Mechanism of Action of a Chinese Herbal Compound Containing Quercetin, Luteolin, and Kaempferol in the Treatment of Vitiligo Based on Network Pharmacology and Experimental Verification

Objective This study aimed to explore the mechanisms of Baishi tablets (BSTs) in the treatment of vitiligo through network pharmacology-based identification and experimental validation. Methods In brief, the compounds and related targets of BST were extracted from the TCMSP database, and disease information was obtained from the OMIM, GeneCards, PharmGkb, TTD, and DrugBank databases. A Venn diagram was generated to visualize the common targets of BST and vitiligo. GO and KEGG analyses were performed to explore the potential biological processes and signaling pathways. The PPI network and core gene subnetwork were constructed using STRING and Cytoscape software. In addition, the measurement of apoptosis in PIG1 cells and intracellular reactive oxygen species were measured using quercetin (QU), luteolin (LU), and kaempferol (KA) to protect melanocytes from oxidative stress. Results A total of 55 compounds with 236 targets and 1205 vitiligo-related genes were obtained from the TCMSP database. GO and KEGG analyses were performed to explore the potential biological processes and signaling pathways, revealing that BST may cure vitiligo by influencing the biological processes of cellular oxidative stress and related signaling pathways. A critical subnetwork was obtained with 13 core genes by analyzing the PPI network, which includes HMOX1, CXCL8, CCL2, IL6, MAPK8, CASP3, PTGS2, AKT1, IL1B, MYC, TP53, IFNG, and IL2. Furthermore, a molecular docking analysis was conducted to simulate the combination of compounds and gene proteins, reflecting that QU, LU, and KA can strongly bind the core genes. Through a series of experimental validations, we found that QU, LU, and KA could attenuate H2O2-induced apoptosis in melanocytes. Further evidence revealed that QU, LU, and KA could enhance the scavenging of intracellular reactive oxygen species (ROS). Conclusion Based on the results of network pharmacology analysis and experimental verification, QA, LU, and KA can be utilized to protect PIG1 cells by inhibiting oxidative stress and reducing the intracellular level of ROS. This may explain the underlying mechanism of BST therapy and provide a novel strategy for the treatment of vitiligo.


Introduction
Vitiligo is an acquired, chronic depigmenting disorder of the skin. It results from the selective destruction of melanocytes [1]. Approximately 0.5%-1% of the global population is affected by vitiligo. Vitiligo prevalence is not associated with sex or ethnicity [1][2][3]. Vitiligo is considered to be a multifactorial disease. In addition to genetic and environmental factors, other factors (autoimmune, neural, and oxidative stress) have been suggested to have an efect on vitiligo [4].
Management of vitiligo includes the topical use of glucocorticoids, calcineurin inhibitors, and phototherapy. A small number of patients who meet the indications can also choose surgical transplantation of pigments or decolorization treatment [5]. However, due to the complexity of the primary pathogenesis of vitiligo, the treatment methods stated above cannot cure it; they provide only short-term benefts, and the long-term efects are often unsatisfactory. Moreover, such treatments often introduce toxic side efects (e.g., skin atrophy, phototoxic reactions, and skin cancer).
Traditional Chinese medicine (TCM) is a powerful and far-reaching system of medical treatment. From the viewpoint of compatibility of TCM and the interaction of various traditional Chinese medicines, TCM formulations have relatively low toxicity and few side efects and have shown curative efects in the treatment of several diseases [6]. Studies have shown that TCM formulations have great potential in vitiligo treatment [7,8].
Baishi tablets (BSTs) consist primarily of Chaihu (bupleurum), Chishao (red peony), Zhixiangfu (rhizome cyperi), Baishao (white peony), and Zhishi (citrus aurantium). BSTcan replenish qi, promote stagnation, aid blood circulation and detoxifcation, and dispel wind. Te dermatology department of our institution has used BST to treat vitiligo for many years, and excellent outcomes have been achieved. Our research group conducted a clinical randomized trial in 2015, and the Vitiligo Disease Activity (VIDA) score of the patients treated with BST was signifcantly reduced [9]. Nevertheless, the components of TCM formulations have a wide range of functions and complex active ingredients, and the targets and mechanisms of their regulation have not been fully studied.
Network pharmacology is a promising method that combines pharmacology and computer science to construct and visualize the interaction network of multiple genes, targets, and signaling pathways. It is highly suitable for researching drugs with complex ingredients (e.g., TCM formulations) and is a cost-efective method of drug development [10][11][12].
Herein, we applied network pharmacology to identify the active ingredients and examine the core targets and signaling pathways of BST for vitiligo treatment. We also undertook molecular docking studies to ascertain how BST binds to its predicted targets. Subsequently, we conducted a series of in vitro experiments on immortalized human melanocytes. Te fowchart of our study is shown in Figure 1.

Identifcation of the Active Compounds in BSTand Related
Target Genes. Te compounds Chaihu, Chishao, Xiangfu, Baishao, and Zhishi were searched and obtained from the Traditional Chinese Medicine System Pharmacology Database (TCMSP; http://tcmspw.com/tcmsp.php/) [13], as were their related gene targets and other biological information. Specifcally, to evaluate the characteristics of absorption, distribution, metabolism, and excretion, we used oral bioavailability (OB) and drug likeness (DL) to flter candidate active compounds with thresholds of OB ≥ 30% and DL ≥ 0.18 [14]. Moreover, the targets of the active compounds were transformed into gene symbols via the UniProt database (http://www.uniprot.org/) by limiting the species to Homo sapiens for further analyses [15].

Establishment of a Compound-Target-Vitiligo Network and Functional Analyses.
Having prepared two sets of target lists for the gene targets of compounds and vitiligo-related targets, screening for drug-disease crossover was carried out. A Venn diagram was generated with R (R Institute for Statistical Computing, Vienna, Austria) using the Venn Diagram package to show the intersection sets. A compound-target-disease network diagram was established using Cytoscape 3.8.0 (https://cytoscape.org/) to show the relationship among vitiligo, BST, and the related gene targets [21].
Subsequently, analyses of functional enrichment and enrichment of signaling pathways were undertaken using the gene ontology (GO; http://geneontology.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www. genome.jp/kegg/) databases, respectively. Te "clusterprofle" and "bioconductor" packages within R were employed to assess the biological process (BP), cellular component (CC), molecular function (MF), and key signaling pathways. Signifcantly enriched terms were identifed, and p < 0.05 and q < 0.05 indicated a strong association with related BPs [22,23].

Protein-Protein Interaction (PPI) Networks and Critical
Subnetworks. Te Search Tool for the retrieval of interacting genes/proteins (STRING) (http://string-db.org/) database [24] was used to identify the biological interactions among the potential gene targets. Intersecting PPIs were obtained through the intersecting gene sets of BST and vitiligo. After importing the results of PPIs into Cytoscape, a diagram of critical subnetworks was established, and core genes were investigated using the CytoNca plugin [25]. Eligible genes were selected if each score was higher than the median value of betweenness, closeness, degree, the eigenvector, the local average connectivity-based method, and network scores. After displaying this analytical process twice, the fnal results of the core genes were utilized to establish a critical subnetwork.

Molecular Docking.
Compounds with the top-three highest numbers of related critical genes and their common core genes were selected for molecular docking. After downloading the two-dimensional (2D) molecular structure of ligands from the PubChem database (https:// pubchem.ncbi.nlm.nih.gov/) [26], the 3D structure with the minimum energy was calculated and exported via ChemBio 3D (http://www.adeptscience.co.uk/products/ lab/chembio3d/). Moreover, the 3D structure of the receptor proteins encoded by the core genes was searched in the UniProt database and downloaded from the Research Collaboratory for Structural Bioinformatics Protein Database (https://www.rcsb.org/) [27].
After preparing the fles for the 3D structure, the receptor proteins were dehydrated, and ligands were removed using PyMOL (https://pymol.org/2/). AutoDock (https:// autodock.scripps.edu/) was utilized to modify the receptor protein as well as carry out the hydrogenation and charging calculations of proteins [28]. Subsequently, the parameters of the docking site of the receptor protein were set to include the sites of the active pocket. Molecular docking between compounds and receptors was investigated via Vina within AutoDock [29].

Cells and Cell
Culture. An immortalized human melanocyte cell line (PIG1) was purchased from the American Type Cell Collection (Manassas, VT, USA). PIG1 cells were cultured in Dulbecco's modifed Eagle's medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Gibco). PIG1 cells were cultured in a 37°C incubator in an atmosphere of 5% CO 2 .

Apoptosis
Measurement. PIG1 cells were cultured in 60 mm petri dishes after treatment with 50 μmol/L QU, LU, or KA for 24 h. Ten, H 2 O 2 (fnal concentration: 1 mmol/L) was added to each well, and incubation was undertaken for an additional 2 h. Simultaneously, we set up simple compound-treatment groups and a control group (without any treatment). After the previous treatment, we digested PIG1 cells with EDTA-free trypsin, collected them in tubes, washed them twice with phosphate-bufered saline (PBS), and resuspended them in PBS. According to the protocol, PIG1 cells were stained with an Annexin V-FITC Apoptosis Detection Kit (Liankebio, Hangzhou, China) and detected by fow cytometry. FlowJo (http://www.fowjo.com/) was used to measure the percentage of apoptotic cells.

Measurement of Levels of Intracellular Reactive Oxygen
Species (ROS). QU, LU, and KA were purchased from Aladdin (purity ≥98.5%; Shanghai, China). QU, LU, and KA were dissolved in dimethyl sulfoxide (DMSO; Millipore Sigma-Aldrich, Burlington, Massachusetts MA, USA) for further use. PIG1 cells were cultured in six-well plates after being treated with diferent concentrations (25 or 50 μmol/L) of QU, LU, and KA for 24 h. H 2 O 2 was added (fnal concentration: 1 mmol/L) to each well and incubated for an additional 2 h. After the corresponding treatment, we washed cells twice with serum-free medium. According to product instructions, cells were stained with dichlorodihydrofuorescein diacetate using a reactive oxygen species kit (Shanghai Biyuntian Biotechnology, Shanghai, China) and photographed under an inverted fuorescence microscope (Olympus, Tokyo, Japan).

Statistical Analyses.
Te data are the mean ± SD. Statistical analyses were carried out using Prism 7.0 (GraphPad, San Diego, CA, USA) or SPSS 22  Evidence-Based Complementary and Alternative Medicine were repeated at least three times. p < 0.05 was considered signifcant.

Active Compounds and Potential Targets.
Using the criteria of DL ≥ 0.18 and OB ≥ 30%, 55 main and efcacious compounds of the fve herbs were retrieved and selected (Supplementary Table 1). Subsequently, the compoundrelated targets were annotated into a gene-symbol set using the UniProt database. After removing duplicates, a set of 1205 vitiligo-related targets (Supplementary Table 2A) Table 2B). Moreover, an intersection of the compound targets and vitiligo-related genes, which contained 71 target proteins (Supplementary Table 2C), was obtained ( Figure 2(b)).

Enrichment Analyses.
Analyses of functional enrichment using the GO database revealed the underlying BPs, CCs, and MFs of the 71 target genes. Using p < 0.05 and q < 0.05 as criteria, 2112 signifcantly enriched GO terms were obtained: 1975 BPs, 26 CCs, and 111 MFs. Te top 10 terms is shown in Figure 4(a).
Te top 10 BPs were responses to metal ions; nutrient levels; lipopolysaccharide; molecules of "bacterial origin," "radiation," "oxidative stress," "aging," "antibiotics," "drugs," and "reactive to oxygen species." Te top 10 CCs were "cyclin-dependent protein kinase holoenzyme complex," "membrane raft," "membrane microdomain," "membrane region," "caveola," "serine/ threonine-protein kinase complex," "transcription factor complex," "nuclear chromatin," "mitochondrial outer membrane," and "plasma membrane raft." Te top 10 MFs were "cytokine activity," "cytokine receptor binding," "receptor ligand activity," "phosphatase binding," "transcription cofactor binding," "RNA polymerase II basal transcription factor binding," "heme binding," "protein phosphatase binding," "tetrapyrrole binding," and "transcription coactivator binding." Using p < 0.05 and q < 0.05 as criteria, analyses of signaling-pathway enrichment using the KEGG database were performed. We found that 165 potential signaling pathways were enriched (Supplementary Table 3), and the top 30 signaling pathways are shown in Figure 4(b). Bubble plots demonstrated that these gene targets afected signaling pathways related to the "biological process of oxidative stress such as lipid and atherosclerosis," "AGE-RAGE signaling pathway in diabetic complications," "fuid shear stress," and "atherosclerosis."  Table 4). After importing the results from the PPI network into Cytoscape and using the CytoNCA plugin, these related genes were identifed twice to establish a coregene subnetwork.  Figure 5(b)). Te results from the analyses of functional enrichment and enrichment of signaling pathways using the GO and KEGG databases, respectively, revealed that the core genes were involved in the "cellular response to oxidative stress" and had critical roles in signaling pathways. Detailed information on these compounds is summarized in Supplementary  Figure 6. A greater absolute value for the docking afnity indicates a stronger binding ability between the active site of a protein receptor and a compound. Te docking results indicated that QU, LU, and KA could enter and bind the active pocket of the four core target proteins, could form hydrogen bonds with amino acid residues, and exhibited high binding afnity.    (Figure 8). Tese results suggested that QU, LU, and KA could reduce ROS in melanocytes and protect melanocytes from oxidative stress to a certain extent.

Discussion
Vitiligo is caused by melanocyte destruction. Vitiligo pathogenesis could be due to heredity, autoimmunity, neurochemical factors, or oxidative stress [1,2]. Oxidative     stress may be the initial event leading to vitiligo development [30]. Melanocytes from patients with vitiligo are inherently defective and susceptible to oxidative stress [31]. ROS generation occurs during melanin synthesis by melanocytes. Te stress status of melanocytes can also lead to an excessive accumulation of ROS, which results in changes to the antioxidant system. Te imbalance of oxidative and antioxidant systems in vitiligo patients increases the sensitivity of melanocytes to external oxidants, thereby resulting in premature aging and apoptosis [32,33]. Excessive accumulation of ROS can also cause cellular DNA damage and lipid peroxidation, which afect cellular function [34,35]. Terefore, reducing oxidative stress in melanocytes should be a rational strategy for vitiligo treatment.
BST has a satisfactory efect in the clinical treatment of vitiligo. However, due to the complex components of TCM formulations, a more accurate and systematic study of their possible targets and mechanisms is needed. We used network pharmacology to explore the mechanism of action of BST for vitiligo treatment. QU, LU, KA, baicalein, nobiletin, and isorhamnetin were screened out as the main active ingredients of BST. QU, LU, and KA had the most extensive targets in vitiligo, so we selected them for experimental verifcation.
QU is a polyphenolic favonoid found widely in onions, cabbage, apples, and tea [36,37]. In recent years, scholars have revealed that QU has powerful antioxidant efects. It has a preventive efect on osteoporosis, certain tumor types, and certain cardiovascular diseases. Te antioxidant activity of QU occurs mainly through the direct induction of glutathione (GSH) synthesis in the body. GSH acts as a hydrogen donor for redox reactions in the body, while superoxide dismutase captures O 2 molecules and transforms them into H 2 O 2 , thereby having an antioxidant efect [38]. In addition, the -OH group in QU can bind to the active sites of oxidative enzymes, such as acetylcholinesterase and butyrylcholinesterase, to inhibit their activity and elicited an antioxidant efect [39]. QU can also regulate the NRFB, 5′ adenosine monophosphate-activated protein kinase, and mitogenactivated protein kinase signaling pathways [40,41]. Studies have shown that QU increases the tyrosinase activity and synthesis of melanoma cells and normal melanocytes to promote melanogenesis [42]. QU has a weakening and protective efect on H 2 O 2 -induced endoplasmic reticulum stress in melanocytes [43]. Here, we demonstrated that a certain concentration of QU reduced the ROS level in human immortalized melanocytes induced by H 2 O 2 and had a certain degree of protection against oxidative stress in melanocytes. Combined with the results of network pharmacology, we showed that QU had the most extensive binding targets in BST.
LU is a favonoid found in vegetables and fruits and is used in Chinese herbal medicines [44,45]. TCM formulations containing LU have been employed to treat high blood pressure, infammation, and cancer [44]. LU can inhibit the release of interleukin (IL)8, a critical proinfammatory chemokine in vitiligo and may have the potential to treat vitiligo [46].
KA is a natural favonoid found in tea, fruits, and vegetables. Lee and colleagues found that KA could inhibit ultraviolet B-induced expression of cyclooxygenase-2 (COX-2) release in mouse skin epidermal (JB6P + ) cells and attenuate ultraviolet B-induced COX-2 release and activator protein-1 transcriptional activity [47]. KA can also improve the skin fbrosis induced by bleomycin by reducing oxidative stress and infammation [48]. Our experimental study revealed that a certain concentration range of KA could reduce ROS levels and protect melanocytes.
Hence, QU, LU, and KA are favonoids that have antiinfammatory and antioxidant efects. Terefore, we speculated that they could afect the oxidative stress of melanocytes. We conducted experiments to verify that under oxidative stress (mimicked by H 2 O 2 use), melanocyte apoptosis and ROS production were reduced under pretreatment by QU, LU, or KA. Terefore, BST may have a specifc protective role in the oxidative stress pathway of melanocytes.
In the PPI network analysis of BST, HMOX1, CXCL8, CCL2, IL6, MAPK8, CASP3, PTGS2, AKT1, IL1B, MYC, TP53, IFNG, and IL2 were screened out, all of which are core targets in the treatment of vitiligo. PTGS2 is also known as COX2, which plays an essential part in producing prostaglandin (PG)E2 and is made by epidermal keratinocytes in response to ultraviolet radiation [49,50]. PGE2 is essential for the proliferation and melanogenesis of melanocytes, the loss of which can lead to vitiligo. In addition, studies have shown that the functional polymorphisms of COX2 afect the risk of vitiligo [51]. Heme oxygenase-1 (HMOX1) is the most highly induced antioxidant gene in H 2 O 2 -treated PIG1 cells. HMOX1 has been demonstrated to protect human melanocytes from oxidative damage via the E2-related factor 2 (Nrf2)-antioxidant response element (ARE) pathway [52]. AKT1 is a RAC-alpha serine/threonine protein kinase. Te phosphorylation of AKT1 could promote the accumulation of β-catenin, thereby activating the microphthalmiaassociated transcription factor and tyrosinase family, eventually leading to melanogenesis of melanocytes [53]. Overexpression of cellular tumor antigen P53 (TP53) could protect the pigmentation around hair follicles in vitiligo patients after ultraviolet-B treatment, change the migration ability of melanocytes, and improve pigmentation in vitiligo patients [54]. IL-6 is a vital immune factor involved in autoimmune infammation in vitiligo. Its increased expression in serum and a skin lesion could trigger an immune response that targets and kills melanocytes and leads to vitiligo [1,55]. Tese results suggest that QA, LU, and KA could protect melanocytes, promote melanogenesis, inhibit melanocyte death, and protect melanocytes from oxidative damage through various mechanisms.
Our study had three main limitations. First, experimentally validated targets were the predicted results of network pharmacology, but there were certain deviations compared with the actual targets. Second, we revealed that the main compounds QU, LU, and KA could protect melanocytes from oxidative stress; however, how they regulate targets and afect downstream signaling pathways to have a role in vitiligo treatment was not tested. Tird, our study Evidence-Based Complementary and Alternative Medicine was based on network pharmacology, so the compounds with the most targets were selected for experimental verifcation according to the results of network pharmacology, but their concentration in BST could not be determined.

Conclusions
Under the prediction obtained using network pharmacology, we clarifed the active compounds in BSTand their main targets in vitiligo treatment. Based on network pharmacology and in vivo experiments, QA, LU, and KA can be utilized to protect PIG1 cells. Tis phenomenon was achieved thanks to the inhibition of oxidative stress by reducing the intracellular level of ROS. Tis may explain the underlying mechanism of action of BST therapy and could provide a novel strategy for the treatment of vitiligo.

Data Availability
Te datasets presented in this study are openly available from TCMSP, GeneCardS, OMIM, TTD, PharmGkb, and DrugBank belong to public databases. Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that there are no conficts of interest.

Authors' Contributions
Ziqian Xu and Yihui Xie contributed equally to this work. WS and JH conceived and designed the study. ZX and YX provided equal contributions to research design, data analysis, and article writing. JS revised the manuscript. All persons designated as the authors have participated sufciently in the work to take public responsibility for the content of the manuscript. All the authors ensure that they all gave substantial contributions.

Supplementary Materials
Supplementary