Introduction

Osteoporosis is a systemic disease characterized by decreased bone mass and deterioration of bone microarchitecture, leading to increased bone fragility and fracture risk. Currently employed pharmacological treatments include bisphosphonates and parathyroid hormone analogues. However, these drugs have certain limitations and adverse effects, such as joint or muscle pain, osteonecrosis of the jaw, and the risk of atypical femoral fractures1. Therefore, the discovery of novel therapeutic agents with enhanced efficacy and favourable safety profiles represents a critical and urgent priority2.

Chinese yam (Rhizoma Dioscoreae), characterized by a sweet flavour and a neutral nature in traditional Chinese medicine, is typically employed as an adjuvant herb to “tonify the liver and kidney, and alleviate knee and shin soreness”. As a medicinal food homology material, its primary bioactive components include polysaccharides, saponins, polyphenols, fatty acids, proteins, amino acids, and trace elements3. In treating osteoporosis, it has been confirmed that the isolated components of Chinese yam, such as yam polysaccharides4, yam-derived exosome-like nanovesicles5, and diosgenin6, can regulate the proliferation and differentiation of osteoblasts by inhibiting the CASP3 protein and blocking the Wnt/β-catenin pathway. However, existing research has predominantly focused on individual active constituents of Chinese yam and their isolated signaling pathways, while the anti-osteoporotic effects of many other potential components within the herb remain to be fully uncovered. Consequently, a comprehensive understanding of the complete set of constituents and the underlying mechanisms by which Chinese yam counteracts osteoporosis is still lacking. There is a pressing need to systematically elucidate its network regulatory characteristics, emphasizing the synergistic effects of multiple components and targets.

With the rapid advancement of computer technology, bioinformatics, and systems biology theories, network pharmacology has emerged as a powerful research strategy for exploring the complex mechanisms of traditional Chinese medicine or functional foods. The key concept of “network target, multicomponent” in network pharmacology can systematically clarify the mechanism of action of medicinal and functional foods on different diseases and promote the development of evidence-based medicine and the discovery of new medicinal or functional foods. This study aimed to investigate the potential mechanisms underlying the therapeutic effects of Chinese yam on osteoporosis through network pharmacology, molecular dynamics simulations, and biological validation (Fig. 1). This research provides a robust scientific foundation for elucidating its therapeutic mechanisms and developing medicinal and functional foods.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Schematic representation of the experimental procedures.

Materials and methods

Screening of active components in Chinese yam and prediction of targets

Using the Latin binomial name “Rhizoma Dioscoreae” as the keyword, the following universal screening criteria were applied to the Traditional Chinese Medicine Systems Pharmacology (TCMSP) database (https://old.tcmsp-e.com/tcmsp.php): oral bioavailability (OB) ≥ 30% and drug-like properties (DL) ≥ 0.18, to identify compounds with good oral absorption and high drug potential, as described in the TCMSP database7. Moreover, the HERB database (http://herb.ac.cn/) was used to retrieve relevant information about Chinese yam not accessible in the TCMSP database, with the screening criterion being a relative molecular mass ≤ 500. The identified active components in the HERB database were further filtered on the basis of the following criteria: a lipid‒water partition coefficient ≤ 5, ≤ 10 hydrogen bond acceptors, and ≤ 5 hydrogen bond donors. The obtained SMILES and chemical structures from PubChem were subjected to reverse target prediction in the Swiss Target Prediction database, using a probability ≥ 0.1 as the screening criterion. The corresponding gene names were subsequently retrieved from the UniProt database to construct the Chinese yam drug target network.

Construction of drug‒active ingredient‒target networks and bioinformatics analysis

Using “Osteoporosis” as the keyword, we conducted searches and screenings across the OMIM (http://www.omim.org/), GeneCards (https://www.genecards.org/), DrugBank (https://go.drugbank.com/), TTD (https://idrblab.org/ttd/), and DisGeNET (https://disgenet.com/) databases to construct a disease target network for osteoporosis. The active targets of Chinese yam and osteoporosis disease-related targets were compared using the Jvenn website (http://bioinfo.genotoul.fr/jvenn/), a venn diagram was generated, and the shared targets between Chinese yam and osteoporosis were obtained. The obtained shared targets were input into the STRING database (https://cn.string-db.org/) search box to construct a protein‒protein interaction (PPI) network diagram. The resulting TSV file was downloaded and visualized using Cytoscape 3.9.1 software, with the top 5 targets ranked by node degree value being selected. Concurrently, the shared targets of Chinese yam and osteoporosis were subjected to Gene Ontology (GO) functional analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis using the DAVID platform (http://david.abcc.ncifcrf.gov/), with visualization performed using the Weishengxing platform (http://www.bioinformatics.com.cn/). We obtained the permission from Kanehisa’s laboratory to use the KEGG software, as described in previous studies8,9,10.

Molecular docking

The target protein for molecular docking was selected from the results above, and its 3D structure was retrieved from the PDB database (https://www.rcsb.org/). The 2D structure of the active component from Chinese yam was downloaded from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). Additionally, the effective osteoporosis treatment drug calcitriol was used as a positive control to compare the efficacy of the docking method and the affinity of related compounds. Molecular docking was performed using Schrodinger Maestro v2021–2 software. Maestro software was used to supplement missing hydrogen atoms in the target protein structure, after which the target protein energy was minimized using the OPLS4 force field. Maestro’s LigPrep function is employed to generate diverse ligand conformations for docking. The potential binding pockets were analysed using the SiteMap feature. Flexible ligand docking was performed with the ligand docking function. Finally, the results were visualized using PyMOL 3.0 and Discovery Studio 2021 Client software.

Molecular dynamics simulation

Molecular dynamics simulation was performed using the Gromacs 2023 software on the most active compound-target combination with the highest binding energy identified in the molecular docking results. The protein parameters were set to CHARMM 36. The specific steps and parameter configurations were as follows: The PDB format of the complex was converted to the “gro” format, and the complex served as the initial structure for molecular dynamics simulations. The Generalized Amber Force Field (GAFF) was applied to the active compound using AmberTools22, while Gaussian 16 W was employed for the hydrogenation of small molecules and calculation of the RESP potential. Both the active compound and protein were solvated in the TIP3P water model, with periodic boundary conditions applied. During the initial phase of the simulation, 100,000 steps of isothermal-isovolumetric (NVT) and isothermal-isobaric (NPT) equilibration processes were conducted to ensure the system reached thermodynamic equilibrium. During the equilibration process, the cut-off radii for van der Waals forces and Coulomb interactions were set to 1.0 nm. Finally, under constant temperature (300 K) and constant pressure (1 bar), 5,000,000 steps of production simulations were completed, with a total duration of 100 ns. The obtained results were analysed for root mean square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (Rg), solution accessible surface area (SASA), hydrogen bonds (H-bonds), and Gibbs free energy calculations.

Preparation of Chinese yam extract

Chinese yam (Batch No.20250102, Dongguan Guoyao Group Traditional Chinese Medicine Decoction Pieces Co., Ltd.) was identified as Chinese yam by Liu Fengying, a licenced pharmacist at the Traditional Chinese Medicine Pharmacy of Dongguan Songshan Lake Central Hospital. The sample complied with the thin-layer chromatography identification standard for Chinese yam in the 2020 edition of the Chinese Pharmacopoeia. 50 g of Chinese yam was soaked in 500 mL of water for 30 min. Afterwards, the mixture was brought to a boil and reduced to a simmer, after which the decoction was maintained for 1 h. After cooling, the filtrate was filtered and collected. Five hundred millilitres of water was added to the residue, and the decoction process was repeated. The two decoctions were combined, and the solution was concentrated to 50 mL using a rotary evaporator. The concentrated solution was placed in a freezer at −20 °C for 12 h and then transferred to a lyophiliser precooled at −80 °C for 2 h. The lyophilisation parameters were as follows: temperature, 80 °C; vacuum, 5 ± 1 Pa; and drying duration, 48 h. The final lyophilised Chinese yam powder weighed 9.004 g. The lyophilised Chinese yam powder was dissolved in zebrafish embryo culture medium or cell culture medium and filtered through a membrane to obtain a Chinese yam stock solution with a concentration of 10 mg/mL.

Zebrafish experiment

AB zebrafish larvae were provided by Hubei Chuangxin Biotechnology Co., Ltd. The AB zebrafish larvae were cultivated in media supplemented with 5 mmol/L NaCl, 0.17 mmol/L KCl, 0.16 mmol/L MgSO4, and 0.33 mmol/L CaCl2. The wild-type zebrafish strain AB were randomly selected at 3 days post-fertilization, distributed into 6‒well plates, and divided into 6 groups with 6 zabrafish per well (per group). Each well contained 3 mL of water. Normal control group: zebrafish treated with aquaculture water. Model group: zebrafish treated with 25 µmol/L dexamethasone (DEX) (Sigma, St. Louis, Missouri, USA) aqueous solution. Test substance groups (3 groups): zebrafish treated with 50 µg/mL, 100 µg/mL, or 250 µg/mL Chinese yam extract (This concentration range had been validated as having no effect on zebrafish survival rates). Positive control group: zebrafish treated with 100 nmol/L aqueous calcitriol (MedChemExpress, New Jersey, USA) solution. The corresponding culture medium was replaced daily. The temperature was set at 28.0 ± 0.5 °C, with a controlled circadian rhythm of 14/10 hours. After 9 days of fertilization, the embryos were stained with carmine red to observe the effects on embryonic development. The experimental protocols were approved by the Ethics Committee of Affiliated Dongguan Hospital, Southern Medical University (Dongguan People’s Hospital) and all methods were carried out in accordance with relevant guidelines (including ARRIVE guidelines) and regulations.

Alizarin red staining

The 9-day-old zebrafish larvae were anaesthetized and euthanized using 100 mg/L MS-222 anaesthetic; the anaesthetic was drained and the larvae were fixed in 4% paraformaldehyde for 2–4 h. The fixative was removed; the larvae were dehydrated with 50% ethanol for 10 min; the dehydrant was discarded; and the larvae were washed in PBST three times for 5 min each. Then, with the larvae were bleached with 1% KOH and 1% hydrogen peroxide for 25–30 min. The bleaching solution was aspirated, and the larvae were washed three times with PBST for 5 min each. Next, the larvae were stained with alizarin red (Sigma, St. Louis, Missouri, USA) for 2.5 h. The dye solution was aspirated, and the larvae were washed with PBST to remove excess dye. The zebrafish larvae were incubated in 0.5% KOH–glycerol (3:1) for 6–8 h, 0.5% KOH–glycerol (1:1) for 6–8 h, and 0.5% KOH–glycerol (1:3) for 12 h, followed by immersion in pure glycerol. The skulls of the zebrafish larvae were examined under a microscope while the stained images were collected. The stained areas of the skulls in zebrafish larvae were quantified using IPP 6.0 software and expressed in pixels.

MC3T3-E1 cell culture

MC3T3-E1 cells (GNM15, National Collection of Authenticated Cell Cultures, Shanghai) were cultured in MEM medium (Gibco, USA) supplemented with 10 mmol/L β-glycerophosphate sodium (Sigma, St. Louis, Missouri, USA) and 50 mg/L vitamin C (Solubio Technology, Beijing, China), all of which were prepared and used immediately.

Assays for cell viability

MC3T3-E1 cells were seeded at a density of 2 × 10³ cells per well in a 96‒well plate. After being cultured until approximately 70% confluence was reached, the cells were treated with aqueous extracts of Chinese yam at different concentrations (5, 10,25,50,75, and 100 µg/mL) and incubated for an additional 48 h. A total of 100 µL of Cells Counting Kit–8 (CCK–8) (NCM Biotech, Suzhou, China) working solution was added per well, and the reaction was carried out at 37 °C under light protection for 2 h. The absorbance was measured using a multifunctional microplate reader (Thermo Fisher Scientific Inc, USA) at a wavelength of 450 nm.

Alkaline phosphatase (ALP) staining and ALP activity

A total of 2.5 × 104 cells per well were inoculated in a 24‒well plate and were divided into the following groups: osteogenic induction medium (OM) group, DEX group (100 µmol/L), DEX + Chinese yam extract (5 µg/mL), DEX + Chinese yam extract (25 µg/mL), and DEX + Chinese yam extract (50 µg/mL). The medium was changed to 500 µL per well, and the medium was replaced every day. ALP staining was performed and ALP activity was assessed on the 7th day using a BCIP/NBT alkaline phosphatase colorimetric kit and an ALP assay kit (Beyotime, Shanghai, China), respectively, according to the instructions.

Western blot analysis

Total protein was collected and extracted, and the protein concentration was determined by the BCA method. Equal amounts of protein samples were separated by SDS‒PAGE electrophoresis and transferred to PVDF membranes. After blocking with 5% skim milk, the membranes were incubated overnight at 4 ℃ with primary antibodies: p-PI3K and PI3K (1:2000; Affinity, Melbourne, Australia), runt-related transcription factor 2 (RUNX2) (1:1000; Abcam, Cambridge, UK), osteocalcin (OCN) (1:1000; Proteintech, Wuhan, China), and β-actin (1:5000; Beyotime, Shanghai, China). After the membranes were washed with TBST, they were incubated with the corresponding secondary antibodies were incubated at room temperature for 1 h. The relative expression levels of target proteins were analysed using enzyme-linked chemiluminescence (Millipore, USA) and scanning photodensity (ChemiDoc XRS1 Systems Bio–Rad, Hercules, USA). Image J software was used to analyse the results, with β-actin used as the internal standard.

Statistical analysis

All experimental data are expressed as the mean ± SD and were processed using GraphPad Prism 8 software. Multiple groups were analysed statistically using one-way ANOVA. When statistically significant differences were present, pairwise comparisons were performed using Tukey’s multiple comparison test. Statistical significance was set at p < 0.05.

Results

Results of screening the active components and drug targets of Chinese yam

In this study, a total of 71 active components from Chinese yam were identified using the TCMSP database. Further selection based on OB ≥ 30% and DL ≥ 0.18 yielded 16 active components, as shown in Table 1. The filter criteria are well-established benchmarks in network pharmacology studies. They are designed to filter for compounds with satisfactory oral bioavailability and pharmacological properties during early screening stage, thereby increasing the likelihood of identifying biologically active and pharmaceutically relevant constituents7. These components were subsequently screened through the Herb database, and then a secondary screening was conducted in the PubChem database with following conditions: lipid-water partition coefficient ≤ 5, hydrogen bond acceptors ≤ 10, and hydrogen bond donors ≤ 5, yielding 57 active components (Table 2). After the two databases were merged and deduplicated, a total of 73 active components of Chinese yam were obtained. The 73 compounds underwent reverse target prediction in the Swiss Target Prediction database. Using a probability ≥ 0.1 as the screening criterion, and after deduplication, a total of 439 drug targets for Chinese yam were identified.

Table 1 Information related to active components of Chinese yam retrieved from the TCMSP database.
Full size table
Table 2 Information related to active components of Chinese yam retrieved from the HERB database.
Full size table

Osteoporosis target screening results

Using the GeneCards, OMIM, TTD, Disgenet, and DrugBank databases, a total of 2189, 45, 29, 1098 and 187 genes associated with osteoporosis, respectively, were identified. On this basis, the data were integrated and deduplicated to obtain 2771 target genes. A venn diagram analysis of the 439 active component targets in Chinese yam and the 2771 osteoporosis disease-related targets revealed 176 shared targets, with detailed results shown in Fig. 2A and B.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Active components of Chinese yam and their key targets in the treatment of osteoporosis. Venn diagram of osteoporosis-related targets (A). Venn diagram of 176 osteoporosis-related targets of Chinese yam (B). Network diagram of active ingredients in Chinese yam‒potential targets (C). The PPI network of potential targets of Chinese yam (D).

Construction of the target network for active components in Chinese yam

Using Cytoscape 3.9.1 software, we constructed an “active component–target” network for Chinese yam. As shown in Fig. 2C, the network consisted of 22 components and 176 targets. The triangles represent the active components of Chinese yam, whereas the targets are shown as rectangles (with brighter colours indicating higher degree values).

Construction and analysis of the PPI network

Using the STRING database, we constructed a PPI network diagram for the target proteins identified above. With the species set to “Homo sapiens” and an exchange score of 0.7, we constructed a PPI network comprising 176 nodes and 2242 edges. This PPI network was imported into Cytoscape 3.9.1 software for analysis, yielding Fig. 2D. Nodes represent protein nodes, with larger nodes and darker colours corresponding to higher degree values. The top 6 core targets were epidermal growth factor receptor (EGFR), proto-oncogene, non-receptor tyrosine kinase (SRC), interleukin 6 (IL6), AKT serine/threonine kinase 1 (AKT1), catenin beta 1 (CTNNB1, encodes β-catenin), and heat shock protein 90 alpha family class A member 1 (HSP90AA1). Table 3 shows the degree values, betweenness centrality (BC) scores, and closeness centrality (CC) scores for the top 10 targets in the PPI network.

Table 3 Key targets.
Full size table

GO and KEGG pathway enrichment analysis

GO enrichment analysis and KEGG pathway enrichment analysis were performed on the shared Chinese yam and osteoporosis-related targets using the DAVID data platform. The cellular component (CC) analysis yielded 81 entries, primarily enriched in the plasma membrane, cytosol, cytoplasm, and receptor complex (Fig. 3A). Biological processes (BP) analysis yielded 757 entries, with the enriched genes involved primarily in the positive regulation of gene expression, the inflammatory response, and the positive regulation of PI3K/Akt signaling (Fig. 3B). Molecular function (MF) analysis yielded 211 items, primarily involving identical protein binding, ATP binding, histone H2AXY142 kinase activity, and protein tyrosine kinase activity (Fig. 3C).

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

GO and KEGG enrichment analyses for the mechanism of Chinese yam in treating osteoporosis. (A) Cellular component (CC). (B) Biological processes (BP). (C) Molecular function (MF). (D) The top 20 KEGG terms.

KEGG pathway analysis yielded a total of 159 pathways. The top 20 results, ranked by p value, are shown in Fig. 3D. These pathways primarily include the pathways in cancer, metabolic pathways, PI3K/Akt signaling pathway, Ras signaling, and the MAPK signaling pathway. The PI3K/Akt signaling pathway has been widely demonstrated to stimulate osteoblast growth and regulate osteoclast apoptosis11. Therefore, Chinese yam may exert its antiosteoporotic effects by modulating the activity of the PI3K/Akt signaling pathway, specifically through the regulation of key genes highlighted in Fig. 4.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Distribution of target proteins related to Chinese yam in the PI3K/Akt pathway. Red markers represent the potential target proteins.

Constructing the “Active Ingredient–Target–Pathway” network

The top 20 pathways and related targets from the KEGG enrichment analysis were subsequently imported into Cytoscape 3.9.1 software to generate a “Chinese yam–Active component–Target–Signaling pathway–Osteoporosis” diagram. Diamonds represent drugs, hexagons represent diseases, rectangles represent targets, circles represent active components, and V‒shapes represent signaling pathways. The network comprises 222 nodes and 916 edges, as shown in Fig. 5.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Drug‒component–target–pathway–disease network diagram of Chinese yam in the treatment of osteoporosis.

Molecular docking

Molecular docking was performed using the top 6 core targets (EGFR, SRC, IL6, AKT1, CTNNB1, and HSP90AA1) ranked by degree values and the top 6 active components (Phytocassane A, Batatasin II, Gibberellic Acid, Dihydroquercetin, Garcinone D, and Diosgenin). Concurrently, we compared these results with those when the osteoporosis drug calcitriol was used as a control, as summarized in Table 4. The binding affinity is presented as a docking score (in kcal/mol), where a more negative value indicates a stronger predicted binding affinity between the ligand and the target protein, with negative value above 4.25, 5.0, and 7.0 suggesting binding, good binding, and strong binding, respectively12. Among these compounds, Dihydroquercetin strongly bound to all the core targets. Calcitriol bound to these targets but exhibited generally weaker affinity than the other 6 active compounds did.

The conformations formed by each core target with its highest-scoring active compound were adopted as the final docking conformations and visualized using PyMOL 3.0 and Discovery Studio 2021 Client software. Garcinone D formed three hydrogen bonds with EGFR (MET‒793, THR‒854, ASP‒855, and LYS‒745) and AKT1 (TYR‒13, THR‒56, and SER‒54), as illustrated in Fig. 6A and D, respectively. Dihydroquercetin exhibited a range of hydrogen-bonding interactions with the target proteins: five bonds with SRC (ASP‒217, SER‒211, GLN‒184, LYS‒263, LYS‒78, and ILE‒261; Fig. 6B), two with IL6 (GLU‒55 and GLU‒59; Fig. 6C), and single bonds with both CTNNB1 (GLN‒203; Fig. 6E) and HSP90AA1 (ASP‒93; Fig. 6F).

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

The docking modes between the active compounds of Chinese yam and the key targets of osteoporosis. Binding mode of EGFR to Garcinone D (A), AKT1 to Garcinone D (B), IL6 to Dihydroquercetin (C), SRC to Dihydroquercetin (D), CTNNB1 to Dihydroquercetin (E) and HSP90AA1 to Dihydroquercetin(F).

Table 4 Molecular docking score.
Full size table

Molecular dynamics simulation

To further evaluate the effects of Chinese yam bioactive compounds on osteoporosis-related targets, the following complexes with high docking scores were selected: EGFR–Garcinone D (Complex 1), SRC–Dihydroquercetin (Complex 2), IL6–Dihydroquercetin (Complex 3), AKT1–GarcinoneD (Complex 4), CTNNB1–Dihydroquercetin (Complex 5), and HSP90AA1–Dihydroquercetin (Complex 6), which subjected to molecular dynamics simulations to investigate their dynamic properties. RMSD data specifically reflect the temporal evolution of structural differences between proteins and small molecules during simulations, directly indicating the stability of molecular docking systems. As shown in Fig. 7A1–7F1, complexes 1, 2, 3, 4, 5, and 6 reached equilibrium at 30 ns, 60 ns, 80 ns, 70 ns, 75 ns, and 60 ns, respectively. Upon stabilization, the average RMSD values for complexes 1, 2, 3, 4, 5, and 6 were 0.20 nm, 0.24 nm, 1.10 nm, 0.29 nm, 0.31 nm, and 1.20 nm, respectively.

RMSF serves as an indicator for assessing protein dynamics and reflects the motion amplitude of amino acid residues. RMSF reveals the dynamic changes in important functional residues within a protein molecule. A high RMSF value indicates increased structural flexibility, suggesting the potential presence of unstable chemical bonds or cyclic structures. Conversely, a low RMSF value reflects enhanced structural stability, suggesting the possible formation of secondary structures such as helices or folds13. As shown in Fig. 7A2, RMSF analysis of the EGFR–Garcinone D complex revealed significant flexibility in the EGFR amino acid residues within the ranges of 30–60, 160–180, and 280–310. Similarly, complexes 3–6 exhibit multiple flexible residue regions (Fig. 7C2–7F2). As depicted in Fig. 7A3–7F3, the results indicate that the number of hydrogen bonds in complexes 1‒6 is 0‒5, 0‒5, 0‒5, 0‒4, 0‒5, and 1‒5, respectively, demonstrating the presence of stable hydrogen bonds in these complexes. Fig.7 should be placed after the result description in the first two paragraphs of "Molecular dynamics simulation".100 ns molecular dynamics simulation of the EGFR to Garcinone D (A1‒A3), SRC to Dihydroquercetin (B1‒B3), IL6 to Dihydroquercetin (C1‒C3), AKT1 to Garcinone D (D1‒D3), CTNNB1 to Dihydroquercetin (E1‒E3) and HSP90AA1 to Dihydroquercetin (F1‒F3)

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.
Full size image

100 ns molecular dynamics simulation of the EGFR to Garcinone D (A1‒A3), SRC to Dihydroquercetin (B1‒B3), IL6 to Dihydroquercetin (C1‒C3), AKT1 to Garcinone D (D1‒D3), CTNNB1 to Dihydroquercetin (E1‒E3) and HSP90AA1 to Dihydroquercetin (F1‒F3).

The SASA is used to evaluate the extent of contact between proteins and surrounding solvent molecules14. As shown in Fig. 8A, SASA analysis revealed mean values of 160 nm², 140 nm², 90 nm², 110 nm², 85 nm², and 108 nm² for complexes 1–6, respectively, with all the systems reaching equilibrium after 10 ns. Rg analysis evaluates the overall conformation of protein complexes, reflecting the degree of compactness or expansion of the protein structure during the simulation15. As shown in Fig. 8B, the mean Rg values were 2.0 nm, 1.9 nm, 1.6 nm, 1.7 nm, 1.5 nm, and 1.7 nm, respectively, indicating that all complex systems maintained equilibrium throughout the simulation.

Fig. 8
Fig. 8The alternative text for this image may have been generated using AI.
Full size image

Solvent accessible surface area(A) and Radius of gyration analysis(B).

Gibbs free energies were calculated from the RMSD and Rg values of complexes 1‒6, with corresponding 3D and 2D Gibbs free energy plots generated. Gibbs free energy plots are widely used to evaluate free energy changes in receptor–ligand binding interactions, with their values directly reflecting complex stability. As shown in Fig. 9A, the Gibbs free energy of the EGFR‒Garcinone D complex exhibits multiple distinct and sharp minimum energy regions, forming three hydrogen bonds involving residues META:793, LYSA:745, and ASPA:855.

Fig. 9
Fig. 9The alternative text for this image may have been generated using AI.
Full size image

Gibbs free energy analysis of the EGFR to Garcinone D (A), SRC to Dihydroquercetin (B), IL6 to Dihydroquercetin (C), AKT1 to Garcinone D (D), CTNNB1 to Dihydroquercetin (E) and HSP90AA1 to Dihydroquercetin (F).

To further investigate the binding stability of all complex systems, the binding free energies of complexes 1–6 were calculated using the MM/GBSA method. As shown in Table 5, the total binding free energies for complexes 1–6 were − 30.24 ± 2.65, −30.72 ± 3.07, −18.52 ± 3.50, −19.70 ± 4.56, −11.88 ± 3.95, and − 32.73 ± 2.50 kcal/mol, respectively. These low-energy states indicate the formation of stable interactions between the active ingredients and their targets16.

It should be noted that the docking and MD simulations in this study were conducted using specific software and force fields. While this approach is standard, the results may be influenced by the particular algorithms employed. Future studies incorporating cross-validation with alternative docking programs or different force fields would further strengthen the confidence in these predictions.

Table 5 MM/GBSA results.
Full size table

Chinese yam extract alleviated DEX-induced osteogenic mineralization inhibition in zebrafish

Owing to the high degree of conservation between zebrafish and human skeletal development, zebrafish are widely used in osteoporosis research17.The development of zebrafish bones was studied by assessing the extent of bone mineralization through alizarin red S staining (Fig. 10A). Compared with the control group, the DEX-induced osteoporosis model group exhibited significantly reduced bone mineralization in the ventral (p < 0.01; Fig. 10B) and lateral (p < 0.01; Fig. 10C). Compared with that in the model group, zebrafish larvae treated with (50 µg/mL, 100 µg/mL, or 250 µg/mL) significantly increased the mineralization area of bones (p < 0.01), suggesting that Chinese yam alleviated suppression of the bone formation induced by DEX in zebrafish larvae.

Fig. 10
Fig. 10The alternative text for this image may have been generated using AI.
Full size image

Effect of Chinese yam extract on DEX-induced bone formation inhibition in zebrafish larvae. (A) Abdominal and lateral views of zebrafish larval heads at 9 days post-fertilization stained with alizarin red, with or without exposure to Chinese yam extract (50 µg/mL, 100 µg/mL, and 250 µg/mL). (B) Mineralized area of the skull (ventral). (C) Mineralized area of the skull (lateral). Data represent the mean ± SD (n = 6 independent zebrafish larvae).** p < 0.01 vs. the Control group; ##p < 0.01, vs. the DEX model group.

Chinese yam extract alleviated DEX-induced ALP activity inhibition in MC3T3-E1 cells

To evaluate the effect of Chinese yam extract on MC3T3-E1 cells, cells were treated with varying concentrations of Chinese yam aqueous extract (5 µg/mL to 100 µg/mL) for 48 h, after which proliferation was assessed using a CCK‒8 assay. Chinese yam extract did not affect cell viability at concentrations ranging from 5 µg/mL to 75 µg/mL (Fig. 11A).

Fig. 11
Fig. 11The alternative text for this image may have been generated using AI.
Full size image

Effect of Chinese yam extract on the DEX-induced inhibition of MC3T3-E1 cell osteogenic differentiation. (A) The cells were treated with Chinese yam extract at different concentrations for 48 h. ALP activity (B) and ALP staining (C) of MC3T3-E1 cells after the stimulation with DEX in the absence or presence of Chinese yam extract (5 µg/mL, 25 µg/mL, and 50 µg/mL) for 7 days. OM, osteogenic induction medium; DEX, dexamethasone. Data represent the mean ± SD (n = 3 cultures per group). ** p < 0.01 vs. the OM group; ##p < 0.01, vs. the DEX model group.

ALP is an early marker of osteogenic differentiation, and the results of ALP staining and activity analysis in MC3T3-E1 cells are shown in Fig. 11B and C. Compared with the OM group, the DEX group exhibited significantly weaker ALP positive staining and markedly decreased ALP activity (p < 0.01). In contrast, compared with the DEX group, the Chinese yam extract (25 µg/mL and 50 µg/mL) significantly increased ALP activity (p < 0.01) and markedly increased the intensity of ALP positive (p < 0.01). These experimental results suggest that Chinese yam extract can significantly alleviate DEX-induced inhibition of osteogenic differentiation in MC3T3-E1 cells within a specific concentration range.

Chinese yam extract alleviated DEX-induced osteoporosis through the PI3K signaling pathway

To further validate the mechanism by which Chinese yam extract affects osteoporosis, western blot analyses were performed, and the results revealed Chinese yam extract (25 µg/mL and 50 µg/mL) significantly abrogated the changes in the p-PI3K/PI3K protein ratio and promoted the inhibition of PI3K protein phosphorylation inhibition in the DEX model group (p < 0.05; Fig. 12A and B). To further clarify the role of PI3K in the antiosteoporosis effect of the Chinese yam extract, the cells were pre-treated with 10 µM LY294002 (a PI3K inhibitor), and then treated with DEX and the Chinese yam extract. As shown in Fig. 12C and D, LY294002 eliminated the promoting effect of the Chinese yam extract on ALP staining and activity (p < 0.01). We further examined the effects of Chinese yam extract on the protein expression of RUNX2 and OCN in MC3T3-E1 cells. The results showed that DEX significantly downregulated the expression of RUNX2 (p < 0.01) and OCN (p < 0.01), whereas Chinese yam extract promoted their expression (p < 0.05, p < 0.01). However, this positive effect was abolished by LY294002 (p < 0.05, p < 0.01; Fig. 12E and F). These experimental results indicated that the PI3K signaling pathway promotes osteogenic differentiation in MC3T3-E1 cells induced by Chinese yam extract.

Fig. 12
Fig. 12The alternative text for this image may have been generated using AI.
Full size image

Effect of Chinese yam extract on the PI3K pathway in DEX-treated MC3T3-E1 cells. (A and B) Immunoblots and the relative expression levels of p-PI3K/PI3K in MC3T3-E1 cells after stimulation with DEX in the absence or presence of Chinese yam extract (5 µg/mL, 25 µg/mL, and 50 µg/mL) for 5 days. β-actin was used as an internal reference. ALP activity (C) and ALP staining (D) of MC3T3-E1 cells pretreated with 10 µM LY294002 for 1 h, followed by DEX and Chinese yam extract treatment for 7 days. (A and B) Immunoblots and the relative expression levels of RUNX2 and OCN of MC3T3-E1 cells pretreated with 10 µM LY294002 for 1 h, and followed by DEX and Chinese yam extract treatment for 5 days. OM, osteogenic induction medium; DEX, dexamethasone. Data represent the mean ± SD (n = 3 cultures per group). ** p < 0.01 vs. the OM group; #p < 0.05, ##p < 0.01, vs. the DEX model group; &&p < 0.05,&&p < 0.01, vs. the DEX + Chinese yam group.

Discussion

Chinese yam has a sweet taste and a neutral nature and can have effects on lung, spleen, and kidney meridians. Its core functions include tonifying the spleen and nourishing the stomach, generating fluids and benefiting the lungs, and tonifying the kidneys to secure essence. This aligns perfectly with the therapeutic principle of “strengthening the spleen and tonifying the kidneys to nourish the bones”. In traditional Chinese medicine, Chinese yam is often combined with other herbs that tonify the kidneys, replenish essence, and strengthen tendons and bones, such as prepared rehmannia root, cornelian cherry, eucommia bark, and goji berries, to treat osteoporosis. However, the specific mechanism through which Chinese yam treats osteoporosis remains unclear and requires further in-depth research.

On the basis of multiple network pharmacology databases and online analysis platforms, this study revealed that Chinese yam primarily treats osteoporosis through components such as Phytocassane A, Batatasin II, Gibberellin, Dihydroquercetin, Garcinone D, Diosgenin, and others. PPI analysis revealed that the therapeutic targets of Chinese yam for osteoporosis mainly involve EGFR, SRC, IL6, AKT1, and CTNNB1. EGFR is a tyrosine kinase that is widely distributed on cell membranes; it participates in the metabolic regulation of chondrocytes and plays crucial regulatory roles in chondrocyte proliferation, differentiation, and osteoclastogenesis18,19. SRC belongs to the SRC family of protein kinases and participates in the activation of multilevel signaling cascades. It effectively inhibits RANK ligand-induced osteoclast formation, with its c-Src kinase serving as a pivotal node in cellular signaling networks that suppress RANKL-induced osteoclastogenesis20. IL-6 is an inflammatory factor that activates macrophages and exacerbates bone destruction via the RANKL pathway, thereby negatively affecting fracture healing21,22. AKT1 is the most extensively studied member of the AKT family. Research has indicated that overexpression of AKT1 in aged bone marrow mesenchymal stromal cells accelerates their senescence process23. AKT1 knockout in mice is correlated with reduced skeletal muscle mass and shortened lifespan24. CTNNB1,which encodes β-catenin, is a core target that primarily regulates osteoblast proliferation, differentiation, and function through the Wnt/β-catenin pathway, thereby disrupting the balance between bone formation and resorption25.

Bone metabolism is a complex process likely involving crosstalk between multiple pathways. KEGG analysis revealed that Chinese yam primarily affects osteoporosis through the pathways in cancer, metabolic pathways, PI3K/Akt signaling pathway, Ras signaling, and the MAPK signaling pathway. The Wnt/β-catenin pathway is not only a tumor-associated pathway but is also recognized as the master switch for osteoblastic differentiation and a key pathway in osteoporosis26. The Ras signaling pathway is a crucial pathway regulating cell death, cell migration, division, or differentiation. It also modulates the proliferation and osteogenic differentiation of bone marrow mesenchymal stem cells, and exerts anti-osteoporosis effects27. The PI3K/Akt signaling pathway has been extensively demonstrated to play a crucial mechanistic role in osteoporosis28. EGFR serves as an upstream regulator in the PI3K/Akt signaling pathway; its activation influences interactions between downstream transcription factors and the PI3K/Akt signaling pathway29. AKT, a central component of the PI3K/Akt signaling pathway, activates multiple downstream effector signaling molecules. KEGG analysis revealed that the MAPK signaling pathway is involved in the treatment of osteoporosis by Chinese yam. Activation of the MAPK pathway can induce chondrocyte hypertrophy and death while upregulating the expression of degradative enzymes such as matrix metalloproteinases and glycosaminoglycan proteases, thereby contributing to the development of osteoporosis30,31.

Elucidating the synergistic or antagonistic mechanisms among multiple active components is crucial for understanding the holistic action mechanisms of traditional Chinese medicine therapies. The primary constituents of Chinese yam may exert synergistic effects by modulating the aforementioned key targets and signaling pathway networks. Specifically, dihydroquercetin has been demonstrated to significantly inhibit p-p38 MAPK expression, thereby suppressing inflammatory responses32,33. Diosgenin further synergistically enhances this effect by reducing levels of p-p38, p-ERK, and p-JNK in the MAPK signaling pathway, thereby mitigating oxidative damage and inflammation34,35, ultimately inhibiting osteoclast proliferation and reducing bone loss. Dihydroquercetin may also inhibit osteoclastogenesis by blocking RANKL-induced NF-κB signaling while promoting osteoblast activity36,37. Glycyrrhizic acid D influences bone cell differentiation and modulates osteoblast function by inhibiting NF-κB and MAPK signaling pathways38, thereby improving bone quality. Research indicates that interactions between the PI3K/Akt signaling pathway and other pathways significantly influence bone metabolism. Simultaneous activation of the PI3K/Akt pathway alongside inhibition of MAPK/p38 and JNK pathways correlates with markedly increased bone mass and suppressed osteoclast apoptosis39,40. These findings reveal the mutual regulatory relationship between the PI3K/Akt and MAPK pathways and their synergistic effects in maintaining bone homeostasis.

The DEX-induced osteoporosis model established in zebrafish serves as a valuable tool for screening compounds that affect bone metabolism41,42. In this study, the treatment with Chinese yam extract significantly improved the bone microstructure in zebrafish. However, the limitations of zebrafish as a model for osteoporosis research must be acknowledged. Despite high conservation in skeletal development with humans, zebrafish cannot accurately predict human-related gene functions and lack a fully developed adaptive immune system, thereby failing to fully recapitulate the complexity of human osteoporosis mechanisms43. Nonetheless, zebrafish remain a valuable preclinical model for initial compound screening and hypothesis generation for further validation in mammalian systems.

To further investigate the anti-osteoporotic mechanism of Chinese yam, we utilized the DEX-treated MC3T3-E1 cell model44. Chinese yam extract significantly upregulated the p-PI3K/PI3K ratio, indicating activation of the PI3K/Akt pathway45. RUNX2 is a core member of the RUNX transcription factor family, playing a pivotal role in osteoblast differentiation and skeletal development. It directly regulates the expression of osteogenic differentiation marker genes, such as ALP and OCN, making it an essential transcription factor for osteoblast maturation17. It has indicated that the PI3K/AKT signaling pathway plays a significant role in regulating RUNX2 activity46. We hypothesized that Chinese yam extract may enhance RUNX2-mediated ALP and OCN expression by activating the PI3K signaling cascade, thereby promoting osteogenic differentiation. To validate this hypothesis, MC3T3-E1 cells were treated with the PI3K inhibitor LY294002. Results demonstrated that LY294002 treatment effectively reversed Chinese yam-induced increases the increases in ALP activity and the expression of RUNX2 and OCN, supporting the involvement of the PI3K/RUNX2 axis. Collectively, these biological findings suggest that Chinese yam may counteract osteoporosis by inhibiting bone resorption and promoting bone formation through the PI3K/Akt pathway.

However, this study has several limitations. First, the network pharmacology and molecular docking analysis are inherently database-driven predictions. Future studies should further identify the bioactive constituents in Chinese yam extract to improve the accuracy and comprehensiveness of the predictions. Second, regarding experimental validation, neither the zebrafish model nor the MC3T3-E1 cell model can fully replicate the complex in vivo bone remodeling microenvironment, which involves interactions among osteoblasts, osteoclasts, and immune cells. Furthermore, the study did not conduct in vivo efficacy validation in classical mammalian osteoporosis models, which is a critical intermediate step toward clinical translation. Finally, the relationship between the doses of the extracts and compounds used in the experiments and actual dietary intake or therapeutic doses in humans, as well as the assessment of long-term safety, requires further investigation. These limitations point the way toward subsequent studies that are more in-depth and closer to physiological and clinical realities.

Conclusions

In summary, this study reveals the multifactorial and multitarget mechanisms of Chinese yam in the treatment of osteoporosis. Through the application of network pharmacology and molecular docking techniques, it was predicted that the active components of Chinese yam primarily exert their effects by targeting EGFR, SRC, IL6, AKT1, and CTNNB1. Molecular dynamics simulations further validated the stable binding of Chinese yam’s active components to these targets. Zebrafish experiments confirmed that Chinese yam effectively alleviated the inhibitory effects of DEX on bone formation in zebrafish larvae. Cell experiments demonstrated that Chinese yam can mitigate the inhibitory effects of DEX on osteogenic differentiation by activating the PI3K/Akt signaling pathway, providing a theoretical foundation and research data for the clinical application of this drug in the treatment of osteoporosis.