Screening and Determination of Biomarkers and Their Associated Active Components for Enhanced Therapeutic Effect of Exocarpium Citri Grandis–Folium Ginkgo Couplet on Liver Injury Caused by Hyperlipidemia via UPLC-TOF-MS and UPLC-MS

Abstract

Introduction

The combination of Exocarpium Citri Grandis–Folium Ginkgo (E–GC) demonstrates superior therapeutic effects on liver injury caused by hyperlipidemia compared with individual use; however, accurate and quantifiable indicators for evaluating the efficacy and quality of the couplet remain lacking. Reliance solely on changes in certain differential metabolites or the bioavailability of chemical components is insufficient to confirm the superior efficacy of the couplet.

Methods

In this study, a rat model of liver injury caused by hyperlipidemia was used to verify the enhanced efficacy of E-GC. Subsequently, network pharmacology combined with molecular biology validation was applied to identify the primary regulatory targets. The results were integrated with UPLC-TOF-MS, UPLC-MS and database analyses to identify biomarkers indicating the enhanced efficacy of E-GC, as well as serum concentrations of active components in E-GC to quantitatively evaluate the effect of the couplet on their bioavailability. Finally, correlation statistical analyses were conducted on the above-mentioned data to confirm that the identified biomarkers and significantly altered active components can characterize the enhanced efficacy of the couplet.

Results

Metabolomics analysis identified 59 related metabolites, among which arachidonic acid, choline, deoxycholic acid, N-acetylserotonin, and cholic acid significantly increased in the liver following E-GC treatment, with similar changes observed in serum levels (p<0.05). A total of 36 components were identified, apigenin, naringin, and ginkgolide A/B/C exhibited significantly increased bioavailability after combination (p<0.05). These changes were closely associated with the improved regulation of PXR-FXR expression by E-GC compared with individual use, and E-GC also exerted strong regulatory effects on downstream targets, such as LXR, FGF1, ABCB1b, and ERK.

Conclusion

Through the above analyses, in order to solve the problem of lacking accurate and quantifiable indexes to evaluate the effect of TCM couplet, a method was successfully established for evaluating the enhanced efficacy of E-GC in treating liver injury caused by hyperlipidemia, and the principal biomarkers and quality control indicators based on active components were identified.

Graphical Abstract

Introduction

Hyperlipidemia causes significant damage to the liver, an endocrine organ closely associated with lipid metabolism, leading to a series of serious complications.1,2 Conventional chemical drugs and dietary supplements often have relatively limited therapeutic targets and do not effectively promote the transport and efflux of intracellular accumulated lipids or the synthesis of important metabolites, such as bile acids and fatty acids.3–5 Traditional Chinese medicine (TCM) formulas have garnered increasing attention in recent years due to their ability to simultaneously regulate lipid metabolism pathways.6,7 For example, the Yinlan Tiaozhi Formula (YL), developed by our research group, has completed Phase II clinical trials. The results confirm its efficacy in treating hypercholesterolemia and hyperlipidemia, with significant improvements observed in the morphology and function of damaged livers in hyperlipidemic model rats following administration. YL is composed of four medicinal ingredients: Exocarpium Citri Grandis (ECG), Folium Ginkgo (GK), Gynostemma pentaphyllum (GYP), and Propolis (PP). Its core efficacy lies in the synergistic combination of ECG and GK, which form a couplet Exocarpium Citri Grandis–Folium Ginkgo (E-GC). The literature indicates that ECG and GK promote the breakdown and excretion of lipid accumulation in the liver while enhancing the metabolism of cholesterol and triglycerides. They alleviate inflammatory responses in the liver, collectively exerting restorative and regenerative effects on hepatocytes.8,9

However, there remains a lack of accurate and quantifiable evaluation indicators for the enhanced efficacy of the couplet. Specifically, after two TCM ingredients form a couplet, it is unclear which disease-related biomarkers exhibit significant changes in the body and whether such changes differ significantly from the sum of the two individual components; similarly, it remains uncertain whether their bioavailability is significantly increased. These issues have been rarely addressed in previous studies.10 The primary reason may be that simply correlating endogenous metabolites and their level changes with efficacy can only demonstrate the multi-target and multi-component additive effects of TCM formulas, rather than confirming superior efficacy induced by the combination of two components. Therefore, it is necessary to obtain multidimensional evidence of “efficacy–target–metabolite–component”, to screen and quantitatively evaluate related biomarkers, and to demonstrate that their variations are caused by E-GC-induced enhancement of target metabolism rather than the accumulation of effects from different targets. In addition, identifying active components with significantly improved bioavailability is also required. The aforementioned metabolites and compounds should be used as standards for evaluating the efficacy and quality of E-GC.

Our previous findings indicated that the primary targets affected by ECG and GK converge on the PXR–CYP7A1–ABCB1/FXR–LXR–CYP7A1 pathways, which intersect at CYP7A1, a critical regulatory site in cholesterol metabolism. The primary mechanism by which ECG and GK alleviate hepatic lipid stasis involves active components binding to multiple targets within the PXR/FXR pathway. This interaction reduces the levels of several metabolites related to cholesterol synthesis, thereby inhibiting cholesterol synthesis and accumulation. Consequently, hepatocyte apoptosis and inflammation associated with hypercholesterolemia are mitigated. The clinical manifestations and changes in key physiological indicators are closely associated with the treatment of phlegm–dampness obstruction syndrome.11,12 Meanwhile, ECG and GK may inhibit lipid synthesis and promote lipid efflux in hepatocytes.13,14 Identifying relevant metabolites and other evaluation indicators, and integrating these findings, could strengthen the scientific basis of this regimen in removing blood stasis and alleviating liver stasis.15,16 However, whether the combination of these two drugs as a paired formulation produces effects comparable to or greater than the sum of their individual effects, whether it represents a significant advancement over monotherapy, and the central role of this couplet in YL require systematic validation through target, pathway, metabolite, and pharmacodynamic studies, along with assessments of their respective contributions.

Several nuclear receptors, such as FXR, can be influenced by alterations in PXR expression following drug administration. PXR regulates drug metabolism and cytochrome enzyme function by binding nonspecifically to exogenous substances. Hepatocyte lipid levels are regulated by changes in CYP7A1 activity, which also affect cholesterol metabolism, transport, and efflux.17,18 By regulating CYP7A1, PXR influences metabolite synthesis and transport, and reduces hepatic cholesterol accumulation.19 Therefore, the PXR–CYP7A1–ABCB1 pathway represents an important entry point for investigating the mechanism of action of E-GC in the treatment of hyperlipidemia.20

FXR and LXR function as ligand-activated receptors that regulate hepatic lipid metabolic homeostasis and control lipid metabolism in hepatocytes by influencing lipid, bile acid, and fatty acid levels.10,21,22 FXR ligands affect pathways involved in cholesterol metabolism, with the classical CYP7A1-mediated pathway being the most prominent. FXR modulates bile acid levels through SHP, a downstream mediator. Alterations in SHP expression affect the concentration of bile acids, which serve as CYP7A1 ligands and regulate lipid metabolism equilibrium.10,23–25 Similarly, LXR promotes the conversion of cholesterol into bile acids and facilitates their excretion. LXR regulates lipid synthesis and accumulation in hepatocytes and influences lipid levels and transport in vivo.26,27 Altered LXR activity modifies lipid exchange between the blood and liver, thereby contributing to the repair of hepatocyte injury.26 LXR and FXR interact with exogenous compounds, including pharmaceutical agents and TCM components, thereby altering LXR activity and downstream gene expression. FXR regulates CYP7A1, SHP, ABCB1, and hepatocyte lipid levels.28

To elucidate the therapeutic mechanisms of TCM, particularly couplets and complex formulas, studies frequently incorporate multidimensional experimental data, including pharmacological effects, molecular mechanisms, and chemical composition analyses. However, the correlation and integration of these data require further improvement, and additional methodological evaluation is needed.29,30 It is essential to assess multidimensional data, assign appropriate weights to efficacy and mechanistic evidence, and integrate these data systematically by constructing metabolic regulatory network maps, thereby enhancing the precision of macro- and micro-level evidence and promoting their coherent integration.31

This study aims to elucidate how E-GC alleviates hepatic stasis by regulating key lipid metabolism targets within the PXR/FXR axis and to establish component–metabolite relationships based on target and pathway variations. The lipid-modulating efficacy and mechanism of action of E-GC are interpreted using micro-level evidence systems, integrated with TCM theory and the biological basis of hepatic lipid metabolism.

Materials, Instruments, and Methods

Materials and Reagents

E-GC comprises two Chinese herbal medicines, namely, Exocarpium Citri Grandis (ECG) and Ginkgo biloba L. leaves. ECG is the dried outer pericarp of the immature or nearly mature fruits of Citrus grandis “Tomentosa” or Citrus grandis (L). Osbeck (family Rutaceae), while GK refers to the dried leaves of Ginkgo biloba L. (family Ginkgoaceae). Both crude drugs were purchased from Zhixin Pharmaceutical Co., Ltd. (Guangzhou, China). The ECG (Lot No. 20230501) and GK (Lot No.20230601) were authenticated according to the Chinese Pharmacopoeia standards. HPLC-grade acetonitrile, formic acid, methanol, and ethanol were provided by Merck (Darmstadt, Germany). Ultrapure water from a Milli-Q water purification system was used throughout all experiments. All of the other chemical reagents were obtained from Guangzhou Chemical Reagents Co., Ltd. (Guangzhou, China) and were of analytical grade. TC (20220922), TG (20220923), HDL-C (20230621), and LDL-C (20230621) assay kits were obtained from the Nanjing Jiancheng Bioengineering Institute. Isoflurane (Lot No. 20230302) was obtained from Shenzhen RWD Life Science Co., Ltd., and simvastatin (Lot No. J20190007) was supplied by Merck Sharp & Dohme (China). Paraformaldehyde fixative (Lot No. 22120600055) was purchased from Wuhan Servicebio Technology Co., Ltd. PMSF inhibitor (Lot No.: 20230210) and RIPA tissue lysate (Lot No.: 20230210) were supplied by Solarbio Technology Co., Ltd. A BCA kit (Lot No.: RB230506) was purchased from Thermo Fisher Scientific. ECL luminescence detection solution (Lot No.: ATWK10081) was purchased from Abbkine. Antibodies against PXR (Proteintech), FXR (A8697), p-ERK1/2 (AF0155), t-ERK1/2 (AF1015), CYP7A1 (DF2612), LXR (DF6864), and GAPDH (AF0911) were obtained from Affinity (USA). Trizol RNA lysate was provided by Dingguo Changsheng Biotechnology Co., Ltd. (Beijing, China). PerfectStart Green qPCR SuperMix was purchased from Full Gold Biotechnology Co., Ltd. (Beijing, China), and the Evo M-MLV RT premixed qPCR reagent was supplied by Acoray BioEngineering Co., Ltd. (Hunan, China). Primers for PXR, LXRα, FXR, SHP, CYP7A1, and ABCB1b were obtained from Wuhan Servicebio Co., Ltd. Choline (Lot No.: C875195) and deoxycholic acid (Lot No.: D806702) were supplied by Macklin. Hippuric acid (Lot No.: SH8790) was obtained from Solarbio Life Sciences. N-acetylserotonin (Lot No.: C5386) was purchased from APExBIO, and arachidonic acid (Lot No.: T4129) was obtained from TargetMol. Kaempferol (Lot No:110861–2021808), Apigenin (Lot No:111901–202205), Chrysin (Lot No:112024–202105), and Quercetin (Lot No:111699–202208) were obtained from the National Institute for the Control of Pharmaceutical and Biological Products (NICPBP) of China (Beijing, China). All reference standards were purchased from commercial suppliers, and their purities (>95%) were confirmed by the suppliers’ certificates of analysis.

Instruments

An AB SCIEX Q-TOF–MS/MS system was used to identify metabolites and obtain metabolomic profiles (AB SCIEX, Framingham, MA, USA). The components were subsequently determined in vivo using an Agilent Quantum UPLC–MS–MS system (Agilent, USA) equipped with an Agilent C18 column (2.1 mm × 150 mm, 1.8 μm). A StepOnePlus^TM Real-Time PCR system (Thermo Fisher, USA) was used for PCR analysis. An electronic analytical balance XS205DU 1/100,000 (METTLER TOLEDO, Switzerland) was also used.

Animal Experiments

A total of 30 male Sprague-Dawley (SD) rats, weighing 200 ± 20 g, were used in this study. All of the animals were obtained from the Guangdong Medical Laboratory Animal Center (Production License No.: SCXK (Yue) 2019–0002; Quality Certificate No.: 44007200105943). The rats were housed in the SPF-grade animal facility of the Fifth Clinical Medical College, Guangzhou University of Chinese Medicine, with the Facility Use License No.: SYXK (Yue) 2020–0059. The study protocol was approved by the Animal Ethics Committee of the Fifth Clinical Medical College, Guangzhou University of Chinese Medicine, under Approval No.: 049157. The rats were maintained under controlled environmental conditions (temperature: 22°C ± 2°C; relative humidity: 55% ± 5%). Following one week of acclimatization, the animals were randomly divided into five groups (n = 6 per group): control group, model control group, positive control group (simvastatin, 7.2 g/kg), E-GC low-dose group (2.57 g/kg crude drug), and E-GC high-dose group (5.12 g/kg crude drug). Except for the control group, all groups were fed a high-fat diet for one month to induce hyperlipidemia, which was confirmed by significantly elevated serum TC and TG levels. The positive control group received simvastatin (7.2 g/kg, dissolved in distilled water) by gavage, while the model control group continued on the high-fat diet.

The dosages of E-GC extracts were derived from the formulation of Yinlan capsules (500 g ECG and 1000 g GK per 2383 g total crude drug, yielding 1000 capsules). Based on the clinical rat-equivalent dose of Yinlan (2.57 g/kg/day), the low and high doses of E-GC were set at 2.57 and 5.12 g/kg crude drug, respectively, corresponding to the clinical equivalent and twice the clinical equivalent dose. All herbal extracts were administered by gavage once daily for 30 consecutive days.

Before the final administration, all rats were fasted overnight (12 h) with free access to water. One hour after the last dose, the rats were weighed and anesthetized with isoflurane (induction: 4%; maintenance: 2% in 100% O₂ at a flow rate of 1 L/min). Blood samples were collected from the abdominal aorta into tubes without anticoagulant. After clotting, the blood was centrifuged at 3000 rpm for 10 min, and the serum supernatant was collected and stored at −80 °C. Following euthanasia by exsanguination under deep anesthesia, tissue samples of the liver, kidney, and small intestine, as well as serum samples, were harvested and stored at −80°C.

Biochemical Parameters and Histopathological Findings

Appropriate assay kits were used to measure TC, TG, HDL-C, and LDL-C levels in serum, as well as TC and TG levels in liver tissue. Liver tissues from each group were subjected to formaldehyde fixation, paraffin embedding, hematoxylin and eosin staining, observation under a light microscope, and photographic documentation.

UPLC–Q-TOF–MS/MS Detection and Metabolite Identification

Serum samples from each experimental group were prepared by protein precipitation. Briefly, 400 μL of methanol/acetonitrile (1:1, v/v) was added to 100 μL of serum in a 1.5 mL EP tube. After centrifugation at 12,000 rpm and 4 °C for 15 min, 400 μL of the supernatant was transferred to a fresh tube and dried under nitrogen gas. The residue was reconstituted in 100 μL of acetonitrile/water (1:1, v/v), and the solution was centrifuged again under the same conditions. Then, 80 μL of the supernatant was collected for UPLC–Q-TOF–MS/MS analysis. The total ion current chromatogram is presented in Figure S1.

The UPLC–Q-TOF–MS/MS chromatographic conditions were as follows: a Waters UPLC BEH C18 column (2.1 mm × 100 mm, 1.7 μm) was used; the autosampler and column temperatures were maintained at 8°C and 35°C, respectively; the injection volume was 1 μL; and the mobile phase flow rate was 0.3 mL/min. Gradient elution was performed using acetonitrile (A) and 0.1% formic acid in water (B) as the mobile phase (0–2 min, 1% A; 2–8 min, 1–55% A; 8–17 min, 55–98% A; 17–20 min, 98% A; 20–20.5 min, 98–1% A; 20.5–25 min, 1% A).

Mass spectrometry was conducted using an electrospray ionization (ESI) source operating in both positive and negative ion modes, with a mass-to-charge ratio scan range of 50–1000 m/z. The ion source parameters were set as follows: nebulizer gas, 55 psi; curtain gas, 35 psi; auxiliary gas, 55 psi; and gas temperature, 500°C. For positive ion mode, the capillary voltage was maintained at 5.5 kV with a fragmentor voltage of 60 V, a collision energy of 35 eV, and a collision energy spread of 15 eV. For negative ion mode, the capillary voltage was −4.5 kV with a fragmentor voltage of −60 V, a collision energy of 35 eV, and a collision energy spread of 15 eV. Metabolites were identified by comparing retention times and mass spectral data with reference standards and literature reports.

Determination of Candidate Compounds by UPLC–MS/MS

The compounds identified in E-GC (UPLC–Q-TOF–MS/MS Detection and Metabolite Identification) were used as indices for content determination. Reference standards of kaempferol, apigenin, chrysin, and quercetin were accurately weighed and dissolved in methanol to prepare mixed reference solutions. For test solution preparation, 0.25 g of mixed ECG and GK powder was accurately weighed, dissolved in 25 mL of methanol, and sonicated for 30 min. The supernatant was collected and filtered through a 0.22-μm membrane prior to UPLC–MS/MS analysis.

The UPLC–MS/MS chromatographic conditions were as follows: an Agilent C18 column (2.1 mm × 150 mm, 1.8 μm) was used; the autosampler and column temperatures were maintained at 8°C and 35°C, respectively; the injection volume was 1 μL; and the mobile phase flow rate was 0.3 mL/min. Gradient elution was performed using acetonitrile (A) and 0.1% formic acid in water (B) as the mobile phase (0–1 min, 2% A; 1–2.5 min, 2–30% A; 2.5–5 min, 30–65% A; 5–7 min, 65–80% A; 7–11 min, 80–95% A; 11–13 min, 95% A).

Mass spectrometry was performed using an electrospray ionization (ESI) source. The capillary voltage was set to 3000 V for both positive and negative ion modes; the drying gas temperature was 200°C with a flow rate of 15 mL/min; the nebulizer pressure was 30 psi; and the sheath gas temperature was 350°C with a flow rate of 11 mL/min. The mass spectrometric parameters for each component are listed in Table 1.

Table 1.

Ion Pairs of the Candidate Compounds

Compound	m/z	Fragment	Tube Length Offset	Collision Energy	Polarity
Apigenin	271.0	153.0	105	33	+
Hesperidin	609.0	300.6	100	20	–
Bilobalide A	407.2	351.0, 229.0, 362.9	100	16, 22, 14	–
Bilobalide B	423.2	367.2, 383.0	100	17, 17	–
Bilobalide C	439.2	125.0, 365.0	100	28, 22	–
Kaempferol	287.0	153.0	105	39	–

Pharmacokinetic Studies by UPLC–MS/MS

Twelve SD rats (6 per group) from the control and E-GC low-dose groups (as described in Animal Experiments) were used for pharmacokinetic analysis. Blood samples were collected from the orbital vein at predetermined time points (0, 0.083, 0.25, 0.5, 1.0, 1.5, 2.0, 3.0, 4.0, 6.0, 12.0, and 24.0 h) after administration. Samples were heparinized, centrifuged at 12,000 rpm and 4°C, and stored at −20°C until analysis.

For sample preparation, 100 μL of thawed plasma was transferred to a 1.5 mL centrifuge tube, and 300 μL of acetonitrile was added. After centrifugation at 12,000 rpm and 4 °C for 10 min, 300 μL of the supernatant was transferred to a fresh tube and dried under nitrogen. The residue was reconstituted in 100 μL of methanol containing the internal standard (piperine, 100 ng/mL). Following centrifugation at 12,000 rpm and 4°C for 5 min, 80 μL of the supernatant was collected for UPLC–MS/MS analysis under the chromatographic and mass spectrometric conditions described in Determination of Candidate Compounds by UPLC–MS/MS. Pharmacokinetic parameters were calculated using WinNonlin 8.1 software.

Pharmacokinetic Methodology Investigations

Quality control samples were prepared by spiking blank serum with standard solutions of the analyzed components. The method validation primarily included assessments of precision, accuracy, stability (including room temperature, benchtop, and long-term conditions), linearity (concentration range 2–1000 ng/mL), sensitivity (LLOQ), and matrix effects (evaluated in both solvent and plasma). The QC samples were prepared at concentrations of 5, 100, and 800 ng/mL. Recovery rates, relative standard deviations (RSD), and concentrations of the analytes in each sample were calculated.

Network Pharmacology Analysis

The E-GC components have been previously reported by our research group.32,33 Following component analysis, E-GC components were entered into the Swiss ADME website (http://www.swissadme.ch/index.php) using SMILES numbers to screen for active ingredients. Components were considered active compounds if they received at least two “yes” evaluations for high gastrointestinal absorption and satisfied the Lipinski, Ghose, Veber, Egan, and Muegge drug-likeness criteria in ADME. The active compounds are listed in Table 2. The Swiss Target Prediction website (http://www.swisstargetprediction.ch/) was used to predict the targets of E-GC active ingredients. To identify potential protein targets of the active compounds, a screening threshold of “probability > 0.1” was applied to entries from Homo sapiens. The keywords “lipid accumulation in hepatocytes” were used to retrieve targets associated with hepatic lipid stasis from the Comparative Toxicogenomics Database (https://ctdbase.com/) and GeneCards (https://www.genecards.org/), and targets with an “Inference Score >20” were selected. Common targets at the intersection of E-GC-treated hepatic lipid stasis were identified using a Venn diagram generated with the Venny 2.1.0 tool (https://bioinfogp.cnb.csic.es/tools/venny/). The STRING database (https://string-db.org) was used to analyze common targets, with the species restricted to Homo sapiens and the confidence level set to >0.4. The herb–component–target network was constructed using Cytoscape 3.7.1.

Table 2.

E-GC Compound Identification

Compound	Formula	Source
Ginkgolide A	C20H24O9	GK
Quercetin	C15H10O7	ECG
Apigenin	C15H10O5	ECG
Kaempferol	C15H10O6	GK
Meranzin hydrate	C15H16O4	ECG
Caffeic acid phenethyl ester	C16H14O4	ECG
Ferulic acid	C10H10O4	GK
Vanillic Acid	C8H8O4	GK
Gallic acid	C7H6O5	GK
Protocatechuic acid	C7H6O4	ECG
4-Hydroxycinnamic acid	C9H8O3	GK
Naringin	C27H32O14	ECG
Meranzin	C15H16O4	ECG
Luteolin	C15H10O6	GK
Naringenin	C15H12O5	ECG
Biochanin A	C16H12O5	GK
Isomeranzin	C15H16O4	ECG
Limonin	C26H30O8	ECG
Phenethyl caffeate	C17H16O4	ECG
Auraptene	C19H22O3	ECG

Common targets were imported into Metascape (https://metascape.org/gp/index.html), an online tool for Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, and relevant KEGG pathway data were obtained. The top 20 KEGG pathways with the lowest P values were selected and imported into the bioinformatics visualization platform (http://www.bioinformatics.com.cn/) to generate a bubble plot.

Integrating Network Analysis with Metabolomics

To further investigate the therapeutic mechanism of E-GC in removing blood stasis and alleviating liver stasis, an integrated approach combining metabolomics and network pharmacology was employed to identify the core targets involved in the treatment of hepatic lipid stasis. Using Cytoscape 3.7.1, a metabolite–response–enzyme–gene network was constructed. Metabolite-associated targets and common targets were identified at the intersection of E-GC-treated hepatic lipid stasis and metabolite data.

Molecular Docking Validation

The RCSB PDB website (https://www.rcsb.org/) was used to obtain PDB files of core target proteins, while the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) provided the three-dimensional structures of the compounds. Hydrogenation and dehydration were performed, followed by docking validation using AutoDock Vina software (version 1.1.2). Finally, PyMOL (2.5.0) software was used to visualize the molecular docking results.

Analysis of Metabolomics Using UPLC-Q-TOF-MS Data

To enhance data analysis, preprocessing of the raw data was first performed, including the removal of quality control samples with RSD values ≥ 50% and samples with ≥ 50% missing data, imputation of missing values, and data normalization. The normalized data were then imported into the SIMCA 14.1 software for subsequent analysis. Relationships among experimental groups and variations in metabolite types and abundances were evaluated using multivariate statistical methods, including principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), and orthogonal partial least squares discriminant analysis (OPLS-DA).

The OPLS-DA model was validated using permutation tests to assess its reliability and applicability. Metabolites with variable importance in projection (VIP) values > 1.0 in the OPLS-DA model and P < 0.05 in the t-test were selected as differential metabolites, while others were excluded based on predefined criteria. Differential metabolites with fold change (FC) values > 1.2 or < 0.8 were further screened as potential biomarkers.

Following analysis, metabolites that exhibited significant differences between the model and E-GC-treated groups, along with a recovery trend relative to the control group, were classified as differential metabolites associated with the treatment of hepatic lipid stasis. Potential biomarkers were identified using the METLIN and Human Metabolome Database, and pathway analysis of differential metabolites was performed using MetaboAnalyst 6.0 (http://www.metaboanalyst.ca/).

Measurement of Target Metabolites in vivo

A UPLC–MS/MS method for in vivo evaluation was established using the metabolites identified by UPLC–Q-TOF–MS analysis (UPLC–Q-TOF–MS/MS Detection and Metabolite Identification) as markers. Analyses were performed using an Agilent Quantum UPLC–MS/MS system (Agilent, Santa Clara, CA, USA) equipped with an Agilent C18 column (2.1 mm × 150 mm, 1.8 μm), under the chromatographic and mass spectrometric conditions described in Determination of Candidate Compounds by UPLC–MS/MS. Plasma sample pretreatment was conducted according to Pharmacokinetic Studies by UPLC–MS/MS. After preparation of standard and sample solutions, the concentrations of metabolites in collected biological samples (serum, kidney, liver, and intestine) were calculated based on the peak areas of each indicator component. The results were subsequently analyzed to determine significant differences among groups. The findings indicated that E-GC regulates hepatic lipid metabolism, transport, and efflux.

Establishment of Differential Metabolite–Target Connection

The results of pharmacological effects and mechanisms were analyzed to identify the primary drug targets. Groups were classified accordingly, and data from the control, model, and treatment groups (positive control, low-dose, and high-dose groups) were randomly divided into training and validation sets at a ratio of 8:2. Differential metabolites in serum, liver, kidney, and small intestine samples were analyzed, and their significance was evaluated.

A grid search algorithm was used to determine the optimal parameters for the random forest model. To prevent overfitting, the number of trees was initialized at 200, and the minimum number of leaf nodes was set to 5 to identify the optimal parameter combination.

Measurement of the Activity of PXR/FXR–LXR-Related Targets in the Liver by RT-qPCR

Liver samples obtained as described in Animal Experiments were processed as follows. Each liver sample (approximately 50 mg) was homogenized in 0.5 mL of TRIzol reagent and centrifuged for 10 minutes at 12,000 rpm at 4°C. After adding 100 μL of chloroform to the supernatant, the mixture was centrifuged again for 10 minutes at 12,000 rpm at 4°C. The supernatant was then mixed with an equal volume of isopropanol and stored at −40°C overnight. The samples were subsequently thawed and centrifuged for 10 minutes at 12,000 rpm (4°C). After discarding the supernatant, the precipitate was washed twice with 500 μL of 75% ethanol and air-dried for 20 minutes at room temperature. The RNA concentration and purity were measured using a nucleic acid/protein analyzer after dissolving the precipitate in 100 μL of ddH2O. First-strand cDNA was synthesized by reverse transcription. Primers for PXR, CYP7A1, FXR, LXRα, FGFR4, and ABCB1b were used for PCR amplification and expression analysis using cDNA as a template. Primer sequences are listed in Table 3. Relative mRNA expression levels (fold changes) were calculated using the 2^−ΔΔCt method after normalization to 18S.

Table 3.

Primer Sequences Used in qRT-PCR

Gene Name	Forward Primer (5′–3′)	Reverse Primer (5′–3′)
PXR	CCTGTTCCTGAAGATCATGGCTGTC	CCGTCCGTGCTGCTGAATAACTC
CYP7A1	GGCATCTCAAGCAAACACCAT	GCTGTGCGGATATTCAAGGAT
ABCB1b	GTGCTGGTGTGCTCATAG	AACTTCTGCCTAATCTTGTGTA
FXR	TTCGTACCAAATTGCTGCTCAT	AACTCAATGGCAGTCACTCCTG
LXRα	GAGACCAGGGAGGCAACACTT	CATACACTGCATAGCTCGTTCCC
SHP	CGATCCTCTTCAACCCAGATG	AGGGCTCCAAGACTTCACACA
18S	CGGCTACCACATCCAAGGAA	GCGGAATTACCGCGGCT

Measurement of the Activity of PXR/FXR–LXR-Related Targets in the Liver by Western Blotting

Rat liver tissue was collected and thoroughly homogenized, followed by the addition of RIPA lysis buffer (containing 1% PMSF inhibitor) to extract total protein. Protein concentrations were determined using the BCA method. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis was performed for each protein sample. Proteins were then transferred to PVDF membranes and blocked with 5% skim milk at room temperature for 1.5 hours. After incubation with the corresponding primary antibodies overnight at 4°C, the membranes were washed three times with tris-buffered saline containing Tween 20 (TBST) for 10 minutes each. The membranes were then incubated with secondary antibodies at room temperature for 1.5 hours with gentle shaking. After an additional TBST wash, protein bands were visualized using chemiluminescence in an ECL detection system. According to previous researches and references, at least 3 replications were made. Images were analyzed using a gel imaging analysis system. The gray values of target protein bands and corresponding internal reference bands were quantified using ImageJ software, and protein expression levels were calculated.

Evaluation of E-GC Active Ingredient Contribution

A partial least squares structural equation model was constructed to evaluate the contributions of E-GC active ingredients to metabolites. Three latent variables were defined: components (kaempferol, apigenin, and chrysin), targets (FXR, LXR, CYP7A1, and ABCB1), and metabolites (choline, arachidonic acid, and hippuric acid). Data from treatment groups were analyzed to obtain latent variable scores. Causal pathway diagrams were generated, and an evaluation model was established using Goodness of Fit (GOF) values.

Data Analysis

All data were statistically analyzed using SPSS 27.0. Normality of distribution was assessed using the Shapiro–Wilk test, and homogeneity of variances was evaluated using Levene’s test. For data that met both normality and homogeneity of variances assumptions, comparisons among multiple groups were performed using one-way analysis of variance (ANOVA), followed by the least significant difference (LSD) post hoc test for pairwise comparisons. For data that did not meet these assumptions, the Kruskal–Wallis test was used for overall group comparisons, followed by the Mann–Whitney U-test with Bonferroni correction for pairwise comparisons. All results are expressed as mean ± standard deviation (SD). Statistical significance was set at P < 0.05.

Results

E-GC Can Significantly Alleviate Symptoms of Liver Damage Caused by Hyperlipidemia

According to the subgroup animal experiments, serum and liver levels of TG and cholesterol in the model group were significantly higher than those in the control group (P < 0.05) (Figure 1). Electron microscopy of pathological sections (Figure 1) revealed disordered cellular arrangement, increased cell size, and indistinct boundaries in hepatocytes. In some cells, nuclei were absent, and diffuse hepatocellular steatosis with an increased number of vacuolated lesions was observed. These findings confirmed successful model establishment and were consistent with the pathophysiology of hepatic lipid stasis. Compared with the model group, both high- and low-dose E-GC administration groups exhibited significant reductions in blood lipid levels (P < 0.05), improved hepatocyte arrangement, decreased fatty vacuoles, improved cellular morphology, and reduced pathological alterations. The results of blood lipid measurements and pathological observations demonstrated a dose-dependent trend. These findings indicate that E-GC effectively reduces hepatic lipid accumulation, inhibits lipid synthesis, and promotes lipid transport and efflux. Compared with the single ECG or GK treatment groups, although both showed reductions in lipid levels relative to the model group, their effects were inferior to those of the E-GC combination.

Figure 1.

After E-GC intervention, the blood lipid levels and multiple biochemical indicators related to lipid metabolism and liver function in rats returned to normal. (A) Biochemical analyses of serum TC and TG. (B) Biochemical analyses of serum HDL-C and LDL-C. (C) Biochemical analyses of liver TC and TG. (D) H&E staining images of the liver (200× and 400×) (E) Oil red staining images of the liver (200× and 400×). Data are expressed as the mean ± SD for each group of rats (n = 6). Compared with the control group, ###P < 0.001, #P < 0.05; compared with the model group,***P < 0.001, **P < 0.01, *P < 0.05.

No significant differences were observed between the positive control group and the high- and low-dose E-GC groups (P > 0.05), indicating that E-GC exhibited effects comparable to simvastatin and superior to single-component treatments. Pathological analysis further demonstrated that E-GC achieved greater improvement in hepatocyte lipid stasis than the positive control drug, suggesting a more comprehensive therapeutic effect. H&E staining of the liver revealed that, compared to the Control group, hepatocytes in the Model group exhibited disordered arrangement, increased cell size, and indistinct cell boundaries, with some nuclei disappearing. Diffuse hepatic steatosis was evident, accompanied by numerous vacuolar lesions. These findings confirm successful model establishment, consistent with the pathological features of hepatic lipid accumulation. After treatment, hepatocyte arrangement became more orderly, the number of fatty vacuoles significantly decreased, and cellular morphology improved, indicating a marked reduction in lesion severity. Following Oil Red O staining, lipid droplets in liver tissue were stained red while cell nuclei appeared blue. As shown in Figure 1, hepatocytes in the Control group were neatly arranged without red lipid droplets. In contrast, the Model group showed disorganized hepatocyte arrangement and abundant red lipid droplets, clearly indicating severe hepatic lipid accumulation. Compared to the Model group, all treated groups demonstrated reduced red lipid droplets, suggesting that drug administration effectively alleviated hepatic lipid accumulation. Quantitative analysis of lipid droplet area using ImageJ software (Figure 1) showed that, compared to the Control group, the liver lipid droplet area in the Model group was significantly increased (P < 0.01). Compared to the Model group, the liver lipid droplet areas in all treated groups were significantly reduced (P < 0.01).

Screening of Metabolites Related to PXR/FXR–LXR Induced Hepatic Lipid Stasis

UPLC-Q-TOF-MS detection and database identification revealed 61 metabolites in blood samples from each experimental group (Figure 2). Subsequent pathway analysis identified seven metabolites associated with the PXR/FXR–LXR axis. Using these metabolites as indicators, reference standards were employed to optimize UPLC-MS detection conditions and ensure accurate quantification. As results indicated, although Docosahexaenoic acid and Hippuric acid were found in serum samples, their signals in liver samples were not detectable; based on detectability by UPLC-MS and their ability to reflect intergroup differences, N-acetylserotonin, choline, arachidonic acid, deoxycholic acid, and cholic acid were selected as metabolic markers. These markers were used to characterize the regulatory effects of E-GC on the PXR/FXR–LXR axis in alleviating blood stasis and hepatic lipid stasis. Further analysis was conducted to determine correlations and the relative contributions of each metabolite to the mechanism of action. Table 4 lists the mass-to-charge ratios and related mass spectrometry parameters for each metabolite, while TOF-MS and LC-MS profiles are shown in Figures S1 and S2.

Figure 2.

The analysis of the target, main active components and metabolic pathways showed that E-GC contained multiple efficacy-related components, and the regulatory sites and pathways mainly focused on liver cell regeneration, lipid metabolism, and inflammatory response. (A) Venn diagram showing the intersection of disease targets and active ingredient targets. (B) Protein-protein interaction (PPI)network of the intersecting targets. (C) Bubble chart of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis (Top 10). (D) Compound-target-pathway network diagram.

Table 4.

The Charge Ratios of Each Metabolic Substance and Related Mass Spectrometry Parameters

Metabolite	Molecular Formula	Rt (min)	m/z	Collision Energy
Choline	C5H13NO	0.45	104.1, 60.2, 58.2	20, 36
N-Acetylserotonin	C12H14N2O2	2.91	219.1, 160, 131.9	16, 30
Hippuric acid	C9H9NO3	2.96	178.10, 134.1, 77.2	8, 16
Arachidonic acid	C20H32O2	5.18	305.24, 257.3	24
Deoxycholic acid	C24H40O4	5.84	437.3, 345.3	48
Cholic acid	C24H40O5	4.96	407.3, 345.4, 289.2	34, 44
Docosahexaenoic acid	C22H32O2	8.55	327.2, 283	10

Selection of Differential Metabolites Based on Metabolomic Data Analysis

According to PCA results, data points within each experimental group showed high clustering, while clear separation was observed between groups. Although differences among certain groups were not statistically significant (P > 0.05), all of the groups exhibited significant differences compared with the model group (P < 0.05), indicating that modeling and drug intervention substantially affected metabolic marker levels. The results of PLS-DA and OPLS-DA were generally consistent (Figures S3–S5). The heatmap (Figure 3) generated from metabolite profiling demonstrated that differentially abundant metabolites identified by UPLC-Q-TOF-MS exhibited significant differences in serum among experimental groups. These metabolites more accurately reflected disease progression and physiological alterations following drug intervention. Pathway analysis indicated that these metabolites were primarily involved in cholesterol, amino acid, glycerophospholipid, and arachidonic acid metabolism (Figure 2 and Table 5). Serum serves as a key medium for the transport of endogenous substances and inter-organ interactions; therefore, its metabolite profile provides important biological insights. Changes in these metabolites were also measured in the liver, kidney, small intestine, and other organs closely associated with lipid metabolism and the PXR/FXR–LXR pathway, and the resulting data were analyzed.

Figure 3.

The determination of metabolic markers in serum and various organs showed that E-GC restored the normal levels of endogenous substances related to lipid metabolism and liver injury in animals, and this efficacy was most significant in the liver. (A) Serum; (B) Liver; (C) Kidney; (D) Small intestine (n=6). Compared with the control group, ###P < 0.001,##P < 0.01, #P < 0.05; compared with the model group, ***P < 0.001, **P < 0.01, *P < 0.05.

Table 5.

Metabolic Pathway Analysis

Pathway Name	Impact
Phenylalanine metabolism	0.36
Phenylalanine, tyrosine and tryptophan biosynthesis	0.50
Primary bile acid biosynthesis	0.03
Glycerophospholipid metabolism	0.03
Tryptophan metabolism	0.04
Arachidonic acid metabolism	0.29

Figure 3 shows that all of the differential metabolites changed significantly following model induction and were markedly influenced by drug treatment. Literature review and TOF-MS analysis indicated that cholic acid is associated with target disorders; however, neither the positive control drug nor E-GC showed a significant regulatory effect on this metabolite across matrices, suggesting limited relevance to therapeutic effects. Serum levels of docosahexaenoic acid varied significantly among groups, and the recovery effect of E-GC was weaker than that of the positive control group. However, no significant differences were observed in liver and kidney tissues among the groups. In contrast, docosahexaenoic acid levels in the small intestine of the E-GC group were significantly higher than those in the positive control group and exhibited a dose-dependent pattern, indicating multi-organ and multi-target regulatory effects of E-GC.

Choline levels showed significant differences after drug intervention only in the liver, where the effect of E-GC was superior to that of the positive control drug. This suggests that E-GC primarily regulates choline metabolism through modulation of targets within the PXR/FXR–LXR pathway in the liver. For other metabolites, E-GC exhibited effects comparable to or greater than those of the positive control drug, further supporting the superior therapeutic efficacy and mechanism of this formulation.

Confirmation of Major Metabolic Markers

Based on UPLC-MS determination results for each metabolite, combined with literature reports and metabolic database analysis, a relevant metabolic mechanism map (Figure 4) was constructed. This map clarifies the positions, roles, and interactions of each metabolite in hepatic lipid metabolism, accumulation, and transport, and describes their regulatory mechanisms within the PXR/FXR–LXR framework. The results of pharmacological effects, molecular mechanisms, and metabolomics were imported into SPSS and other statistical software to analyze physiological processes related to phlegm and blood stasis described in TCM theory. The optimized random forest algorithm using grid search showed that the optimal number of trees was 240, the minimum number of leaf nodes was 6, and the maximum number of splits per tree was 10. The prediction accuracy of the training and test sets was 0.917 and 0.75, respectively. The importance ranking and confusion matrix of metabolites are presented in Figures S3–S5. Based on these analyses, arachidonic acid, choline, deoxycholic acid, N-acetylserotonin, and cholic acid were identified as major metabolic markers.

Figure 4.

Metabolomics combined with pathway analysis identified 59 differential metabolites, enriched in 4 metabolic pathways, mainly involving lipid metabolism, energy metabolism, and inflammatory response physiological processes. (n=6). (A) Bubble chart of metabolic pathway enrichment analysis. (B) Heatmap of cluster analysis.

Results of Hepatic Lipid Metabolism Network Analysis

A total of 36 E-GC components were screened using the SwissADME tool (Table S1). Target prediction identified 327 potential targets corresponding to these components. Additionally, 1142 targets related to hepatic lipid stasis and hyperlipidemia were obtained from relevant databases. Venn analysis identified 238 overlapping targets as potential targets of E-GC against hepatic lipid stasis (Table S2 and Figure 2). After importing the 238 common targets and 36 components into Cytoscape 3.7, a herb–component–target network was constructed. Degree analysis identified the top key components, including apigenin, hesperidin, kaempferol, and ginkgolide A/B/C, as major contributors to the therapeutic effects of E-GC.

KEGG pathway and Gene Ontology (GO) enrichment analyses were performed on the identified common targets using the Metascape database. Based on KEGG analysis, 20 representative signaling pathways were identified, among which the mitogen-activated protein kinase (MAPK), phosphoinositide 3-kinase–protein kinase B (PI3K-AKT), and Ras signaling pathways were predominant. GO analysis categorized results into biological processes, molecular functions, and cellular components, including protein phosphorylation responses, receptor complexes, and protein kinase activity.

Combined Analysis of Network Pharmacology and Metabolomics

A metabolite–response–enzyme–gene network was constructed using Cytoscape to further elucidate the therapeutic mechanism of E-GC in hepatocyte lipid stasis and to identify key targets (Figure 4). A total of 115 metabolite-associated targets were identified. Subsequently, 49 overlapping targets between network pharmacology and metabolomics were obtained through Venn analysis (Table S3). These 49 targets were imported into Cytoscape to construct a protein–protein interaction (PPI) network. Degree analysis identified five core targets with high connectivity: MAPK3, PTGS2, CYP3A4, MAPK1, and FGFR1. MAPK3 and MAPK1 correspond to ERK1 and ERK2, respectively, which are associated with lipid synthesis and inflammation. PXR regulates lipid metabolism and inflammatory responses, and upon activation, phosphorylated ERK is stimulated, further modulating these processes. LXR and FXR are downstream of ERK and regulate CYP7A1 expression, thereby influencing lipid synthesis and metabolism. A component–target–metabolite network was subsequently constructed using Cytoscape (Figure 5).

Figure 5.

The “metabolite-reaction-enzyme-gene” network indicates that E-GC mainly intervenes in pathways such as amino acid metabolism, bile acid biosynthesis, arachidonic acid metabolism, and glycerophospholipid metabolism, with the main biomarkers being CA, DCA, Cho, AA, and N-Ace, related to 115 genes, reflecting the biological information process from gene to metabolite.

Results of Molecular Docking Validation

As described in Screening of Metabolites Related to PXR/FXR–LXR Induced Hepatic Lipid Stasis, molecular docking was performed between five key components and five core targets. Docking scores and binding conformations are presented in Table 6 and Figure 6. The core active components exhibited binding affinities generally lower than −5.0 kcal/mol, indicating stable interactions. These results suggest that the key components may exert therapeutic effects by competitively binding to target receptors, thereby playing an important role in the treatment of hepatic lipid stasis.

Table 6.

Molecular Docking Parameter

Crucial Component	Chemical Formula	Target Protein	PDB ID of Target Protein	Binding Energy (-Kal/mol)
Apigenin	C15H10O5	MAPK1	6SLG	−3.35
		MAPK3	2ZOQ	−3.11
		PXR	4SOS	−6.01
		FXR	3L1B	−5.15
		LXR	3KFC	−4.11
Naringin	C15H10O4	MAPK1	6SLG	−3.79
		MAPK3	2ZOQ	−3.84
		PXR	4SOS	−6.14
		FXR	3L1B	−5.03
		LXR	3KFC	−6.08
GA	C20H24O9	MAPK1	6SLG	−6.49
		MAPK3	2ZOQ	−8.55
		PXR	4SOS	−10.24
		FXR	3L1B	−7.77
		LXR	3KFC	−8.42
		MAPK1	6SLG	−5.97
		MAPK3	2ZOQ	−4.81
		PXR	4SOS	−6.63
GB	C20H24O10	FXR	3L1B	−5.76
		LXR	3KFC	−5.80
		MAPK1	6SLG	−8.4
		MAPK3	2ZOQ	−8.9
GC	C20H24O11	PXR	4SOS	−10.43
		FXR	3L1B	−8.19
		LXR	3KFC	−9.2

Notes: PDB ID: MAPK1(6SLG); MAPK3(4QTB); PXR(4XHD); FXR(3FLI); LXR(6K9G).

Figure 6.

The molecular docking results of the main components of E-GC showed that each active component could bind to more than one core target, and covered all core targets. (A) PXR; (B) FXR; (C) MAPK1; (D) MAPK3; (E) LXR; (1) Apigenin; (2) Chrysin; (3) Kaempferol; (4) Quercetin.

Results of the Determination of Candidate Compounds in E-GC

UPLC-TOF-MS analysis identified 66 compounds in serum samples of E-GC, including prototype compounds and metabolites. UPLC-MS/MS was used for quantitative determination of candidate compounds screened in Results of Hepatic Lipid Metabolism Network Analysis (Table S4). The concentrations of these components in E-GC met the requirements for therapeutic efficacy and were significantly higher than those observed in single-component treatments reported in previous studies.9,10

Results of Pharmacokinetic Studies

Figure S6 illustrates the blood concentration–time curves for each component. Based on these data, pharmacokinetic analysis was conducted, and the relevant parameters are summarized in Table 7. The candidate compounds exhibited appropriate pharmacokinetic properties and favorable bioavailability. Compared with previous studies on single administration of ECG or GK, these parameters showed significant improvement, indicating that combined administration enhances therapeutic efficacy.

Table 7.

Results of Pharmacokinetic Studies

Component	Cmax (ng/mL)	Tmax (h)	T1/2 (h)	MRT (h)	AUC0-t (hng/mL)	AUC0-∞ (hng/mL)
Naringin	231.493	1.00	1.572	4.046	526.328	556.312
Apigenin	170.407	1.000	0.955	3.453	395.000	405.539
Ginkgolide A	70.286	1.500	1.026	3.197	195.796	199.261
Ginkgolide B	227.847	1.50	1.040	3.554	543.014	557.545
Ginkgolide C	45.432	2.00	3.119	5.502	185.676	209.493

Key Targets of E-GC’s Regulatory Effect on Hepatic Lipid Metabolism

PCR analysis of liver samples showed that PXR expression in the model group was significantly higher than that in the control group (P < 0.05; Figure 7), indicating marked upregulation following model induction. This expression level was reduced in the positive control, ECG, GK, and E-GC treatment groups, with no significant differences compared to the control group. Further, PXR expression in the high-dose group was higher than that in the low-dose group, indicating a dose-dependent regulatory effect. FXR, LXR, CYP7A1, and ABCB1 expression exhibited similar trends (Figure 7). After modeling, expression levels of these targets changed significantly compared with the control group, and were restored to near-normal levels following E-GC treatment in a dose-dependent manner. E-GC demonstrated regulatory effects on target expression that were comparable to or greater than those of ECG, GK, and the positive control drug. Western blot analysis further confirmed these findings, showing that expression levels of all of the measured targets in the model group differed significantly from those in the control group, while E-GC-treated groups showed expression levels comparable to or better than those of the positive control group (Figure 8). Combined PCR and Western blot results confirm that E-GC exerts significant regulatory effects on targets associated with hepatic lipid stasis within the PXR/FXR–LXR axis, providing a biological basis for the therapeutic efficacy of E-GC.

Figure 7.

E-GC could restore the mRNA expression of each core target to the normal level, and the effect was better than that of the positive drug or similar. Data are expressed as the mean ± SD for each group of rats (n = 6). (A) CYP7A1; (B) FXR; (C) LXR; (D) ABCB1b; (E) PXR. Compared with the control group, ##P < 0.01, #P < 0.05; compared with the model group, **P < 0.01, *P < 0.05.

Figure 8.

The protein expression results of each core target were consistent with the mRNA trend, further verifying the good regulatory effect of E-GC. Data are expressed as the mean ± SD for each group of rats (n = 3). (A) CYP7A1; (B) FXR; (C) LXR; (D) ABCB1b; (E) PXR. Compared with the control group, ##P < 0.01, #P < 0.05; compared with the model group, ***P < 0.001, **P < 0.01, *P < 0.05.

Contribution Evaluation Results of E-GC Active Ingredients

The optimal model with the highest Goodness of Fit (GOF) value was identified by evaluating seven different metabolite combinations. A pathway framework illustrating the relationships among components, targets, and metabolites was constructed (Figure 4). Heatmaps were used to display the cross-loadings of observed variables on latent variables, enabling the derivation of influence values between metabolites. Score values of each sample across different groups were visualized in three dimensions. While the score density plot indicated that the high-dose E-GC group outperformed both the positive control and low-dose groups (Figure 9). By integrating the above-mentioned results, contribution of each compound in E-GC on the therapeutic effect and be evaluated. Which is (from highest to relatively lower) apigenin>naringin>ginkgolide C>ginkgolide A>ginkgolide B.

Figure 9.

There is a certain degree of mutual influence among biomarkers, indicating that E-GC has a strong overall linkage effect, exerting therapeutic efficacy through multiple points, multiple pathways, and multiple metabolites. (n=6). (A) Serum, (B) Liver, (C) Kidney, (D) Small intestine.

Discussion

Significant Improvement in Liver Lipid Stasis by E-GC Intervention

The results of blood lipid measurements in the E-GC-treated groups were consistent with pathological findings. The synthesis of TG and cholesterol was significantly inhibited following treatment, and this effect was more pronounced than that observed with ECG or GK alone. Accumulated lipids in the body, particularly in the liver, were decomposed, excreted, or transported to the intestine. Hepatocyte morphology gradually returned to a state comparable to that of the control group, likely due to the alleviation of lipid stasis, reduced liver injury, and enhanced recovery. This effect may be associated with decreased inflammatory responses and reduced accumulation of inflammatory factors mediated by regulation of the PXR/FXR–LXR axis.11,32 However, no significant differences were observed in pathological sections of the intestine among experimental groups, possibly due to the relatively mild intestinal lipid stasis under hyperlipidemic conditions. The effects of hyperlipidemia on the small intestine require further investigation using molecular and metabolomic approaches.33,34

Overall, E-GC exhibited a stronger recovery effect on liver function than the positive control drug, indicating its superior capacity to promote lipid transport and efflux while inhibiting lipid synthesis. The enhanced efficacy of the couplet reflects the holistic and comprehensive therapeutic characteristics of TCM. The biochemical indicators and histopathological observations in the E-GC treatment groups generally showed consistent trends: lipid synthesis was significantly suppressed after administration, while lipid accumulation, particularly in the liver was promoted for degradation and excretion. Meanwhile, hepatocyte morphology gradually recovered to levels close to those of the control group, likely due to alleviation of hepatic lipid accumulation, which may help prevent liver damage and promote liver recovery. This improvement could also be associated with YL’s effects on relevant targets and signaling pathways, leading to reduced inflammatory responses and decreased accumulation of pro-inflammatory factors. H&E staining results indicated that although simvastatin reduced vacuolar degeneration in hepatocytes, its effect was weaker than that of the E-GC group. Oil Red O staining revealed that YL effectively alleviated intracellular lipid accumulation; compared to the Model group, red lipid droplets were significantly reduced (P < 0.01) after treatment, and the number of red lipid droplets in the E-GC treated group was lower than in the Positive group. Histopathological analysis further demonstrated that E-GC improved hepatic lipid accumulation more effectively than the positive control drug, indicating a more comprehensive therapeutic effect of E-GC. Overall, E-GC exhibited stronger liver recovery effects than the positive control, suggesting that E-GC not only inhibits lipid synthesis but also promotes lipid metabolism, transport, and excretion, along with notable anti-inflammatory activity. These findings reflect E-GC’s ability to alleviate lipid accumulation in hepatocytes and exert functions of removing blood stasis and regulating liver function, highlighting the holistic and multifaceted therapeutic potential of traditional Chinese medicine. The effects were particularly evident in the E-GC High group, thus the non-targeted metabolomics analysis subsequently focused on this group.

Effects of E-GC on Hepatic Lipid Stasis and Metabolism via Regulation of the PXR/FXR–LXR Axis

Previous studies by our group have explored the role of PXR in lipid metabolism regulation and the effects of active components in E-GC. In addition to the previously described mechanisms, PXR activation can indirectly influence FXR and LXR activity, primarily through regulation of CYP3A4 downstream of PXR. However, this indirect mechanism alone may not be sufficient to exert strong regulatory effects on the FXR–LXR axis and its downstream targets, suggesting that certain E-GC components may directly interact with both receptors. These components may enhance FXR activity or directly bind to FXR, thereby influencing LXR activity, regulating blood lipid levels, and promoting the efflux of accumulated hepatic lipids.35–37 Following E-GC intervention, upregulation of FXR expression was more pronounced than that of LXR, as demonstrated by PCR and Western blot analyses. This suggests that a greater number of E-GC components act directly on FXR compared with LXR. These findings indicate that FXR plays a more prominent role in regulating fatty acid and cholesterol metabolism and their related pathways. Consequently, E-GC exerts superior therapeutic effects in alleviating hepatic lipid stasis through coordinated regulation of PXR and FXR.

Pharmacological Substances Indicating the Key Chemical Components of E-GC

In this study, core components of E-GC were identified using network pharmacology and validated through molecular docking, demonstrating strong binding affinities between these components and their respective targets. Through activation or inhibition of these targets, the expression of PXR, FXR, CYP7A1, and other related genes is modulated, confirming that E-GC alleviates blood stasis and hepatic lipid stasis through the PXR/FXR–LXR axis. Apigenin, naringin, and kaempferol are flavonoids that regulate lipid synthesis, metabolism, transport, and inflammation, thereby influencing multiple physiological processes. Flavonoids such as naringin and apigenin, as well as terpenoids including ginkgolide A/B/C, regulate lipid metabolism through modulation of LXR-related pathways.38

The observed metabolite changes, component content, and pharmacokinetic profiles collectively contributed to the overall therapeutic effects. Drug concentration and systemic exposure are critical determinants of efficacy, as sufficient accumulation of active components is required to achieve therapeutic outcomes. These components must reach adequate concentrations in vivo to ensure effective biological activity.

Relationship Between Metabolites and the PXR/FXR–LXR Axis and Their Changes Following E-GC Intervention

Metabolomics analysis, combined with literature review and pharmacodynamic results (including blood lipid levels, pathological observations, and target expression changes within the PXR/FXR–LXR axis), demonstrated that E-GC significantly regulates endogenous metabolites and precursors associated with lipid metabolism, particularly cholesterol metabolism, as well as inflammatory mediators related to lipid accumulation. The overall effect of E-GC is to reduce cholesterol synthesis and accelerate cholesterol efflux, indicating its significant role in lipid and energy metabolism. This enhances functional activity in the liver, intestine, and other organs, thereby promoting lipid transport. Choline reduces cholesterol synthesis from fatty acids and counteracts the decrease in CYP7A1 and FXR expression induced by hyperlipidemia.37 Arachidonic acid, a major hepatic fatty acid, inhibits PXR and ABCB1 expression.3 Its derivatives promote inflammation, and inflammatory cytokines such as TNF-α and IL-1 inhibit CYP7A1 mRNA expression.39 Cholic acid and deoxycholic acid, as free bile acids, activate FXR receptors, upregulate FXR expression, reduce hepatic lipid accumulation, and decrease TC and TG levels.40,41 N-acetylserotonin, a metabolite of the tryptophan pathway, participates in cellular energy metabolism,42 while hippuric acid, derived from phenylalanine metabolism, is involved in energy metabolism and the tricarboxylic acid cycle.43

The changes in arachidonic acid levels differed across organs, with more pronounced variation observed from the kidney to the small intestine than from blood to liver. This may be associated with intestinal inflammation and increased lipoxin production in the kidney. Arachidonic acid, as a precursor of lipoxin analogs, has been reported to reduce blood TC and LDL-C levels. In the intestine, resolution of inflammation involves the conversion of prostaglandin E2 into lipoxins derived from arachidonic acid via COX1/COX2 and lipoxygenase pathways.44 Leukotriene B4 (LTB4), synthesized from arachidonic acid via leukotriene A4 hydrolase and related enzymes, promotes lipid synthesis in hepatocytes through the LTB4/Ltb4r1 pathway, contributing to nonalcoholic fatty liver disease.45,46 These mechanisms may explain the observed changes in arachidonic acid levels.

This study identified five core targets through integrative analysis: MAPK3, PTGS2, MAPK1, and CYP3A4. PTGS2, also known as prostaglandin-endoperoxide synthase 2 (COX-2), is a key factor in inflammation and an important target in the arachidonic acid metabolic pathway, catalyzing the conversion of arachidonic acid into prostaglandins and promoting the production of inflammatory mediators.47 CYP3A4 is a cytochrome P450 enzyme and a major drug-metabolizing enzyme involved in Phase I metabolism of numerous drugs. Additionally, CYP3A4 is regulated by its upstream gene PXR and influences multiple physiological and pathological processes. CYP3A4 can metabolize certain drugs into active compounds that affect related genes and targets within the body, thereby modulating physiological functions. Studies have shown that when lipid metabolism is disrupted, PXR becomes activated, subsequently regulating CYP3A4 expression and thus controlling lipid metabolism. MAPK3 and MAPK1 correspond to ERK1 and ERK2, respectively, both of which are associated with lipid formation and inflammatory responses.48 PXR regulates lipid metabolism and inflammatory reactions; upon receptor activation, phosphorylated ERK is activated, further modulating inflammation and lipid metabolism.49 Research indicates that LXR and FXR are located downstream of ERK, and their expression is regulated via ERK, which in turn controls the expression of downstream genes such as CYP7A1, thereby influencing fat synthesis and metabolism.40 Based on KEGG enrichment analysis showing involvement in the MAPK signaling pathway and MAPK1 as a central target, it is predicted that E-GC may alleviate hepatic lipid accumulation by modulating the ERK signaling pathway within the MAPK cascade. Furthermore, by targeting upstream and downstream genes of ERK—PXR, FXR, LXR, and CYP7A1—E-GC may regulate lipid metabolism in the body. Therefore, it is inferred that E-GC likely alleviates hepatic lipid accumulation through the PXR-FXR-ERK signaling pathway.

Metabolic Markers Regulate Hepatic Lipid Accumulation, Transport, and Efflux

Compared with previous studies on Yinlan and its individual components, the differential metabolites identified in this study exhibited more pronounced changes in species, in vivo concentrations, and responses to pharmacological intervention.7 Serum is an important component of the circulatory system in the body, capable of reflecting the overall metabolic status of the entire body. It contains metabolites from various tissues and organs and covers multiple metabolic pathways. Therefore, qualitative analysis is conducted on serum. The liver is the site of lipid accumulation in liver cells and an important site for lipid metabolism. Through quantitative analysis in the liver, specific changes in metabolites can be more accurately studied, revealing the occurrence mechanism of diseases and the action targets of drugs.50–52 Therefore, quantitative analysis of the liver is selected. This study conducted quantitative analysis of 5 metabolites in the liver, namely CA, DCA, Cho, AA, and N-Ace. The analysis results showed that compared with the Control group, although there was no statistical difference (P > 0.05) in AA in the Model group, the content of arachidonic acid in the Model group showed an increasing trend compared with the Control group. The content of CA and DCA in the Positive group did not have a statistical difference compared with the Model group (P > 0.05), indicating that simvastatin has a very small impact on the levels of these metabolites. The recovery of metabolite content levels in the E-GC High group was stronger than that in the E-GC Low group, indicating that E-GC has a dose-dependent effect. Based on the qualitative and quantitative results, these 5 metabolites were finally selected as biomarkers for E-GC to alleviate liver cell lipid accumulation. These biomarkers serve as the biological basis for E-GC to alleviate liver cell lipid accumulation and exert the effect of removing blood stasis and soothing the liver.39–43,50–52

Overall, E-GC exerts extensive regulatory effects on multiple physiological processes, including lipid metabolism, hepatic lipid transport, and inflammation, through modulation of the PXR/FXR–LXR axis. It promotes cholesterol and fatty acid degradation, regulates glucose, lipid, and energy metabolism, and reduces inflammatory responses. The multi-target and multi-mechanism actions of E-GC in alleviating hyperlipidemia-induced hepatic lipid accumulation and liver injury are difficult to achieve with conventional lipid-lowering drugs.

Conclusions

This study established methods for the discovery and evaluation of active components in E-GC, as well as biomarkers representing the therapeutic advantages of the couplet. These findings were supported by analyses of mechanisms, key targets, and metabolomics related to PXR/FXR–LXR regulation of lipid synthesis, transport, and accumulation in the liver. Through network pharmacology, molecular docking, component content determination, and pharmacokinetic screening, six active components of E-GC that contribute to reducing hepatic lipid deposition were identified. These components, namely, naringin, kaempferol, ginkgolide A/B/C, and apigenin, were ranked according to their degree of contribution. The functional mechanisms of E-GC and its components were further elucidated through their effects on in vivo levels of hippuric acid, arachidonic acid, docosahexaenoic acid, and choline. These components also influence the levels of other metabolites that reflect the progression of lipid deposition in hepatocytes. The hierarchical and integrative framework of efficacy–component–mechanism–metabolism was applied to clarify the therapeutic effects of E-GC in treating hepatic lipid stasis, providing important evidence for its clinical application, targeted efficacy, and quality evaluation.

Funding Statement

This work was supported by the National Natural Science Foundation of China (No. 81903761).

Declaration of Generative AI and AI-Assisted Technologies in the Manuscript Preparation Process

During the preparation of this work, the authors used an AI-based proprietary language-editing tool built by LetPub in order to improve the readability and language of the manuscript. After using this service, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Abbreviations

NAFLD, Non-alcoholic fatty liver disease; NASH, Non-alcoholic steatohepatitis; TCM, Traditional Chinese Medicine; E-GC, Exocarpium Citri Grandis–Folium Ginkgo couplet; LXR, Liver X Receptor; FGF1, Fibroblast Growth Factor 1; ABCB1b, ATP-binding cassette, sub-family B (MDR/TAP), member 1B; ERK, Extracellular signal-Regulated Kinase; HMG-CoA, 3-hydroxy-3-methylglutaryl-coenzyme A; LPL, Lipoate-protein ligase; PPAR, Peroxisome proliferator-activated receptor; FXR, Farnesoid X Receptor; SHP, Src homology phosphatase; CYP7A1, Cytochrome P450, family 7, subfamily A, polypeptide 1; PXR, Pargnane X Receptor; TC, Total cholesterol; TG, Triacylglycerol.

Data Sharing Statement

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Ethics Approval and Consent to Participate

A total of 30 male Sprague-Dawley (SD) rats, weighing 200 ± 20 g, were used in this study. All of the animals were obtained from the Guangdong Medical Laboratory Animal Center (Production License No.: SCXK (Yue) 2019-0002; Quality Certificate No.: 44007200105943). The rats were housed in the SPF-grade animal facility of the Fifth Clinical Medical College, Guangzhou University of Chinese Medicine, with the Facility Use License No.: SYXK (Yue) 2020-0059. The study protocol was approved by the Animal Ethics Committee of the Fifth Clinical Medical College, Guangzhou University of Chinese Medicine, under Approval No.: 049157. All animal experiments were conducted in accordance with the Guidelines for the Ethical Review of Laboratory Animal Welfare (GB/T 35892-2018). The animals were anesthetized with isoflurane and euthanized by exsanguination under deep anesthesia. All procedures were performed in accordance with relevant veterinary guidelines. Consent to participate is not applicable, as this study did not involve human subjects. Consent to participate is not applicable, as this study did not involve human subjects.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no competing interests. This paper has been uploaded to papers.ssrn.com as a preprint: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7100978