treated with 10 µM G-Rk1 (Ju et al., 2012). They used the colorimetric COX inhibitor screening assay to measure the inhibitory effects of G-Rk1 on COX-1 and COX-2. It was found that G-Rk1 inhibits both COX-1 and COX-2 activities. However, at a concentration of 20 µM, G-Rk1-derived inhibition was higher on COX-2 than on COX-1 (Ju et al., 2012). Lee et al. (2009) explained in his study that the effect of G-Rk1 on adenosine diphosphate (3-4 µM) induced platelet aggregation was monitored turbidimetrically by using ASA as a positive control. Both ASA and G-Rk1 showed the dose-dependent inhibitory effect on collagen, AA, and U46619 (9,11-dideoxy-11a,9a-epoxymethanoprostaglandin F2a) (thromboxane A2 mimetic drug)-induced platelet aggregation. However, they showed a negligible effect on adenosine diphosphate-induced aggregation. G-Rk1 exhibited the strongest inhibitory effect on collagen, AA, and U46619-induced platelet aggregation. In particular, it presented a 22-fold activity of ASA on AA-induced aggregation (Lee et al., 2009). G-Rk1 was found to be a potent inhibitor of AA and U46619 -induced platelet aggregation (Table 3).

Anti-inflammatory activity
G-Rk1 was found to have an anti-inflammatory effect by inhibiting NF-κB levels in the in vitro models (Lee, 2014). These results were assessed using luciferase assay. HepG2 cells were seeded at 1 × 10 5 cells/well in a 12-well plate and grown for 24 h. While G-Rk1 was pretreated with dimethyl sulphoxide for 1 h and then it was treated with tumor necrosis factor-α (10 ng/mL), the sulfasalazine was used as positive control. Their data demonstrated the strong inhibitory activity of G-Rk1 on NF-κB expression with 50% (IC50) value from 0.75 µM. However, the results revealed that G-Rk1 had cytotoxic effects, which occur in concentrations higher than 10 µM. Another evaluation of G-Rk1 anti-inflammatory activity (Kim et al., 2010) was its suppressing effect on 12-O-tetradecanoyl-phorbol-13-acetate

Effect of G-Rk1 on vascular leakage
A study evaluated the G-Rk1 effect on VEGF (Maeng et al., 2013) by treating primary human retina microvascular endothelial cells with G-Rk1 at a concentration of (10 µg/ml) for 40 min then stimulating it with 20 µg/ml of VEGF to disrupt the cell membrane. Sucrose permeability assay was used to evaluate the endothelial permeability and the results showed that G-Rk1 inhibited VEGF-induced retinal endothelial permeability. They used reverse-transcription polymerase chain reaction (RT-PCR) and densitometric analysis was used to assess translocation of tight junctions (TJ) proteins, and immunostaining was used to evaluate disruption of TJ proteins after the cells were stained with anti-ZO-1, anti-ZO-2, and anti-occludin antibodies. The authors found that G-Rk1 inhibited VEGF effect on TJ protein localization but it did not affect the transcription of TJ proteins ( Table 3).

Effect of G-Rk1 on lipid accumulation
Ginseng is known to have effects on obesity . In vitro treatment of mouse 3T3-L1 fibroblast cells with G-Rk1 resulted in reducing lipid accumulation, in which these cells differentiated into adipocytes after being treated with various G-Rk1 concentrations (10, 50, 100 µM) for 2 h at 490 nm optical absorbance   (Table 3).

Neuroprotective effect of G-Rk1
The combination of G-Rg5/G-Rk1 had a pronounced effect on the excitotoxic and oxidative stress-induced neuronal cell damage that was tested in primary cultured rat cortical cells (Bao et al., 2005). These cells were cultured in vitro for 12-20 days, then exposed to 100 µM glutamate or N-methyl-D-aspartate for 15 min in the absence or presence of G-Rg5/G-Rk1. The cell damage was assessed after 20-24 h by measuring LDH activity in the culture media. Data were calculated from cells exposed to the respective excitotoxic insults without ginsenosides. Data presented that approximately 70-80% of the cells were damaged by glutamate or N-methyl-D-aspartate compared to vehicle-treated control cells. The excitotoxic effect was significantly inhibited by G-Rg5/G-Rk1 in a concentrationdependent manner, in which 50% inhibition was achieved at 14.7 µg/mL of G-Rg5/G-Rk1.
In previous work, Bao et al. (2005) used a passive avoidance test to evaluate the effect of G-Rg5/G-Rk1. The latency in seconds was used to measure the cognitive performance of ethanol-induced amnesia in mice. The mice were orally treated with saline as vehicle and the ratio of G-Rg5/G-Rk1 equal 1:1 with a concentration of 10 mg/kg once a day for 4 days. The latency period of the mice administrated with ethanol was 24.9% less than the one of control mice (without ethanol-treatment), but it was significantly enhanced by the oral administration of G-Rg5/G-Rk1 with 1.2-fold increase than that of the control. The same steps were done, but this time after inducing amnesia with a single injection of scopolamine (3 mg/kg), also G-Rg5/G-Rk1 (10 mg/kg) provided the same enhancing significant result (p < 0.01). In another work, Jing et al. (2006) did the same tests of ethanol-induced amnesia in mice, which were given water as the control and ratio of G-Rg5/G-Rk1 equal 1:1 in the concentration of 10 mg/kg. They found that G-Rg5/G-Rk1 could significantly prolong the latency period by 2.97 folds more than that of the control. These two studies presented that G-Rg5/G-Rk1 would give beneficial results in the memory function of the normal, ethanol or scopolamine-induced amnesia in brains. G-Rk1 was reported to have a significant neurogenic activity in Epidermal growth factor-responsive neurosphere stem cells (erNSCs). However, this activity was less than G- Rg5 (Liu et al., 2007) (Table 3). Park et al. (2015) examined the effect of the G-Rg5/G-Rk1 combination on cisplatininduced nephrotoxicity in mice at cisplatin concentration 25 µM and G-Rg5/G-Rk1 concentrations of (0, 50, 100, 250 µg/ml). Results with EZ-cytotoxic cell viability assay kit showed a significant reduction in cisplatin and induced a reduction in cell viability. This effect was higher than that of Epigallocatechin gallate at the same concentrations as G-Rk1 (Table 3). Siddiqi et al. (2014) evaluated the osteogenic activity of G-Rg5/G-Rk1. MC3T3-E1 cells were treated with differentiation medium (either with or without G-Rg5/G-Rk1) for 12 days at different concentrations in which different substances were added to the culture medium in order to evaluate various effects of G-Rg5/G-Rk1 on differentiated fibroblast. The extent of calcium deposition, which is an indicator of osteoblasts mineralization, was measured by MTT assay. Data were expressed as a percentage of control, which showed that G-Rg5/G-Rk1 protected the extracellular matrix mineralization from antimycin A devastating effects. Besides, it turned out that alkaline phosphatase (ALP) activity evaluated by Smart BCA protein assay kit, increased by two folds after treatment with G-Rg5/G-Rk1 (30-50 µg/mL). et al., 2014)) has a greater effect on cell death than using G-Rg5 or G-Rk1 alone. Besides, it was also proved that co-administration of G-Rg5/G-Rk1 with a ratio 1:1 have various effects such as improving the cognitive performance in ethanol-induced amnesia in mice (Bao et al., 2005;Jing et al., 2006), inhibiting the exotoxic and oxidative stress-induced neuronal cell damage (Bao et al., 2005), and stimulating the mineralization of the extracellular matrix of osteoblasts (Siddiqi et al., 2014).

A (Siddiqi
In this systematic review, we found two studies presenting the antiplatelet aggregation activities with the results indicating that G-Rk1 (10 µM) can be stronger than ASA (100 µM) regarding the antiplatelet aggregation (Ju et al., 2012). Lee et al. (2009) also showed that G-Rk1 inhibited the effects of collagen, AA, and U46619-induced platelet aggregation. G-Rk1 was also indicated as one of the effective anti-inflammatory agents through the inhibition of both COX1 and COX2 activities and NF-κB levels (Ju et al., 2012;Lee, 2014).
Although, more than ten of our included studies reported that G-Rk1 has an anti-cancer effect against different cancer cell lines, all of them were in vitro studies with no in vivo or clinical studies. Unlikely, it was reported that G-Rg3 has an anti-cancer effect in both in vitro and in vivo (Shan et al., 2014). A recent meta-analysis of randomized clinical trials revealed that G-Rg3 combined with chemotherapy for non-small-cell lung cancer could enhance the overall survival rate and alleviate the chemotherapy-induced side effects (Xu et al., 2016). The shortage of in vivo or clinical studies to assess the G-Rk1 anti-cancer effect may raise many questions regarding the effect of G-Rk1 in patients and whether it differs from its in vitro action. In addition, what alterations that may occur in the patients. Therefore, there is a need for in vivo experiments to confirm the G-Rk1 anti-cancer activity and its mechanism.
Regarding the methodological approaches, several limitations were encountered. One of them is that we could not find any clinical study that used G-Rk1 in patients or healthy people. Out of 317 studies, we included 28 studies using our criteria, they were in vitro studies and in vivo animals. Based on the GRADE method, seven studies remained because of indirectness of evidence (Ahn et al., 2016;Bao et al., 2005;Jing et al., 2006;Kim et al., 2013b;Park et al., 2015;Ponnuraj et al., 2014;Siddiqi et al., 2014) and inability to explain heterogeneity in results (Bao et al., 2005). To date, there is a shortage of literature regarding clinical studies and the clinical use of G-Rk1 to treat some diseases in patients, and it consequently prohibits the clinical analysis.

CONCLUSIONS
In general, G-Rk1 has a significant anti-tumor effect on liver cancer, melanoma, lung cancer, cervical cancer, colon cancer, pancreatic cancer, gastric cancer, and breast adenocarcinoma against in vitro cell lines. Furthermore, In vivo experiments are necessary to confirm these effects. Additionally, G-Rk1 has demonstrated several pharmacological effects such as antiplatelet aggregation, anti-inflammatory, anti-oxidant, antimicrobial, anti-insulin resistance, neuroprotective, nephroprotective, and anti-lipid accumulation effects. All of these results support the clinical effects of G-Rk1 and demonstrate the promising possibility to develop the G-Rk1-based treatments, either alone or in combination with G-Rg5, for the previously mentioned conditions.