Dose-Dependent Neuroprotective Effect of Standardized Bee Venom Phospholipase A2 Against MPTP-Induced Parkinson’s Disease in Mice

Parkinson’s disease (PD) is a chronic progressive neurodegenerative movement disorder characterized by the selective loss of dopaminergic neurons within the substantia nigra (SN). While the precise etiology of dopaminergic neuronal demise is elusive, multiple lines of evidence indicate that neuroinflammation is involved in the pathogenesis of PD. We have previously demonstrated that subcutaneous administration of bee venom (BV) phospholipase A2 (bvPLA2) suppresses dopaminergic neuronal cell death in a PD mouse model. In the present study, we established standardized methods for producing bvPLA2 agent isolated from crude BV at good manufacturing practice (GMP) facility. The therapeutic efficacy of purified bvPLA2 agent was examined in MPTP-induced PD mice. Importantly, administration of purified bvPLA2 in a dose-dependent manner reversed motor deficits in PD mice as well as inhibited loss of dopaminergic neurons within the SN of PD mice. The concentration-dependent action of standardized bvPLA2 appeared to be related to the induction of CD4+CD25+Foxp3+ regulatory T cells (Tregs), which, in part, inhibits T helper 1 (Th1) and Th17 polarization and suppresses microglial activation in PD mice. Taken together, these results suggest that standardized bvPLA2 purified from BV shows a neuroprotective effect against PD and thus has a potential target for treatment of PD.


INTRODUCTION
Parkinson's disease (PD) is known as the second most common neurodegenerative disease to affect about 3% of the population over the age of 65 (Lang and Lozano, 1998;de Lau et al., 2004). It is generally characterized by the selective loss of dopaminergic neurons as significant neuropathological hallmarks (Bertram and Tanzi, 2005). Until recently, several animal models have been developed to discover PD pathogenesis as well as search for neuroprotective therapeutic targets (Wang et al., 2015). Among all of these models, MPTP model has been widely used to initiate PD in animal models by penetrating blood-brain barrier (BBB) and destroying dopaminergic neurons in the substantia nigra (SN; Sakaguchi et al., 2008).
While the etiology of PD is unclear, a considerable body of research suggests that inflammatory responses play an important role in the development and progression of PD (Tufekci et al., 2012). Emerging evidence indicates the enhanced inflammatory responses, infiltration of T cells into the brain and glial cell activation are prominent features of PD (Wang et al., 2015). Recently, significant efforts recently have been focused on developing novel anti-inflammatory agents for PD treatment (Samii et al., 2009).
Bee venom (BV) consists of a complex mixture of peptides, enzymes, lipids and bioactive amines. Accumulating evidence suggests a wide range of pharmaceutical properties of BV. Accordingly, BV therapy has been developed to treat various diseases, including inflammatory diseases (Lee et al., 2014) and neurodegenerative diseases (Hossen et al., 2017).
One of the major components of BV is phospholipase A 2 (PLA 2 ), comprising approximately 10%-12% of the dry weight of the venom in the European honeybee, Apis mellifera (Habermann, 1972). The PLA2 derived from BV (bvPLA 2 ) belongs to group III secretory PLA 2 (sPLA 2 ) that has been implicated in diverse cellular responses, such as signal transduction, host defense, blood coagulation, and pain relief (Hossen et al., 2016). Interestingly, several lines of evidence have indicated the therapeutic effect of bvPLA 2 in neurodegenerative diseases, including prion disease (Jeong et al., 2011). Consistent with this finding, we previously reported the neuroprotective effect of bvPLA 2 against neurodegenerative diseases including Alzheimer's disease (Ye et al., 2016) and PD (Chung et al., 2015). The cellular action of bvPLA 2 appeared to mainly suppress immune responses via stimulation of dendritic cells, ultimately enhancing the function of regulatory T cells which play an essential role in maintaining immune tolerance (Sakaguchi et al., 2008).
It appears that generating clinical-grade and sterile pharmaceutical products from BV is challenging, mostly related to identification, isolation, and purification of bioactive components from BV (Ameratunga et al., 1995;Lee et al., 2014). In the present study, we developed an effective strategy for therapeutic components based on bvPLA 2 which were isolated and purified from BV at good manufacturing practice (GMP) facility. To evaluate the translational relevance, we tested the neuroprotective effects of the standardized bvPLA 2 in MPTP-induced mouse model of PD. Additionally, dose-dependent effect of bvPLA 2 isolated from BV on MPTP-induced PD mice was investigated to determine an optimal dose. Thus, the present study may shed new light on developing new therapeutic targets for PD to provide a basis for standardization and GMP of bvPLA 2 drug.

Animals
All experiments were performed in accordance with the approved animal protocols and guidelines established by Kyung Hee University. Briefly, 7-to 8-week-old male C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). All mice were maintained under pathogen-free conditions on a 12-h light/dark cycle and temperature-controlled conditions, with food and water ad libitum.

BvPLA 2 Isolation, Preparation, Manufacturing, and Quality Management
A standardized BV PLA2 was prepared by Inist St Co. Limited (Eumseong-gun, South Korea). For isolation and purification, raw BV was purchased from Bee Venom Lab LLC (Tbilisi, GA, USA) and dissolved in high performance liquid chromatography (HPLC) grade water at a concentration of 1 mg/ml. Then the diluted samples were applied to PTFE membrane filter (pore size 0.45 µm; Sigma-Aldrich, St Louis, MO, USA). To reduce the volume for the subsequent steps, the filtered mixtures were further concentrated by a tangential flow filtration (TFF) system, fitted with Pellicon 3 devices with Ultracel-10kDa membrane (Merck Millipore, Billerica, MA, USA). For manufacturing, the purified bvPLA 2 was dried with freeze-drying and was collected as a white powder. The bvPLA 2 content was determined using HPLC system and then diluted to a concentration of 0.1 mg/ml. Undesired substances including allergen were removed by membrane filters (pore size 0.22 µm PVDF sterile membrane filter; Jet Bio-Filtration Co., Ltd, Guangzhou, China). The separation and detection were carried out on reversed-phase (RP)-HPLC system on a C18 column (pore size: 180 Å; Sigma-Aldrich, St Louis, MO, USA) using a Waters 2695 liquid chromatograph and a Waters 2489 UV-visible detector (Waters Corporation, Milford, MA, USA). The sample was chromatographed at 25 • C at a flow rate of 2 ml/min. The elution was performed with a linear gradient of 0%-80% acetonitrile in 0.1% trifluoroacetic acid (TFA; Sigma-Aldrich, St Louis, MO, USA) and the elution profile was monitored at 220 nm. The area of the peak detected was used to measure the recovery of bvPLA 2 and the separation profiles of purified bvPLA 2 were compared with those of commercial standard bvPLA 2 (Sigma-Aldrich, St Louis, MO, USA). The bioactivity of purified bvPLA 2 was compared with inactive mutated recombinant bvPLA 2 -H34A (Lee et al., Manuscript submitted for publication). In order to measure bioactivity of purified bvPLA 2 , PLA 2 activity was measured with EnzCheck PLA2 Assay Kit (Invitrogen, Carlsbad, CA, USA) according to the supplier's instructions. All these procedures were carried out at an aseptic GMP facility. For quality management, the purity test was performed to ensure that there was no detectable heavy metals, insoluble particulate matter, endotoxins, or microbes. A quality control using endotoxin assays was performed by Charles River Laboratories Korea (Incheon, South Korea). Certificate of Analysis, which specifies the pyrogenicity of the endotoxin assessed by kinetic turbidimetric assay, was supplied. The commercial standard endotoxin was dissolved and diluted with Limulus amebocyte lysate (LAL) reagent water (LRW) and tris aminomethane buffer in order to determine the endotoxin level of purified standard bvPLA 2 . Changes in the bvPLA 2 content were examined for 3 months in stability test. Hence, based on these observations, the purified bvPLA 2 appeared to be appropriate as standardized bvPLA 2 .

MPTP-Induced PD Mouse Model
1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP; 20 mg/kg; Sigma-Aldrich, St Louis, MO, USA) was intraperitoneally (i.p.) administered to mice four times a day at 2 h intervals, inducing severe and persistent depletions of dopamine as previously described (Jackson-Lewis and Przedborski, 2007). During the experiment, mice were monitored for their physical condition and weight loss. The mortality rate of mice after MPTP injection was 0%-30% in each group and the animals surviving with <20% weight loss were included in the analysis.

BvPLA 2 Treatment
One day after the last MPTP injection, MPTP-injected mice were received with either bvPLA 2 or phosphate buffered saline (PBS). For administration, bvPLA 2 was dissolved in PBS and administered by once daily subcutaneous (s.c.) injections for six consecutive days in the concentration range of 0.01-0.5 mg/kg. Mice treated with commercial standard bvPLA 2 (Sigma-Aldrich, St Louis, MO, USA) by s.c. injection at a dose of 0.5 mg/kg for 6 days was used to compare the effect of purified bvPLA 2 .

Unbiased Stereological Estimation
Unbiased stereological estimation of the total number of TH-, ED1-, or H&E-positive cells was made using an optical fractionator as described previously, with minor modifications (West, 1993). In brief, the sections used for counting covered the entire SN from the rostral tip of the SN pars compacta (SNpc) to the caudal end of the SN pars reticulate (SNr). The counting was carried out using the Olympus CAST-Grid system (Olympus, Ballerup, Denmark). The counting frame was placed randomly on the first counting area and moved systematically over all counting areas until the entire delineated area was sampled. The total number of cells was calculated according to the optical fractionator equation (West et al., 1991).

Flow Cytometry
Flow cytometric detection of regulatory T cells was performed using fluorescein isothiocyanate (FITC)-conjugated anti-mouse CD4 (clone GK1.5; eBioscience, San Diego, CA, USA), phycoerythrin (PE)-conjugated anti-mouse CD25 (clone PC61.5; eBioscience, San Diego, CA, USA), and Alexa Fluor 647 anti-mouse Foxp3 (clone MF23; BD Pharmingen TM , San Jose, CA, USA). The single-cell splenocytes were washed with PBS and stained with FITC-conjugated anti-CD4 and PE-labeled anti-CD25 antibodies in staining buffer. The cells were subsequently fixed and stained with Alexa Fluor 647 anti-Foxp3 antibody overnight at 4 • C in the dark. After washing, the cells were stored at 4 • C in the dark for subsequent detection.

Enzyme-Linked Immunosorbent Assay (ELISA)
To assess CD4 + T helper (Th) cell subsets from splenocytes, we measured cytokine production by enzyme-linked immunosorbent assay (ELISA). First, the splenocytes at a concentration of 1 × 10 7 cells/ml were prepared from mice. Then, splenic CD4 + T cells were isolated by positive selection using anti-CD4 (L3T4) MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions. The purified splenic CD4 + T cells were stimulated for 12 h with 50 ng/ml of phorbol myristate acetate (Sigma-Aldrich, St Louis, MO, USA) and 1,000 ng/ml of ionomycin (Sigma-Aldrich, St Louis, MO, USA). The supernatants were taken and the levels of IFN-γ, IL-4, and IL-17A were measured from these supernatants by ELISA kits (BD Biosciences, San Jose, CA, USA) according to the supplier's instructions.

Pole Test
The degree of bradykinesia of the mouse was measured by the pole test with a slight modification of a previous protocol (Ogawa et al., 1985). Briefly, a tube of ∼50 cm in length and ∼1 cm in diameter was wrapped in gauze and a wooden pole was attached to the top. The time at which the mice turned completely downward and the total time to climb down the pole were measured with a cut-off limit of 30 s. Each mouse was given five trials, and the average of the best three measurements was used as the result. Trials, where the mouse jumped or slid down the pole, were excluded.

Statistical Analysis
Statistical analysis was performed using GraphPad Prism (v5.0; GraphPad) software. Each data was compared between conditions using unpaired T-test or one-way analysis of variance (ANOVA), or Kruskal-Wallis test followed by post hoc group comparisons. Data are expressed as the means ± standard error of the mean (SEM); P < 0.05 was considered significant.

Standardization of the Manufacturing Process of Bee Venom PLA 2
In the present study, we developed standardized methods for preparing bvPLA 2 from active components of European honeybee, Apis mellifera ( Figure 1A). To purify the crude BV, the extract was ultra-filtrated and subsequently concentrated to increase the yield of bvPLA 2 , as well as to remove undesirable products. Purified standard bvPLA 2 were identified and separated using RP-HPLC system with a C18 column ( Figure 1B). Commercial standard bvPLA 2 was also used to identify the separation profiles of purified bvPLA 2 and determine the content of purified of bvPLA 2 . Additionally, the bioactivity test was carried out to test specific PLA 2 activity (Figures 1C,D). To ensure that purified bvPLA 2 was safe, we performed the quality control tests, resulting in that it had an allowable endotoxin level approved by US Food and Drug Administration (FDA ; Table 1). Therefore, based on these observations, we considered this purified bvPLA 2 as properly manufactured and standardized agent.

BvPLA 2 Improves Motor Activity in a Dose-Dependent Manner in MPTP-Induced PD Mice
We first examined whether purified standard bvPLA 2 protected behavioral deficits in PD mice. As shown in Figure 2A, purified bvPLA 2 was given to MPTP-treated mice at a dose of 0.01, 0.1, or 0.5 mg/kg for six consecutive days, beginning 1 day after the last MPTP injection. On day 6 after MPTP treatment, MPTP-challenged mice took much longer than the control mice to turn downward ( Figure 2B) and to descend the pole (Figure 2C), indicating basal ganglia-related movement disorders in MPTP-treated mice. Treatment of purified bvPLA 2  0.5 mg/kg), inactive recombinant mutant bvPLA 2 (bvPLA 2 -H34A; 0.5 mg/kg), or standard melittin purified from BV (0.5 mg/kg) was administered to the mice for 6 days, beginning 1 day after MPTP injection. (B,C) The motor ability of mice was evaluated on a pole test on day 6 post-MPTP. Time to orient downward (B) and total time to descend (C) were measured. The data are expressed as the means ± SEM, n = 5-8 per group; * p < 0.05, * * p < 0.01, * * * p < 0.001, ns, not significant.
significantly shortened the time to turn and to down in MPTP-treated mice. However, inactive mutant bvPLA 2 (H34A) treatment induced no significant difference compared with Frontiers in Aging Neuroscience | www.frontiersin.org MPTP-treated mice. No significant differences were detected between purified standard bvPLA 2 group and commercial standard bvPLA 2 group treated with the same dose (0.5 mg/kg). As observed, administration of purified standard bvPLA 2 improved motor deficits induced by MPTP in a concentrationdependent manner.

BvPLA 2 Is a Regulator of Peripheral Regulatory T Cell Differentiation
Growing body of evidence supported the role of regulatory T cells in the disease progression of PD both in human PD patients (Chen et al., 2015) and animal models of PD (Reynolds et al., 2010). Here, we speculated that the action of purified standard bvPLA 2 on improving sensorimotor function in PD mice may be related to the induction of Tregs in the periphery.
To address our hypothesis, we first measured the cellular proportions of Tregs in MPTP mice after treatment with different doses of bvPLA 2 (Figure 3). In fact, no significant difference in CD4 + CD25 + Foxp3 + Treg populations was detected between control and MPTP mice on 7 days after MPTP challenge ( Figure 3B). Importantly, administration of bvPLA 2 at a concentration of 0.1 and 0.5 mg/kg induced an increase in the number of splenic Tregs on MPTP-treated mice, when compared with mice treated with MPTP only. Whereas lower doses of bvPLA 2 (0.01 mg/kg) induced no significant differences in the population of Treg cells. Similarly, commercial bvPLA 2 (0.5 mg/kg) stimulated Treg induction in the periphery with no significant difference compared with purified standard bvPLA 2 (0.5 mg/kg). All these findings support the notion that bvPLA 2 induces the expansion of CD4 + CD25 + Foxp3 + Treg in the periphery in PD, which can suppress the inflammatory response.

BvPLA 2 Inhibits the Expansion of Th1 and Th17 Effector Cells That Are Associated With PD
Since the balance of CD4 + T cell subsets is highly correlated with disease activity in PD (Chen et al., 2015), we further FIGURE 3 | Dose-dependent effect of purified standard bvPLA 2 on inducing differentiation of Treg cells in MPTP-treated mice. (A,B) On day 7 after MPTP treatment, CD4 + CD25 + Foxp3 + regulatory T cells in splenocytes were analyzed by flow cytometry. The percentage of CD25 + Foxp3 + cells in CD4 + T cells were assessed in MPTP-treated mice with bvPLA 2 treatment (purified bvPLA 2 : 0.01, 0.1, 0.5 mg/kg; commercial (Com.) bvPLA 2 : 0.5 mg/kg) for consecutive 6 days after 1 day of MPTP challenge. Combined results from two independent experiments with at least three mice per group were plotted. MPTP+bvPLA 2 group: standard bvPLA 2 , purified from crude BV, was injected from Day 1 to Day 6 post-MPTP at a concentration of 0.5 mg/kg. The data are expressed as the means ± SEM; * p < 0.05, * * p < 0.01, * * * p < 0.001, ns, not significant.

BvPLA 2 Protects Dopaminergic Neurons Against MPTP Neurotoxicity in a Concentration-Dependent Manner
We further evaluated whether purified standard bvPLA 2 exhibited a dose-dependent neuroprotection against neuronal degeneration of PD. Expectedly, the immunohistochemistry data revealed apoptotic cell death in H&E stained and loss of dopaminergic neurons in TH stained sections of MPTP-challenged mice (Figure 5). In contrast, there was a dramatic neuronal protection in SN of MPTP mice following bvPLA 2 treatment in a dose-dependent manner. When administered at a comparative lower dose (0.01 mg/kg), bvPLA 2 administration did not significantly rescue MPTP-induced neuronal loss. Similar results were observed in MPTP mice given with commercial bvPLA 2 at a dose of 0.5 mg/kg. These data indicate that purified bvPLA 2 effectively attenuated the loss of dopaminergic neurons associated with PD in a concentrationdependent way.

BvPLA 2 Suppresses Microglial Activation Caused by MPTP in Mice in a Dose-Dependent Manner
Mounting evidence has demonstrated that microglial activation may play a key role in amplifying the neuroinflammatory response, which can exacerbate dopaminergic neuronal death (Kim and Joh, 2006). Hence, we examined the expression levels of microglia markers, such as Iba1 and ED1, to explore whether purified standard bvPLA 2 could directly suppress the activation of microglia. As shown in Figure 6A, the level of Iba1-positive microglia was markedly increased in the SN of MPTP mice at 7 days after MPTP post-injection, whereas bvPLA 2 administration clearly reduced the expression of Iba1 in PD mice a concentration-dependent manner. Similarly, MPTP caused a significant increase in ED1-positive neuronal cells compared with the control group, while administration of purified bvPLA 2 markedly reduced the number of ED1 + microglia in the brains of MPTP-challenged mice (Figures 6B,C). The effect of commercial standard bvPLA 2 (0.5 mg/kg) appeared to be similar to that of 0.5 mg/kg of purified bvPLA 2 , showing no difference between the two groups. All these findings support the notion that purified bvPLA 2 inhibits microglial activation in MPTP-treated mice in a dose-dependent manner.

DISCUSSION
In the present study, we developed an effective method of bvPLA 2 purification to create a valid and safe standardized medicine. Using MPTP animal models, we further demonstrated the therapeutic potential of purified standard bvPLA 2 against PD. Additionally, we identified the dose-dependent neuroprotective effect of purified standard bvPLA 2 on PD mice. Growing evidence suggests that neuroinflammation is involved in the development and progression of PD (Wang et al., 2015). It is generally suggested that nerve damage in the pathogenesis of PD can be aggravated by the activated microglia producing an excess of pro-inflammatory cytokines. In contrast, some recent studies have proposed that activated microglia, especially M2 microglia, can be beneficial, at least in the early phase of the neurodegeneration process (Hu et al., 2012). Additionally, a growing body of evidence suggests that neuroinflammatory processes can act as a double-edged sword in diverse pathologic conditions linked to neurodegenerative diseases (Wang et al., 2015). Hence, when identifying a novel anti-inflammatory drug for treating neurodegenerative disease, though evaluation to designate a therapeutic window at a minimum effective dose is necessary.
To our knowledge, the present study was the first to present manufacturing process of bvPLA 2 from crude BV, as well as to assess its therapeutic efficacy for treatment of PD. We developed effective strategies to isolate and purified bvPLA 2 from active components of crude BV at GMP facilities. The separated bvPLA 2 displayed bioactivity with showing a specific activity of about 2,000 unit/mg. In order to evaluate translational relevance and therapeutic potential of purified bvPLA 2 , we examined the therapeutic effect of purified bvPLA 2 by injecting three different doses (0.01, 0.1, 0.5 mg/kg) on MPTP-challenged PD mice.
Indeed, it is known that high doses of bvPLA 2 (2.5 mg/kg) can cause a variety of allergic reactions, while relatively low doses of bvPLA 2 (0.25 mg/kg) induce protective effects against inflammatory conditions (Palm et al., 2013). Previously, we also observed that relatively low doses of bvPLA 2 (0.2-0.5 mg/kg) could induce anti-inflammatory effects in murine models of atopic dermatitis (Jung et al., 2017) and asthma (Park et al., 2015). Hence, we assumed the biological function of bvPLA 2 may differ in a dose-dependent manner, reflecting that at low doses bvPLA 2 act more generally to reduce neuroinflammation. Further studies are required to clarify the paradoxical effect and dose-dependent activity of bvPLA 2 .
In this study, we found a dose-dependent neuroprotective effect of purified standard bvPLA 2 containing ∼78% of PLA 2 and ∼15% of melittin. In line with the experiments with purified bvPLA2, we also examined whether melittin caused a neuroprotective effect against PD by isolated from crude BV with high purity (98.5% of melittin). Administration of purified melittin did not induce any neuroprotective effect on PD mice. Additionally, the inactive mutant of bvPLA 2 (bvPLA 2 -H34A) caused an inhibitory effect on the neuronal death and motor deficits in MPTP-challenged mice. Melittin is a major component of BV, comprising around 50% of dry venom. Recently, many studies have reported the therapeutic effects of melittin against cancers, neurodegenerative diseases and chronic inflammatory diseases including rheumatism (Park et al., 2004;Yang et al., 2011;Rady et al., 2017). It has been suggested that the biological activity of bvPLA 2 can be enhanced by melittin (Mingarro et al., 1995;Shen et al., 2010). Certainly, more detailed sets of experimental investigations are needed to verify its potential usefulness in the future applications.
The initial study regarding the involvement of inflammation in PD progression over 30 years ago clearly reported activated microglia in the SN of a PD patient (McGeer et al., 1988). Since then accumulated studies support the role of activated microglia and increased inflammatory mediators in the pathology of PD (Collins et al., 2012). In particular, accumulating evidence suggests the important role of adaptive immune system in the development and progression of PD (Ferrari and Tarelli, 2011). Importantly, the cytotoxic CD4 + T cells have been discovered in the SN of PD patients and animal PD models (Brochard et al., 2009;Stone et al., 2009). Indeed, the adaptive immune system would modulate neuroinflammatory response via regulation of microglia activation (Tansey and Goldberg, 2010). In this study, we observed the inhibitory effect of bvPLA 2 isolated and purified from BV on dopaminergic cell loss of MPTP mice. Impressively, induction of immunosuppressive regulatory T cells was observed in MPTP mice treated with purified standard bvPLA 2 . These results suggest that bvPLA 2 may exert anti-inflammatory action via activation of Treg cells, which can suppress the inflammatory process by targeting cytotoxic CD4 + T cells.
In summary, we standardized the purification process to provide standard bvPLA 2 from active component of crude BV and observed dose-dependent therapeutic effect of standard bvPLA 2 on MPTP-induced PD mice. Based on our results, it appears that bvPLA 2 is a potent drug that can promote the survival of dopaminergic neurons, suggesting a novel therapeutic candidate as an add-on to conventional dopaminergic substitution treatments.

ETHICS STATEMENT
All experiments were performed in accordance with the approved animal protocols and guidelines established by Kyung Hee University.

AUTHOR CONTRIBUTIONS
KK and HB conceived and designed the project. KK, SL, JS, and HJ performed the experiments and analyzed the data.
J-TH was in charge of standardization of the manufacturing process of bvPLA 2 from crude BV. The manuscript was written by KK and HB.