The components of P. corylifolia that enter the bloodstream were obtained as potential active ingredients using the SymMap database (http://www.symmap.org/). The non-hemolytic constituents of psoralen were meticulously investigated within the TCMSP database (https://old.tcmsp-e.com/tcmsp.php) using the criteria of druglikeness ≥0.18 and oral bioavailability ≥30% as the selection parameters. The active ingredients obtained were searched for in the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) to determine their chemical formulae. The Swiss Target Prediction tool (http://swisstargetprediction.ch/) was then used to determine the target of the active ingredient.

The GSE75819 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=gse75819) transcriptome data (13), downloaded from the GEO database (http://www.ncbi.nlm.nih.gov/geo/), contains data from 15 vitiligo patients and 15 healthy individuals. Differential analysis was conducted using the ‘limma’ package (https://bioconductor.org/packages/release/bioc/html/limma.html) in the R programming language. Differential genes were identified using a significance level of P<0.05 and |log2FC|>1 as the criteria. The findings were then depicted via volcano plots and heat maps to provide a comprehensive visualization. The differential genes and vitiligo genes from the GeneCards (https://www.genecards.org/) and DisGeNET databases (https://disgenet.com/) were brought together as vitiligo target genes using a Venn diagram (http://www.bioinformatics.com.cn). With the help of the online tool Venny 2.1.0 (http://liuxiaoyuyuan.cn/gongju/weientu.html), the vitiligo targets were intersected with the targets of the P. corylifolia active ingredient and the intersecting targets were identified as potential vitiligo interventions.

Network and tape files consisting of P. corylifolia and its common targets with vitiligo were produced. These files were then be imported into Cytoscape 3.10.1 (https://apps.cytoscape.org/) application to construct the active ingredient-target-disease network.

The possible targets of P. corylifolia for vitiligo were searched for in the STRING database (https://cn.string-db.org). Free nodes were eliminated and the resulting TSV files were downloaded, visualized and analyzed using Cytoscape 3.10.1.

GO and KEGG pathway data obtained from the intersecting genes of P. corylifolia and vitiligo were retrieved and downloaded with the help of the DAVID database (https://david.ncifcrf.gov). GO and KEGG enrichment analyses were conducted to investigate the biological pathways and potential functions of the intersecting genes, with the aim of analyzing their biological processes and metabolic pathways. Finally, the data were visualized using the MicroBioInformatics online tool (https://www.bioinformatics.com.cn).

The 3D crystalline structure of the target protein was retrieved from the Protein Data Bank (PDB; http://www.rcsb.org/). It was then downloaded and processed for dehydration and hydrogenation using AutoDockTools (https://vina.scripps.edu/). This structure was ultimately selected as the receptor and preserved in the PDBqt file format. Active constituents were retrieved from the TCMSP database, hydrogenated using AutoDockTools and selected as ligands before being exported in the PDBQT file format. Molecular docking simulations were then performed using AutoDock Vina to determine the binding affinity of each compound to the target protein. Subsequent analysis of these affinities was conducted using the ‘pheatmap’ (https://cran.r-project.org/web/packages/pheatmap/index.html) and ‘circlize’ R packages (https://cran.r-project.org/web/packages/circlize/index.html), complemented by visualization in PyMOL (http://www.pymol.org/pymol).

P. corylifolia (10 g) was crushed and added to 100 ml of distilled water/75% ethanol. The mixture was left to soak for 1 h and then subjected to water bath heating extraction for a further 1 h. The extract was filtered and collected. Add a further 100 ml of distilled water/75% ethanol and perform a water bath heating extraction for a further 30 min. Filter and collect the extract, then combine the two extracts and concentrate them.

AB wild-type zebrafish (Shanghai FishBio Co., Ltd.), B16F10 cells (Wuhan Pricella Biotechnology Co., Ltd.; Elabscience Bionovation Inc.), 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; MTT; tetrazolium salt), N-phenylthiourea (PTU), levodopa (L-DOPA), laboratory biochemical incubator (Tianjin Laboratory Instrumentation Co., Ltd.), K-400L stereo microscope (Motic Incorporation, Ltd.), SpectraMax iD3 multifunctional enzyme labeler (Molecular Devices, LLC.).

The aqueous and ethanol extracts of P. corylifolia were subjected to UPLC-Q-TOF/MS analysis to ascertain the total ion mobility map, in compliance with the chromatographic and mass spectrometric parameters as detailed in Tables SI and SII. Drawing upon the multistage mass spectrometric data derived from the samples, subsequent processing and comprehensive analysis were conducted in synergy with the high-resolution mass spectrometry database for natural products and pertinent scholarly literature.

Preliminary experiments determined that 200 µM PTU would be used to generate a pigment-deficient zebrafish model. Zebrafish embryos were cultured in a constant-temperature incubator at 28°C for 24 h, designated as 24hpf/1dpf. PTU-induced modeling was then performed for 48 h, at which point the embryos were designated as 72hpf/3dpf. Drug administration commenced immediately afterward for 72 h, with the final time point designated as 144hpf/6dpf. Experimental groups and procedures are detailed below: At 1 dpf, healthy zebrafish embryos were selected and placed into 6-well plates at 15 zebrafish embryos per well. Groups included: Blank control group cultured in fresh culture water; modelling group cultured in fresh culture water containing 200 µM PTU; and groups with different concentrations of aqueous and ethanol extracts (1, 2.5 and 5 µg/ml). Following drug administration, anesthesia was induced using 0.016% MS-222 until zebrafish ceased swimming or exhibited ventral positioning (flipping onto their backs), indicating full anesthesia (14). Anesthetized zebrafish were observed under a microscope within 5 min to assess pigment development, with images captured for documentation. After the experiment, zebrafish were immersed in a high-concentration 0.08% MS-222 solution (typically 5X) until gill movement ceased (~5 min). Anesthesia was maintained for an additional 10–15 min to ensure the mortality of experimental fish (15). Finally, they were placed in a −20°C freezer for centralized disposal of all experimental animals (16). All zebrafish experiments were reviewed and approved by the Animal Ethics Committee of Bozhou University, with ethics approval number DFDW/BZUU-2025-ZF-031.

B16F10 melanocytes in the exponential growth phase were carefully selected and inoculated into 96-well plates at a concentration of 5×10⁴ cells/ml in 180 µl of volume per well. Following inoculation, the cells were placed in a 5% CO₂ incubator at 37°C and allowed to proliferate for 24 h. Subsequently, the cells were treated with aqueous and ethanol extracts of P. corylifolia at concentrations of 0.01, 0.1, 1, 10 and 100 µg/ml, followed by an incubation at 37°C period of 48 h. The solution was then evacuated and discarded. Then, 30 µl of MTT reagent was added to each well and the reaction was carried out at 37°C for 4 h. DMSO (150 µl) was added to each well and the solution was vigorously agitated for 10 min. After a 48-h incubation period, 30 µl of MTT reagent was added to each well. The reaction was then conducted at 37°C for 4 h. Then, 150 µl of DMSO was added to each well and shaken for 10 min to facilitate the reaction. Absorbance readings for each well were taken at a wavelength of 570 nm to evaluate the effect of aqueous and alcoholic extracts derived from tonic acid on melanocyte viability and functional integrity.

B16F10 melanocytes were cultured according to the procedure in the aforementioned Measurement of melanocyte viability. After 24 h, 0.01, 0.1, 1 and 10 µg/ml concentrations of the aqueous and ethanol extracts of P. corylifolia were added. After 72 h of incubation at 37°C, the cells were washed twice with PBS, following which 100 µl of a non-denaturing lysis buffer containing 1 nM PMSF were added. The cells were lysed at 4°C for a duration of 20 min, following which they were collected. Upon reaching 72 h of culture at 37°C, the cells underwent a washing process with PBS, repeated twice. Thereafter, a volume of 100 µl of non-denaturing lysate, which contained 1 nM PMSF, was introduced. The cells were subsequently lysed once again at 4°C for 20 min and then collected. The collected cells were centrifuged at 4°C for 10 min at 21,756 × g. The resulting supernatant was used to determine the protein content via the BCA assay and the protein concentration was subsequently calculated. The volume of the supernatant containing 10 µg of total protein was measured and transferred to a 96-well plate. The volume was then brought up to 100 µl using PBS (0.1 M, pH 6.8). Then, 100 µl of a 0.01% L-DOPA solution was added and the plate was incubated at 37°C in the dark for 3 h. OD475 was used to evaluate the catalytic activity of tyrosinase. The melanin precipitate pellet was then suspended in a 100 µl NaOH solution containing 10% DMSO to ensure full solubilization of the melanin within each experimental group. The melanin precipitate with the lower concentration was supplemented with 100 µl of a NaOH solution containing 10% DMSO. Then, 100 µl of the melanin solution was pipetted into each well of a 96-well plate. OD405 was determined to assess melanocyte content.

Data were analyzed exclusively using GraphPad Prism 5 software (Dotmatics), employing one-way analysis of variance (followed by Tukey's post hoc test) and two-tailed paired t-tests. P<0.05 was considered to indicate a statistically significant difference.

Searching the SymMap and TCMSP databases, the present study obtained nine constituents of P. corylifolia, including isobavachin, bavachin, stigmasterol, bavachalcone, bakuchiol, psoralidin and angelicin (Table I).

A total of 259 effective targets of the nine active components of P. corylifolia were obtained through the Swiss Target Prediction database. The GeneCards and DisGeNET databases screened 1,357 and 395 potential vitiligo targets, respectively. Microarray data from GSE75819 detected a total of 25,402 genes, including 691 upregulated and 220 downregulated genes (Fig. 1A and B). Using Venny, the intersection between the microarray differential genes related to vitiligo and the active ingredient targets of P. corylifolia, were identified obtaining a total of 70 intersecting genes (Fig. 1C). These included TNF, AKT1, PPARG, PTGS2, HSP90AA1, ESR1, BCL2, NFKB1, MTOR, GSK3B, PARP1, MDM2, MAPK1 and PIK3CA.

The Cytoscape 3.10.1 application was used to create the ‘active ingredient-target-disease’ network illustrated in Fig. 2A, comprising 80 nodes and 140 edges. The Network Analyzer, an integral component of the software, was used to calculate the degree value for each node. A higher degree value indicates a stronger association between the active components of P. corylifolia and vitiligo-related targets. The active ingredients of P. corylifolia based on degree value were isobavachin, bavachin, stigmasterol, bavachalcone, bakuchiol, isobavachalcone, psoralidin and angelicin, with degree values of 29, 26, 16, 16, 16, 13, 10 and 8, respectively. It suggested that these constituents may be crucial in the psoralen treatment of vitiligo.

The curated list of 70 intersecting genes was seamlessly integrated into the STRING database, giving rise to a sophisticated PPI network, specifically curated under the criterion of ‘Homo sapiens’. This intricate network encompasses a total of 70 nodes interconnected by 461 edges, exhibiting an average node degree of 13.2, as illustrated in Fig. 2B. The derived protein interaction data were meticulously integrated into the Cytoscape 3.10.1 software suite, wherein subsequent manipulation via the Analyze Network module facilitated the generation of an exhaustive network visualization. This visualization underscored the pivotal targets of P. corylifolia in the context of vitiligo therapeutic intervention, as depicted in Fig. 2B. In the figure the size of the circular node indicates the degree value, the larger node degree value, the redder node indicates that the target is more critical, the closer the relationship between the active ingredient of P. corylifolia and the target of vitiligo. The top 15 target proteins are TNF, AKT1, PPARG, PTGS2, HSP90AA1, ESR1, BCL2, NFKB1, MTOR, GSK3B, PARP1, MDM2, MAPK1, PIK3CA and MAPK14, with degree values of 86, 84, 80, 72, 68, 66, 66, 64, 54, 54, 50, 48, 46, 44, 44, respectively. The top 15 targets were further calculated by applying the MCC algorithm using the cytoHubba plug-in in Cytoscape software (Fig. 2C) and the intersection of the targets calculated by the MCC algorithm and Degree was taken to obtain a total of AKT1, HSP90AA1, BCL2, MTOR, NFKB1, MAPK1, ESR1, GSK3B, TNF, PTGS2 ten key targets. The identified targets may represent the principal therapeutic targets of P. corylifolia in the context of vitiligo treatment and they are concurrently the focus of molecular docking validation as presented in the present study.

The 70 overlapping genes were uploaded to the DAVID database for enrichment analysis, yielding results for 272 biological processes (BP), 41 cellular components (CC) and 66 molecular functions (MF). Following rigorous screening, the top ten candidates in each category were selected based on their P-values: BP, CC and MF. These selections were then depicted graphically via bar charts to facilitate a comprehensive visual comparison. The results showed that BP was mainly enriched in the regulation of apoptosis, protein phosphorylation and the inflammatory response; CC was mainly enriched in the mitochondria, the cytoplasm and the phosphatidylinositol 3-kinase complex; and MF was mainly enriched in homodimeric protein binding, enzyme binding and protein serine/threonine kinase activity, as shown in Fig. 3.

The 70 intersecting genes obtained were imported into the DAVID database for enrichment analysis. The results of the enrichment analysis revealed 140 KEGG pathways and a KEGG functional enrichment bar graph was drawn according to the results of the top 20 P-values (Fig. 3B). The KEGG pathway enrichment results were categorized (Fig. 3C). The enrichment outcomes indicate that the genes are predominantly associated with multiple facets of the PI3K-Akt signaling pathway and oncogenic pathways, as well as pathways relevant to hepatitis.

The eight active ingredients of P. corylifolia obtained (Isobavachin, Bavachin, Stigmasterol, Bavachalcone, Bakuchiol, Isobavachalcone, Psoralidin and Angelicin) and the 10 key targets (AKT1, HSP90AA1, BCL2, MTOR, NFKB1, MAPK1, ESR1, GSK3B, TNF and PTGS2) were molecularly docked (Table II and Fig. 4A). The binding energy, measured in kilocalories per mole (kcal/mol), was found to be <-5.0 kcal/mol, indicating a strong and stable interaction between the drug components and the key target proteins. It should be noted that a more negative binding energy denotes a stronger and more stable binding effect. The binding energy was found to be <-5.0 kcal/mol for all components, signifying robust and stable interactions with the key target proteins. The resultant binding energies for each constituent with respect to the target were visualized using a clustered heatmap generated via the ‘pheatmap’ R package (Fig. 4B).

UPLC-Q-TOF/MS analysis was employed to detect and comprehensively analyze the main components of the aqueous and ethanol extracts of P. corylifolia. UPLC-Q-TOF/MS technology was employed to detect and analyze the extracts and the results indicated that the main constituents in the aqueous and ethanolic extracts of P. corylifolia were angelicin, psoralidin, stigmasterol, bavachin and bakuchiol, among others (Fig. S1; Table III).

Compared with the control group, the aqueous and ethanol extracts of P. corylifolia markedly affected pigment synthesis in depigmented zebrafish, with increased melanin synthesis as the administered drug concentration increased (Fig. 5).

Compared with the control group, the results revealed that 0.01–1 µg/ml of the aqueous and ethanol extracts of P. corylifolia had no effect on cell viability, 10 µg/ml promoted cell viability and 100 µg/ml markedly inhibited cell viability (Fig. 6A and D). Therefore, 0.01–1 µg/ml of the aqueous and ethanol extracts of P. corylifolia were selected for subsequent experiments on melanin synthesis and tyrosinase activity. Compared with the control group, 0.01–10 µg/ml of the aqueous extracts of P. corylifolia markedly promoted melanin synthesis, with the effect becoming more pronounced at concentrations of 0.01–1 µg/ml (Fig. 6B). Compared with the control group, 0.01 µg/ml of the ethanol extract of P. corylifolia had no significant effect on melanocyte synthesis, 0.1–10 µg/ml markedly promoted melanocyte melanin synthesis (Fig. 6E). Compared with the control group, 0.1–10 µg/ml of the aqueous extract markedly increased tyrosinase activity, with a more pronounced effect at higher concentrations (Fig. 6C and F).