Bilobalide attenuates steroid-induced osteonecrosis of the femoral head by upregulating the ERK/HIF-1α signaling pathway and promoting angiogenesis-osteogenesis coupling

Abstract

Steroid-induced osteonecrosis of the femoral head (SONFH) is a severe bone disease associated with long-term glucocorticoid use, characterized by impaired bone metabolism and vascular insufficiency. Bilobalide (BB), a natural sesquiterpene from Ginkgo biloba, exhibits anti-apoptotic, antioxidant, and pro-angiogenic properties, yet its role in SONFH remains unclear. We integrated network pharmacology and molecular docking to predict the targets and pathways of BB in SONFH. Key targets were validated using molecular docking software. For in vivo experiments, a rat SONFH model was established using methylprednisolone (MPS), and BB was administered orally. Micro-CT, H&E staining, TUNEL assay, and immunohistochemistry were employed to evaluate bone microstructure, apoptosis, and the expression of osteogenic and angiogenic markers. Immunofluorescence was used to assess HIF-1α expression in rat femoral head tissues. For in vitro experiments, MC3T3-E1 osteoblasts were treated with dexamethasone(DEX) and BB. Cell viability was detected using the CCK-8 assay, and the protein levels of the HIF-1α and ERK pathways were examined by Western blot. Network pharmacology identified 94 common targets between BB and SONFH, with enrichment in HIF-1 and ERK signaling pathways. Molecular docking confirmed strong binding affinities between BB and core targets. In MPS-induced rats, BB treatment significantly improved bone mineral density, trabecular microstructure, and reduced osteocyte apoptosis. BB also upregulated HIF-1α, Runx2, OCN, CD31, and VEGF expression, indicating enhanced osteogenesis and angiogenesis. In vitro, BB rescued dexamethasone-induced suppression of osteoblast viability and upregulated the ERK/HIF-1α pathway. Bilobalide attenuates SONFH progression by activating the ERK/HIF-1α signaling pathway, promoting osteogenesis and angiogenesis, and reducing osteocyte apoptosis. These findings highlight BB as a promising candidate for SONFH prevention and support the utility of network pharmacology in mechanistic natural product research.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-37128-3.

Introduction

Steroid-induced osteonecrosis of the femoral head (SONFH) is a severe osteoarticular disorder triggered by long-term or high-dose glucocorticoid use, accounting for a substantial proportion of non-traumatic ONFH cases and being particularly prevalent among patients undergoing prolonged steroid therapy¹. To date, the underlying mechanisms of glucocorticoid-induced ONFH remain incompletely elucidated, primarily involving dysregulated bone metabolism, disrupted lipid homeostasis, and impaired blood microcirculation within the femoral head^2,3. Although surgical interventions such as total hip arthroplasty are available for advanced-stage ONFH, concerns regarding postoperative complications and the high revision rates associated with prosthetic longevity pose significant challenges, particularly for young and middle-aged patients⁴. Current clinical management of SONFH primarily focuses on symptom palliation and delaying disease progression, rendering the identification of effective early intervention strategies and molecular targets a pivotal direction in fundamental and translational SONFH research. A critical objective in clinical management is the timely early-stage identification and intervention to decelerate disease progression and enhance patients’ quality of life.

In recent years, natural products have demonstrated considerable promise for the prevention and treatment of osteonecrosis, owing to their multifaceted pharmacological properties encompassing anti-inflammatory, antioxidant, and anti-apoptotic activities^5,6. Bilobalide (BB), a characteristic sesquiterpene lactone constituent of Ginkgo biloba extracts, has been documented to possess diverse biological activities, including anti-apoptotic, pro-angiogenic, and oxidative stress-modulating effects⁷. Notably, the distinctive bicyclic C8 skeleton of BB suggests its potential for modulating hypoxic stress responses⁸. Research shows that BB protects human melanocytes from oxidative damage caused by hydrogen peroxide by enhancing antioxidant enzyme activity and reducing endoplasmic reticulum stress⁹. Hypoxia-inducible factor-1alpha (HIF-1α) serves as a primary transcriptional regulator for cellular adaptation to low oxygen conditions, playing a vital role in sustaining the metabolic equilibrium of osteocytes and facilitating both angiogenesis and bone formation¹⁰. Additionally, BB has been shown to prevent apoptosis and encourage angiogenesis by activating the PI3K/Akt signaling pathway¹¹. In the context of bone metabolism, BB was shown to influence osteoclastogenesis through modulation of the SIRT3/NF-κB axis¹². However, its role and underlying mechanisms in SONFH have not been systematically investigated.

Network pharmacology, conceptualized in 2007, has become essential for understanding drug mechanisms of action¹³. By integrating systems biology and multi-dimensional databases, network pharmacology enables the comprehensive dissection of potential drug targets, key signaling pathways, and their associative networks with diseases, thereby providing a theoretical foundation for drug discovery and mechanistic studies^6,14. This study employed a network pharmacology approach, integrating multiple databases including TCMSP, SwissTargetPrediction, GeneCards, and DisGeNET, to identify and construct an interaction network of BB-SONFH targets, and performed KEGG pathway enrichment analysis to uncover the potential core pathways involved. The binding affinity and interaction mode between BB and the key target proteins were evaluated through molecular docking simulations using Autodock Vina. Subsequent in vitro and in vivo experiments were performed to validate the predicted mechanisms.

To substantiate the biological relevance of the predictions, the protective mechanism of BB against SONFH was further investigated using an animal model via in vivo experiments. The investigation focused on its efficacy in modulating the HIF-1α pathway and preserving osteocyte function. This research seeks to clarify the possible mechanisms by which BB attenuates SONFH through comprehensive analysis and experimental validation, offering a basis for its future use as a preventive strategy.

Materials and methods

Network Pharmacology analysis

Targets associated with BB and SONFH were identified

BB-related targets were collected from the SuperPRED, HERB, STITCH, CTD, and TargetNET databases, and supplemented by retrieving published reports. Targets associated with osteonecrosis pathways and SONFH were obtained from the DrugBank, DisGeNET, GeneCards, and OMIM databases. The UniProt database was utilized to standardize target protein nomenclature and eliminate redundancies across platforms and databases. Targets associated with steroid-induced osteonecrosis of the femoral head were obtained from the GeneCards and PharmGKB databases. The searches utilized the keywords “Steroid-induced osteonecrosis of the femoral head”, “femoral head steroid necrosis”, and “SONFH“¹⁵. The integrated sources combine various data dimensions connecting genes to diseases, guaranteeing that the retrieved target data is reliable and comprehensive. Targets without UniProt IDs were excluded, and disease-associated targets were consolidated and deduplicated to create the final disease target set. The intersection between the compound (BB) targets and the disease (SONFH) targets was identified, and a Venn diagram was constructed using the Venny 2.1 platform to visualize the common targets.

Constructing protein-protein interaction network

The STRING online platform was utilized to analyze overlapping targets, identifying common targets of BB and SONFH. The study focused on ‘Homo sapiens’ and applied a confidence score threshold of 0.400 to ensure high-quality protein-protein interaction data, identifying interactions among the target proteins. The network generated by STRING was exported and visualized using Cytoscape 3.10.1¹⁵. Targets were prioritized according to their degree centrality values, creating a list of key targets that may mediate BB’s effects on SONFH.

KEGG and GO enrichment analysis

The shared targets of the compound and disease were analyzed for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment using the DAVID database. GO analysis was divided into Biological Process (BP), Cellular Component (CC), and Molecular Function (MF) to annotate the roles of the drug’s target proteins in gene function related to the disease. KEGG analysis primarily involved pathway analysis, aiming to elucidate the major signaling pathways through which the drug exerts its protective effects^16–18.

Molecular Docking analysis

The small molecule ligand (BB) was sourced from the PubChem database, converted to a 3D structure using Chem Office 20.0, and saved in.mol2 format. Protein target crystal structures were obtained from the RCSB PDB database for use as molecular docking receptors. Protein structures were prepared by eliminating water molecules, small ligands, and heteroatoms using PyMOL 2.6.0.MOE 2019 software was employed for energy minimization of the compound, preprocessing of the target proteins, and identification of active binding pockets. Molecular docking was performed using MOE 2019, configured for 50 runs. This was succeeded by a post-docking analysis to assess the binding free energy. Docking results were analyzed and visualized using PyMOL and Discovery Studio Visualizer (2019).

Animals and grouping

All experimental procedures complied with the Guide for the Care and Use of Laboratory Animals and were approved by the Animal Ethics Committee of Shaanxi University of Chinese Medicine (Approval number SUCMDL20241124001). Thirty-two male Sprague-Dawley rats, aged 6 weeks and weighing 190–220 g, were sourced from Xi’an Jiaotong University’s Animal Center and randomly assigned to three groups. One group functioned as the control, while the other two groups were used to create a glucocorticoid-induced SONFH model. The groups were designated as the model group (MPS) and the BB treatment group (MPS + BB).

Model setup and Pharmacological intervention

The glucocorticoid-induced ONFH model was established as follows. Rats were kept under consistent, specific pathogen-free (SPF) conditions (23 ± 2 °C, 12-hour light/dark cycle) with ad libitum access to food and water, enabling unrestricted movement. Rats were weighed before administering the drug. Lipopolysaccharide (LPS, 20 µg/kg, Sigma) was administered intraperitoneally each day for the first three days. LPS was administered to induce a systemic inflammatory and hypercoagulable state, thereby enhancing the susceptibility of bone tissue to corticosteroids and increasing the incidence of osteonecrosis¹⁹. Methylprednisolone (60 mg/kg, Pfizer, USA) was given intramuscularly once daily for four days. MPS injections were alternated between the left and right gluteal muscles. Penicillin sodium (80,000 U) was administered intraperitoneally twice weekly to prevent infection. Rats in the BB treatment group received BB (MCE, China) via oral gavage at a dose of 20 mg/kg. This dosage was selected based on a previous study²⁰, which demonstrated that BB effectively protects against ischemic injury and rescues mitochondrial aerobic respiration in rats. At the 5-week endpoint, euthanasia was performed via an intraperitoneal overdose of Pentobarbital Sodium (150 mg/kg). Sample collection proceeded only after the absence of respiration and heartbeat was verified.

Micro-CT analysis

Bone mineral density, bone mass, and trabecular microstructure in the femoral head were evaluated using a high-resolution micro-CT system (Bruker AXS GmbH, Germany). The parameters measured were trabecular bone mineral density (Tb.BMD), trabecular number (Tb.N, 1/mm), thickness (Tb.Th, mm), separation (Tb.Sp, mm), and bone volume fraction (BV/TV, %). The scanning process was executed under specified parameters: a voltage of 60 kV, a current of 400 µA, a full 360-degree rotation in 2-degree angular increments, and an acquisition duration ranging from 20 to 30 min for each specimen. The entire femur was scanned at a 12 μm resolution and subsequently reconstructed in three dimensions. A specified area within the distal femoral epiphysis was captured across a minimum of 100 slices. The two-dimensional images were subsequently assembled into three-dimensional reconstructions utilizing MicroView V.2.1 (GE Healthcare, USA) software.

Histological and H&E staining

Femoral head samples were fixed in 4% paraformaldehyde for 48 h and then decalcified in 10% ethylenediaminetetraacetic acid (EDTA, Servicebio, China) for 4 weeks. The samples were embedded in paraffin and sliced into 6 μm sections. Hematoxylin and eosin (H&E) staining was conducted to assess general morphology and trabecular bone structure. Assessments were conducted by two independent observers.

IHC staining, TUNEL staining, and IF staining

Sections were deparaffinized and rehydrated with a graded alcohol series to aid in antigen retrieval. Following this, sections underwent incubation with primary antibodies, followed by the application of corresponding secondary antibodies. The results were evaluated by two independent observers. Chromogenic reaction was developed using a DAPI Kit (Beyotime, China). Cellular apoptosis was detected and quantified using a TUNEL kit (Roche, Switzerland), following the manufacturer’s protocol for labeling DNA strand breaks. Sections were permeabilized before the application of reagents. A 10-minute nuclear counterstain using DAPI was performed in a dark environment. Subsequently, the sections were coverslipped with an anti-fade medium. The fluorescence images showed TUNEL-positive nuclei, identified by bright green fluorescence, co-localizing with blue DAPI-stained nuclei. Immunofluorescence (IF) staining was also employed to elucidate the expression levels and localization of specific targets. Fluorescent secondary antibodies were employed alongside target-specific primary antibodies. Tissue sections were observed under a fluorescence microscope. Fluorescence-labeled positive cells were subsequently quantified with ImageJ software (NIH, USA). For each experiment, cells in at least five random fields were counted, and three independent experiments were performed.

Cell culture and viability assessments

MC3T3-E1 mouse pre-osteoblastic cells (Shanghai ZhongQiaoXinZhou Biotechnology, China) were cultured in MEM-α medium with 10% fetal bovine serum. Cells were cultured at 37 °C with 5% CO2 in a humidified incubator. The viability of MC3T3-E1 cells treated with dexamethasone (DEX, Beyotime, China) was evaluated using a Cell Counting Kit-8 assay.MC3T3-E1 cells were seeded into 96-well plates. Cells at 80% confluence were exposed to 100 µM DEX for 24 h to induce cytotoxicity. This concentration was selected based on previously established protocols²¹, which confirmed that 100 µM DEX effectively mimics glucocorticoid-induced osteoblast injury in vitro. Subsequently, cells were treated with different concentrations of BB for an additional 24 h. The concentration of 10 µM BB was determined based on our preliminary dose-response screening, which demonstrated optimal cytoprotective effects without cytotoxicity. The control group was administered an equal volume of DMSO without the addition of DEX or BB. Cell viability was assessed with the Cell Counting Kit-8 (CCK-8, Beyotime, China) following the manufacturer’s guidelines. In summary, each well received 10 µL of CCK-8 solution and was incubated at 37 °C for 2 h. Absorbance at 450 nm was recorded using a microplate reader.

Western blot analysis

Protein expression levels in osteoblasts were evaluated using western blotting techniques. Protein extraction utilized RIPA lysis buffer (Beyotime, China), enhanced with protease and phosphatase inhibitors (1 mM PMSF, Beyotime, China). Protein concentrations were measured using a BCA assay kit (Beyotime, China) according to the manufacturer’s instructions. After that, equal volumes of the samples underwent SDS-PAGE (Beyotime, China) and were transferred to PVDF membranes (Beyotime, China) via electrophoresis. The membranes were subsequently blocked using a 5% non-fat milk solution in TBSTS (Servicebio, China). They were incubated overnight at 4 °C with specific primary antibodies. Subsequently, the membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. After conducting comprehensive washes, the immunoreactive bands were identified utilizing an enhanced chemiluminescence (ECL) system. GAPDH functioned as the internal loading control for normalization purposes, and band intensities were quantified utilizing ImageJ software. Detailed information regarding the antibody sources, catalog numbers, and dilution factors is listed in Supplementary Table S1.

Statistical analysis

All experiments were performed with at least three independent replicates. Data are expressed as mean ± standard deviation (SD) and analyzed with GraphPad Prism 8.0 software. A one-way ANOVA was used to assess statistical significance. A p-value below 0.05 was deemed statistically significant.

Results

Network Pharmacology analysis

BB’s structural data were evaluated across five databases. Following the standardization of target names and removal of duplicates via the UniProt database, 101 unique potential targets were identified. Disease-associated targets for SONFH were obtained from the GeneCards and PharmGKB databases, resulting in 2184 unique targets after removing duplicates. Using the Venny 2.1 online tool, a Venn diagram revealed 94 common targets shared by both the drug and the disease (Fig. 1a). A network representing the “disease-target-component” was constructed using Cytoscape software. This network includes BB as a green node, the common targets as blue nodes, and the disease as a yellow node (Fig. 1b). The 94 common targets were input into the STRING database to create a protein-protein interaction (PPI) network, using “Homo sapiens” as the organism and a minimum interaction score of 0.4. Disconnected nodes have been removed from the network. The PPI network was visualized and analyzed topologically using Cytoscape 3.10.1, where node size and color intensity indicated degree centrality. The Network Analyzer tool was utilized to identify core targets, defined as those exhibiting degree centrality values exceeding the average (Fig. 1c). Gene Ontology (GO) analysis was performed on the 94 common targets, covering the main categories of Biological Process (BP), Cellular Component (CC), and Molecular Function (MF). The Gene Ontology analysis revealed substantial enrichment of the intersecting genes across 1812 Biological Process terms, 67 Cellular Component terms, and 152 Molecular Function terms. The top eight significantly enriched terms from each category were visualized using bar graphs and bubble charts (Fig. 1d). KEGG pathway enrichment analysis identified 162 pathways linked to the 94 targets. The top 20 pathways, organized by p-value, are presented in Fig. 1e. These enrichment analyses provided a critical “roadmap” for our subsequent validation. Specifically, the significant enrichment of the “HIF-1 signaling pathway” and “ERK1/ERK2 cascade” indicated that these pathways might be the key mechanisms by which BB attenuates SONFH. This prediction directly guided our decision to validate the ERK/HIF-1α axis in the subsequent molecular docking and biological experiments.

Fig. 1.

Network pharmacology analysis of BB and SONFH. (a) Venn diagram of BB and SONFH-related targets, comprising 101 BB-related targets (left), 2184 SONFH-related targets (right), and 94 common targets (intersection). (b) Disease-target-component network: The green node indicates BB, the blue nodes denote common targets, and the yellow node signifies SONFH. (c) PPI network of common targets (left). Node size and color intensity (darker blue indicates a higher degree) represent connectivity; core targets are highlighted in the inner circle. Bar graph of core targets generated using R software (right). (d) GO enrichment analysis. (e) KEGG pathway enrichment analysis. The KEGG pathway map was reproduced with permission from Kanehisa Laboratories.

Molecular docking

Molecular docking was conducted to explore the protective mechanism of BB against SONFH by analyzing the binding modes and affinities between BB and key SONFH-related targets. The candidate targets for docking were selected based on a combined strategy of topological importance and pathway relevance. First, we identified core targets with high degree centrality from the PPI network (Fig. 1c). Second, guided by the KEGG enrichment results (Fig. 1e), we narrowed down the list to key targets specifically enriched in the HIF-1 signaling pathway and its associated MAPK cascade. A docking energy value lower than 5 kcal/mol is generally considered to indicate good binding activity. As shown in Table 1, the binding energies of BB with four target proteins were all below − 5 kcal/mol, suggesting strong binding capabilities. Further analysis of the binding modes revealed distinct interaction patterns between BB and different target proteins (Table 1; Fig. 2): with TNF, interactions primarily involved hydrogen bonds; with ESR1, hydrophobic interactions dominated; with HIF1A, both hydrogen bonds and hydrophobic interactions were present; with MAPK3, hydrogen bonds, hydrophobic interactions, and carbon-hydrogen bonds were observed. In the docking models (Fig. 2), the BB molecule is represented as green sticks, and the binding residues of the target proteins are shown in blue. These results reveal, at an atomic level, how BB stably binds to target proteins through diverse non-covalent interactions, providing a structural basis for its mechanism against SONFH.

Table 1.

Docking results of SONFH with core target molecules.

Targets	Binding Energy(kJ/mol)	Hydrogen Bonds	Hydrophobic Interactions	Carbon-Hydrogen Interaction
TNF	−6.9136	Cys101、Gln102、Ser99、Lys98
IL-1β	−4.9973	Thr144	Tyr121	Thr144、Glu111
ESR-1	−5.4133		Leu428、Met388、Ile424、Met421、His524、Leu525、Phe404
PTGS2	−4.9152	Thr212、His386、His214、His388	His207、Val447、His388	Lys211
MAPK3	−5.0122	Asn171、Tyr53、Glu50	Tyr53	Ser170、Gly49
FOS	−4.3746	Asn147	Arg144、Lys148、Met138
HIF-1α	−5.3056	Lys191、His230、 Asn233、Thr189	His230、Lys191

Fig. 2.

Molecular docking models of BB with potential protective targets for SONFH. (a) HIF-1α, (b) TNF, (c) IL-1β, (d) ESR-1, (e) PTGS2, (f) MAPK3, (g) FOS.

BB alleviates MPS-induced SONFH in rats

The protective effect of BB on MPS-induced ONFH was evident in rat femoral heads. Micro-CT scanning, performed to assess morphological changes, provided visualization of both cortical and trabecular bone. While the control group displayed femoral heads with smooth, uniformly colored cartilage surfaces, the model group displayed significant surface irregularities and atypical coloration in the weight-bearing region. The administration of BB treatment, especially at elevated dosages, resulted in the reduction of these lesions (Fig. 3a). Three-dimensional reconstructions from sagittal, coronal, and transverse views confirmed the protective role, showing severe trabecular bone loss and structural collapse in the MPS group, which was significantly mitigated by BB administration. Quantitative analysis of a weight-bearing region of interest (ROI) revealed that the MPS group had a significantly lower average BMD than the normal group.BB treatment mitigated this effect, enhancing bone microarchitecture parameters such as BMD, BV/TV, Tb.N, and Tb.Th (Fig. 3b, c, e, f), while notably reducing BS/BV and Tb.Sp compared to the model group (Fig. 3d, g).

Fig. 3.

BB reduced necrosis of the femoral head in a rat model. (a) Representative micro-CT images. (b–g) Quantitative assessment: BMD, BV/TV, BS/BV, Tb.N, Tb.Th and Tb.Sp. MPS: Methylprednisolone; BB: Bilobalide; BMD: Bone Mineral Density (g/cm³); BV/TV: Bone Volume/Tissue Volume (%); BS/BV: Bone Surface/Bone Volume (1/mm); Tb.N: Trabecular Number (1/mm); Tb.Th: Trabecular Thickness (mm); Tb.Sp: Trabecular Separation (mm). (n = 5 per group, data presented as mean ± SD, significance levels: ***p < 0.001, **p < 0.01, *p < 0.05).

Histological analysis further confirmed that BB alleviated MPS-induced osteocyte impairment. In H&E staining, the MPS + PBS group exhibited sparse and collapsed trabeculae along with a higher number of empty lacunae. In contrast, the BB group showed minimal bone tissue necrosis, limited trabecular replacement by necrotic tissue, and a significantly lower percentage of unfilled lacunae. The H&E staining results aligned with the micro-CT findings, demonstrating a protective effect of BB against MPS-induced trabecular bone loss in the femoral head (Fig. 4a, b). Decalcification generally leads to a low incidence of empty lacunae, with normal bone tissue showing less than 10% empty lacunae, while necrotic tissue can exhibit an incidence of 40–80%. TUNEL staining revealed that BB reduced osteocyte apoptosis (Fig. 4c, d).

Fig. 4.

Analysis of BB-mediated attenuation of SONFH severity in rats. (a, b) Representative H&E-stained images and quantitative analysis of the empty lacunae rate from different indicated treatment groups. (c, d) TUNEL staining and quantitative analysis of apoptotic cells in the femoral head across different groups. (n = 3 per group, data are shown as mean ± SD, ***p < 0.001, **p < 0.01, *p < 0.05).

BB promotes osteogenesis and angiogenesis to counteract SONFH progression in rats

Immunofluorescence (IF) labeling was employed to detect the expression of markers associated with osteogenesis and angiogenesis. The results showed that compared to the model group, the BB treatment group exhibited significantly enhanced fluorescence intensity for the key osteogenic transcription factor Runx2 (Fig. 5a, b) and the vascular endothelial marker CD31 (Fig. 5e, f). This finding indicates that BB not only promotes angiogenesis but also enhances osteogenic differentiation, exerting a dual promotive effect on the attenuation of the necrotic femoral head area. Immunohistochemistry (IHC) results further corroborated these findings. The positive expression areas for Osteocalcin (OCN) (Fig. 5c, d) and Vascular Endothelial Growth Factor (VEGF) (Fig. 5g, h) were significantly larger in the BB treatment group compared to the model group. VEGF is a key downstream target gene of HIF-1α and plays a central role in angiogenesis²², whereas OCN is a marker of late osteoblast differentiation involved in bone matrix mineralization²³. The upregulation of both expressions suggests that BB concurrently promotes the angiogenesis-osteogenesis coupling process within the femoral head, a key mechanism for the attenuation and repair of necrotic bone tissue.

Fig. 5.

BB promotes osteogenesis and angiogenesis to alleviate SONFH progression in rats. (a, b) IF staining and quantitative analysis of Runx2 in the femoral head across different groups. (c, d) Immunohistochemical analysis and quantification of OCN in the femoral head. (e, f) IF staining and quantitative analysis of CD31 in the femoral head across different groups. (g, h) Immunohistochemical analysis and quantification of VEGF in the femoral head. (n = 3 per group, data are shown as mean ± SD, ***p < 0.001, **p < 0.01, *p < 0.05).

BB stimulates the ERK/HIF-1α signaling pathway in the management of SONFH

To further investigate the molecular mechanism of BB’s action, IF staining was performed on the rat samples, showing that the BB treatment group had significantly increased HIF-1α fluorescence intensity compared to the model group (Fig. 6b, c). Subsequently, in vitro experiments were conducted using MC3T3-E1 osteoblastic cells. CCK-8 assay results showed that DEX significantly inhibited osteoblast viability, and BB pretreatment markedly alleviated this damaging effect, indicating a protective role of BB against glucocorticoid-induced injury in osteoblasts (Fig. 6a). Western blot analysis (Fig. 6d) revealed that dexamethasone treatment decreased the protein expression levels of HIF-1α and phosphorylated ERK (p-ERK) in osteoblasts (Fig. 6e, g), while the level of total ERK protein remained largely unchanged (Fig. 6f). BB intervention reversed this trend, significantly restoring the expression levels of HIF-1α and p-ERK. This suggests that BB mitigates the inhibitory effects of dexamethasone on osteoblasts by activating the ERK/HIF-1α signaling pathway. It is noteworthy that the ERK signaling pathway is known to be involved in regulating cell proliferation, differentiation, and survival²⁴. Our findings suggest a potential crosstalk between the ERK and HIF-1α pathways, jointly regulating osteoblast function and angiogenesis.

Fig. 6.

BB upregulates the ERK/HIF-1α signaling pathway. (a) Cell viability was assessed by the CCK-8 assay in MC3T3-E1 cells. (b, c) IF staining and quantitative analysis of HIF-1α in the femoral head from different rat groups. (d) Representative western blot bands. (e-g) Relative protein expression levels of HIF-1α, ERK, and p-ERK. (n = 3 per group, data are shown as mean ± SD, ***p < 0.001, **p < 0.01, *p < 0.05).

Discussion

This study comprehensively explored the protective potential of BB for SONFH through a combination of network pharmacology, molecular docking, and both in vivo and in vitro experiments. Our research demonstrates that BB alleviates glucocorticoid-induced osteocyte apoptosis and microstructural damage, while promoting osteogenesis and angiogenesis coupling, primarily through HIF-1α signaling pathway activation. These findings furnish both theoretical and empirical data supporting BB as a prospective protective agent against SONFH.

Analysis of network pharmacology revealed that the potential targets of BB are concentrated in pathways associated with oxidative stress, mitochondrial function, and angiogenesis. Molecular docking revealed that BB exhibits strong binding affinity for targets including TNF, IL1B, ESR1, PTGS2, MAPK3, and HIF1A, indicating its potential to influence SONFH progression through multi-target synergistic effects. Among these, HIF1A, a master transcriptional regulator of the hypoxic response, has established regulatory relationships with several other targets^25,26. This implies that the HIF-1α pathway may serve as a pivotal hub mediating the protective effects of BB. Extended glucocorticoid use impairs femoral head microcirculation and causes local hypoxia, leading to osteocyte apoptosis and bone tissue damage. HIF-1α, a crucial transcription factor in the hypoxic response, is pivotal in SONFH pathogenesis^27,28.

In a rat SONFH model induced by MPS, BB treatment was shown to significantly mitigate the progression of femoral head necrosis. H&E staining showed that BB reduced the empty lacunae rate and osteocyte apoptosis index, and improved bone microstructure. TUNEL staining further confirmed that BB significantly decreased osteoblast apoptosis, supporting its role in maintaining osteocyte homeostasis. These histological improvements were supported by molecular evidence; immunofluorescence and immunohistochemistry results showed that BB treatment significantly increased HIF-1α protein levels, suggesting that BB might exert its protective effects by inhibiting HIF-1α degradation or promoting its stabilization. HIF-1α is a key regulator in cellular adaptation to hypoxia, playing a significant role in skeletal angiogenesis and osteogenesis²⁹. Research indicates that the HIF-1α pathway significantly contributes to angiogenesis and the differentiation of bone marrow mesenchymal stem cells³⁰. These processes are essential for bone regeneration and may be beneficial in the context of SONFH.

The repair process of femoral head necrosis requires precise coordination between angiogenesis and osteogenesis, a process termed “angiogenesis-osteogenesis coupling“³¹. Our immunofluorescence and immunohistochemistry findings indicated that BB treatment enhanced HIF-1α expression, alongside elevated levels of angiogenic markers CD31 and VEGF, as well as osteogenic markers Runx2 and OCN.BB potentially enhances angiogenesis and bone formation in the femoral head simultaneously through the activation of the HIF-1α signaling pathway. This coupling effect is crucial for the repair of osteonecrosis. Vascular invasion is a prerequisite for the coupling of angiogenesis and osteogenesis; nascent blood vessels not only serve as conduits for bone tissue metabolism and nutrient supply but also provide a homeostatic microenvironment for osteoprogenitor cell self-renewal and a niche source for hematopoietic stem cell differentiation in the bone marrow^32,33. The H-type vessel, a distinct vascular subtype, is crucial for linking osteogenesis and angiogenesis. H-type vessels are inherent osteogenic healing structures in bone³⁴. In bone defects, impaired vasculature induces H-type vessel formation through the HIF-1α signaling pathway³⁵. Immunofluorescence staining of rat tissue sections showed a notable reduction in HIF-1α expression levels in the BB treatment group relative to the model group.

Furthermore, in our MPS-induced osteoblast model, dexamethasone significantly inhibited the viability of MC3T3-E1 osteoblasts, and Bilobalide effectively reversed this effect. Concurrently, Western Blot results indicated that dexamethasone suppressed the protein expression of HIF-1α and the p-ERK/ERK ratio. This suggests that the protective effects of Bilobalide are associated with the upregulation of the HIF-1α and ERK signaling pathways. This indicates that Bilobalide may enhance bone repair capacity by promoting osteoblast proliferation and differentiation, a process associated with the upregulation of HIF-1α and ERK signaling. The ERK pathway has been extensively demonstrated to be involved in regulating cell proliferation, differentiation, and apoptosis^24,33. HIF-1α acts as a bridging factor in ERK1/2-dependent pathways for long-term hypoxia-induced apoptosis²⁵. A positive feedback loop is present between ERK and HIF-1α: ERK enhances HIF-1α transcriptional activity through phosphorylation, which in turn promotes the expression of downstream target genes such as VEGF^35,36. Our findings suggest that Bilobalide may concurrently promote osteoblast survival and functional activity by concurrently enhancing HIF-1α stability and ERK phosphorylation. From a protective strategy perspective, targeting the HIF-1α signaling pathway has emerged as a promising direction for preventing or managing femoral head necrosis. For instance, PHD inhibitors can promote angiogenesis and bone formation by stabilizing HIF-1α³⁷. Our study indicates that BB may function as a natural activator of the ERK/HIF-1α pathway with analogous effects.

Although these findings are encouraging, certain limitations remain. A limitation of our study is that while we observed upregulated HIF-1α and p-ERK levels, we did not employ specific inhibitors or gene knockout models to definitively prove that the anti-osteonecrotic effect is solely dependent on this pathway. However, the consistency between our network pharmacology predictions and experimental phenotypes strongly supports this mechanism. Future studies should employ specific inhibitors or genetic silencing strategies to definitively validate the necessity of the HIF-1α pathway. Simultaneously, other potential pathways involved warrant further exploration. Moreover, this study focused on osteoblast protection and angiogenesis; the potential impact of BB on osteoclast activity, a key component of bone remodeling in SONFH, was not examined. Future investigations incorporating assessments such as tartrate-resistant acid phosphatase staining would provide a more comprehensive view of bone turnover. The pharmacokinetic profile, bioavailability, and long-term safety of BB in bone tissue also require further investigation. Future research should focus on elucidating the multi-pathway mechanisms of BB, optimizing its dosage, and conducting pre-clinical and clinical evaluations to advance its protective application. Nevertheless, in this study, we adopted a research paradigm integrating multi-omics strategies with experimental validation: initial prediction of BB’s potential targets and pathways via network pharmacology, verification of its binding activity through molecular docking, and finally, confirmation of its function and mechanism in animal and cell models. This comprehensive evidence chain, from prediction to validation, significantly enhances the reliability of our conclusions and provides a referable framework for the systematic investigation of the mechanisms of natural product active ingredients.

Conclusion

This study, which integrates network pharmacology, molecular docking, and in vivo experiments, demonstrates the potential of BB in attenuating SONFH by activating the ERK/HIF-1α pathway and improving bone microstructure. These findings not only provide a scientific basis for BB as a candidate protective agent but also validate the utility of network pharmacology in orthopedic drug discovery. Future studies are warranted to explore additional pathways and to confirm the efficacy and safety of BB in clinical settings.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, Qi Chen and Yangquan Hao; Data curation, Qi Chen and Yangquan Hao; Formal analysis, Qi Chen, Bo Wang, Hu Liang, and Hanbo Xu; Funding acquisition, Yangquan Hao; Investigation, Qi Chen, Bo Wang, Hanbo Xu, Hu Liang, and Kun Zhang; Methodology, Qi Chen; Project administration, Yangquan Hao; Resources, Yangquan Hao; Software, Qi Chen, Bo Wang and Yangquan Hao; Supervision, Kun Zhang and Yangquan Hao; Validation, Qi Chen, Bo Wang, Hu Liang and Hanbo Xu; Visualization, Qi Chen and Bo Wang; Writing – original draft, Qi Chen and Bo Wang; Writing – review & editing, Yangquan Hao. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by the Clinical Collaboration and Innovation Project of Integrated Traditional Chinese and Western Medicine by the Shaanxi Provincial Administration of Traditional Chinese Medicine (No.2020ZXY010) (to Yangquan Hao).

Data availability

The authors affirm that all data supporting the findings are completely accessible and can be acquired by submitting a request to the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.

Institutional review board statement

All animal experiments were conducted in accordance with the ARRIVE guidelines 2.0. The study received approval from the Ethics Committee of the Institutional Animal Care and Use Committee at Shaanxi University of Chinese Medicine (Approval number SUCMDL20241124001, approved on 24 November 2024).

Informed consent statement

The study did not involve humans.

Third-party material

This manuscript includes results and/or pathway maps generated from the KEGG database. Formal permission to publish this material under a CC BY license has been obtained from Kanehisa Laboratories.