A Systematic Study of Yiqi Qubai Standard Decoction for Treating Vitiligo Based on UPLC-Q-TOF/MS Combined with Chemometrics, Molecular Docking, and Cellular and Zebrafish Assays

The Yiqi Qubai (YQ) formula is a hospital preparation for treating vitiligo in China that has had reliable efficacy for decades. The formula consists of four herbs; however, the extraction process to produce the formula is obsolete and the active ingredients and mechanisms remain unknown. Therefore, in this paper, fingerprints were combined with the chemometrics method to screen high-quality herbs for the preparation of the YQ standard decoction (YQD). Then, the YQD preparation procedure was optimized using response surface methodology. A total of 44 chemical constituents, as well as 36 absorption components (in rat plasma) of YQD, were identified via UPLC-Q-TOF/MS. Based on the ingredients, the quality control system of YQD was optimized by establishing the SPE-UPLC-Q-TOF/MS identification method and the HPLC quantification method. Network pharmacological analysis and molecular docking showed that carasinaurone, calycosin-7-O-β-d-glucoside, methylnissolin-3-O-glucoside, genkwanin, akebia saponin D, formononetin, akebia saponin B, and apigenin may be the key active components for treating vitiligo; the core targets associated with them were AKT1, MAPK1, and mTOR, whereas the related pathways were the PI3K-Akt, MAPK, and FoxO signaling pathways. Cellular assays showed that YQD could promote melanogenesis and tyrosinase activity, as well as the transcription and expression of tyrosinase-associated proteins (i.e., TRP-1) in B16F10 cells. In addition, YQD also increased extracellular tyrosinase activity. Further efficacy validation showed that YQD significantly promotes melanin production in zebrafish. These may be the mechanisms by which YQD improves the symptoms of vitiligo. This is the first systematic study of the YQ formula that has optimized the standard decoction preparation method and investigated the active ingredients, quality control, efficacy, and mechanisms of YQD. The results of this study lay the foundations for the clinical application and further development of the YQ formula.


Introduction
Vitiligo is the most common hypopigmentation disorder.It is an acquired disease characterized by progressive loss of melanocytes.Its global prevalence is 0.5-2%, with Pharmaceuticals 2023, 16, 1716 2 of 29 no racial or sex differences [1].Vitiligo is a complex disease whose pathogenesis remains unknown but may be related to genetic variation, autoimmunity, oxidative stress, the autophagy of melanocytes, and/or neuroendocrine function [2].The goal of vitiligo treatment is to suppress depigmentation and stimulate repigmentation in the patients.Currently, the available treatments for vitiligo mainly include glucocorticoids, immunosuppressants, phototherapy, and surgical methods [3][4][5][6].Unfortunately, the available therapies are not effective for all patients, and some have wide-ranging side effects that suppress further treatment [7].Therefore, increasing research has focused on finding new multi-target therapeutic medicine with fewer side effects to fit the complex treatment mechanisms of vitiligo and the need for long-term medication.
There is a long history of using traditional Chinese medicine (TCM) to treat vitiligo in China, with which there have been positive results and few adverse effects [8].The theory of using TCM in the treatment of vitiligo is that it benefits "qi" by activating "blood" and dispelling pathologic "wind" [9]."Qi" is a theory of Chinese medicine that represents the substance that constitutes the human body, maintains human life activities, and promotes Yang, the driving force of metabolism and physiological function of the human body [10].The Yiqi Qubai (YQ) formula is a TCM that has been utilized in the dermatology department of Shanghai General Hospital for decades [11] and has had remarkable curative results [12,13].It consists of four herbs, Astragali radix (Huangqi, HQ), Akebiae fructus (Yuzhizi, YZZ), Leonuri fructus (Chongweizi, CWZ), and Caragana sinica roots (Jinquegen, JQG), in a weight ratio of 1:1:0.75:1.5.Each of the four herbs can soothe the liver and regulate "qi", activate "blood", and dispel "wind" [14][15][16].The upregulation of mushroom tyrosinase (TYR) activity and melanogenesis of YQ was demonstrated using serum pharmacological methods [17].Although YQ has been used clinically for many years, there is a lack of systematic research on it.Firstly, the decoction process does not follow the standard decoction principles, which may affect the uniformity of the preparation quality.Secondly, the composition of the decoction is unclear, which is not conducive to an in-depth study of its material basis and mechanisms.In addition, the identification methods in the existing quality standards are simple and lack quantitative methods to ensure quality.
Essentially, modern TCM preparations rely on the decoction, which is the oldest Chinese medicinal compound that can reflect the material basis of a prescription; therefore, the quality of the decoction is crucial.The Chinese Pharmacopoeia Commission formally proposed the "standard decoction" in 2016, which is characterized by the authenticity of the herbs, the uniformity of the extraction process, and the rigor of quality control [18].The preparation principle of the standard decoction is based on the theory of TCM, following the boiling principle of TCM decoction, adopting standardized technology, and decoction with water.The preparation method is recommended to be optimized with reference to "the management standard of traditional Chinese medicine decoctions in medical institutions" published by the State Drug Administration of China [19].The standard decoction of TCMs has a uniform process and a product with consistent quality; this will assist in the preparation of granules and allow preparations to be widely used in clinical practice.In order to ensure the accuracy of the medication and the consistency of the administered dose, it is necessary to establish a method for the preparation of a YQ standard decoction (YQD) that could be the basis for the clinical use of the medication.
As the raw material of a standard decoction, the quality of the TCM herbs is affected by origin, harvesting season, storage time, processing method, etc., which largely determines the efficacy of TCM preparations.The TCM fingerprint is a comprehensive, quantifiable technology that can provide more detailed information regarding the type and contents of the TCM components [20,21].As data mining methods have developed, chemometrics has gained more and more attention because it can simplify complex data and uncover hidden patterns.The combination of chemometric methods such as hierarchical cluster analysis (HCA) and principal component analysis (PCA) with fingerprints is often used to classify and evaluate the quality of TCM herbs [22,23].In recent years, mouse B16F10 melanoma cells have been used as a sensitive and reliable model for the quantitative analysis of melanin [24].In addition, zebrafish have a high degree of chromosome homology with humans and have a translucent body that allows direct visualization of the development of melanin streaks, making it a suitable model organism for studying melanogenesis in vivo [25].Melanocyte-specific enzymes, such as TYR and the TYR-associated proteins TRP-1 and TRP-2, are involved in melanin synthesis.TYR is the main rate-limiting enzyme in melanin synthesis, and any factor that activates TYR activity can effectively promote the production of melanin [26].
In this study, HPLC fingerprints were combined with chemometrics to establish a quality control system of raw herbs to screen high-quality herbs for the preparation of YQD.The preparation process of YQD was then optimized via response surface methodology (RSM).A UPLC-Q-TOF/MS technique was used to analyze the chemical constituents of YQD and the absorption components (in rat plasma).Next, the SPE-UPLC-Q-TOF/MS identification method and HPLC quantification method were developed to improve the quality control system.Network pharmacology analysis and molecular docking based on the identified components were used to predict the active ingredients, targets, and mechanisms of vitiligo treatment.Then, we investigated the effects of YQD on melanogenesis, TYR activity, and TRP-1 expression in B16F10 cells, as well as extracellular TYR activity, to explore the possible mechanisms.In addition, the efficacy was investigated by studying its effect on melanogenesis in zebrafish.To the best of our knowledge, this is the first systematic study on the preparation, composition, quality control, efficacy, and mechanism of YQD.The results of this study lay the foundations for the clinical application and further development of YQD.The whole study process is summarized in Figure 1.

HPLC-DAD Analysis and Similarity Evaluation
Biological and abiotic factors affected the chemical quality of plants and showed wide geographical variation.In order to ensure the stability of the raw materials, it is necessary to study the origins before establishing a standard decoction preparation process.Using HPLC-DAD, 15 batches of HQ (H-1-H-15), YZZ (Y-1-Y-15), CWZ (C-1-C-15), and JQG (J-1-J-15) from different origins were detected and their chromatograms were recorded.The precision, repeatability, and stability of the analytical method were validated.The results indicated that the fingerprint analysis method was stable and reliable.The similarity of the fingerprint chromatograms was assessed.Table S1 shows the similarity matching results for the fingerprints.A total of 11, 10, 4, and 9 common peaks were identified on the fingerprint spectra for HQ, YZZ, CWZ, and JQG, respectively (Figure 2A).The intragroup similarities of HQ, YZZ, CWZ, and JQG were 0.919-0.998,0.773-0.997,0.932-0.999,and 0.883-0.999,respectively.As shown in the results, each of the four herbs, when obtained from different production areas, differed within the group, especially YZZ and JQG; these differences may be due to the different geographies, harvesting times, and sources of the herbs.Biological and abiotic factors affected the chemical quality of plants and showed wide geographical variation.In order to ensure the stability of the raw materials, it is necessary to study the origins before establishing a standard decoction preparation process.Using HPLC-DAD, 15 batches of HQ (H-1-H-15), YZZ (Y-1-Y-15), CWZ (C-1-C-15), and JQG (J-1-J-15) from different origins were detected and their chromatograms were recorded.The precision, repeatability, and stability of the analytical method were validated.The results indicated that the fingerprint analysis method was stable and reliable.The similarity of the fingerprint chromatograms was assessed.Table S1 shows the similarity matching results for the fingerprints.A total of 11, 10, 4, and 9 common peaks were identified on the fingerprint spectra for HQ, YZZ, CWZ, and JQG, respectively (Figure 2A).The intragroup similarities of HQ, YZZ, CWZ, and JQG were 0.919-0.998,0.773-0.997,0.932-0.999,and 0883-0.999,respectively.As shown in the results, each of the four herbs, when obtained from different production areas, differed within the group, especially YZZ and JQG; these differences may be due to the different geographies, harvesting times, and sources of the herbs.

Hierarchical Cluster Analysis (HCA)
In order to evaluate the similarity of the samples based on their fingerprint data, HCA was used to categorize them [22].As shown in Figure 2B, HQ samples were divided into four clusters, of which H-11 (Zhangye, Gansu) and H-15 (Shanxi) were classified as group I and group II, respectively, H-12 (Zhangye, Gansu) and H-14 (Neimeng) were classified as group III, and the remaining samples (Dingxi, Gansu) were classified as group IV.The YZZ samples were divided into three clusters, of which Y-11 (Yichang, Hubei) was a single group.CWZ samples were divided into three clusters, C-11 (Nanyang, Henan) and C-14 (Anhui) were classified as group I and group II, respectively, and the other samples (Nanyang, Henan; Chifeng, Neimeng; and Fuxin, Liaoning) were classified as group III.The JQG samples were divided into three clusters, J-15 (Sichuan) was alone in group I, J-10 (Yichang, Hubei) and J-11 (Yichang, Hubei) were in group II, and the remaining samples (Jingmen/Xiangyang/Yuan'an, Hubei) were classified as group III.The findings reveal that samples from similar origins were more likely to be placed into the same group, meaning the content of the ingredients was comparable.Nevertheless, some samples from the same origin were categorized independently, probably because of batch variations.There was not a clear regional clustering of YZZ, which may be related to the fact that it is collected from a variety of sources (YZZ is the dry fruit of Akebia quinata (Thunb.)Decne., A. trifoliata (Thunb.)Koidz.var.australis (Diels) Rehd., and A. trifoliata (Thunb.)Koidz.).Differences in the chemical makeup of the herbs can affect their quality; therefore, the quality of herbs from various locations needs to be assessed to determine the best source of raw materials.

Principal Component Analysis (PCA)
As a statistical method of dimension reduction, PCA recombines the original variables into independent comprehensive variables [23].Calculating the composite score of principal components (PCs) and sorting could help to determine the relative quality of each sample.In this work, PCA was performed using SIMCA (ver.14.1).The eigenvalue and variance contribution rate of the correlation coefficient are shown in Table S2.With an eigenvalue >0.99, the first three PCs in HQ had the cumulative variance contribution rate of 81.609%; the first two PCs in YZZ had the cumulative variance contribution rate of 85.467%; the first one PC in CWZ was 73.803%; and the first three PCs in JQG had the cumulative variance contribution rate of 88.242%.These PCs contained the most information on the variables and fully reflected the original data.The factor load matrix and loadings plot (Figure S1) revealed that peaks 5 (calycosin-7-O-D-glucoside), 7, 8, and 11 in HQ; peaks 5 (cryptochlorogenic acid), 4, 7, and 2 in YZZ; peaks 1, 3, and 2 (chlorogenic acid) in CWZ; and peaks 7, 1, and 9 in JQG all demonstrated a higher contribution to PC1 than the other peaks.These peaks might be the primary chemical characteristic peaks that can be utilized to differentiate the quality of the different varieties of the four herbs.
The PC scores and comprehensive score values of the 15 batches of HQ, YZZ, CWZ, and JQG were calculated using the initial eigenvalue and weights of the PCs.A higher score means better quality.As shown in Table 1, the comprehensive score of H-12 (Zhangye, Gansu), Y-11 (Yichang, Hubei), C-13 (Chifeng, Neimeng), and J-11 (Yichang, Hubei) were highest for the HQ, YZZ, CWZ, and JQG samples, respectively.Similar to the HCA, samples with the highest and lowest comprehensive scores were individually clustered in the PCA score chart (Figure 2C).The ecological settings and herb quality are closely connected, and standard decoctions should be made from suitable and stable herbs.Our results showed that H-12, Y-11, C-13, and J-11 were better quality and should be used for the subsequent preparation of YQD.RSM is a collection of statistical and mathematical methodologies for developing and optimizing procedures in which many variables impact a desired response; it has the aim of optimizing that response [27].The Box-Behnken design (BBD), a type of RSM, reduces the number of tests and investigates the interactions between the influencing factors to obtain the optimal process parameters quickly and effectively [28].The experimental parameters were optimized using BBD based on the results of the single-factor studies of the solid-liquid ratio, decoction time, and concentration process.Three independent variables, namely, the first solid-liquid ratio (X 1 ), the first extraction time (X 2 ), and the second extraction time (X 3 ), were selected to be included in the response surface modeling, and a three-factor, three-level (−1, 0, +1) BBD was carried out.Synthesizing all indicators (calycosin-7-O-β-d-glucoside and chlorogenic acid contents of the extract and its total solid content), the overall desirability (OD) value was used to express the overall impact results.According to the Hassan method [29], the OD value was calculated as follows: The di value of each indicator was used to calculate the total OD as follows [30]: A total of 17 experiments were conducted (n = 17) and the BBD test design and response values are shown in Table 2. To determine the predicted response values, the following equations were fitted using a quadratic mode: where X 1 is the first liquid-solid ratio, X 2 is the first extraction time, and X 3 is the second extraction time.A further ANOVA was performed on the regression model (Table S3).The table showed that the model was statistically significant (p < 0.05) and that the linear terms (X 1 , X 2 , and X 3 ) had a significant (p < 0.05) effect on extraction.The interaction surfaces and contour plots of X 1 , X 2 , and X 3 are shown in Figure 3A-D.The plots showed that when the first extraction time (X 2 ) or the second extraction time (X 3 ) was fixed at a certain level, the OD value increased rapidly alongside the increase in the first liquid-solid ratio (X 1 ) and then decreased slightly.This indicated that the increase in the first liquid-solid ratio was not always conducive to the response and that saturation may have occurred.By adjusting the liquid-solid ratio appropriately, two compounds could be yielded more efficiently.The interaction surfaces and contour plots between X 2 and X 3 are shown in Figure S2.In cases where the second extraction time was fixed at a certain level, the OD score increased with the first extraction time; however, this increase was small.Perhaps the increased extraction time resulted in a more efficient destruction of the plant cells, resulting in a higher extraction rate; however, this interaction was not the primary cause.The optimal extraction conditions for YQD were calculated by solving the model's extreme values and analyzing the contours.The first liquid-solid ratio was 8.086, the first extraction time was 59.998 min, and the second extraction time was 39.998 min.The optimal extraction process resulted in an OD value of 1.06.Due to actual manufacturing considerations, the optimum conditions of YQD were as follows: a liquid-solid ratio of 8, the first extraction time was 60 min, and the second extraction time was 40 min.

Experimental Validation of the Model
The optimized YQD extraction process was as follows: 100 g of the prescription amounts of the HQ, YZZ, CWZ, and JQG pieces were blended together and extracted with eight times the amount of water.After being soaked for 30 min, fire-boiled, and decocted on a small fire for 60 min, the mixture was immediately filtered through a 200-mesh screen.The filter residue was then extracted with seven times the amount of water, fireboiled, decocted on a small fire for 40 min, and immediately filtered again.The twice-

Experimental Validation of the Model
The optimized YQD extraction process was as follows: 100 g of the prescription amounts of the HQ, YZZ, CWZ, and JQG pieces were blended together and extracted with eight times the amount of water.After being soaked for 30 min, fire-boiled, and decocted on a small fire for 60 min, the mixture was immediately filtered through a 200-mesh screen.The filter residue was then extracted with seven times the amount of water, fire-boiled, decocted on a small fire for 40 min, and immediately filtered again.The twice-combined extraction decoction was concentrated to 0.2 g/mL under reducing pressure at 60 • C.
Six validation experiments were conducted to verify the above optimal solution.The OD value was 0.96 and the average contents of calycosin-7-O-β-d-glucoside and chlorogenic acid were 0.107 mg/g and 0.338 mg/g, respectively.Table S4 contains the detailed extraction rates.

Identification of Chemical Constituents
UPLC-Q-TOF/MS was used for qualitative analysis of YQD.First, the MS data were matched with the Natural Products HR-MS/MS Spectral Library 1.0 database.Compounds not included in the database were identified in the literature and mass spectrometry pyrolysis laws.Finally, 44 chemical compositions (C1-C44) were identified in YQD (Figure 4A,B and Table 3).Among them, C12 from HQ and C13, C24, C30, and C31 from YZZ had higher response values.Typical compounds of flavonoids and triterpenoid saponins were used to analyze the cleavage pattern as follows:      [31,32], C24 was identified as abekia saponin.The MS 2 spectrum and fragmentation pathway of C12 and C24 are shown in Figure 5.All of the above results are beneficial for the further discovery of the absorbed ingredients after oral administration.

Identification of Absorbed Components in Rat Plasma
A total of 36 components of YQD were identified from rat plasma using background reduction, baseline correction, and molecular characterization extraction methods, of which, 4 were prototype components (P1-P4) and 32 were metabolites (M1-M32).Results are shown in Figure 4C and Table 4.The prototype components were stachydrine (P1), carasinaurone (P2), calycosin-7-O-β-d-glucoside (P3), and akebia saponin B (P4).Among them, P1 and P3 are the quality control indicators of HQ and CWZ, respectively, in the Chinese Pharmacopoeia (2022) [33].P2 and P4 are the signature active ingredients of JQG and YZZ, respectively [34,35].Due to low oral bioavailability and a short half-life, the components with high responses in vitro (C13, C24, C30, and C31) were not found as prototypes or metabolites in vivo [36][37][38].By studying the metabolic pathways in detail, we found that the metabolites were mainly from glucuronidation, sulfation, hydroxylation, and demethylation reactions of the flavonoids in YQD.

Identification of Absorbed Components in Rat Plasma
A total of 36 components of YQD were identified from rat plasma using background reduction, baseline correction, and molecular characterization extraction methods, of  Glucuronidation: M2 showed a quasi-molecular ion at m/z 621.1487 in the negative ion mode that was 176 Da higher than calycosin-7-O-β-d-glucoside at m/z 447.1286, suggesting that calycosin-7-O-β-d-glucoside had undergone a glucuronidation reaction.The fragment ions of M2 at m/z 283.0620 and 268.0344 were like calycosin-7-O-β-d-glucoside; therefore, it was speculated that M2 was a glucuronidation of calycosin-7-O-β-d-glucoside.Similarly, the metabolites M10, M23, M6, and M27 were speculated to be glucuronide products of apigenin, formononetin, genkwanin, and methylnissolin, respectively.
Sulfation: M32 showed a quasi-molecular ion at m/z 347.0230 in the negative ion mode that was 80 Da higher than formononetin at m/z 267.0663, indicating that a sulfation reaction had occurred in formononetin.Fragment ions of M32 at m/z 267.0650 and 252.0431 were similar to formononetin; therefore, it was speculated that M32 was a sulfated metabolite of formononetin.
Glucuronidation + Sulfation: in the negative ion mode, M5 showed a quasi-molecular ion at m/z 525.0359, a fragment ion at m/z 269.0465, and was speculated to be the sulfated and glucuronidated product of apigenin because the quasi-molecular ion was 176 and 80 Da higher than that of apigenin at m/z 269.0455, and the fragment ion was similar to apigenin.
Demethylation: M24 was tentatively identified as a demethylated product by comparing the quasi-molecular ion between M24 and formononetin; the gap was 14 Da in negative ion mode.
In addition, M30 showed a quasi-molecular ion at 303.1212 and fragment ions at m/z 303.1024, 167.0699, 133.0677, and 123.0434 and so was hypothesized to be isoflavan transformed by methylnissolin [39].The [M − H] − m/z of M9 was 541.0301, which was 176, 80, and 17 Da higher than that of apigenin.It was speculated that M9 was the hydroxylation, sulfation, and glucuronidation product of apigenin.The possible metabolic pathways are shown in Figure 6.
Pharmaceuticals 2023, 16, x FOR PEER REVIEW 18 of 36 therefore, it was speculated that M2 was a glucuronidation of calycosin-7-O-β-d-glucoside.Similarly, the metabolites M10, M23, M6, and M27 were speculated to be glucuronide products of apigenin, formononetin, genkwanin, and methylnissolin, respectively.Sulfation: M32 showed a quasi-molecular ion at m/z 347.0230 in the negative ion mode that was 80 Da higher than formononetin at m/z 267.0663, indicating that a sulfation reaction had occurred in formononetin.Fragment ions of M32 at m/z 267.0650 and 252.0431 were similar to formononetin; therefore, it was speculated that M32 was a sulfated metabolite of formononetin.
Glucuronidation + Sulfation: in the negative ion mode, M5 showed a quasi-molecular ion at m/z 525.0359, a fragment ion at m/z 269.0465, and was speculated to be the sulfated and glucuronidated product of apigenin because the quasi-molecular ion was 176 and 80 Da higher than that of apigenin at m/z 269.0455, and the fragment ion was similar to apigenin.
Demethylation: M24 was tentatively identified as a demethylated product by comparing the quasi-molecular ion between M24 and formononetin; the gap was 14 Da in negative ion mode.
In addition, M30 showed a quasi-molecular ion at 303.1212 and fragment ions at m/z 303.1024, 167.0699, 133.0677, and 123.0434 and so was hypothesized to be isoflavan transformed by methylnissolin [39].The [M − H] -m/z of M9 was 541.0301, which was 176, 80, 17 Da higher than that of apigenin.It was speculated that M9 was the hydroxylation, sulfation, and glucuronidation product of apigenin.The possible metabolic pathways are shown in Figure 6.

Establishment of the SPE-UPLC-Q-TOF/MS Identification Method
Identification is an important part of the quality standard of TCM.Traditional TLC or HPLC methods often have low sensitivity, many interfering factors, and complicated pretreatment methods that cannot meet the identification needs of TCM compound preparations.In this study, the SPE-UPLC-Q-TOF/MS technique was used to identify the characteristic peaks of YQD; the separation of the components was better after SPE extraction.Asperulosidic acid (C3) and carasinaurone (C11) in the H 2 O eluting phase; neochlorogenic acid (C5), chlorogenic acid (C7), and cryptochlorogenic acid (C8) in the 30% methanol elution phase; and calycosin-7-O-β-d-glucoside (C12) and calceolarioside B (C13) in the 60% methanol elution phase were selected as the characteristic components because they were from four separate herbs with high response values, complete separation, and good peak shapes.
The results of sample identification showed that, in the H 2 O elution section, the retention times of C3 and C11 were 21.81 and 36.12 min, respectively; in the 30% methanol elution section, the retention times of C5, C7, and C8 were 23.46, 27.17, and 27.91 respectively; and in the 60% methanol elution section, the retention times of C1 and C13 were 36.41 and 38.38 min, respectively (Figure 7I).
Identification is an important part of the quality standard of TCM.Traditional TLC or HPLC methods often have low sensitivity, many interfering factors, and complicated pretreatment methods that cannot meet the identification needs of TCM compound preparations.In this study, the SPE-UPLC-Q-TOF/MS technique was used to identify the characteristic peaks of YQD; the separation of the components was be er after SPE extraction.Asperulosidic acid (C3) and carasinaurone (C11) in the H2O eluting phase; neochlorogenic acid (C5), chlorogenic acid (C7), and cryptochlorogenic acid (C8) in the 30% methanol elution phase; and calycosin-7-O-β-d-glucoside (C12) and calceolarioside B (C13) in the 60% methanol elution phase were selected as the characteristic components because they were from four separate herbs with high response values, complete separation, and good peak shapes.
The results of sample identification showed that, in the H2O elution section, the retention times of C3 and C11 were 21.81 and 36.12 min, respectively; in the 30% methanol elution section, the retention times of C5, C7, and C8 were 23.46, 27.17, and 27.91 min, respectively; and in the 60% methanol elution section, the retention times of C1 and C13 were 36.41 and 38.38 min, respectively (Figure 7I).

Establishment of HPLC Quantitative Method
An HPLC method was established for the determination of calycosin-7-O-β-d-glucoside (CAG, C12) and chlorogenic acid (CA, C7).CAG is a representative component of HQ [33].CA may promote melanogenesis in melanoma cells [40] and had a higher response value in the UPLC/MS chromatogram.Therefore, these two substances were selected as quantitative components to compensate for the lack of quantitative analysis in the existing quality standards.

Establishment of HPLC Quantitative Method
An HPLC method was established for the determination of calycosin-7-O-β-d-glucoside (CAG, C12) and chlorogenic acid (CA, C7).CAG is a representative component of HQ [33].CA may promote melanogenesis in melanoma cells [40] and had a higher response value in the UPLC/MS chromatogram.Therefore, these two substances were selected as quantitative components to compensate for the lack of quantitative analysis in the existing quality standards.
As can be seen from Figure 7II, the response value of the target peak was high, the separation degree was good, and the other peaks did not interfere with the measurement.The standard curves were Y = 0.292X-0.3234,r = 0.9998 (CAG) and Y = 0.099X-0.0567,r = 1.000 (CA).The linear ranges for CAG and CA were 5.21-208.4µg/mL and 10.32-412.8 µg/mL, respectively.The CAG and CA control solutions were taken separately, and the peak areas were measured via successive injections six times according to the established HPLC conditions.The results showed that the RSDs of CAG and CA were 0.25% and 1.63%, respectively (n = 6), which indicated that the precision of the instrument was good.For the evaluation of repeatability, the RSDs of the peak areas from six samples of the same batch were calculated and were 0.77% and 3.31%, respectively (n = 6), indicating that the repeatability of the method was good.The stability of the samples was determined by re-analyzing the RSDs of the peak area from one prepared sample at 0, 2, 4, 6, 8, 12, and 24 h.The RSDs of CAG and CA were 0.54% and 3.97%, showing that the sample solutions were stable within 24 h at room temperature.The recoveries were calculated using three concentrations (equivalent to 0.5, 1.0, and 1.5 times the original amounts in the samples), and the recovery rate of CAG was 92.25-98.63%and RSD was 0.87-2.30%; the recovery rate of CA was 97.78-105.30% and the RSD was 2.97-3.46%,which indicated that the method had a high recovery rate and good accuracy.
Using the newly established method, ten batches of YQD samples were evaluated.The results (Table S5) show that the content of CAG and CA were 0.333-0.484mg/g and 0.275-0.383mg/g, respectively.The RSD was less than 15%, indicating that the method could conduct the quantitative analysis of the two components.Under the analytical conditions, the response value and resolution of CAG and CA met the quantitative requirements well.

Bioactive Ingredients and Potential Targets of YQD
Network pharmacology emphasizes the multi-level and multi-angle discovery of potential targets and signaling pathways.This includes identifying active ingredients, their targets, disease targets, and screening the core genes and then constructing networks within the compound-targetdisease-signaling pathway and evaluating the effects of compounds on these networks [41,42].The ingredients comprising TCM preparations are numerous; however, the components present at the highest amounts or those that can be readily absorbed by the body may be the active components.Therefore, the compounds found in rat plasma were subjected to further network pharmacology analysis.As the metabolites of daudzein and equol (M1, M3, M4, M7, M8, M14, M17, M19, M21, and M22) are derived from diverse sources [43], they were not specific metabolic components of YQD and were therefore excluded.Additionally, apigenin (C42), a prototype component that produces many metabolites, and four components (C13, C24, C30, and C31) with high response values in the UPLC/MS chromatogram were added to the pharmacological prediction model.The 31 compounds used for network pharmacology analysis are shown in Table 5.In total, 466 targets related to the compounds were found (Table S6), including AKT1, TNF, ALB, mTOR, MAPK1, IL2, PIK3CA, IGF1, and others.

GO and KEGG Enrichment Analysis
According to GO classification enrichment analysis, the vitiligo targets were categorized as being involved in the stress response, activation of cytokines, biological regulation, and cell communication (Figure 8II(A)).The top 20 vitiligo-related pathways included the TNF, FoxO, PI3K-Akt, and mTOR signaling pathways according to KEGG enrichment (Figure 8II(B) and Table S10).
To validate that the ingredients could regulate the targets through the aforementioned pathways to produce the therapeutic effects, we constructed a compound-target-pathway (CTP) network using 41 core target genes (in the top 20 pathways) (Figure 8II(C)).The network had 92 nodes, including 31 components, 20 pathways, and 41 targets.M30, C42, M10, M15, M24, M13, M6, and P3 had higher than average degree values in the network, whereas the targets with higher degree values were AKT1, MAPK1, PIK3CA, and mTOR.The results were consistent with the PPI and CTD network analyses.The PI3K-Akt, TNF, MAPK, and FoxO signaling pathways were enriched for more than 10 targets (Table S11) that have been reported in the literature as potentially critical pathways for vitiligo treatment [52][53][54].

Molecular Docking Study
Molecular docking is an alternative technique for predicting the binding affinity of molecules to the active site of a particular receptor without the use of animals; it enables high-throughput screening of components [55,56].It is generally accepted that the lower the energy when the conformation of the ligand binding to the receptor is stable, the great-er the probability of action.Target prediction in network pharmacology is heavily reliant on the predictive capacity of network prediction tools, and the data are primarily generated from experiments [57].However, because of the restricted number of experimentally confirmed drug targets, certain less-studied TCMs projected fewer targets, causing their constituents to be excluded from the active ingredients.
In this study, the predicted active compounds with high (M30, M24, and C42), medium (M32, M10, M15, M6, and C31), and low (P4, C30, and P2) degree values in network pharmacology were docked with potential core targets (AKT1, MAPK1, and mTOR) to verify the predicted results and explore the possibility of compounds with a low degree value participating in the binding of targets.As shown in Figure 9A, the binding energies of the 11 ligands to the core targets ranged between −4.7 and −11.8 kcal/mol, which implies that all of them had good binding activities to the proteins except for P2/MAPK1 (−4.7 kcal/mol).We found that the strongest binding ability to AKT1 was with M6 and M30, which formed hydrogen bonds with residues ALA-58 (Figure 9B) and TYR-272 (Figure 9C) that had binding energies of −10.2 and −9.0 kcal/mol, respectively.The component with the strongest binding capacity to MAPK1 was M15, and the hydrogen bond binding sites were at ARG-359, VAL-21, GLN-355, ARG-91, and ASP-100 (Figure 9D).M10/MAPK1 ranked second and formed hydrogen bonds with residues ASP-167, GLY-169, and TYR-36 (Figure 9E).The binding energy of the 11 ligands to mTOR ranged between −8.2 and −11.8 kcal/mol, with C31/mTOR having the lowest binding energy and forming hydrogen bonds with five residues (Figure 9F).In addition, P4 is predicted to bind strongly to mTOR, forming hydrogen bonds with GLU-2032, ASP-2102, and ARG-73 (Figure 9G).The results suggest that the components with lower predicted activity determined using network pharmacology may also have a strong binding capacity to core proteins and thus exert therapeutic effects.

Effects of YQD on Melanogenesis In Vitro and In Vivo 2.5.1. Effects of YQD on Melanogenesis in B16F10 Cells
The viability of the murine B16F10 melanoma cell line was examined in the presence of different concentrations of YQD using the CCK-8 assay.The results showed that B16F10 cells treated with YQD at concentrations of 1-1000 µg/mL for 48 h did not display any changes in cell viability when compared with the control (Figure 10A).
To determine the effects of YQD on pigmentation, we examined the melanin content of B16F10 cells treated with different concentrations (1-1000 µg/mL) of YQD and a positive control drug (8-MOP) for 48 h.As shown in Figure 10B, YQD increased the melanin content in B16F10 cells in a dose-dependent manner.
To investigate its effect on the melanogenic pathway in B16F10 cells, we explored whether YQD affected TYR activity and the expression of the melanin biosynthetic gene (TRP-1).As shown in Figure 10C, the activity of cellular TYR increased significantly in a concentration-dependent manner following incubation with YQD.The results of the qRT-PCR experiments showed that the mRNA level of TRP-1 was elevated after YQD treatment (Figure 10D); moreover, Western blotting showed that the protein level was also increased significantly in a dose-dependent manner (Figure 10E).These results indicate that YQD induced pigmentation by enhancing TYR activity and upregulating TYR family genes.

Effects of YQD on Extracellular TYR Activity
An in vitro TYR activation assay showed that YQD could activate mushroom TYR and that the activation rate was concentration dependent (Table S12).The TYR activation rate for 8 mg/mL of YQD reached 79.66%, which was equivalent to the activation rate for 0.5 mg/mL of 8-MOP.

Effects of YQD on Melanogenesis in Zebrafish
After observing the effects of YQD in mammalian cells, we wanted to determine whether a model organism had a similar response.Zebrafish embryos were used to investigate the role of YQD in promoting melanogenesis in vivo.Firstly, the acute toxicity and dose were examined by  S13, and the condition of zebrafish 72 hpf in the presence of YQD is depicted in Figure S3A.The results show that concentrations of YQD below 1000 µg/mL had no influence on zebrafish embryos.However, a considerable number of juvenile fish died in both groups at 2000 µg/mL (the mortality rates were 50% and 100%).All embryos had died 72 hpf when the concentration was increased to 4000 µg/mL.The calculated LC 50 was 1997 µg/mL (Figure S3B).mTOR, forming hydrogen bonds with GLU-2032, ASP-2102, and ARG-73 (Figure 9G).The results suggest that the components with lower predicted activity determined using network pharmacology may also have a strong binding capacity to core proteins and thus exert therapeutic effects.In conclusion, combining the above computational results and considering the metabolic pathways of the metabolites, we hypothesized that C11 (carasinaurone, JQG), C12 (calycosin-7-O-β-d-glucoside, HQ), C16 (methylnissolin-3-O-glucoside, HQ), C21 CWZ), C31 (akebia saponin D, YZZ), C32 (formononetin, HQ), C40 (akebia saponin B, YZZ), and C42 (apigenin, HQ) may be the pharmacodynamic markers of YQD that deserve further in-depth study.Next, we explored the effects of YQD on melanogenesis.Zebrafish embryos were separated into eight groups and constantly treated with blank culture medium, varying doses of YQD (200-1200 µg/mL), or a positive control drug (8-MOP) from 24 hpf to 84 hpf.The density and distribution of the melanin granules were observed and recorded every day.We found that the melanin density in the head and tail gradually increased as the embryos grew.The difference between groups peaked at 84 hpf; melanin production in the head, dorsal spine, and abdomen of zebrafish treated with YQD was more noticeable than in other parts of the body, whereas the survival and development of juvenile fish were not affected in the high-concentration treatment group (Figure 11A).We measured the melanin density in the tails of zebrafish and found that the melanin density in the YQD-treated groups (400-1200 µg/mL) was significantly higher than that in the control group (Figure 11B).
of different concentrations of YQD using the CCK-8 assay.The results showed that B16F10 cells treated with YQD at concentrations of 1-1000 µg/mL for 48 h did not display any changes in cell viability when compared with the control (Figure 10A).
To determine the effects of YQD on pigmentation, we examined the melanin content of B16F10 cells treated with different concentrations (1-1000 µg/mL) of YQD and a positive control drug (8-MOP) for 48 h.As shown in Figure 10B, YQD increased the melanin content in B16F10 cells in a dose-dependent manner.To investigate its effect on the melanogenic pathway in B16F10 cells, we explored whether YQD affected TYR activity and the expression of the melanin biosynthetic gene (TRP-1).As shown in Figure 10C, the activity of cellular TYR increased significantly in a concentration-dependent manner following incubation with YQD.The results of the qRT-PCR experiments showed that the mRNA level of TRP-1 was elevated after YQD treatment (Figure 10D); moreover, Western blo ing showed that the protein level was also increased significantly in a dose-dependent manner (Figure 10E).These results indicate that YQD induced pigmentation by enhancing TYR activity and upregulating TYR family genes.

Effects of YQD on Extracellular TYR Activity
An in vitro TYR activation assay showed that YQD could activate mushroom TYR and that the activation rate was concentration dependent (Table S12).The TYR activation rate for 8 mg/mL of YQD reached 79.66%, which was equivalent to the activation rate for 0.5 mg/mL of 8-MOP.

Effects of YQD on Melanogenesis in Zebrafish
After observing the effects of YQD in mammalian cells, we wanted to determine whether a model organism had a similar response.Zebrafish embryos were used to investigate the role of YQD in promoting melanogenesis in vivo.Firstly, the acute toxicity and dose were examined by observing pericardial edema, delayed yolk absorption, and mortality in zebrafish embryos within 96 h postfertilization (hpf) in the presence of YQD.The statistical results are shown in Table S13, and the condition of zebrafish 72 hpf in the presence of YQD is depicted in Figure S3A.The results show that concentrations of YQD below 1000 µg/mL had no influence on zebrafish embryos.However, a considerable number of juvenile fish died in both groups at 2000 µg/mL (the mortality rates were 50% and 100%).All embryos had died 72 hpf when the concentration was increased to 4000 µg/mL.The calculated LC50 was 1997 µg/mL (Figure S3B).
Next, we explored the effects of YQD on melanogenesis.Zebrafish embryos were separated into eight groups and constantly treated with blank culture medium, varying doses of YQD (200-1200 µg/mL), or a positive control drug (8-MOP) from 24 hpf to 84 hpf.The density and distribution of the melanin granules were observed and recorded every day.We found that the melanin density in the head and tail gradually increased as the embryos grew.The difference between groups peaked at 84 hpf; melanin production in the head, dorsal spine, and abdomen of zebrafish treated with YQD was more noticeable than in other parts of the body, whereas the survival and development of juvenile fish were not affected in the high-concentration treatment group (Figure 11A).We measured the melanin density in the tails of zebrafish and found that the melanin density in the YQD-treated groups (400-1200 µg/mL) was significantly higher than that in the control group (Figure 11B).

Materials and Regents
HQ, YZZ, CWZ, and JQG were collected from different parts of China.The origins and batch numbers are shown in Table S14.Methanol and acetonitrile (both liquid chromatography grade) were purchased from Merck (Darmstadt, Germany) and formic acid was acquired from ANPEL Lab Tech.(Shanghai, China).8-Methoxypsoralen (8-MOP) was purchased from Macklin Biochemical Co., Ltd.(Shanghai, China).The solid-phase extraction (SPE) cartridges (Oasis HLB, 3 mL) were purchased from Waters (Milford, MA, USA).Pure water (18.2MΩ/cm) was deionized using a Milli-Q system (Millipore, Bedford, MA, USA).

Statistical Analysis of the HPLC Fingerprints
Chromatographic data from 15 different batches of each herb were mined and analyzed.Similarity analysis was performed using the "Chinese Materia Medica chromatographic fingerprint similarity evaluation system" (ver.2012) [20].The time window width was set to 0.1 min, and the median method was selected for full-spectrum peak matching.The simulated average fingerprint R was generated, the similarity between the sample and the average fingerprint R was calculated, and then the common peaks that existed in each herb were observed [21].A chemometric analysis including HCA and PCA was used to examine the similarities and differences between the samples [22].SIMCA software (ver.14.1) was used to categorize the samples, and dendrograms were drawn to characterize the classification results.Significant PCs were extracted under the condition that the corresponding eigenvalues were greater than 0.99.A score plot was used to assess sample variation and PC comprehensive scores were calculated to evaluate the sample quality.

Preparation of the HQ, YZZ, CWZ, and JQG Decoction
An amount of 100 g of the HQ, YZZ, CWZ, or JQG pieces was extracted using 10 times the amount of water.After being soaked for 30 min and decocted for 60 min, the mixture was immediately filtered through a 200-mesh screen.The filtrate residue was then extracted with 8 times the amount of water, decocted for 40 min, and immediately filtered again.The twice-combined extraction decoction was concentrated to 0.2 g/mL through reducing pressure [19].

Optimization and Validation of the YQD Preparation Conditions
A three-factor with three levels (−1, 0, +1) BBD was applied to optimize the critical parameters, such as the first solid-liquid ratio, first extraction time, and second extraction time [28].The three responses used to evaluate the influence of the extraction parameters were the calycosin-7-O-βd-glucoside and chlorogenic acid contents of the extract and its total solid content, and the OD value was used to express the overall impact results.Multiple regression analysis was used to develop the best-fit polynomial quadratic model [22].The p-value, lack of fit value, and coefficient of determination (R 2 ) from the ANOVA were used to assess the model adequacy.Experimental design, statistics, response surfaces, and contour plots were obtained using Design-Expert software (ver.12.0).
After optimizing the preparation conditions of YQD, six validation experiments were conducted for the optimal solution, with a first liquid-solid ratio of 8, a second liquid-solid ratio of 7, a first extraction time of 60 min, and a second extraction time of 40 min.The experimental values were compared with the predicted ones to validate the model.

Preparation of the Sample and Standard Solution
HPLC and UPLC-TOF/MS samples: an amount of 1 mL of each decoction sample (0.2 g/mL) was accurately extracted, centrifuged at 13,000 rpm for 10 min, and then the supernatant was filtered through a 0.22 µm membrane before use.SPE-UPLC-TOF/MS samples: an amount of 500 µL of YQD was added to 500 µL of deionized water, mixed thoroughly, and centrifuged at 12,000 rpm for 15 min.Then, 1 mL of the supernatant was subjected to the activated SPE column and eluted successively with H 2 O, 30%, 60%, and 100% methanol (1 mL for each).The flow rate was controlled uniformly (about 10 drops/min).The eluent was collected and dried, redissolved in 100 µL of 50% methanol, vortexed for 3 min, sonicated for 3 min, and centrifuged at 12,000 rpm for 15 min, and the supernatant was then taken for analysis.
Standard solutions: stock solutions of calycosin-7-O-β-d-glucoside and chlorogenic acid were separately prepared in methanol.The stock solutions of each analyte were diluted in 60% methanol to prepare calibration solutions and filtered through 0.22 µm membranes before use.

Animals and Drug Administration
Sprague-Dawley rats (180-200 g, 6-7 weeks old) were provided by Shanghai Silaike Experimental Animal Co., Ltd.(Shanghai, China, approval number: 2021-0009).The rats were kept in a room (24 ± 2 • C with 60% ± 5% humidity) with a 12 h light/dark cycle for one week and had free access to water and conventional laboratory food to acclimate.The rats were randomly assigned to the control group or YQD group (n = 6 each).The rats in the YQD group were orally administered with YQD (dissolved in purified water) once a day at a dose of 15 g/kg (decoction pieces weight/body weight) for 5 consecutive days.The same volume of saline solution was given to the rats in the control group.Blood was collected 1 h after the last dosing.
All experiments were conducted in accordance with the Guiding Principles for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Jiao Tong University (approval number: A2021145, approval date: 23 December 2021).All work was performed following the recommendation of the ARRIVE guidelines.

Plasma Collection and Preparation
The blood samples were collected and centrifuged at 4000 rpm for 10 min at 4 • C. The plasma samples were then transferred to Eppendorf tubes and kept at −80 • C for further analysis.To precipitate protein, 200 µL of plasma was treated with 800 µL of methanol, vortex mixed for 1 min, and centrifuged at 12,000 rpm for 15 min at 4 • C. The supernatant was transferred to a new tube and evaporated to dryness.For further analysis, the residue was redissolved in 80 µL of 50% methanol and centrifuged at 12,000 rpm for 15 min.

Cell Cultures and Viability Assay
Mouse B16F10 melanoma cells (Chinese Academy of Sciences, Shanghai, China) were cultured in RPMI 1640 (Gibco, MA, USA) medium, supplemented with 10% FBS and a penicillin-streptomycin antibiotic mix in 5% CO 2 at 37 • C.
The cell viability was measured using CCK-8 solution (Macklin, Shanghai, China) [24].Briefly, B16F10 cells were seeded in 96-well plates (1 × 10 4 cells/well) for 24 h.The cells were then treated with 8-MOP (100 µmol/L) as a positive control and different concentrations of YQD for 48 h.Then, the culture medium was removed and 10 µL of CCK-8 solution was added to each well.After incubation for 1 h at 37 • C, the absorbance was measured at 450 nm using a Varioskan Flash spectrometer (Thermo Scientific, Waltham, MA, USA).All the assays were performed in triplicate and the cell viability was calculated as (A sample /A control ) × 100%.

Melanin Measurement and TYR Activity Assay
Melanin release from B16F10 cells was measured using a previously described method with slight modifications [24].B16F10 cells were seeded at a density of 1 × 10 4 cells/well in a 6-well plate overnight.Then, the cells were treated with different concentrations of YQD and 8-MOP (100 µmol/L) for 48 h.Following the treatment, the cells were washed with ice-cold PBS and dissolved in 200 µL of 1 mol/L NaOH (with 10% DMSO) at 70 • C for 1 h.The absorbance at 490 nm was measured via a multiplate reader.The protein concentration of each sample was determined using the BCA Protein Assay Kit (Biomed, Beijing, China).The melanin content of the YQD-treated cells was calculated relative to the untreated cells.
The assay for TYR activity was carried out as previously described [26], with a slight modification.B16F10 cells were incubated with test samples for 48 h and lysed with 1% Triton X-100 solution containing 1% sodium deoxycholate for 15 min at 0 • C. Each lysate was centrifuged at 12,000 rpm for 10 min to obtain the supernatant.After protein quantification and adjustment, 100 µL of the cell lysate was mixed with 100 µL of 0.1% L-DOPA (Yuanye Bio-Technology, Shanghai, China) and incubated at 37 • C for 30 min.The absorbance was measured at 475 nm using a multiplate reader, and the TYR activity in YQD-treated cells was presented relative to the activity in untreated cells.

Western Blot Analysis
B16F10 cells were treated with different concentrations of YQD for 48 h.Cells were then lysed in cold RIPA buffer for 30 min on ice.The lysates were centrifuged at 12,000 rpm for 15 min at 4 • C and the supernatants were collected.The proteins were separated by 10% SDS-PAGE and then transferred onto PVDF membranes.The membranes were blocked with 5% skim milk, incubated with appropriate antibodies (TRP-1, ABclonal, Wuhan, China) at 4 • C overnight, and then incubated with the corresponding secondary antibodies for 2 h at room temperature.The targeted proteins were detected using the ECL detection system (GE Healthcare, Beijing, China) and visualized with the ChemiDoc XRS + Imaging System (Bio-Rad, Hercules, CA, USA).

Extracellular TYR Activity
The Tyrosine Assay Kit (Yuanye Bio-Technology, Shanghai, China) was used to measure the extracellular TYR activity.L-DOPA, TYR, PBS, and different concentrations of YQD were added to the tubes according to the manufacturer's instructions.The mixture was incubated for 10 min at 37 • C in a water bath; then, TYR was added and incubated for 2 min at 37 • C. The absorbance was determined at 475 nm and the activation rate was calculated.

In Vivo Melanogenic Effect in Zebrafish
Wild-type zebrafish embryos (AB strain) were collected from Shanghai Jiao Tong University.The embryos were incubated with E3 solution as a control and different concentrations of YQD from 24 h postfertilization (hpf) to 96 hpf at 28 • C. Two replications were set up for all groups (n = 30 in each group).The pericardium edema rate, yolk absorption delay rate, and mortality rate of zebrafish within 96 hpf were observed and recorded to investigate the toxicity of YQD.The LC 50 was calculated using GraphPad Prism (ver.8.0.2).
In the melanin measurement assay [24], the embryos were incubated with 8-MOP (100 µmol/L) as positive control or variable concentrations of YQD from 24 hpf to 84 hpf.Three replications were set up for all groups (n = 10 in each group).A stereomicroscope (Olympus Instruments, Tokyo, Japan) was used to observe and record the melanin in the heads and tails of the zebrafish daily.The melanin density in the tail was measured using Image-Pro Plus (ver.6.0).

Statistical Analysis
All data were reported as the mean and standard deviation.The results were analyzed and illustrated using GraphPad Prism (ver.8.0.2) software.The level of statistical significance is described in the respective figure legends (* p < 0.05, ** p < 0.01, and *** p < 0.001, as indicated by a oneway ANOVA).

Conclusions
In this study, a combination of fingerprints and chemometrics was used to select the optimal herbs for the preparation of YQD.The preparation process for YQD was optimized, and the chemical constituents and the absorbed constituents were identified.The SPE-UPLC-Q-TOF/MS identification method and HPLC quantification method were established to improve the quality standard of YQD.It was predicted that several signature components and metabolites from the four herbs might be important active components, reflecting the rationality of the YQ formula.In addition, B16F10 cell and zebrafish assays confirmed the role of YQD in promoting melanogenesis.The possible mechanism is that YQD can increase the expression and transcription of TRP-1 and enhance the activity of TYR, which is meaningful for the treatment of vitiligo.The results of this study fill many gaps in the research of YQ formulations and lay a good foundation for the clinical application and further development of YQD.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/ph16121716/s1.Tables S1: Similarity analysis of the fingerprint chromatograms; Table S2: Eigenvalue and variance contribution rate; Table S3: ANOVA table for the OD (Box-Behnken design); Table S4: Experimental validation results; Table S5: The content of CAG and CA in 10 batches of YQD; Table S6: Compounds and target used for network pharmacology analysis; Table S7: Overlapping of drug-related targets and disease-related targets; Table S8: Data of PPI network diagrams; Table S9: The degree value in the CTD network; Table S10: Data of KEGG analysis; Table S11: The degree value in the CTP network; Table S12: Activation rate of YQD to mushroom tyrosinase.Table S13: The toxicity of YQD to zebrafish embryos; Table S14: Origin and batch information of HQ, YZZ, CWZ, and JQG.

Figure 1 .
Figure 1.Schematic diagram of the study process.

Pharmaceuticals 2023 , 36 Figure 3 .
Figure 3.The interaction surface (A) and the contour diagram (B) of the first liquid-solid ratio (X1) and the first extraction time (X2); the interaction surface (C) and the contour diagram (D) of the first liquid-solid ratio (X1) and the second extraction time (X3).

Figure 3 .
Figure 3.The interaction surface (A) and the contour diagram (B) of the first liquid-solid ratio (X 1 ) and the first extraction time (X 2 ); the interaction surface (C) and the contour diagram (D) of the first liquid-solid ratio (X 1 ) and the second extraction time (X 3 ).The optimal extraction conditions for YQD were calculated by solving the model's extreme values and analyzing the contours.The first liquid-solid ratio was 8.086, the first extraction time was 59.998 min, and the second extraction time was 39.998 min.The optimal extraction process resulted in an OD value of 1.06.Due to actual manufacturing considerations, the optimum conditions of YQD were as follows: a liquid-solid ratio of 8, the first extraction time was 60 min, and the second extraction time was 40 min.

36 Figure 4 .
Figure 4. (A) Base peak chromatogram (BPC) of the YQD sample in negative ion mode; (B) base peak chromatogram (BPC) of the YQD sample in positive ion mode; and (C) extracted ion chromatograms (EIC) of absorbed components and metabolites after oral administration of YQD at 1 h.

Figure 4 .
Figure 4. (A) Base peak chromatogram (BPC) of the YQD sample in negative ion mode; (B) base peak chromatogram (BPC) of the YQD sample in positive ion mode; and (C) extracted ion chromatograms (EIC) of absorbed components and metabolites after oral administration of YQD at 1 h.

Figure 8 .
Figure 8. I: (A) Venn diagram of targets related to YQD and vitiligo.(B) Visualization analysis of the PPI diagram-the size of the circle represents the target degree.(C) Adjacent nodes of the overlapping targets between YQD and vitiligo-the X-axis represents the number of adjacent proteins of the target and the Y-axis represents different targets.(D) Compound-target-disease (CTD) network-31 components (blue) were connected to 78 targets (green) and YQD (red), and 78 target genes were connected to 31 components and vitiligo (red).II: (A) analysis.(B) Top 20 enriched KEGG pathways.(C) Compound-target-pathway (CTP) network-31 components (blue) were connected to 41 targets (orange), and the 41 targets were connected to 20 pathways (yellow) and 31 components.

Figure 8 .
Figure 8. I: (A) Venn diagram of targets related to YQD and vitiligo.(B) Visualization analysis of the PPI diagram-the size of the circle represents the target degree.(C) Adjacent nodes of the overlapping targets between YQD and vitiligo-the X-axis represents the number of adjacent proteins of the target and the Y-axis represents different targets.(D) Compound-target-disease (CTD) network-31 components (blue) were connected to 78 targets (green) and YQD (red), and 78 target genes were connected to 31 components and vitiligo (red).II: (A) GO analysis.(B) Top 20 enriched KEGG pathways.(C) Compound-target-pathway (CTP) network-31 components (blue) were connected to 41 targets (orange), and the 41 targets were connected to 20 pathways (yellow) and 31 components.
edema, delayed yolk absorption, and mortality in zebrafish embryos within 96 h postfertilization (hpf) in the presence of YQD.The statistical results are shown in Table Pharmaceuticals 2023, 16, x FOR PEER REVIEW 25 of 36
Figure S1: Loading plot of HQ (A); YZZ (B); CWZ (C); and JQG (D); Figure S2: The interaction surface (A) and the contour diagram (B) of the first extraction time (X 2 ) and the second extraction time (X 3 ); Figure S3: (A) The condition of zebrafish 72 hpf in the presence of YQD; (B) The calculated LC 50 of YQD in zebrafish.

Table 1 .
Rank of principal component scores and comprehensive scores for HQ, YZZ, CWZ, and JQG.

Table 2 .
Three-factor Box-Behnken design (BBD) with factorial levels shown in actual values for the YQD.
The neutral loss of glucose (C 6 H 10 O 5 , −162 Da) in the MS 2 spectrum formed a fragment ion m/z 285.0750, which further lost a methyl (•CH 3 , −15 Da), and the methoxy group (CO, −28 Da) formed the fragment ions m/z 270.0517 and m/z 242.0541.Fragment ion m/z 285.0750 may also lose methanol (CH 3 OH, −32 Da) and further lose the methoxy group (CO, −28 Da) to obtain the fragment ion m/z 225.0522.Fragment m/z 137.0217 may be obtained via the RDA reaction of the compound or the deglycosylated glycoside element.C12 was identified as calycosin-7-O-β-d-glucoside by comparing its MS 2 cleavage behavior and retention time with those of the reference compounds., with a molecular formula of C 53 H 86 O 23 and fragment ions at m/z 959.5186, 813.4651, 651.4084, and 489.3562, which were generated by the quasi-molecular ion peaks losing one molecule of xylose, one molecule of deoxymannose, and one molecule of glucose in turn; the fragment ions at m/z 489.3562 was arjunolic acid sapogenin.Using data from reference 22O 10 , 30.86 min) showed a quasi-molecular ion at m/z 447.1269.C24 (eluted at 42.79 min) showed a quasi-molecular ion at m/z 1091.5812[M + H] +

Table 4 .
Identification of absorbed prototype components and metabolites in rat plasma by UPLC-Q-TOF/MS (1 h after the last administration).

Table 5 .
Information table on active components of YQD.