Polydatin Prevents Lipopolysaccharide (LPS)-Induced Parkinson's Disease via Regulation of the AKT/GSK3β-Nrf2/NF-κB Signaling Axis

Parkinson's disease (PD) is a common neurodegenerative disease characterized by selective loss of dopaminergic neurons in the substantia nigra (SN). Neuroinflammation induced by over-activation of microglia leads to the death of dopaminergic neurons in the pathogenesis of PD. Therefore, downregulation of microglial activation may aid in the treatment of PD. Polydatin (PLD) has been reported to pass through the blood-brain barrier and protect against motor degeneration in the SN. However, the molecular mechanisms underlying the effects of PLD in the treatment of PD remain unclear. The present study aimed to determine whether PLD protects against dopaminergic neurodegeneration by inhibiting the activation of microglia in a rat model of lipopolysaccharide (LPS)-induced PD. Our findings indicated that PLD treatment protected dopaminergic neurons and ameliorated motor dysfunction by inhibiting microglial activation and the release of pro-inflammatory mediators. Furthermore, PLD treatment significantly increased levels of p-AKT, p-GSK-3βSer9, and Nrf2, and suppressed the activation of NF-κB in the SN of rats with LPS-induced PD. To further explore the neuroprotective mechanism of PLD, we investigated the effect of PLD on activated microglial BV-2 cells. Our findings indicated that PLD inhibited the production of pro-inflammatory mediators and the activation of NF-κB pathways in LPS-induced BV-2 cells. Moreover, our results indicated that PLD enhanced levels of p-AKT, p-GSK-3βSer9, and Nrf2 in BV-2 cells. After BV-2 cells were pretreated with MK2206 (an inhibitor of AKT), NP-12 (an inhibitor of GSK-3β), or Brusatol (BT; an inhibitor of Nrf2), treatment with PLD suppressed the activation of NF-κB signaling pathways and the release of pro-inflammatory mediators in activated BV-2 cells via activation of the AKT/GSK3β-Nrf2 signaling axis. Taken together, our results are the first to demonstrate that PLD prevents dopaminergic neurodegeneration due to microglial activation via regulation of the AKT/GSK3β-Nrf2/NF-κB signaling axis.


BACKGROUND
Parkinson's disease (PD), which is featured by the selective death of dopaminergic neurons in the substantia nigra (SN) of midbrain, is the second most common neurodegenerative disease, affecting up to 1% of people over the age of 60 worldwide (1). PD is associated with main manifestations, including rigidity, resting tremor and postural gait disorder, and is accompanied by progressive non-motor symptoms such as cognitive impairment, mood disturbance, sleep dysfunction, gastrointestinal problems, and dysautonomia (2)(3)(4). While previous researches have proved that the pathogenesis of PD is associated with oldage factor, environmental factor, genetic factor, oxidative stress and free radical formation, accumulating evidence indicates that the neuroinflammatory response plays a critical role in the progression of PD (5,6).
Lipopolysaccharide (LPS), extracted from the outer membranes of Gram-negative bacteria, can effectively activate immune cell microglial cells in the brain. Previous studies have demonstrated that intranigral injection of LPS selectively induces the death of dopaminergic neurons (12,13). Thus, considering the relationship between neuroinflammation and PD, intranigral injection of LPS is often used to induce animal models of PD. Given the potential role of inflammation in the pathogenesis of PD, the cellular/molecular mechanisms leading to the death of dopaminergic neurons in LPS-induced models of PD must be elucidated.
The researchers reveal that there is evidence that the incidence of idiopathic PD among chronic anti-inflammatory drug users is relatively low (23,24). Considering the connection neuroinflammation with PD, dugs or ingredients with antiinflammatory activity are highly appreciated (25). Moreover, several studies have reported that natural products have neuroprotective effect on the prevention and treatment of PD (10,26,27). Polydatin (3,4 ′ ,5-trihydroxystilbene-3-β-Dglucoside; PLD), a natural resveratrol glucoside, is extracted from the roots of Polygonum cuspidatum and found in cocoacontaining products, red wine, and peanuts, among other foods (28). Previous researches have revealed that PLD exhibits various biological effects, such as anti-inflammatory activity and antioxidant activity (29,30). Moreover, PLD has been reported to cross the blood-brain barrier and prevent motor function degeneration in multiple animal models of PD (31). However, no studies have investigated whether PLD could prevent or relieve the pathogenic process of PD by inhibiting microglial activation. In our research, we explored the neuroprotective properties of PLD in an LPS-induced rat model of PD, as well as the potential anti-inflammatory mechanisms underlying the effects of PLD.

Animal Models
Nine-to eleven-week-old male Wistar rats (Baiqiuen Medical College of Jilin University, Changchun, China) were acclimated in in microisolator cages under at 22 • C with a 12/12-h light/dark cycle and were provided ad libitum access to food and water. All rats were kept under these conditions for 2 weeks before each research. All rats were randomly allocated to the following five groups (n = 18 in each group): sham group, LPS group, and PLD (25, 50, or 100 mg/kg) + LPS groups. Rats were treated and injected to be LPS-induced PD models, the protocols were performed as previously described (11,32). PLD [25, 50, or 100 mg/kg, suspended in a 0.5% sodium carboxymethylcellulose (CMC-Na) solution and then dissolved with 1 mL of PBS, once daily] were administered orally on the 3 days prior to operation. Rats in the shamoperated group were performed orally an equal volume of the vehicle solution. After LPS-injected, rats were given by gavage once a day for 4 weeks with PLD. After rats were determined by the last behavioral test, the SN of rats was separated to investigate the effect of PLD on dopaminergic neurons and microglial activation by immunohistochemical analysis and western blotting. Dopaminergic neurons were labeled with tyrosine hydroxylase (TH) (1:1,000; Abcam, Cambridge, CA, USA), and microglia were labeled with ionized calcium binding adaptor molecule-1 (IBA-1) (1:200, Proteintech, Chicago, IL, USA). The SN of remaining rats was rapidly obtained to determine the release levels of the pro-inflammatory mediators, and the protein expression levels of TH and IBA-1.

Open-Field Test
The open-field apparatus was a 100 × 100 × 40 cm. The bottom arena of the box was carved by a 20 × 20 cm black grid. Two and four weeks after LPS was injected, rats were subjected to an openfield test to measure the effect of the PLD treatment on motor activity. Rats were tested in a quiet, low-light environment and were allowed to adapt to the environment for 5 min. PD rats were treated with an open-field test at 2 and 4 weeks after LPS injection to investigate the effect of PLD treatment on motor activity. The bottom arena of the box was washed with a 5% water-ethanol solution prior to open-field testing to avoid the effect of previous rats.

Rotarod Motor Function Test
Accelerating rotation tests are commonly performed to measure the coordination and motor balance of rats with PD (33). Two and four weeks after LPS was treated, rats were performed to a rotational test to assess the effect of PLD treatment on rats' motor dysfunction. It has been reported that the apomorphine-induced rotational test is a classical and comom method to investigate the damage of the dopaminergic system and assess the behavioral Frontiers in Immunology | www.frontiersin.org dysfunction in PD model rats (34). PD rats were put onto the cylinder for a training session (10 rpm for 10 min) to adapt this test. Injected 0.5 mg/kg apomorphine, rats were putted into the cylinder for 30 min to measure the functional motor activity. The number of turns was recorded throughout the test.

Cell Culture and Treatment
A murine microglia cell line, BV-2 cells, was purchased from the Cell Culture Center in Chinese Academy of Medical Sciences (Beijing, China). BV-2 cells were grown in DMEM supplemented with 10% FBS in 5% CO 2 at 37 • C relative humidity and passaged by trypsin digestion (0.05%). The medium was changed to serumfree DMEM at least 6 h prior to the treatment of PLD or LPS. BV-2 cells were pretreated with various concentrations of PLD (dissolved in PBS) for 1 h and then stimulated with LPS (100 ng/mL) for specific time.

No Assay
Nitrite, a soluble oxidation product of NO, was measured with the Griess reagent in accordance with the manufacturer's instructions. Firstly, BV-2 cells (6 × 10 5 cells/mL) were seeded in 24-well plates and grown overnight. After BV-2 cells were treated with PLD for 1 h and stimulated with LPS for 24 h, the supernatant was collected and mixed with Griess reagent (part 1 and part 2) in a 96-well plate. The absorbance values were obtained after incubated in the dark for 20 min using a microplate reader (Synergy HT, BioTek, USA). The nitrite concentrations were calculated by the standard curve for sodium nitrite.

Immunofluorescence Assay
Immunocytochemistry-immunofluorescence (ICC-IF) experiments were performed to detect the nuclear translocation of Nrf2 and the NF-κB p65 subunit. BV-2 cells were seeded onto poly-L-lysine-coated slips in 24-well plates and cultured overnight, following which they were treated with PLD (400 µM) or LPS (100 ng/mL). The nuclear translocation of Nrf2 and the NF-κB p65 subunit were analyzed with anti-Nrf2 (1:50 Abcam, Cambridge, UK, USA) and anti-NF-κB p65 antibodies (1:100 Cell Signaling Technology, MA, USA), the procedure as Frontiers in Immunology | www.frontiersin.org previously described (11). Representative images were showed from 9 fields-of-view per treatment group.

Immunohistochemical Staining
Dopaminergic neurons in rats' SN were determined with rabbit anti-TH polyclonal antibody (1:1,000; Abcam, Cambridge, CA, USA). Microglial activation in rats' SN were detected with rabbit anti-IBA-1 polyclonal antibody (1:200, Proteintech, Chicago, IL, USA), the immunohistochemistry procedure as previously described (35). The numbers and ratios of IBA-1 and TH-positive cells were counted and then analyzed by three researchers not involved in the experimental treatments.

ELISA Experiments
The rat SN was quickly dissected, placed in ice-cold PBS and then washed with ice-cold PBS. After dried and weighed, the SN of PD rats was transferred into liquid nitrogen. The SN was grinded and then diluted by PBS at a tissue weight/volume ratio of 1:5. After disrupted using an ultrasonic homogenizer, the SN tissue was centrifuged at 12,000 rpm for 20 min at 4 • C. The supernatant was analyzed on the basis of manufacturer protocols of rat ELISA kits.
BV-2 cells were performed to method of 0.05 % trypsinization and seeded onto 24-well plates (2.5 × 10 5 cells/well). Cells were pretreated with BT (200 nM) for 2 h or PLD (100, 200, or 400 µM) for 1 h, following which they were exposured to LPS (100 ng/mL) for 24 h. After the media were obtained and centrifuged, the release levels of PGE2, TNF-α, IL-1β, and IL-6 were assessed by manufacturer protocols of mouse ELISA kits.

Statistical Analysis
All data were collected from repeated experiments and presented as means ± SEM. To perform the statistical test, SPSS 19.0 software (IBM) was used. The differences between experimental groups were analyzed with one-way ANOVA, while multiple comparisons were performed with the least significant difference (LSD) method. Differences were set statistically significant at p < 0.05 or p < 0.01. FIGURE 4 | PLD treatment suppresses the production of pro-inflammatory mediators in the SN. Four weeks after LPS injection, the SN was isolated from the brains of PD rats to measure the production of pro-inflammatory mediators (iNOS, COX-2, TNF-α, IL-1β, and IL-6). (A) The protein levels of iNOS (B) and COX-2 (C) were measured via Western blotting. β-actin was utilized as an internal control. The release levels of TNF-α (D), IL-1β (F), and IL-6 (E) were measured via ELISA. Values are mean ± SEM (n = 6 in each group). ## p < 0.01, vs. sham-operated control group; *p < 0.05, **p < 0.01 vs. LPS group.

PLD Treatment Ameliorates Behavioral Dysfunction in Rats With LPS-Induced PD
To explore the effect of PLD treatment on motor dysfunction of LPS-induced PD rats, rats were performed to behavioral tests at 2 and 4 weeks after LPS injection. The openfield test is a classic and common behavioral test used to comprehensively evaluate spontaneous behavior in PD rats. PLD treatment prominently ameliorated impairments in locomotor activity as determined by the number of squares crossed (Figures 1A,B) and rearing behavior (Figures 1C,D) during the 5-min open-field test. Considering that apomorphine treatment affects dopaminergic system and causes rotational behavior toward the side of damage, apomorphine-induced rotation is common used to assess the effect of drug on the dopaminergic system. Our results indicated that PLD treatment visibly decreased the number of apomorphine-induced rotations (Figures 1E,F). These findings suggest that PLD treatment exerted beneficial effects on motor dysfunction in rats with LPS-induced PD.

PLD Treatment Increases the Survival Rate of Dopaminergic Neurons in the SN
It has been reported that unilateral intranigral injection of LPS is common used to research the damage effects of neuroinflammation on the dopaminergic system. TH is the ratelimiting enzyme involved in the synthesis of catecholamines and participates in dopamine synthesis. In PD animal experiments, PBS or 10 µg of LPS was injected into the right SN of rats to establish animal models. Rats were sacrificed 4 weeks after LPS injection to obtain the SN. Immunohistological staining for TH in SN was conducted to further assess the protective effect of PLD on dopaminergic neurons in PD rats. We found that the number of TH-positive neurons was significantly lower in LPSinjected rats than in sham-operated rats. However, results showed that PLD treatment (25,50, and 100 mg/kg/day) markedly suppressed the death of TH-positive neurons (Figures 2A,B). Meanwhile, we also investigated the effect of PLD on TH protein levels in the SN of rats using western blot analysis. Results showed that PLD treatment markedly increased the levels of TH protein in SN of rats ( Figure 2C). Taken together, these findings indicate that PLD treatment attenuated LPSinduced damage to TH-positive neurons, suggesting that PLD increases the survival of dopaminergic neurons in PD animal model.

PLD Treatment Alleviates the LPS-Induced Activation of Microglia in the SN
In order to determine whether the protective effect of PLD on dopaminergic neurons is associated with an antineuroinflammatory response, we investigated the effect of PLD on microglial activation in LPS-induced PD rat. From 3 days prior to LPS injection to 28 days post-LPS injection, PLD was implemented orally at doses of 25, 50, and 100 mg/kg once per day. Four weeks after LPS injection, immunohistochemical and Western blot assays were performed to examine the effect of PLD on microglial activation. It has been revealed that the expression level of IBA-1 protein is a specific marker of microglial activation. We found that immunohistochemical staining experiments revealed that PLD treatment significantly inhibited microglial activation in a dose-dependent manner (Figures 3A,B). And our results indicated that PLD treatment also reduced levels of IBA-1 protein in a concentration-dependent manner ( Figure 3C). Overall, these findings indicate that PLD treatment remarkably suppresses microglial activation in the SN of LPS-induced PD rats.

PLD Treatment Suppresses the Production of Pro-inflammatory Mediators in the SN
To further investigate the effect of PLD on neuroinflammation in the SN, we performed Western blot and ELISA experiments to examine the production of pro-inflammatory mediators in the SN. Our results revealed that LPS injection remarkably up-regulated the production of pro-inflammatory mediators in the SN, while PLD treatment significantly suppressed the production of these pro-inflammatory mediators in a dosedependent manner (Figure 4). These results demonstrate that PLD inhibits the production of pro-inflammatory mediators in the SN of rats with LPS-induced PD.

PLD Treatment Increases the Expression of p-AKT, p-GSK-3β Ser9 , and Nrf2, and Downregulates the Phosphorylation of NF-κB p65 in Rats With LPS-Induced PD
To explore the anti-inflammatory mechanisms underlying the effects of PLD in an LPS-induced model of PD, we performed Western blot analyses for protein levels of p-AKT, p-GSK-3β Ser9 , Nrf2, and p-NF-κB p65. Previous studies have reported that pharmacological targeting of GSK-3 and Nrf2 exerts neuroprotective effects (20), and that AKT, GSK-3β, and Nrf2 activity is responsible for improvements in PD symptoms. Our results indicated that PLD treatment increased protein levels of p-AKT, p-GSK-3β Ser9 , and Nrf2, and downregulated the phosphorylation of NF-κB p65 in rats with LPS-induced PD. These findings suggest that the anti-inflammatory effects of PLD in LPS-induced PD may be associated with increases in protein levels of p-AKT, p-GSK3β Ser9 , and Nrf2, as well as decreases in the phosphorylation of NF-κB p65 (Figure 5).

PLD Attenuates the Inflammatory Response in LPS-Activated BV-2 Cells
Similar to primary microglia, the murine microglial BV-2 cell line produces NO, PGE2, and various cytokines following LPS stimulation. To validate whether PLD regulates microgliamediated neuroinflammation, we first investigated the effect of PLD on the inflammatory response in BV-2 cells. We have found that less than 500 µM of PLD has no cytotoxicity on BV-2 ( Figure S2). Our findings indicated that PLD suppressed the production of proinflammatory mediators (iNOS, COX-2, TNF-α, IL-1β, and IL-6) in LPS-activated BV-2 cells in a concentration-dependent manner (Figure 6). These results indicate that PLD may inhibit the inflammatory response in LPS-activated BV-2 cells.

PLD Enhances the Intracellular Localization of Nrf2 and Inhibits the Intracellular Localization of the NF-κB p65 Subunit in LPS-Activated BV-2 Cells
In our in vivo model of LPS-induced PD, PLD enhanced the activation of Nrf2 and inhibited the activation of NF-κB. To further demonstrate whether the effect of PLD on neuroinflammation is associated with these two signaling pathways, BV-2 cells were pretreated with PLD (400 µM) for 1 h, followed by LPS for 1 h (100 ng/mL). Previous studies have indicated that activation of Nrf2 and NF-κB signaling pathways may cause nuclear translocation of Nrf2 (37) and the NF-κB p65 (38) subunit. Because the nuclear translocation of Nrf2 plays a key role in inhibiting the inflammatory response, we investigated whether PLD upregulates the nuclear translocation of Nrf2. Indeed, our results indicated that PLD efficiently promoted the nuclear translocation of Nrf2 (Figures 7A,B). Furthermore, LPS treatment exerted little-to-no effect on the nuclear translocation of Nrf2 in BV-2 cells (Figures 7A,B). These findings demonstrate that activation of the NF-κB signaling pathway upregulates inflammation in microglia. The activation of the NF-κB signaling pathway mainly represents the nuclear translocation of the NF-κB p65 subunit. Our results indicated that LPS treatment induced the nuclear translocation of NF-κB p65, while pretreatment with PLD dramatically suppressed LPS-induced nuclear translocation of NF-κB p65 in BV-2 cells (Figures 7C,D).

PLD Activates the AKT/GSK3β-Nrf2
Signaling Axis in BV-2 Cells PLD upregulated the AKT/GSK-3β phosphorylation levels and Nrf2 expression level in the SN in our in vivo model of LPSinduced PD. Several previous studies have indicated that Nrf2 expression is regulated by AKT and GSK-3β (20,37,39,40). And AKT is the upstream of GSK-3β and can regulate the activation of GSK-3β. To investigate whether the effect of PLD on Nrf2 is regulated by AKT and GSK-3β, BV-2 cells were pretreated with MK2206 (an inhibitor of AKT) or NP-12 (an inhibitor of GSK3β). After microglia were treated with PLD for various durations (0, 0.5, 1, 3, and 6 h), total levels of Nrf2 protein expression were measured via Western blotting. Our results indicated that PLD treatment upregulated the total protein expression of Nrf2 in BV-2 cells. Because Nrf2 expression was remarkable in BV-2  9 | PLD activates the AKT/GSK3β-Nrf2 signaling axis in BV-2 cells. BV-2 cells were pretreated with MK2206 (an inhibitor of AKT, 10 µM) for 2 h, following which they were treated with PLD for 1 h. After cells were harvested, the protein expression of AKT, p-AKT (A), GSK3β, p-GSK-3β Ser9 (B), and Nucleosol-Nrf2 (C) was measured via Western blotting. BV-2 cells were pretreated with NP-12 (an inhibitor of GSK-3β, 2.5 µM) for 4 h, following which they were treated with PLD for 1 h. After cells were harvested, the protein expression of GSK-3β, p-GSK-3β Ser9 (D), and Nucleosol-Nrf2 (E) was measured via Western blotting. BV-2 cells were pretreated with BT (Brusatol: an inhibitor of Nrf2, 200 nM) for 6 h, following which they were treated with PLD for 1 h. After cells were harvested, the protein expression of Nucleosol-Nrf2 (F) was measured via Western blotting. β-actin was utilized as an internal control, while PCNA was utilized as a nucleosol internal control. Similar results were obtained from three independent experiments. Values are presented as the mean ± SEM (n = 4 in each group), ## p < 0.01 vs. control group, **p < 0.01 vs. PLD group. cells treated with PLD for 1 h, we also investigated the effect of different PLD treatment durations on p-AKT and p-GSK-3β Ser9 expression in microglia (0, 0.5, 1, 3, and 6 h). Similarly, p-AKT and p-GSK-3β Ser9 expression was notable in BV-2 cells treated with PLD for 1 h. Consistent with our in vivo findings, these results indicate that PLD may increase the phosphorylation of AKT and GSK-3β to upregulate the expression of Nrf2 in BV-2 cells (Figure 8). Subsequently, we measured the effect of MK2206 on AKT/GSK-3β Ser9 phosphorylation and Nrf2 expression, observing that PLD-induced Nrf2 activation and GSK-3β Ser9 phosphorylation were effectively blocked by treatment with MK2206 (Figures 9A-C). We also measured the effect of NP-12 on the GSK-3β Ser9 phosphorylation and Nrf2 expression. Our results indicated that NP-12 treatment substantially increased PLD-induced GSK-3β Ser9 phosphorylation and Nrf2 activation (Figures 9D,E). Taken together, these results suggest that PLD treatment activates the AKT/GSK3β-Nrf2 signaling axis in BV-2 cells.

PLD Suppresses the Phosphorylation of NF-κB p65 and the Release of Pro-inflammatory Mediators in LPS-Activated BV-2 Cells via Nrf2 Activation
To determine whether the anti-inflammatory effect of PLD is related to Nrf2 activation, BV-2 cells were pretreated with BT (an inhibitor of Nrf2) (Figure 9F). Our findings indicated that BT pretreatment attenuated the inhibitory effect of PLD on the phosphorylation of NF-κB p65. The release level of NO and PGE2 is regulated by iNOS and COX-2, respectively. NO and PGE2 are the important components of neuroinflammation and participate in cell survival and death in the pathological processes of neurodegenerative diseases. We found that PLD memorably suppressed the release level of neurotoxic factors (NO, PGE2, TNF-α, IL-1β, and IL-6) in LPS-treated BV-2 cells. However, BT pretreatment attenuated the inhibitory effect of PLD on the release of pro-inflammatory mediators. These results suggest that the pretreatment regimen effectively suppressed the phosphorylation of NF-κB p65 and the release of proinflammatory mediators in LPS-activated BV-2 cells via Nrf2 activation (Figure 10).

DISCUSSION
In the present study, we aimed to determine whether PLD protects against dopaminergic neurodegeneration by inhibiting the activation of microglia. Our findings demonstrated that PLD treatment ameliorates behavioral dysfunction in rats with LPS-induced PD. Furthermore, our results indicated that PLD treatment prevented the loss of dopaminergic neurons in the SN. While LPS administration significantly upregulated the expression of IBA-1 and proinflammatory mediators in the SN, treatment with PLD significantly attenuated these effects, suggesting that PLD suppresses neuroinflammation due to overaction of microglia in a rat model of PD. And we have found that PLD suppressed M1 microglia phenotype and enhanced M2 microglia phenotype in activated microglia ( Figure S1). To further investigate the neuroprotective mechanisms of PLD, we measured the production of proinflammatory mediators as well as levels of p-AKT, p-GSK-3β Ser9 , Nrf2, and p-NF-κB p65 expression in BV-2 cells. Our findings indicated that PLD enhanced the production of p-AKT, p-GSK-3β Ser9 , and Nrf2 while suppressing NF-κB p65 activation in microglia. These results suggest that PLD prevented dopaminergic neurodegeneration due to microglial activation via activation of the AKT/GSK3β-Nrf2 signaling axis.
PD is a global neurodegenerative movement disorder whose primary cause remains elusive. Nonetheless, previous studies have indicated that the inflammation and immune activation in brain contribute to the pathophysiology of PD (41). Activated microglia and astrocytes could produce reactive oxygen intermediates, NO, and inflammatory cytokines, which lead to neuroinflammatory activities resulting in neurodegeneration. Therefore, an understanding of the neuroinflammatory mechanisms and key biomolecules that control microglial activation is indispensable for developing a novel therapeutic strategy for the prevention of dopaminergic neurodegeneration in patients with PD.
In PD research, various PD models are established and used to explore the pathogenesis of PD. For instance, 6-hydroxydopamine (6-OHDA) is used to establish a PD model through oxidative stress, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) and rotenone through mitochondrial complex I inhibition, and LPS is used to establish a PD model through its glial cell activation. It has been reported that unilateral stereotaxic injection of LPS into the rat's SN leads to microglial over-activation, which selectively produces lasting degeneration of dopaminergic neurons resulting in the pathological and clinical features of PD (42). Therefore, LPS-induced PD model is performed to mimic the effect of neuroinflammation on brain.
Microglia, resident macrophages of the nervous system, represents the first line of defense against infection or injury to the nervous system (43). It has been summarized that the excessive release of these pro-inflammatory mediators causes damage of dopaminergic neurons, which is then toxic to neighboring neurons and lead to the death of neurons, representing a perpetual cycle of neuronal death (44). Therefore, the inhibition of microglia-mediated neuroinflammation presents a feasible approach for the prevention and treatment of PD. In the present study, microglia is replaced by microglial line BV-2 cells to explore the anti-neuroinflammatory effects and mechanisms of PLD. Although BV-2 cells are not a complete replacement for microglia, BV-2 cells possess many features of microglia and are usually used to research neuroinflammation induced by activated microglia.
NF-κB, a transcription factor, regulates the expression of pro-inflammatory enzymes and cytokines, which contribute to amplification of inflammation response leading to neuronal damage (45). Activation of the NF-κB signaling pathway may lead to the phosphorylation and translocation of NF-κB p65, in turn upregulating the inflammatory response, which may be associated with the pathogenesis of PD (46). Such findings suggest that the inhibition of NF-κB plays a key role in the prevention and treatment of PD. In the present study, PLD suppressed the activation of NF-κB, thereby downregulating neuroinflammatory responses in both a rat model of PD and activated microglia.
Nrf2 plays an integral role in microglia-mediated protection of neurons from inflammatory responses (47,48). Moreover, previous studies involving Nrf2-knockout mice have demonstrated that loss of Nrf2 can exacerbate neurodegenerative phenotypes (49)(50)(51). Additional studies have revealed that activation of Nrf2 downregulates neuroinflammatory responses in microglia (52). A previous study has demonstrated that 100 mg/kg/d of PLD treatment alone have little or no effect on animal studies (31). In the present study, PLD upregulated the expression of Nrf2 in a rat model of PD and in activated microglia. To evaluate the potential link between the neuroprotective effects of PLD and Nrf2 activation, we investigated whether pretreatment with the Nrf2 inhibitor BT attenuates the inhibitory effect of PLD on the release of proinflammatory mediators in activated microglia. As detailed in Figure 10, our findings indicated that PLD treatment inhibited the production of proinflammatory mediators via the activation of Nrf2.
Next, we further explored the mechanisms involved in PLDinduced Nrf2 activation. Previous studies have reported that the effect of GSK-3β on the inflammatory response is associated with the expression and nuclear localization of Nrf2 (37,53). AKT activation is necessary for the phosphorylation of GSK-3β Ser9 , which serves as a potential modulator of the inflammatory response (39,40). Moreover, several studies have demonstrated that Nrf2 may be upregulated via activation of AKT and permanent inactivation of GSK-3 (54)(55)(56). Therefore, in the present study, BV-2 cells were pretreated with inhibitors of AKT, GSK-3β, and Nrf2 to determine the effect of PLD on AKT/GSK-3β Ser9 phosphorylation and Nrf2 expression. Our findings indicated that PLD treatment increased the phosphorylation of AKT and GSK-3β Ser9 , as well as the expression of Nrf2, in BV-2 cells (Figure 8). Moreover, BV-2 cells were pretreated with the AKT inhibitor MK2206 in order to determine whether the expression of p-GSK-3β Ser9 and Nrf2 is induced via AKT activation in PLD-treated microglia. Our results indicated that PLD upregulated the phosphorylation of GSK-3β Ser9 and the intracellular protein expression of Nrf2 via the activation of AKT (Figures 9A-C). To investigate whether the intracellular protein expression of Nrf2 is regulated by GSK-3β inactivation, BV-2 cells were pretreated with the GSK-3β inhibitor NP-12. Our results demonstrated that PLD upregulated the intranuclear level of Nrf2 via the inactivation of GSK-3β (Figures 9D,E), suggesting that PLD treatment exerts anti-inflammatory effects by activating the AKT/GSK3β-Nrf2 signaling axis in BV-2 cells.
In our study, we only investigate the effect of PLD on murine microglial line BV-2 cells. Astrocytes are also the most abundant glial cells in the mammalian brain and play both beneficial and detrimental roles in PD. Activated astrocyte also play important roles in neuroinflammation in the pathophysiology of PD. Maybe PLD can also inhibit the activation of astrocyte to exert neuroprotective effect in PD. Although we found that PLD inhibited microglia-mediated neuroinflammation, the precise molecular mechanisms and protein targets remain elusive. A significant challenge with quantitative chemical proteomics approaches will be determining the cellular molecular target of PLD.

CONCLUSIONS
In conclusion, the present study is the first to demonstrate that PLD effectively prevented dopaminergic neurodegeneration from inflammation-mediated damage by inhibiting the production of pro-inflammatory mediators and activation of the NF-κB signaling pathway via activation of the AKT/GSK3β-Nrf2 signaling axis (Figure 11). Taken together, these findings support the notion that PLD can be applied as a preventative treatment for PD. Considering the extensive biological activities and food source of PLD, future clinical studies are proposed to determine whether PLD can be researched and developed as a new food supplement for the prevention or intervention of PD.

ETHICS STATEMENT
This work was performed in accordance with approved animal protocols and guidelines established by the Institutional Animal Care and Use Committee of Jilin University (Changchun, China) (approved on February 27, 2015;Protocol No. 2015047). We done our best to minimize animal suffering and to reduce the number of animals used.

AUTHOR CONTRIBUTIONS
JL and DL conceived and designed the study. BH, SF, and WW performed the experiments, analyzed the data, and wrote the paper. YL, DH, XR, TM, GC, WG, and XK collected the samples and information. All authors reviewed the manuscript. In addition, all authors have read and approved the manuscript.