The present study used The Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP; http://tcmsp-e.com/) for data analysis (14). The bioactive compounds isolated from F. arctii were identified via ‘Fructus Arctii’ as a keyword and the drug-likeness screening criterion had to have a minimum value of 0.18, while oral bioavailability screening had to have a minimum value of 30%. The number of targets for the active ingredients were recorded. The UniProt ID was acquired from the UniProt database for the target (https://www.uniprot.org/) (15). Using ‘keloid’ as a keyword, targets related to keloid is searched and obtained from the Therapeutic Target Database (TTD; http://db.idrblab.org/ttd/) (16), DrugBank (https://www.drugbank.ca/) (17), GeneCards (https://www.genecards.org/) (18) and Online Mendelian Inheritance in Man (OMIM; http://www.omim.org/) (19). Duplicates were removed from the merged data sets originating from each database. Afterward, a Venn diagram was created to illustrate the overlap between the targets of F. arctii and keloids (Fig. 1).

Following the data entry requirements of Cytoscape software (version 3.7.0; http://cytoscape.org), the data from the previous steps were organized and imported into Cytoscape to generate a network depicting the relationships among active ingredients, targets and pathways (20). Following this, the outcomes of the analysis pertaining to F. arctii and common targets were saved from the network. Following this, visual representations of the connections between drug-component-targets were achieved by modifying the shape and color of each node.

The common targets of F. arctii and keloid were uploaded to the STRING database (21). Homo sapiens was designated as the species, a high confidence level of 0.700 was established as the minimum interaction threshold and all other parameters were maintained at their default values. This allowed the construction of a PPI core network of keloids, with F. arctii acting upon it (22). The Network Analyzer and cytoHubba plug-ins in Cytoscape 3.7.0 software were used to calculate the topological properties of the PPI network and the top 10 Hub genes were selected.

The aforementioned predicted primary targets were entered into the Metascape database (http://metascape.org), with the species defined as Homo sapiens. Following database retrieval and transformation operations, enrichment analyses via Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) were conducted. The significant GO functions and KEGG signaling pathways were screened on the basis of the P-value. This allowed for the prediction of the mechanism of action of F. arctii in the treatment of keloids.

The structures of SIT and PTEN (PDB ID:7JVX) were obtained via the PubChem database (https://pubchem.ncbi.nlm.nih.gov) (23) and RCSB Protein Data Bank (RCSB PDB; http://www.rcsb.org/structure) (24) for molecular docking studies. They were then reconstructed using PyMOL software (25). The AutoDock Tools (version 1.5.6; http://autodock.scripps.edu) were used for protein-ligand docking analysis. The 3D interactions were generated using PyMOL software. The results of successful molecular docking were visualized with PyMOL 2.5.4 software. The binding affinity energy (kcal/mol) was computed via CB-Dock2 (https://cadd.labshare.cn/cb-dock2/php/index.php) to determine the optimal docking model with the lowest energy. Molecular visualization plays a crucial role in analyzing and communicating modeling studies. Therefore, the details of the ligand-receptor interactions were revealed via PyMol2 software (Free version; http://pymol.org) and BIOVIA Discovery Studio Visualizer software (version 2021; Dassault Systèmes) (26).

The KEL FIB (CRL-1762) human keloid fibroblast (KF) cell line was obtained from ATCC. The cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin/streptomycin (Beyotime Institute of Biotechnology) at 37°C in a humidified atmosphere with 5% CO₂. To investigate the effects of SIT, KF cells (KFs) were treated with various concentrations (15, 30, 60 and 120 µM) of SIT (MedChemExpress) prior to functional analyses. KFs were treated with YS-49 (10 µM) (cat. no. HY-15477; MedChemExpress), a PI3K/AKT pathway activator, for 24 h to assess cell replication (27). SIT and YS-49 were dissolved in dimethyl sulfoxide (DMSO) and the cells in the corresponding control groups were incubated with an equal volume of DMSO.

To establish an initial concentration range for SIT, 5×10³ cells per well were seeded in a 96-well plate. The cells were subsequently exposed to a gradient of SIT concentrations (15, 30, 60 and 120 µM) for 12, 24 and 48 h. Cell viability was subsequently assessed via a CCK-8 assay. After the addition of the CCK-8 reagent, the cells were incubated for 2 h, after which the absorbance (OD value) at 450 nm was measured via a microplate reader (Molecular Devices, LLC). The obtained absorbance values were normalized to those of the control group.

A Cell-Light EdU Apollo567 In Vitro Kit (C10310-1, RiboBio, Guangzhou, Guangdong, China) was used for EdU staining. Cell proliferation imaging experiments were conducted using 96-well cell culture plates. The manufacturer's guidelines were followed to detect EdU-labeled positive cells. Cells were incubated with Hoechst 33342 dye (10 µg/ml) for 30 min at room temperature in the dark. After staining, they were washed twice with PBS. EdU assays were used to assess the activity of DNA replication as an indicator of changes in the proliferative capacity of KFs.

KFs were seeded in a six-well plate at a density of 1×10⁵ cells per well and grown to 90% confluence. Subsequently, the cells were scraped with a 1-ml sterile pipette and the resulting cell pellets were rinsed with PBS three times. After incubation in serum-free medium, the cells were treated with different concentrations of SIT. The samples were observed under a microscope at 0 and 24 h after drug treatment. Cell migration was assessed by quantifying the percentage of the wound closure area in three independent experiments via ImageJ version 1.53t software (National Institutes of Health).

In 24-well Transwell chambers, cell invasion and migration were evaluated. To analyze invasion, the plate was precoated with Matrigel (Beijing Solarbio Science & Technology Co., Ltd.) for 4 h at 37°C before cell seeding; for migration, chambers devoid of Matrigel were used. A total of 1×10⁴ cells were seeded into the upper compartment of each Transwell chamber, which contained 100 µl of DMEM (Corning, Inc.). The lower chamber was filled with 500 µl of complete medium containing 10% FBS as a chemoattractant. The cells were fixed and incubated for 20 min with 0.1% crystal violet after 24 h of SIT treatment. A cotton swab was subsequently used to delicately remove the cells that were still present on the upper surface of the membrane insert. The quantity of cells that had successfully migrated to the lower surface was then determined.

Western blot analysis was used to determine protein expression in KFs in accordance with previously described procedures (28). Protein lysate was extracted using RIPA buffer (Beijing Solarbio Science & Technology Co., Ltd.) and protein concentration was determined using the BCA Protein Assay Kit (Beyotime Institute of Biotechnology). Proteins (30 µg/lane) were separated by SDS-PAGE on 8–10% gels, transferred onto PVDF membranes, blocked with 5% non-fat dry milk (MilliporeSigma) at room temperature for 1 h and were then incubated with primary antibodies overnight at 4°C. The primary antibodies used were as follows: anti-PI3K (4249T; 1:1,000; CST Biological Reagents Co., Ltd.), anti-phosphorylated (p-)PI3K (17366; 1:1,000; CST Biological Reagents Co., Ltd.), anti-PTEN (AF5447; 1:1,000; Affinity), anti-AKT (4691; 1:1,000; CST Biological Reagents Co., Ltd.), anti-p-AKT (4060; 1:1,000; CST Biological Reagents Co., Ltd.), anti-E-cadherin (3195; 1:1,000; CST Biological Reagents Co., Ltd.), anti-Vimentin (5741; 1:1,000; CST Biological Reagents Co., Ltd.), anti-Snail (3879; 1:1,000; CST Biological Reagents Co., Ltd.), anti-zonula occludens-1 (ZO-1; 8193; 1:1,000; CST Biological Reagents Co., Ltd.) and anti-β-actin (4967; 1:1,000; CST Biological Reagents Co., Ltd.). Subsequently, the blots were probed with HRP-conjugated anti-rabbit and anti-mouse secondary antibodies (1:5,000; cat. nos. RGAR001 and RGAM001; Proteintech Group, Inc.), which were diluted in TBS-0.1% Tween 20 solution, at room temperature for 2 h. The blots were observed with an enhanced chemiluminescence Western blot detection reagent (MilliporeSigma). ImageJ version 1.53t software (National Institutes of Health) was used to conduct a statistical analysis of the grayscale values of the proteins. After normalization to β-actin, the internal control, the relative expression of the target protein in each group of cells was computed.

A total of 10 male SPF-grade BALB/c nude mice, aged six weeks and weighing 20±2 g, were procured from the Experimental Animal Center of Yanbian University (Yanji, China) and were randomly assigned to two groups (n=5 mice/group). Nude mice were housed in a specific pathogen-free environment with a temperature of 25°C and humidity of 30%. The mice were exposed to a 12-h light/dark cycle, and were given free access to adult mouse feed sterilized with cobalt-60 and water sterilized by autoclaving. Following a 1-week acclimatization period with standard feeding, a nude mouse keloid model was established. This was achieved by injecting a keloid fibroblast suspension with a density of 5.0×10⁷ cells/ml diluted in Matrigel (Corning, Inc.) under aseptic conditions. The mixture was then kept on ice. Subsequently, 0.1 ml of the prepared keloid fibroblast suspension was subcutaneously injected into the axilla of the right forelimb of each nude mouse. The keloid volume, weight and growth of the mice were monitored every three days. The average tumor volume was calculated via the following formula: tumor volume=(length × width²)/2. After 7 days, keloid formation became visible under the skin in the right axilla of the nude mice. Upon reaching a size of approximately 50 mm³, the SIT-treated group of nude mice received intraperitoneal injections of SIT at a safe therapeutic dose of 50 mg/kg, as determined in a previous study (29). The treatments were repeated every two days for a total of 10 cycles. The control group received equivalent saline injections. Throughout the course of the experiment, the mice were closely monitored for any signs of poor health, including weight loss, abnormal posture, changes in activity levels, labored breathing and signs of distress. No signs of severe distress were observed in the animals prior to sacrifice. At the end of the experiment, all mice were sacrificed by cervical dislocation following anesthesia with an injection of sodium pentobarbital at a dose of 70 mg/kg body weight and death was confirmed by the absence of respiration and heartbeat for more than 5 min. The fibroproliferative tissues were collected for measurement and further experiments.

The subcutaneous tumors were fixed in 10% formalin at room temperature for 24 h, dehydrated in graded ethanol and embedded in paraffin. The paraffin-embedded tumors were then sectioned into 4-µm slices, which were oven-baked at 56°C overnight and stained using the H&E Stain kit (Beijing Solarbio Science & Technology Co., Ltd.), according to the manufacturer's protocol. For IHC, the slides were placed in 10 mM sodium citrate buffer (pH 6.0) and boiled, then simmered for 10 min, for antigen retrieval, before being cooled for 30 min. To remove endogenous peroxidase activity, slides were incubated with 3% hydrogen peroxide aqueous solution for 10 min at room temperature (20–25°C). Furthermore, blocking was performed using TBS-0.1% Tween with 5% normal goat serum for 1 h at room temperature (Cell Signaling Technology, Inc.). The sections were then incubated overnight at 4°C with primary antibodies against PTEN (cat. no. AF5447; 1:200 dilution; Affinity Biosciences). The next day, a biotinylated Goat Anti-Rabbit IgG H&L (Biotin) secondary antibody (1:100; cat. no. ab207995; Abcam) was added, and the slides were incubated at room temperature for 30 min. Streptavidin-horseradish peroxidase (cat. no. SA10001; Invitrogen; Thermo Fisher Scientific, Inc.) was then used to incubate the slides at room temperature for 30 min and 200 µl DAB was added to each section as the chromogen. Hematoxylin was used to counterstain the sections at room temperature for 1 min. Finally, images of the slides were captured using a Nikon Eclipse Ti-S/L100 inverted phase contrast fluorescence microscope with a 20× objective.

Blood biochemical analysis was conducted to evaluate biosafety parameters, including creatinine (CREA), alanine aminotransferase (ALT), serum aspartate aminotransferase (AST) and blood urea nitrogen (BUN) (27). The serum was separated by centrifugation at 4,000 × g for 10 min at 4°C. ALT, AST, BUN and CREA levels were determined following the instructions provided with the kit (Nanjing Jiancheng Bioengineering Institute).

Data analysis was performed via GraphPad Prism 7.0 software (Dotmatics). Group variables were characterized by the number of cases, scores, measurements and percentages. Data that followed a normal distribution are presented as the means ± standard deviations. Multiple group comparisons were conducted via one-way ANOVA with either Tukey's or Dunnett's post hoc tests to determine significant differences between groups. P<0.05 was considered to indicate a statistically significant difference. Each experiment was performed in triplicate.

The present study used the TCMSP screening criteria to identify eight chemical components (Table I). Among these components were neoarctin A, arctiin, SIT, kaempferol, aupraene, β-carotene, (3R,4R)-3,4-bis[(3,4-dimethoxyphenyl)methyl]oxolan-2-one and cynarine. By setting the species to Homo sapiens, the genes of each target pair were cross-referenced with the UniProt protein database to obtain the standardized symbols. A total of 99 target genes associated with the active ingredients and 740 targets related to keloid formation were identified (Table SI). A Venn diagram depicting the overlapping targets between F. arctii and keloid formation showed 29 shared targets (Fig. 2). The details of these targets are provided in Table II.

The present study used Cytoscape 3.7.0 software to analyze the drug-active ingredient-target-pathway network for F. arctii and keloids. Fig. 3 shows the exported analytical data. A total of eight KEGG pathways linked to keloid development were selected to highlight the ones with targets from F. arctii. It was found that five primary active ingredients potentially interact with 99 targets across eight signaling pathways, suggesting that F. arctii may inhibit keloid development through multiple chemical components, targets and pathways. In the network, the light blue ovals denoted active ingredients, the green diamonds indicated pathways and the red circles represented common targets among components and keloids and between the pathways and keloids.

A PPI network for common genes was constructed using the STRING database. The network comprised 29 nodes and 151 edges, with an average node degree of 10.4 (Fig. 4A; Table SII). Fig. 4B and Table III show the nodes, including MYC, CASP3, PTEN, PTGS2, NOS2, CASP9, CASP8, BAX, CTNNB1 and AKT1, arranged in descending order of their degrees. Research shows that SIT significantly inhibits the progression of various malignant tumors by regulating PTEN expression (30,31). However, investigations on the combined mechanism of action of SIT and PTEN in keloids have yet to be conducted. The authors aim to explore the potential roles and mechanisms of action of SIT and PTEN in keloids. More network connections indicate stronger associations between the proteins. The PPI network suggested that F. arctii treatment affected keloids via multiple pathways, components and targets.

The results of the GO functional enrichment analysis for molecular functions (MFs), cellular components (CCs) and biological processes (BPs) are illustrated in Fig. 5A and Table SIII. The interaction targets were associated with 198 BPs, 27 CCs and 27 MFs. A histogram depicting the first 10 GO terms with P<0.01 was prepared. The primary BP terms included ‘positive regulation of vasoconstriction’, ‘response to estradiol and the apoptotic process’, ‘response to lipopolysaccharide’, ‘response to xenobiotic stimuli’, ‘response to hypoxia’, ‘positive regulation of smooth muscle cell proliferation’, ‘angiogenesis’, ‘inflammatory response’ and ‘MAPK cascade’. The primary CC terms were ‘caveolae’, ‘caspase complex’, ‘cytoplasm’, ‘macromolecular complex’, ‘membrane raft’, ‘plasma membrane’, ‘extracellular matrix’, ‘nucleus’, ‘death-inducing signaling complex’ and ‘mitochondrion’. The enriched MF terms included ‘enzyme binding’, ‘alpha1-adrenergic receptor activity’, ‘identical protein binding’, ‘peptidase activity’, ‘heme binding’, ‘macromolecular complex binding’ and ‘serine-type endopeptidase activity’.

The findings of pathway enrichment analysis indicated significant enrichment of the 28 core targets in 99 pathways (Table SIV). A bubble plot was prepared displaying the top 20 pathways with P<0.01 (Fig. 5B). The top 10 significant pathways included the PI3K-AKT signaling pathway, neurodegeneration pathways, the MAPK signaling pathway, the IL-17 signaling pathway, fluid shear stress and atherosclerosis, AGE-RAGE signaling in diabetic complications, the relaxin signaling pathway, the TNF signaling pathway, focal adhesion and apoptosis. Among these pathways, the PI3K-AKT signaling pathway may exhibit the closest association with keloids.

The molecular model revealed that SIT strongly interacts with PTEN, as depicted in Fig. 6. The space-filling model clearly demonstrated the perfect embedding of SIT within the crystal pocket of PTEN. Similarly, both the ribbon model image and local 2D binding model revealed the specific interactions between SIT and the associated protein residues. After simulating the components using AutoDock tools, a high affinity of −8.3 kcal/mol was computed. The receptor-ligand docking hypothesis suggests an inverse relationship between the docking energy and binding affinity. A lower docking energy indicates stronger binding affinity. SIT engages in van der Waals interactions with the PTEN residues PHE279, VAL275, ILE280, LEU320, THR319, GLN149, ASP324 and ASN323. The stable binding of SIT to PTEN involves the formation of alkyl and pi-alkyl interaction bonds with residues LEU318, TYR176, TYR177, ARG172 and ARG173.

Fig. 7A displays the chemical structure of SIT. To assess the effect of SIT on KFs proliferation and determine the optimal drug concentration for cellular experiments, the changes in cellular activity at 12, 24 and 48 h post-SIT treatment were compared to that in the control group using the CCK-8 assay. Fig. 7B shows that SIT significantly reduced the viability of KFs in a concentration- and time-dependent manner. The treatment of KFs with SIT of various concentrations for 48 h decreased the cell viability from 70 to 5%. The IC₅₀ values for KFs at 12, 24 and 48 h were 70.09, 33.78 and 33.75 µM, respectively. KFs treated with 30 µM SIT for 48 h showed a drug concentration of 33.75 µM, which was the 48 h IC₅₀ value. The inhibitory effects observed at other drug concentrations and time points deviated significantly from the corresponding IC₅₀ values. Subsequently, a drug gradient study on keloid cell cultures was conducted using 15 and 30 µM SIT as well as a drug reversion analysis using 30 µM SIT, both for 48 h.

Evidence from previous studies has shown that SIT inhibits the migration of various cancer cells (32). To assess the effect of SIT on the behavior of KFs, scratch and Transwell assays were performed and cell migration and invasion evaluated. The results revealed that in the control group, the scratch gap gradually closed over 24 h. SIT effectively impeded KFs migration in a concentration- and time-dependent manner (Fig. 7C and D). Migration was significantly reduced in cells treated with 15 and 30 µM SIT compared with that in control cells. Furthermore, to assess the migratory and invasive potential of cells, they were treated with SIT at various concentrations over a 24 h treatment period (Fig. 7E and F). The Transwell assay results revealed that treatment with SIT at various concentrations effectively inhibited keloid cell migration and invasion (Fig. 7G and H).

The present study assessed DNA synthesis by measuring the incorporation of EdU, a thymidine nucleotide analog, during the S phase of the cell cycle. The EdU incorporation rate indicates the DNA synthesis rate. Fig. 7I and J show the effective inhibition of KFs growth by SIT. Compared with the control treatment, treatment with 15 and 30 µM SIT significantly reduced the number of EdU-labeled cells (P<0.01).

E-cadherin, Vimentin and Snail are established key biomarkers of EMT (33). To determine the importance of EMT in cell migration and invasion (34), the present study evaluated the effect of SIT on EMT regulation in KFs. It was found that the inhibitory effect of SIT on migration is correlated with the attenuation of EMT. The morphological changes in KFs after 48 h of SIT treatment were assessed. Post treatment, the KFs transitioned from a multi-protruding spindle shape to a more regular triangular or short fusiform shape (Fig. 8A). The expression of EMT markers, including E-cadherin, ZO-1 and Vimentin, in SIT-treated KFs were analyzed using western blotting. The expression levels of E-cadherin and ZO-1 increased in a concentration-dependent manner in the SIT-treated groups, whereas the Snail and Vimentin levels decreased in a dose-dependent manner (Fig. 8B and C).

After it was established that SIT regulates EMT and the expression of associated markers, the present study investigated its inhibitory mechanisms in KFs. SIT treatment significantly increased PTEN protein expression (Fig. 8D and E). Given that PI3K and AKT are crucial downstream targets of PTEN (35), the expression levels of PI3K, p-PI3K, p-AKT and AKT were assessed concurrently. Total cellular proteins were extracted from cells treated with 15 and 30 µM SIT and control cells. SIT treatment reduced p-PI3K and p-AKT expression; however, it did not alter the total PI3K and AKT protein levels. Moreover, the changes in the p-PI3K and p-AKT levels were inversely correlated with the PTEN levels. These findings suggest that SIT may suppress growth of KFs via the PTEN/PI3K/AKT pathway.

SIT may suppress KFs growth via the PTEN/PI3K/AKT pathway. To confirm this, rescue experiments were conducted with YS-49, a PI3K/AKT activator (27). Findings from CCK-8 assays indicated that YS-49 treatment significantly reversed the inhibition of proliferation in SIT-treated KFs (Fig. 9A). Additionally, the results of the EdU, scratch and Transwell assays indicated that YS-49 reversed the inhibitory effects on the proliferation, migration and invasion of cells caused by SIT (Fig. 9B-H). Following YS-49 treatment, the PTEN, E-cadherin and ZO-1 levels decreased in SIT-treated KFs, whereas the Snail, Vimentin, p-PI3K and p-AKT levels increased (Fig. 10A-D). Collectively, these findings confirmed that SIT effectively suppresses the proliferation, migration and invasion and EMT progression of KFs by inhibiting the PTEN/PI3K/AKT signaling pathway.

After confirming the in vitro effects of SIT on cell proliferation, migration and invasion, its effect on keloid formation in vivo was assessed using a nude mice model with KFs xenografts. Fig. 11A shows the flowchart of the animal experiments, detailing the model establishment and treatment procedures. Keloid growth was markedly slower in the SIT group than in the saline group, which indicated the inhibitory potential of SIT (Fig. 11D). The results for keloid measurements (Fig. 11B), tumor weight (Fig. 11C) and nude mice weight were consistent (Fig. 11E). The saline group showed the lowest weight, whereas the treatment group showed normal growth trends. Nude mice in the SIT group began losing weight after the 15th day of treatment, probably due to keloid reduction. This indicated effective in vivo anti-keloid effects. To further elucidate the mechanism underlying the inhibition of keloid growth by SIT in vivo, PTEN expression in fibroproliferative tissues was evaluated using immunohistochemistry. SIT treatment markedly increased the PTEN protein levels (Fig. 11F). Notably, H&E staining revealed no significant effects on the heart, liver, kidney, or spleen, indicating the absence of adverse effects of SIT treatment on these vital organs (Fig. S1). No evidence was observed of myocardial fiber disarray, necrosis, or inflammatory infiltration. Cardiac muscle structure and cellular integrity remained intact. No signs of hepatocyte degeneration, necrosis, or inflammatory cell infiltration were observed. There was no indication of glomerular damage, tubular necrosis, or interstitial inflammation. No evidence of hemorrhage or structural disorganization was detected. These results, which are consistent with the in vitro findings, indicated that SIT inhibits keloid growth by modulating PTEN expression in vivo.

The present study assessed formulation safety via blood biochemical analyses, including those for BUN, CREA, ALT and AST. The levels of BUN, CREA, ALT and AST remained consistently low across groups (Fig. 12). The findings indicate that SIT has a favorable safety profile for keloid treatment in nude mouse.