Veratrilla baillonii Franch Could Alleviate Lipid Accumulation in LO2 Cells by Regulating Oxidative, Inflammatory, and Lipid Metabolic Signaling Pathways

Based on the pathological theory of lipid metabolism and using network pharmacology, this study was designed to investigate the protective effect of water extract of Veratrilla baillonii (WVBF) on non-alcoholic fatty liver disease (NAFLD) model using LO2 cells and to identify the potential mechanism underlying the effect. The components of V. baillonii were identified from the public database of traditional Chinese medicine systems pharmacology database (TCMSP). Cytoscape software was used to construct the related composite target network. Then, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were carried out for critical nodes. The BioGPS database was used to determine the distribution of the target in tissues and organs. Moreover, the inhibitory effect of V. baillonii was further investigated using an in vitro hepatocyte NAFLD model. Fourteen active components were then selected from the 27 known compounds of V. baillonii. The targets of gene enrichment analysis were mainly distributed in the lipid catabolism-related signaling pathway. Network analysis revealed that five target genes of TNF, MAPK8, mTOR, NF-ĸB, and SREBP-1c were key nodes and played important roles in this process. Organ localization analysis indicated that one of the core target site of V. baillonii was liver tissue. The results of the in vitro study revealed that WVBF can alleviate the inflammatory response and lipid accumulation in LO2 hepatocytes by inhibiting oxidative stress and the adipocytokine signaling pathway. Genes and proteins related to the lipid synthesis, such as SREBP-1C, acetyl-CoA carboxylase (ACC), and fatty acid synthase (FASN), were significantly decreased, and PPARα expression is significantly increased with WVBF administration. In conclusion, V. baillonii may regulate local lipid metabolism and attenuate oxidative stress and inflammatory factors through the PPARα/SREBP-1c signaling pathway. The present study also indicates that multiple components of V. baillonii regulate multiple targets and pathways in NAFLD. The findings highlight the potential of V. baillonii as a promising treatment strategy for nonalcoholic fatty liver injury.


INTRODUCTION
Non-alcoholic fatty liver disease (NAFLD), the most common chronic liver disease, is a metabolic syndrome characterized by hepatocellular steatosis and fat storage without a history of excessive alcohol consumption . Globally, NAFLD accounts for 25.2% of global burden of disease epidemics, with approximately 30% of adults having underlying lipid metabolic disorders (Zobair et al., 2016). NAFLD affects nearly one-third of the world's population and the prevalence of this disorder is increasing every year (Hu et al., 2019). Excessive accumulation of triglyceride (TG) due to an imbalance in fatty acid uptake, synthesis, transportation, and oxidation is the leading cause of liver injury (Du et al., 2016). Based on the underlying pathological conditions, NAFLD can further develop into simple fatty liver, non-alcoholic steatohepatitis (NASH), fatty liver fibrosis, or cirrhosis, which significantly increases the associated mortality (Perumpail et al., 2017;Ipsen et al., 2018). Since no pharmacological therapy has been approved for the treatment of NAFLD so far, there is an urgent necessity to define multiple components, multiple targets and pathways of traditional Chinese medicine (TCM) based strategies of future interventions.
Therefore, this study explored the main components, key targets, and signaling pathways of V. baillonii in NAFLD. Further, the efficacy network and pharmacodynamic mechanism of WVBF were investigated using the network pharmacology method. Additionally, LO2 liver cells were exposed to mixed free fatty acids (FFA), to establish an in vitro NAFLD model for investigation of the protective effect of V. baillonii on the liver.

Compound Profiling and Disease Target Identification
The following databases were searched to identify the compounds in V. baillonii: traditional Chinese medicine systems pharmacology database (TCMSP) (http://lsp.nwu.edu.cn/tcmsp.php), a unique pharmacology platform that captures the relationships between herbal ingredients, targets, and diseases: SymMap (http://www. symmap.org/search/). The components were filtered by integrating oral bioavailability (OB ≥ 30%) and drug-likeness (DL ≥ 0.18), as suggested by the TCMSP and SymMap databases. OB and DL were used for candidate active ingredient screening based on a computer integrated model of absorption, distribution, metabolism, and excretion (ADME) (Xu et al., 2014).
Non-alcoholic fatty liver-related targets were retrieved from the online human Mendelian genetics (OMIM, https://omim. org/), DisGeNET (http://www.disgenet.org/), and GeneCards (https://www.genecards.org/) databases. "Non-alcoholic fatty liver disease" was used as the keyword and duplicate values were deleted. The Bioconductor package in the software package R was used to standardize the disease gene targets obtained (Zeng et al., 2019). Finally, Cytoscape 3.2.1 was used to perform visual network analysis of the "disease-target."

Screening of Candidate Targets for the Treatment of Non-Alcoholic Fatty Liver Disease
Based on the previous steps, two sets of target data files were prepared: drug-related component targets and disease-related targets. Cross genes were screened using the Venn Diagram software package in R. Intersecting protein interactions (PPIs) were analyzed using the String 11.0 database and the common targets were counted using R software. Finally, Cytoscape 3.2.1 was used to perform visual network analysis of the "drug-active ingredient-target-disease" network.

Network Construction and Central Network Topological Analysis
PPIs for each target were generated from a string database that provides experimental and predictive interaction information based on systematic co-expression analysis, shared selection signal detection across genomes, and automated text mining of scientific literature . The central network analysis was performed according to the topological method. Three topological parameters, degree centrality (DC), betweenness centrality (BC), and closeness centrality (CC), were calculated to evaluate the central attributes of the nodes in the network . In the target network of WVBF and NAFLD, DC ≥ 2 × median DC, BC ≥ median BC, and CC ≥ median CC were used as the screening criteria to obtain the key targets. The key target interaction network was also depicted using Cytoscape 3.2.1.

Bioinformatics Annotation Analysis
Bioinformatics annotation using Bioconductor R language was used to evaluate genes with high expression patterns. In this study, the PANTHER classification system (http://www.pantherdb.org/), Gene Ontology (GO) annotation database website (http://www. geneontology.org), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis (http://www.genome.jp/ kegg/), and PPIs analysis (http://string-db.org/) were used to analyze the effect of potential targets of active ingredients in V. baillonii on gene function and signaling pathways.

Prediction of Target Organ Recognition
In order to understand the intervention function of V. baillonii in NAFLD, the distribution of targets in tissues and organs was analyzed. To gain information about key targets function, the tissue specific probes of the key targets core were identified using the BioGPS Gene Expression Database (Novartis Research Foundation; http://biogps.org). The tissue distribution of targets was determined, specific methods have been described as the following aspects. Firstly, all the core targets collected above were inputted as keywords into BioGPS for screening, and species of human was selected. Secondly, the detailed information of the core targets was queried and click to retrieve the target to show its expression in different tissues. Finally, the obtained distribution and expression of each target in different tissues were imported into the SPSS, and the gene expression in different tissues was plotted (Pritchard et al., 2012;Wang et al., 2017).

Extract Preparation
The dried roots of V. baillonii ( Figure 1A) were bought from Kunming city, Yunnan province in China. The plant materials were authenticated by Dr. Liu Xinqiao, who is affiliated to the School of Pharmaceutical Sciences, South-Central University for Nationalities, China. The voucher specimens (No. S20140710) were deposited at the Herbarium situated in the College of Pharmaceutical Sciences at South-Central University for Nationalities, Wuhan, Hubei, P. R. China. The WVBF was extracted as described previously (Ge et al., 2016;Yu et al., 2016;Li et al., 2019). Briefly, WVBF was prepared by extracting the plant (200 g) crushed into small pieces. The mixture was  refluxed with water (1:10, w/v) for 2 h. The filtrates were collected and the residues were then refluxed in water (1:10, w/v) for 1.5 h. Two batches of filtrates were combined. Afterwards, the concentrated extract was dried by vacuum concentration to obtain the WVBF extract with a yield of 25.6% (w/w, dried extract/crude herb).

Preparation of Sample Solution
The FFA mixture was preparation as reference described (Zhang et al., 2015). Non-fatty acid BSA was dissolved in DMEM medium to prepare a BSA solution with a concentration of 18.4%. OA (9 mmol·L −1 ) and PA (9 mmol·L −1 ) were each dissolved in medium containing BSA. An FFA/BSA mixture with mixture of (OA: PA, 2:1) was produced. The FFA/BSA mixture was stirred at 37°C for 6 h and the pH of the medium was adjusted to 7.4 with a pH regulator. The medium was filtered aseptically and stored at −20°C for later use.
To detect the cell viability, AST and ALT were used as the basis indices for selecting the optimal time and the concentrations of FFA (Mao et al., 2015;Zhang S. R. et al., 2018).

Assessment of Cell Viability
Cell viability was evaluated by the methylcyclopentadienyl manganese tricarbonyl (MTT) method. After the indicated treatment, MTT was added at a working concentration of 5 mg·ml −1 and the solution was incubated for 4 h. Then, the MTT solution was removed and 150 ml well −1 of dimethyl

Flow Cytometry Analysis of Apoptosis
After treatment, the LO2 cells were collected after centrifugation at 2,000 r·min −1 for 5 min. The cells were treated according to the instructions of the Annexin V-FITC Cell Apoptosis Assay Kit. Then, a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA) was used to evaluate cell apoptosis and necrosis.

Cell Oil Red O Staining
The levels of lipids were assessed by oil red O stains on cells. Firstly, the cells grown in six-well plates were washed with phosphate buffer saline (PBS) three times, then 4% paraformaldehyde was used to fix the cells for 20 min. Secondly, the cells attached to the coverslip were stained with the oil red O (ORO) fixative for 15 min, and soaked with 60% isopropanol for 5 min. Then, hematoxylin stain was added onto the slides for 1 min after the slides were washed with distilled water according to the manufacturer's instructions. The lipid droplets stained with oil red O were visualized with a CKX41 inverted microscope equipped with a DP72 microscope digital camera. Then 100% isopropanol was used to extract intracellular cholesterol. The liquid absorbance was measured at OD490 with spectrophotometer.

Biochemical Analyses
The cell lysates were homogenized and the total lipids were extracted by a mixture of chloroform and methanol (2:1). The protein concentrations of lysis buffer were measured using an enhanced bicinchoninic acid (BCA) protein assay kit. The concentrations of TG, total cholesterol (TC), total superoxide dismutase (T-SOD), glutathione peroxidase (GSH-Px), MDA, and LPO in cell homogenate were determined with commercial kit. All experimental results were normalized according to total protein levels of the samples. The levels of AST and ALT in cell culture supernatant were also determined using a fully automatic biochemical analyzer (Mindray, BS-600, China), according to the manufacturer's protocol.

Analysis of Intracellular Reactive Oxygen Species Generation
Fluorescent dye 2,7-dichlorofluorescein-diacetate (DCFH-DA, Beyotime, China) was used to detect intracellular ROS generation. After the cells were exposed to FFA for 24 h, and treatment of BT, AG, PC, and WVBF, the LO2 cells were washed with DMEM and stained with 10 mm of DCFH-DA for 30 min at 37°C. Subsequently, cells were harvested, rinsed twice with DMEM, and then resuspended in 0.5 ml of DMEM and analyzed for DCF fluorescence by flow cytometry.
Each sample was measured three times. The expression levels of the above genes were normalized to those of glyceraldehyde 3phosphate dehydrogenase (GAPDH) and measured by the comparative 2 −DDCt method.

Western Blotting
Total protein was extracted from frozen LO2 cells by adding protein lysates (Beyotime Biotechnology Co, Ltd, Shanghai, China). Protein was quantified using the Bradford method (Thermo Fisher Scientific Inc, Rockford, IL). Total protein was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (50 mg), and was then transferred to Millipore Corp, Billerica, MA membranes and sealed with 5% skim milk powder at 37°C for 1 h. Subsequently, rabbit anti-mouse monoclonal antibodies against PPARa, SREBP-1c, FASN, Nrf2, and GAPDH (1:1,000; Abcam, Cambridge, MA) were added to the membrane and it was shaken at 4°C overnight. The membrane was washed with phosphate buffered saline with Tween (PBST) three times, for 5 min each time. Then, horseradish peroxidase-labeled goat anti-rabbit secondary antibody (1:4,000; Cell Signaling Technology, Danvers, MA) was added to the membrane and it was incubated at room temperature for 2 h. The membrane was washed twice using tris-buffered saline with Tween (TBST) for 10 min and was then treated with

Statistical Analysis
The results are expressed as mean ± standard error (SE). Multiple comparisons were performed using one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. For all analyses, P values below 0.05 were considered to indicate statistical significance. Analyses were performed using SPSS 18.0 software for Windows.

Active Ingredient Targets and Disease Targets
As shown in Table 3, 14 active ingredients were collected from V. baillonii and 496 target proteins were found. Although the number of targets of each compound was different, the targets of the 14 active ingredients significantly overlapped. Topological analysis of protein-interaction network nodes, including screening of key nodes to eliminate duplicates, revealed a total of 287 target proteins. The constructed "composite target" network was used to evaluate the relationships between 287 targets, which had as many as 2,672 nodes. Targets associated with NAFLD were identified in the OMIM, DisGeNET, and GeneCards databases. In this study, 1,200 target genes of disease were retrieved, and 872 key nodes were obtained through screening.

Target Network Analysis
All active compound protein targets and disease-related proteins were classified into two independent groups. The set and its relations were expressed in a closed-loop form with a fixed position to obtain a Venn diagram and 31 interacting proteins, as shown in Figure 2A. Proteins were obtained using String 11.0 and topological data analysis was conducted through CytoNCA, a plugin for Cytoscape, to obtain a network PPI of 17 key targets, as shown in Figure 2B. Through network prediction, it was found that TNF, MAPK8, mTOR, HRAS, and PTGS2 were important target genes in the first few degrees.

Gene Ontology Function, Kyoto Encyclopedia of Genes and Genomes Pathway, and Localization Analysis of Key Targets
The GO functional analysis revealed that V. baillonii affected 17 potential key target genes of five main biological processes, including regulation of lipid metabolic process, positive regulation of nitric oxide biosynthetic process, positive regulation of fever generation, regulation of monooxygenase activity, and regulation of fatty acid metabolic process, as shown in Figures 3A, B. In addition, a visual analysis was performed on the top 20 signal pathways enriched by KEGG analysis ( Figure 3C). KEGG enrichment analysis demonstrated that many target genes of V. baillonii were closely related to oxidative stress indices, inflammatory factors, the TNF signaling pathway, lipid metabolism, and other signaling pathways such as   the AMPK signaling pathway and PI3K-Akt signaling pathway, among others. Moreover, to study the relationship between V. baillonii and vital organs, the BioGps database was used. The above mentioned 17 key targets were imported in the BioGps database for organ positioning. The network was divided into several tissue modules, including liver (12 targets), heart (11 targets), and kidney (7 targets) ( Figure 3D). The results of this analysis indicated that the targets of the active components of V. baillonii were closely related to lipid metabolism disorder. The majority of targets of V. baillonii were located in liver tissues ( Figure 3E). Lipid metabolic disorders are considered to be one of the vital causes of NAFLD (Ziamajidi et al., 2013). Important target genes of TNF, MAPK8, mTOR, NF-ĸB, and PPARa were mainly distributed in the "adipocytokine signaling pathway". The AMPK pathway and insulin signaling pathway were the other important pathways. These results suggest that the molecular mechanisms of the most active compounds of V. baillonii are related to inflammation and lipid metabolism.
Based on these predictions and the available literature, experiment at the cellular level was conducted then to verified our hypothesis.

Free Fatty Acids Reduced Cell Viability and Increased Lipid Accumulation in LO2 Cells
As shown in Figure 4Aa, FFA (0.125-2.0 mM) dose-dependently reduced the cell viability of LO2 cells after incubation for 24 or 48 h. A 0.5 mmol·L −1 dose of FFA inhibited cell growth by more than 50% after 24 or 48 h. Furthermore, ALT and AST leakage were markedly increased in the FFA group compared with the control group (Figures 4Ab, Ac). As the concentration of FFA increased, the accumulation of red-stained lipid droplets in LO2 hepatic cytoplasms increased in a dose-dependent manner (Figures 4Ad, B). Then, a low cytotoxic dose of 0.5 mmol·l −1 FFA was used and the shorter induction group was exposed for 24 h in the following experiment. Effects of Water Extract of Veratrilla baillonii on Free Fatty Acids-Induced LO2 Cells

Water Extract of Veratrilla baillonii Elevated Cell Viability and Reduced Free Fatty Acids-Induced Cell Apoptosis
After 24 h of FFA treatment, ALT, AST levels in LO2 cells were significantly elevated, cell viability were significantly decreased. ALT, AST levels significantly decreased, the cell viability rates significantly increased in cells treated with WVBF, BT, AG, and PC ( Figures 5A, B). Flow cytometry analysis showed that the increased cell apoptosis and cell necrosis rate induced by FFA could be significantly reduced by WVBF treatment in a dosedependent manner ( Figure 5C), indicating the protective effect of WVBF treatment on FFA-induced LO2 liver injury.

Water Extract of Veratrilla baillonii Attenuates Lipid Metabolic Risk Factors and Accumulation in Cells
FFA resulted in a steady increase in TG, TC levels in liver cells, and increased accumulation of fat droplets in the cytoplasm after incubated for 24 h. And the TG, TC levels could be significantly attenuated by WVBF, BT, AG, and PC ( Figures 6A, B). Moreover, WVBF significantly reduced the FFA-induced accumulation of fat droplets in the cytoplasm after incubated for 24 h ( Figure 6B). The high-dose of WVBF group exhibited the best efficacy and showed a specific dose-dependent enhancement. The above results indicate that WVBF regulate the lipids metabolism in cells, thus remission lipotoxicity induced hepatocyte injury.  The exposure of the cells to FFA resulted in a significant decrease in T-SOD and GSH-PX level, increase in MDA and LPO levels, and produced stronger DCF signals compared to the control group. However, after WVBF, BT, AG, and PC treatment resulted in a marked reduction in DCF fluorescence, indicating an inhibitory effect of WVBF on FFA-induced intracellular ROS production. The above results show that WVBF has certain roles in antioxidation and regulation of oxidation( Figures 7A, B).

Water Extract of Veratrilla baillonii Regulated the PPARa/SREBP/NF-kB Pathway
The key targets and important signaling pathways, including TNF, MAPK8, mTOR, and NF-ĸB targets and the adipocytokine signaling pathway, which were obtained from the network pharmacology analysis, were further explored and verified at the cellular level. As shown in Figure 8A, FFA significantly up-regulated the mRNA levels of SREBP-1c, FASN, ACC, NF-kB, and TNF-a and down-regulated the mRNA levels of Nrf2 and PPARa. The same trends were observed at the protein level ( Figures 8B, C). However, those imbalances were significantly recovered by WVBF in a dose-dependent manner. The results were in accordance with the network analysis to some extent and provide evidence to support the view that activation of oxidative, inflammation, and lipid metabolism stress pathways are involved in a NAFLD cell model.
medicine (Hopkins, 2008). Many studies have shown that the effectiveness of TCM in the complex diseases depends on the synergistic effects of multiple compounds and their targets. Network pharmacology and system biology can explain the effects of drugs on biological network destruction from the perspective of macro or global regulation and provide new research ideas and technical means for studying the mechanism of ethnic drug compounds (Muhammad et al., 2018;. In this study, the complexity of the active ingredients in V. baillonii and the diversity of potential regulatory targets in diseases were studied through network pharmacology analysis. The results revealed that the main active ingredients of V. baillonii, such as amarogentin, 1-hydroxy-2,3,7-trimethoxyxanthone, and 1,7-dihydroxy-3methoxyxanthone, possess anti-inflammatory, antioxidant, and liver protection effects, among others. These findings are consistent with those reported in the literature. Clinical and pharmacological studies have found that V. baillonii has a variety of anti-inflammatory, antioxidant, antiviral, and liver protective effects in autoimmune diseases. It has remarkable efficacy in the treatment of pulmonary fever, enteritis, cholecystitis, hepatitis, and other diseases of the digestive system (Lian et al., 2010).
Based on the present results, a NAFLD cell model establishment by FFA-induced LO2 was preliminarily confirmed (Zhang et al., 2015;Mao et al., 2015). There are significant increases in ALT and AST levels in patients with fatty liver disease. In NAFLD patients, the levels of TG, and TC, indicators of blood lipids, are significantly increased (Chavez-Tapia et al., 2011). A previous study (Ge et al., 2016) reported that WVBF can decrease Aconitum brachypodum-induced acute toxicity in KM mice and can protect liver tissue by decreasing the release of ALT, AST, and TG in serum. The present study also found that the levels of ALT, AST, TG, and TC in NAFLD model cells were significantly reduced after WVBF treatment, suggesting that WVBF can reduce FFAinduced steatosis injury and the accumulation of lipids in hepatocytes.
Oxidative stress plays an essential role in the regulation of NAFLD from steatosis to steatohepatitis, liver fibrosis, and cirrhosis (Izdebska et al., 2017). Oxidation of fatty acids is considered to be an important source of ROS in fatty liver. ROS can attack a variety of unsaturated fatty acids and induce lipid peroxidation in cells, results in the formation of aldehyde byproducts, such as MDA (Takaki et al., 2013). These molecules can spread inside and outside the cell and accelerating the effects of oxidative stress. Our previous study (Yu et al., 2016) found that WVBF has antioxidant effects which are related to the amarogentin components in WVBF. The current study also confirmed that WVBF has a significant antioxidant effect.
The Nrf2 signaling pathway is an important mechanism of cell resistance to oxidative stress. It can regulate the expression of liver detoxification and antioxidant defense genes (Kathirvel et al., 2010). Preliminary results indicate that WVBF can regulate the NRF2 pathway, WVBF treatment of NAFLD cells increased intracellular Nrf2 expression in a dose-dependent manner, suggesting that the antioxidant potential of WVBF could be an important mechanism for improving liver function of NAFLD cells, which is similar to the results of our previous study (Dai et al., 2018).
PPARa plays an essential role in fat synthesis (Qiao et al., 2013). The SREBP-1c subtype is a key regulatory element in the  process of lipid synthesis. Its downstream factors include acetyl-CoA carboxylase (ACC) and fatty acid synthase (FAS), which are essential enzymes for fat production (Kanuri and Bergheim, 2013). FFA treatment enhanced the expression of PPARa and decreased the expression of SREBP-1c in FFA-exposed NAFLD model cells, indicating that WVBF activated the PPARa pathway and inhibited the expression of the SREBP-1c/ACC/FAS pathway. Inflammation is another important factor mediating the pathogenesis of NAFLD (Zhou et al., 2018). TNF-a, a proinflammatory cytokine, is involved in the pathogenesis of NAFLD (Parafati et al., 2018). Continuous accumulation of fatty acids activates the NF-kB pathway, promoting the release of inflammatory cytokines and causing inflammation . In the current study, WVBF inhibited the increased expression of NF-kB and TNF-a induced by FFA, thereby reducing the liver inflammation reaction in the cell. Together, the above findings indicate that WVBF treatment can significantly reduce FFA-induced lipid metabolism disorder, oxidative stress, and the inflammatory response in LO2 cells by activating PPARa and inhibiting fat synthesis of the SREBP-1c/ FASN/ACC pathway. This ameliorates the symptoms of NAFLD and plays a role in protecting the liver (Figure 9). These results are consistent with our predicted target and signaling network analysis.

CONCLUSION
This study revealed the potential active compounds and the protective action of V. baillonii in a NAFLD cell model. WVBF inhibited oxidative stress and the inflammatory response, and reduced lipid metabolism disorder through various signaling pathways, suggesting that V. baillonii may be a candidate drug n=3 (n, the number of experiments), ( ## P < 0.01, ### P < 0.001 vs. control group; *P < 0.05, **P < 0.01, and ***P < 0.001 vs. FFA group).
for the treatment of NAFLD. Since NAFLD is a very complex condition in which many factors are involved, lipid accumulation in LO2 liver cells induced by fatty acids can only partially represent NAFLD condition. Further studies are needed to determine the regulatory effect and mechanism of WVBF in animal models of NAFLD.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

AUTHOR CONTRIBUTIONS
X-JH and C-JH conceived the study, participated in its design and coordination, and helped draft the manuscript. C-JH and SL n =3 (n, the number of experiment), ( ## P < 0.01, ### P < 0.001, vs. control group; *P < 0.05, **P < 0.01, and ***P < 0.001 vs. FFA group). carried out the experiments and drafted the manuscript. JW carried out the network study. JL and G-ZY carried out the chemical analysis of the WVBF, and ZZ participated in the design of the study and performed the statistical analysis.