Network Pharmacology and Rat Caudal-disc Puncture Validation of Yi Qi Huo Xue Decoction For Lumbar Disc Degeneration

Summary

This study used network pharmacology, molecular docking, and a rat caudal-disc puncture model to investigate the potential preclinical effects of Yi Qi Huo Xue Decoction on lumbar disc degeneration, with a focus on candidate targets, including AKT1, BCL2, and STAT3, and pathways such as PI3K-AKT signaling.

Abstract

In this study, network pharmacology, molecular docking, molecular dynamics simulation, and experimental validation were used to investigate the pharmacological basis and potential molecular mechanisms of Yi Qi Huo Xue Decoction (YQHXD) in lumbar disc degeneration (LDD). A total of 678 active compound-related targets of YQHXD and 296 potential LDD-associated therapeutic targets were identified. Protein-protein interaction network analysis and topological analysis identified several candidate targets, including AKT1, IL6, TNF, BCL2, and STAT3. Enrichment analyses using Gene Ontology and Kyoto Encyclopedia of Genes and Genomes indicated that the candidate targets were linked to functions including protein binding and ATP binding, with pathway enrichment that included PI3K-AKT signaling. Molecular docking predicted interactions between selected active compounds, such as quercetin, Tetraneurin A, and isorhamnetin, and the core targets. Molecular dynamics simulations further suggested that the lowest-energy complexes remained stable. Seventy male Sprague-Dawley rats were randomly assigned to the control, sham-operated, model, positive control, and low-, medium-, and high-dose YQHXD groups. After the establishment of a coccygeal intervertebral disc puncture model in the model, positive control, and YQHXD groups, rats were administered YQHXD or celecoxib by oral gavage for 8 weeks, whereas the control and sham-operated groups received normal saline. Histopathological evaluation showed that YQHXD treatment was associated with reduced intervertebral disc degeneration in this model. Enzyme-linked immunosorbent assay, polymerase chain reaction, and western blot analyses showed changes in the mRNA and protein expression levels of selected targets, including AKT1, BCL2, and STAT3. These findings suggest that YQHXD may exert preclinical protective effects in LDD through multi-component and multi-target regulation involving inflammation- and apoptosis-related mechanisms. This study provides preliminary preclinical evidence supporting further investigation of YQHXD for LDD.

Introduction

Lumbar disc degeneration (LDD) is a common age-related orthopedic condition characterized by late onset, difficulty of treatment, high recurrence, and disability. LDD involves degenerative pathological changes associated with an imbalance between catabolic and anabolic metabolism of the extracellular matrix (ECM) in the intervertebral disc (IVD) and is closely associated with low back pain (LBP)¹. Epidemiological studies show that LBP is one of the leading causes of disability in clinical practice².

From 1990 to 2017, the number of global cases increased from 42.5 million to 64.9 million³, affecting daily life and imposing a substantial social and economic burden. Therefore, effective treatment of LDD remains clinically important. In clinical practice, LDD is often managed with conservative or surgical symptomatic treatments, which can alleviate symptoms but generally do not reverse intervertebral disc degeneration (IVDD). Commonly used oral nonsteroidal anti-inflammatory drugs for conservative treatment may also be associated with renal, gastrointestinal, and cardiovascular adverse effects.

Traditional Chinese Medicine (TCM) has long been used in disease management⁴. Herbal compounds and individual constituents have been reported to delay IVDD progression and promote tissue regeneration⁵. In clinical practice, TCM formulas are also used as part of conservative management strategies for degenerative bone and joint diseases. The Wu family medical school in the Suzhou area has used Yi Qi Huo Xue Decoction (YQHXD) as a core prescription for lumbar IVD disorders; however, further mechanistic evidence is needed to clarify its potential role in LDD.

Previous studies have suggested that YQHXD may delay IVDD by promoting cellular autophagy and exerting anti-aging effects^6,7. However, the pharmacological mechanisms by which YQHXD may affect LDD remain unclear, and a systematic analysis of its mode of action is still lacking. In this study, network pharmacology was used to analyze the multi-component and multi-target characteristics of YQHXD in relation to LDD. Molecular docking was used to evaluate interactions between compounds and target proteins, followed by molecular dynamics simulations. These predictions were then assessed using a rat caudal-disc puncture model to investigate potential mechanisms of YQHXD in LDD⁸. As shown in Figure 1, this study provides preclinical evidence that YQHXD may affect LDD-related processes by modulating genes and proteins such as AKT1, BCL2, and STAT3, with pathway enrichment involving PI3K-Akt signaling and other disease-related pathway categories.

Protocol

All experiments were approved by the Institutional Review Board of Nanjing University of Chinese Medicine (Protocol Code: 202501A018) and were conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. All experimental procedures adhered to the ARRIVE guidelines.

Experimental drugs
The YQHXD sample solution contained 20 g of Astragalus membranaceus root, 20 g of Angelica sinensis root, 15 g of Ligusticum chuanxiong rhizome, 6 g of Stephania tetrandra root, 15 g of Chaenomeles speciosa fruit, and 10 g of Sinapis alba mature seeds, with a total crude drug weight of 86 g per prescription. All herbal materials were authenticated according to the Chinese Pharmacopoeia (2020 edition) by the Department of Pharmacy, Ruikang Hospital Affiliated to Guangxi University of Chinese Medicine. The YQHXD sample solution was prepared using a water decoction method. Sterile distilled water at 8x the total crude drug volume was added, and the herbs were soaked for 30 min until fully submerged. The mixture was then boiled and decocted over low heat for 30 min before filtration. The residues were extracted again with the same volume of sterile distilled water following the same procedure. The filtrates from the two extractions were combined and concentrated to 55.5 mL under reduced pressure using rotary evaporation to obtain the YQHXD stock solution. Based on the total crude drug weight, the final crude drug concentration was 1.55 g/mL, with an extraction yield of approximately 64.5%. The dosage administered to experimental rats was converted from the clinical adult dose using the body surface area normalization method. The conversion coefficient between humans and rats was approximately 6.3; therefore, the rat dosage (mg/kg) was calculated as the human dosage (mg/kg) × 6.3. The high-dose group received 15.48 g/kg/day as the reference dose, while the medium- and low-dose groups received 7.74 g/kg/day and 3.87 g/kg/day, respectively. For a gavage volume of 10 mL/kg, the corresponding working concentrations were 1.548, 0.774, and 0.387 g/mL for the high-, medium-, and low-dose groups, respectively. The stock solution was diluted accordingly to achieve the target dosage for each experimental group before administration. After preparation, the decoctions were cooled, aliquoted under sterile conditions, and stored at −40 °C until use.

Network pharmacology analysis

Prediction of active ingredient target proteins
On November 24, 2025, active compounds were retrieved from the TCMSP database using the search terms “Huangqi, Angelica, Chuanxiong, Stephania, Mugua, and Jiezi.” Compounds were screened based on oral bioavailability (OB ≥ 30%) and drug-likeness (DL ≥ 0.18)⁹. The rationale for these thresholds is as follows: the TCMSP database integrates compound data from authoritative sources such as PubChem, ChEMBL, and ChemSpider and provides ADME-related parameters, including oral bioavailability. Oral bioavailability refers to the proportion of an orally administered dose that reaches systemic circulation unchanged, and high OB is essential for identifying bioactive molecules. Drug-likeness is an indicator used to evaluate the druggability of compounds and to predict their solubility and stability. In this study, a more stringent OB threshold (≥30%), exceeding the database-recommended value (≥20%), was applied to identify candidate compounds with favorable oral absorption and pharmacokinetic profiles. Meanwhile, DL was used to estimate the drug-like potential of compounds, and a cutoff of DL ≥ 0.18 was adopted based on the average DL value observed in traditional Chinese medicine compounds. The selected active compounds were retrieved in SMILES format from the PubChem database and imported into the SwissTargetPrediction database for target prediction, with the species set to Homo sapiens. Target proteins were standardized using official gene symbols, and only targets with probability values greater than 0 were retained to improve prediction reliability. Compounds with no predicted targets were excluded from subsequent network analysis. After merging all targets, duplicate entries were removed using spreadsheet software to ensure that each target was represented only once in the final dataset.

Collection of key targets
Disease-related genes were retrieved from the GeneCards¹⁰, OMIM¹¹, and TTD¹² databases using the keyword “lumbar disc degeneration.” The collected gene targets were merged, and duplicate entries were removed to obtain a final set of disease-associated targets.

Potential targets of YQHXD for treating LDD
The Venny 2.1.0 platform was used to determine the potential targets of YQHXD in LDD. A network was then generated to show the associations between the active components and their predicted targets. Core active components were selected according to their degree values in the network.

Network construction of drug components and disease intersection targets
Active ingredients, target proteins, and LDD-related genes from the TCM formula YQHXD were imported into the Cytoscape software¹³ to construct a drug-component-target-disease network diagram.

Construction of the PPI network and identification of key core targets
Intersecting targets were submitted to the STRING platform¹⁴ for analysis, with the organism set to Homo sapiens. To build the protein-protein interaction (PPI) network, the interaction score threshold was set at >0.4, disconnected nodes were hidden, and all other settings were left as default. The resulting data were imported into the software, where the NetworkAnalyzer tool was used to evaluate topological properties and calculate node degrees. The top five nodes ranked by degree were identified as key hub targets.

Enrichment analysis of core targets
Intersecting targets were uploaded to the DAVID database¹⁵ for Gene Ontology (GO) functional annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. The organism background was set as Homo sapiens, with the enrichment cutoff defined according to the default DAVID parameters, including an EASE threshold < 0.1 and a Count threshold ≥ 2. The gene universe was based on the default whole-genome background provided by DAVID. Multiple testing was corrected using the Benjamini-Hochberg method to control the false discovery rate (FDR), and the DAVID results were adjusted on this basis. Enriched pathways were ordered by Count value in descending order. GO and KEGG enrichment bar plots were produced using an online bioinformatics platform.

Molecular docking
To predict interactions between the key active components of YQHXD and the core targets of LDD, molecular docking analysis was performed in this study. The top five key active compounds, ranked according to network topological degree values and target relevance, were selected as ligands, while the corresponding core target proteins were used as receptors. Ligand SDF files were retrieved from PubChem, and three-dimensional structures of the target proteins were downloaded from the Protein Data Bank (PDB). Docking analysis was performed on the CB-Dock2 platform using its fully automated blind docking protocol. After uploading the protein structures, the platform automatically removed heteroatoms, including small molecules, ions, and water molecules, and subsequently repaired missing atoms and hydrogen atoms. Auto Blind Docking mode was used. In this mode, CB-Dock2 identifies potential protein-surface binding pockets with a curvature-based cavity detection algorithm and then performs docking calculations using the integrated AutoDock Vina engine¹⁶. During the docking process, the proteins were treated as rigid structures, whereas the ligands were considered flexible, and all docking parameters were set to the platform default values. The docking results were subjected to three-dimensional visualization and interaction analysis using the built-in tools of CB-Dock2¹⁷, and binding energies were calculated to estimate the predicted binding affinities between the active compounds and target proteins.

Molecular dynamics simulation
Molecular dynamics simulations were carried out for the target protein-ligand complex using GROMACS 2022.2. The protein was modeled with the Amber14SB force field, and the solvent system was represented using the TIP3P water model. Ligand atomic charges were calculated using the AM1-BCC method implemented in Antechamber, followed by assignment of GAFF2 atom types. The resulting topology files were converted into a GROMACS-compatible format using ACPYPE. Ion parameters consistent with the TIP3P water model were assigned based on the Joung-Cheatham parameter set. The starting complex was placed in a truncated dodecahedral simulation box with at least 1.2 nm between the protein surface and the box edge. TIP3P water molecules were added to solvate the system, and Na⁺ and Cl⁻ ions were introduced to neutralize the system charge and set the ionic strength to 0.15 M. Energy minimization was performed with the steepest descent algorithm until the maximum force was below 1,000 kJ·mol⁻¹·nm⁻¹. The system was then equilibrated for 100 ps under the NVT ensemble at 298 K and 1 bar. After equilibration, a 100 ns production simulation was run under the NPT ensemble with a 2 fs time step. Non-bonded interactions were handled with the Verlet cutoff scheme. Electrostatic interactions were calculated using the particle mesh Ewald (PME) method, and a 1.2 nm cutoff was used for both van der Waals and Coulomb interactions. Hydrogen-containing bonds were constrained with the LINCS algorithm. Temperature was controlled with the Nosé-Hoover thermostat, and pressure was controlled with the Parrinello-Rahman barostat, maintaining the system at 298 K and 1 bar. Trajectories were saved every 10 ps. Structural dynamics and intermolecular interactions were analyzed using built-in GROMACS tools, VMD, and PyMOL. Binding free energy was estimated with gmx_MMPBSA where applicable.

Animal studies

Animal rearing
Seventy male Sprague-Dawley (SD) rats aged 3–8 weeks were randomly assigned to seven experimental groups (n = 10 per group) using computer-generated random numbers, ensuring no statistically significant differences in baseline body weight among groups. All animals were housed under controlled conditions (22 ± 1 °C, 55 ± 10% relative humidity, 12 h light/dark cycle) with free access to food and water. Cage positions were randomized at each cage change to minimize environmental bias. After 1 week of acclimatization, rats in the model groups underwent caudal IVD puncture to induce degeneration. Local infiltration anesthesia with bupivacaine was administered preoperatively to reduce pain. Animals were closely monitored during surgery and throughout postoperative recovery. Postoperative observations included recovery time, restoration of spontaneous activity, food and water intake, and body weight changes. One week after model establishment, two rats from each group were randomly selected for postoperative digital radiography (DR) imaging and hematoxylin and eosin (HE) staining. Radiographic analysis showed reduced caudal IVD height, while HE staining revealed a marked reduction in nucleus pulposus (NP) notochordal cells and disorganization of the annulus fibrosus lamellar structure¹⁸, confirming successful model establishment. Eight weeks after surgery, rats were euthanized by inhalation of an overdose of isoflurane, and tail tissue and blood samples were collected. After preliminary validation, three independent samples per group that met predefined quality criteria were used for the final histology, ELISA, PCR, and western blot analyses. Histological sections were evaluated under single-blind conditions, with slides labeled only by animal identification numbers. For ELISA, PCR, and western blot assays, blinded coding was applied during sample aliquoting and labeling, and the testing order of all samples was randomized prior to experiments. Raw data, including absorbance values, C_t values, and band intensities, were decoded into group identities by a statistician not involved in experimental grouping or sample processing and subsequently analyzed by statistical personnel. All statistical analyses were performed by statisticians who were not involved in experimental grouping or sample processing. Any adverse events, including wound infection, dehiscence, bleeding, or severe lethargy, were immediately recorded and evaluated. Mild conditions were closely monitored with enhanced supportive care. Humane endpoints were applied if rats exhibited body weight loss exceeding 20% of baseline, severe pain unrelieved by analgesics, loss of mobility, inability to access food or water, infection, necrosis, other severe postoperative complications, respiratory distress, or any life-threatening condition. In such cases, animals were humanely euthanized by inhalation of an overdose of isoflurane.

Experimental modeling
Rats were anesthetized with inhaled isoflurane and positioned prone. The Co6/7, Co7/8, and Co8/9 tail IVDs were located under X-ray fluoroscopic guidance. A 21 G needle was advanced vertically into the IVD until it reached the opposite side of the skin, after which the needle tip was withdrawn toward the center of the tail. A second needle served as a positional reference. To induce sustained IVD degeneration, the needle was rotated 360°, kept in place for 30 s, and then removed. In the Sham group, the puncture depth through the skin and muscle was approximately 2 mm, without damaging the IVD. In the other groups, the needle was inserted 5 mm toward the center of the IVD, avoiding blood vessels. After insertion, the needle was rotated 360° and left in place for 30 s. The puncture site was disinfected, marked, and protected with sterile gauze. One week after modeling, X-ray imaging of the caudal spine showed a decrease in IVD height. HE staining revealed a significant reduction in NP notochordal cells and disorganization of the annular lamella structure, confirming successful model establishment. After successful modeling, the groups receiving TCM treatments and the positive control (PC) group were administered the appropriate doses, calculated based on the body surface area ratio between humans and rats.

Grouping and intervention
A total of 70 male SD rats aged 3–8 weeks were used in this experiment and divided into seven groups:
① Control group (Ctrl): Rats in this group were randomly assigned at the same time and under the same conditions as the experimental groups. They were housed in individual cages in an SPF-level animal facility, with free access to standard chow and drinking water.
② Sham group: The rats were anesthetized by inhalation of isoflurane. After anesthesia was established, the tail skin was disinfected with povidone-iodine, and the rats were positioned prone. The Co6/7 and Co7/8 IVDs were located under X-ray fluoroscopic guidance. The needle was advanced approximately 2 mm through the skin and muscle without injuring the IVD. Distilled water (10 mL/kg) was administered once daily by gavage for 8 weeks.

③ Model group: The rats were anesthetized with inhaled isoflurane. After anesthesia was established, the tail skin was disinfected with povidone-iodine, and the rats were positioned prone. Under X-ray fluoroscopic guidance, the Co6/7, Co7/8, and Co8/9 IVDs were located. A 21 G needle was advanced vertically into the IVD until it reached the opposite side of the skin, and the needle tip was then withdrawn toward the center of the tail.

A second needle was used as a positional reference. The needle was rotated 360° and held in place for 30 s to induce continuous degeneration of the IVD before removal. Distilled water (10 mL/kg) was administered daily by gavage for 8 weeks, and the rats' recovery was closely monitored after surgery.
④ PC group: The positive control group was subjected to the same modeling procedure as the Model group and subsequently received oral gavage of celecoxib (20 mg/kg/day), prepared as a suspension in 0.5% sodium carboxymethylcellulose (CMC-Na), for 8 consecutive weeks, with a gavage volume of 10 mL/kg body weight.
⑤ YQHXD low-dose group [YQHXD(L)]: Modeling was performed as in the Model group. The rats received YQHXD at 3.87 g/kg/day by gavage for 8 weeks, with a gavage volume of 10 mL/kg and a working concentration of 0.387 g/mL.
⑥ YQHXD medium-dose group [YQHXD(M)]: Modeling was performed as in the Model group. The rats received YQHXD at 7.74 g/kg/day by gavage for 8 weeks, with a gavage volume of 10 mL/kg and a working concentration of 0.774 g/mL.
⑦ YQHXD high-dose group [YQHXD(H)]: Modeling was performed as in the Model group. The rats received YQHXD at 15.48 g/kg/day by gavage for 8 weeks, with a gavage volume of 10 mL/kg and a working concentration of 1.548 g/mL.

Pathological changes
The rat caudal IVD tissue was surgically isolated, and changes in the IVD and surrounding muscle tissues were observed. The collected tissue was then processed into pathological sections, with specimens prepared in longitudinal sections for morphological observation. The caudal vertebral IVD, fixed in formaldehyde for 96 h, was rapidly decalcified using a decalcifying solution, dehydrated through a graded alcohol series, and cleared with xylene. After embedding in paraffin, 5 µm paraffin sections were prepared. After deparaffinization with xylene and dehydration through a series of alcohol gradients, the sections were stained with hematoxylin and eosin. They were then dehydrated, cleared, and mounted. The samples were observed under an optical microscope to examine morphological changes in the caudal IVD structure across the groups. Subsequently, Masson’s trichrome staining was performed. The sections were stained sequentially with potassium dichromate solution and hematoxylin for 3 min each, followed by differentiation with acidic ethanol, and then stained with Picrosirius Red for 10 min. After treatment with phosphotungstic acid for 10 min, the sections were stained with toluidine blue for 5 min. After dehydration, the sections were cleared with xylene, mounted, and observed under a microscope to examine changes in the caudal IVD.

ELISA detection of serum target proteins
Eight weeks after surgery, rats were anesthetized with inhaled isoflurane, and abdominal aortic blood was collected. The peripheral blood samples were kept at room temperature for 1 h and centrifuged at 1,400 × g for 10 min. Serum was separated and stored at −80 °C until use. Serum AKT1, STAT3, and Bcl-2 levels were quantified by enzyme-linked immunosorbent assay (ELISA). Assays were performed according to the manufacturers’ protocols. Absorbance was read at 450 nm, and concentrations were determined from standard curves. The detection ranges for AKT1, STAT3, and BCL-2 were 0.625–40 ng/mL, 31.25–1000 pg/mL, and 0.16–10 ng/mL, respectively, with sensitivities of 0.1 ng/mL, 10 pg/mL, and 0.1 ng/mL. All serum samples were analyzed without dilution, and all results fell within the corresponding standard curve ranges. All samples and standards were assayed in duplicate, and normalization was performed based on equal serum volumes.

PCR detection of gene expression levels
Intervertebral disc tissues were dissected from rat caudal vertebrae, and approximately 20 mg was collected from each sample. The tissues were pulverized in liquid nitrogen and lysed in 500 µL of non-enzymatic lysis buffer for 15 min. Total RNA was then extracted, and its concentration and purity were measured with a spectrophotometer. Samples with A₂₆₀/A₂₈₀ values of 1.8–2.0 were accepted for analysis, and RNA concentrations were adjusted to 500 ng/µL. Reverse transcription was carried out according to the manufacturer’s protocol. Quantitative PCR (qPCR) was performed on a real-time PCR system to measure Akt1, Bcl-2, and Stat3 mRNA expression. Primer sequences are provided in Table 1. The PCR program consisted of initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. Melting curve analysis was conducted after amplification to verify product specificity. Gapdh was used as the reference gene, and all samples were analyzed in triplicate. Primer amplification efficiencies were 90%–110%. Relative gene expression was calculated using the 2^−ΔΔCt method.

Western blotting for protein expression levels
Total protein was isolated from rat caudal intervertebral disc tissues from each group. Tissues were lysed in RIPA buffer supplemented with protease inhibitors. Protein concentration was measured using a BCA protein assay kit and adjusted to 2 µg/µL. Equal protein amounts (15 µg per sample) were separated by 10% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes at a constant current of 300 mA for 90 min using a wet transfer system. The membranes were then blocked with 5% non-fat milk for 1 h at room temperature and incubated overnight at 4 °C with primary antibodies against AKT1 (1:5,000), STAT3 (1:6,000), Bcl-2 (1:1,500), and β-actin (1:4,000). After TBST washing, the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies at a dilution of 1:5,000 for 1 h at room temperature. Protein bands were detected using an enhanced chemiluminescence (ECL) reagent, and images were acquired with a gel imaging system under identical exposure conditions. Band intensities were measured using ImageJ and normalized to β-actin to determine relative protein expression. Each experiment was performed independently in triplicate, with each sample obtained from a different animal.

Statistics
Results are expressed as mean ± standard deviation (SD). Differences among groups were evaluated by one-way analysis of variance (ANOVA), followed by the least significant difference (LSD) post hoc test. Statistical significance was defined as P < 0.05.

Results

Conclusions derived from the study
This study integrated network pharmacology, molecular docking, molecular dynamics simulations, and in vivo experiments to investigate the potential protective effects of YQHXD against LDD. The results suggest that YQHXD may act, at least in part, through regulation of the AKT1/STAT3/BCL-2 signaling axis, characterized by reduced STAT3 expression and increased AKT1 and Bcl-2 expression. Histological and molecular analyses showed that YQHXD treatment was asso...

Discussion

Traditional Chinese herbal medicines have been reported to exert pharmacological effects including anti-inflammatory, analgesic, and blood circulation–promoting activities¹⁹. Previous studies have suggested that Xiaosui Huahe Decoction, the precursor formula of YQHXD, may promote resorption of massive IVD herniation by affecting nucleus pulposus (NP) resorption and IVD absorption²⁰; however, the underlying molecular mechanisms remain unclear. In the present study, net...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
0.5% sodium carboxymethylcellulose (CMC-Na)	Not specified		Vehicle used to prepare the celecoxib suspension for gavage.
21 G needle	Not specified		Used for caudal intervertebral disc puncture.
2× Universal Blue SYBR Green qPCR Master Mix	Servicebio Technology Co., Ltd., Wuhan	G3326	qPCR amplification reagent.
30% acrylamide/bis-acrylamide gel solution (29:1)	Solarbio Co., Ltd.		Gel preparation solution for SDS-PAGE; listed in original materials table as 30%/29:1 gel solution.
3D frozen grinding machine	Servicebio Technology Co., Ltd., Wuhan		Used for tissue homogenization/grinding.
4× SDS-PAGE concentrating gel buffer (pH 6.8)	Solarbio Co., Ltd.		Western blot gel preparation reagent.
Acidic ethanol	Not specified		Used for histological staining differentiation.
ACPYPE	ACPYPE project		Used to convert ligand topology files into a GROMACS-compatible format.
AKT1 ELISA kit	Hangzhou UpingBio Technology Co., Ltd.	SYP-R1928	Used to measure serum AKT1 levels; detection range 0.625–40 ng/mL.
AM1-BCC charge method	Antechamber/AMBER Tools		Used for ligand atomic charge calculation.
Amber14SB force field	AMBER force field		Protein force field used in molecular dynamics simulation.
Angelica sinensis root	Jianshou Pharmacy, Nanning, Guangxi, China	Batch No. 20250503	YQHXD botanical component; 20 g per prescription.
Anhydrous ethanol	Tianjin Aopusheng Chemical Co., Ltd.		Used in tissue processing/dehydration.
Antechamber	AMBER Tools		Used for ligand charge calculation and atom typing.
Anti-AKT1 antibody	Wuhan Sanying Co., Ltd.		Primary antibody for Western blotting; dilution 1:5,000.
Anti-Bcl-2 antibody	Wuhan Sanying Co., Ltd.		Primary antibody for Western blotting; dilution 1:1,500.
Anti-STAT3 antibody	Wuhan Sanying Co., Ltd.	10253-2-AP	Primary antibody for Western blotting; dilution 1:6,000.
APS reagent	Beyotime/Biyuntian Co.		Western blot gel polymerization reagent.
Astragalus membranaceus root	Jianshou Pharmacy, Nanning, Guangxi, China	Batch No. 20250405	YQHXD botanical component; 20 g per prescription.
Autoclave	SANYO, Japan		Sterilization equipment.
AutoDock Vina engine	AutoDock/Vina project		Docking engine used within CB-Dock2.
BCA protein assay kit	Solarbio Co., Ltd.		Used to determine protein concentration before Western blotting.
Bcl-2 ELISA kit	Wuhan Fine Biotech Co., Ltd.	PMK1545	Used to measure serum Bcl-2 levels; detection range 0.16–10 ng/mL.
Bupivacaine	Not specified		Local infiltration anesthetic used preoperatively.
CB-Dock2	CB-Dock2 web server		Automated blind molecular docking platform.
Celecoxib	Sichuan New Ston Pharmaceutical Co., Ltd.	H20233911	Positive-control drug; 20 mg/kg/day by gavage.
Chaenomeles speciosa fruit	Jianshou Pharmacy, Nanning, Guangxi, China	Batch No. 231001	YQHXD botanical component; 15 g per prescription.
Cytoscape	Cytoscape Consortium	Version 3.9.0	Used to construct and analyze network diagrams.
DAB chromogenic reagent kit (20×)	Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd.		Listed in existing materials table.
DAVID	National Cancer Institute	Version 6.8	Database for Annotation, Visualization and Integrated Discovery; used for GO and KEGG enrichment analysis.
Decalcification solution	Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd.		Used for decalcification of caudal vertebral IVD tissue.
Decolorization shaker	Liu Yi Biotechnology Co., Ltd., Beijing		Laboratory shaker.
DEPC water	Wuhan Sanying Co., Ltd.		RNA work/reagent preparation.
Digital radiography imaging system	Not specified		Used for postoperative radiographic model validation.
Distilled water	Not specified		Vehicle/control gavage and YQHXD decoction preparation.
ECL chemiluminescent substrate	Biosharp Co.		Western blot signal detection reagent.
EDTA antigen retrieval solution (pH 8.0)	Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd.		Listed in existing materials table.
Electrophoresis apparatus	Not specified		Used for SDS-PAGE electrophoresis.
Electrophoresis buffer	Servicebio Co.		Western blot electrophoresis solution.
FM40 ice maker	Changliu Scientific Instruments Co., Ltd., Beijing		Ice maker.
Formaldehyde	Not specified		Fixative used for caudal vertebral IVD tissue.
GAFF2 atom types	AMBER Tools		Ligand atom types used for molecular dynamics simulations.
Gel imaging system	Tianen Scientific, Shanghai		Used to capture Western blot images.
GeneCards	Weizmann Institute of Science	Version 5.26	Human gene database used to retrieve disease-associated targets.
gmx_MMPBSA	gmx_MMPBSA project		Used for MM-PBSA binding free-energy estimation where applicable.
GraphPad Prism	GraphPad Software	Version 8.0	Statistical/graphing software listed in the authors' earlier manuscript version.
Grinding beads	Servicebio Technology Co., Ltd., Wuhan		Used for tissue grinding/homogenization.
GROMACS	GROMACS development team	Version 2022.2	Used for molecular dynamics simulations.
Hematoxylin and eosin staining reagents	Not specified		Used for HE staining of caudal IVD sections.
High-speed centrifuge	Eppendorf, Germany		Laboratory centrifuge.
High-speed low-temperature tissue grinder	Servicebio, China		Used for low-temperature tissue grinding.
Horizontal shaker	Chirinbell Company		Laboratory shaker.
HRP-labeled goat anti-rabbit antibody	Servicebio Co.		Secondary antibody for Western blotting; dilution 1:5,000.
HyperScript III RT SuperMix for qPCR with gDNA Remover	Servicebio Technology Co., Ltd., Wuhan		Reverse transcription reagent listed in existing materials table.
Ice maker	SIMAG, China		Ice maker.
ImageJ	National Institutes of Health		Used for Western blot band-intensity quantification.
Inverted biological microscope	Olympus, Japan		Used for microscopy of stained sections.
Isoflurane	Not specified		Inhalation anesthetic/euthanasia agent.
Isopropanol reagent	China National Pharmaceutical Group Chemical Reagent Co., Ltd.		RNA extraction reagent.
Joung-Cheatham ion parameters	Not applicable		Ion parameter set used for molecular dynamics simulations.
Ligusticum chuanxiong rhizome	Jianshou Pharmacy, Nanning, Guangxi, China	Batch No. 20250301	YQHXD botanical component; 15 g per prescription.
Liquid nitrogen	Not specified		Used for tissue grinding before RNA extraction.
Low-speed centrifuge	Eppendorf, Germany		Used for serum separation and other low-speed centrifugation steps.
Masson's trichrome staining solution	Not specified		Used for Masson staining of caudal IVD sections.
Microplate reader (TECAN)	TECAN, Switzerland		Used for absorbance measurement at 450 nm; listed in existing materials table.
Microplate reader (Thermo Fisher Scientific)	Thermo Fisher Scientific, USA		Multifunctional microplate reader listed in existing materials table.
Microsoft Excel/spreadsheet software	Microsoft		Used to remove duplicate target entries and manage tabular data.
NanoDrop 2000 spectrophotometer	Thermo Fisher Scientific, USA		Used to measure RNA concentration and purity.
Non-enzymatic lysis buffer	Not specified		Used for tissue lysis before RNA extraction.
OMIM	Johns Hopkins University School of Medicine		Online Mendelian Inheritance in Man; used to retrieve disease-associated targets.
Online bioinformatics plotting platform	Not specified		Used to generate GO and KEGG bar plots.
PAGE gel polymerization agent	Solarbio Co., Ltd.		Western blot gel preparation reagent.
Paraffin	Not specified		Embedding medium for histological sections.
PCR primers for Akt1, Bcl-2, Stat3, and Gapdh	Wuhan Servicebio Technology Co., Ltd.		Primer sequences are listed in Table 1.
PDB	RCSB PDB		Protein Data Bank; used to obtain 3D structures of target proteins.
Phosphorylated/protease inhibitors	Servicebio Co.		Added to RIPA lysis buffer for protein extraction.
Phosphotungstic acid	Not specified		Used in Masson staining.
Picrosirius Red	Not specified		Used in Masson staining.
Polyvinylidene difluoride membrane (0.45 μm)	Servicebio Co.		PVDF membrane used for Western blot transfer.
Positive-control drug vehicle	Not specified		0.5% CMC-Na suspension vehicle for celecoxib.
Povidone-iodine	Not specified		Used for surgical site disinfection.
Protein marker	Thermo Fisher Scientific, USA		Molecular weight marker for Western blotting.
Protein transfer apparatus	Bio-Rad, USA		Wet transfer equipment for Western blotting.
Protein transfer module (wet transfer)	Bio-Rad, USA		Western blot wet transfer module.
PubChem	National Center for Biotechnology Information (NCBI)		Used to obtain SMILES/SDF files of compound structures.
Pure water preparation system	Le Feng Biotechnology Co., Ltd., Shanghai		Used for pure water preparation.
PyMOL	Schrödinger, LLC		Used for visualization and analysis of molecular dynamics/docking outputs.
qPCR real-time PCR system	Applied Biosystems QuantStudio, USA		Used for quantitative PCR amplification.
RIPA tissue/cell lysis buffer	Solarbio Co., Ltd.		Used for total protein extraction.
RNA dissolution solution	Servicebio Technology Co., Ltd., Wuhan		RNA handling reagent.
Rotary evaporator	Not specified		Used to concentrate the combined YQHXD filtrates under reduced pressure.
Sinapis alba mature seeds	Jianshou Pharmacy, Nanning, Guangxi, China	Batch No. 20240801	YQHXD botanical component; 10 g per prescription.
Sprague-Dawley rats	Shanghai Slack Experimental Animal Co., Ltd.	SCXK (Shanghai) 2022-0004	Seventy male rats, aged 3–8 weeks, used for the caudal-disc puncture model.
SPSS	IBM, USA	Version 21.0	Statistical software listed in the authors' earlier manuscript version.
STAT3 ELISA kit	Hangzhou UpingBio Technology Co., Ltd.	SYP-R1555	Used to measure serum STAT3 levels; detection range 31.25–1000 pg/mL.
Stephania tetrandra root	Jianshou Pharmacy, Nanning, Guangxi, China	Batch No. 230301	YQHXD botanical component; 6 g per prescription.
Sterile gauze	Not specified		Used to protect the puncture site after modeling.
STRING	STRING Consortium	Version 12.0	Protein-protein interaction network platform.
SweScript All-in-One RT SuperMix for qPCR	Servicebio Technology Co., Ltd., Wuhan	G3337	Reverse transcription reagent listed in existing materials table.
SwissTargetPrediction	SIB Swiss Institute of Bioinformatics		Used to predict potential targets of compounds.
TBST buffer	Not specified		Used for membrane washing during Western blotting.
TCMSP	Center for Bioinformatics, College of Life Science, Northwest A&F University	Version 2.3	Traditional Chinese Medicine Systems Pharmacology Database; used to obtain active ingredients of Chinese herbal compounds.
TIP3P water model	Not applicable		Water model used in molecular dynamics simulations.
Toluidine blue	Not specified		Used in Masson staining.
Transfer buffer	Servicebio Co.		Western blot transfer buffer.
TriQuick reagent	Servicebio Technology Co., Ltd., Wuhan		RNA extraction reagent.
TRIzol reagent	Not specified		Total RNA extraction reagent mentioned in the manuscript.
TTD	The Chinese University of Hong Kong (IDRBLab)		Therapeutic Target Database; used to retrieve disease-associated targets.
Venny	BioinfoGP, CNB-CSIC	Version 2.1.0	Used to identify overlapping targets between YQHXD and LDD.
Verlet cutoff scheme	GROMACS		Non-bonded interaction scheme used in molecular dynamics simulations.
Vertical electrophoresis apparatus	Bio-Rad, USA		Used for SDS-PAGE.
VMD	University of Illinois at Urbana-Champaign		Used for visualization and analysis of molecular dynamics outputs.
Wet transfer system	Bio-Rad, USA		Used for protein transfer in Western blotting.
X-ray fluoroscopy system	Not specified		Used to identify caudal IVD levels during puncture modeling.
Xylene	Not specified		Used for deparaffinization/clearing during histology.
β-actin primary antibody	Wuhan Sanying Co., Ltd.		Loading-control primary antibody for Western blotting; dilution 1:4,000.