Ginseng-containing Shentao Ruangan (STR) formula alleviates hepatic fibrosis via hepatocellular Nrf2-ROS-TGF-β axis regulating hepatocyte-HSC crosstalk: Integrating single-cell RNA sequencing and multi-tiered validation Authors: Shangyi Huang, Ruisheng Zhou, Jing Li, Yi Zheng, Lunwei Yang, Tianqi Gao, Yi Yuan, Dailin Wu, Yafei Shi, Dayong Zheng, Daihan Zhou, Ying Tang Published: 2026 ABSTRACT AbstractBackgroundThe ginseng-containing Shentao Ruangan (STR) formula is clinically used for chronic liver conditions, including hepatic fibrosis. Guided by traditional Chinese medicine (TCM) principles, it targets the “deficiency leading to accumulation (Xu Sun Sheng Ji)" pathology in hepatic fibrosis via the “strengthening healthy qi and dispelling pathogenic factors (Fu Zheng Qu Xie)" mechanism.AimsTo elucidate how STR attenuates hepatic fibrosis, focusing on its regulation of the fibrotic liver microenvironment.MethodsSingle-cell RNA sequencing profiled the fibrotic liver microenvironment; GEO analysis confirmed reactive oxygen species (ROS) as a central fibrotic driver. In vitro/in vivo experiments assessed STR's effects on ROS and fibrosis. Mechanisms were explored via RNA sequencing, network pharmacology, molecular docking, Nrf2 inhibitors and Nrf2-knockout (Nrf2−/−) mice.ResultsSingle-cell sequencing identified hepatocyte-hepatic stellate cell (HSC) crosstalk as a core fibrotic feature, with hepatocellular metabolic dysfunction and elevated ROS promoting fibrosis (GEO-validated). STR alleviated fibrosis and suppressed excessive hepatic ROS in vitro and in vivo. Integrated analyses identified ROS modulation as STR's key anti-fibrotic mechanism, with Nrf2 as the critical target. STR activated hepatocellular Nrf2, alleviating oxidative stress, reducing TGF-β expression and secretion, and thus inhibiting HSC activation and extracellular matrix deposition. Molecular docking, Nrf2 inhibitors and Nrf2−/− mouse studies confirmed Nrf2's essential role.ConclusionThis study links hepatocyte-HSC crosstalk/ROS-TGF-β-driven activation in hepatic fibrosis to TCM's “Xu Sun Sheng Ji”, while STR's Nrf2-mediated antioxidant and anti-fibrotic actions embody “Fu Zheng Qu Xie”, integrating TCM with modern medicine. STR, a ginseng-containing formula, is a promising targeted fibrosis therapy, advancing TCM-modern medicine integration.Keywords: Ginseng-containing Chinese medicine formula, Hepatic fibrosis, Nrf2-ROS-TGF-β axis, Hepatocyte-HSC crosstalk, Single-cell RNA sequencing, TCM Introduction Hepatic fibrosis, the common pathological endpoint of chronic liver injury, is characterized by persistent hepatic stellate cell (HSC) activation, excessive extracellular matrix (ECM) deposition, and progressive disruption of hepatic parenchyma, with progression leading to hepatocellular carcinoma (HCC) and hepatic decompensation [[1], [2], [3]]. A recent large biopsy cohort study showed advanced fibrosis patients exhibit significantly higher all-cause mortality [4], with no FDA-approved therapies to directly inhibit hepatic fibrosis. During hepatic fibrosis progression, perturbed liver microenvironmental homeostasis impairs hepatocyte function, elicits multiple cellular responses, alters cell-cell interactions, and ultimately drives fibrotic advancement [5]. Thus, restoring this homeostasis has emerged as a key strategy to promote fibrosis resolution [5,6]. Oxidative stress is a well-recognized driver of fibrotic progression, mediates hepatocellular injury and activates HSC proliferation, migration, and ECM deposition [[7], [8], [9]]. Nuclear factor erythroid 2-related factor 2 (Nrf2) is a core regulator of cellular redox homeostasis [10]. Under oxidative stress, Nrf2 dissociates from its inhibitor Keap1, translocates to the nucleus, and induces the expression of cytoprotective genes such as heme oxygenase-1 (HMOX1/HO-1) [11]. While Nrf2 activation is a potential antifibrotic strategy, existing studies primarily target its role in HSCs, neglecting its ability to regulate cell-HSC crosstalk in the fibrotic microenvironment. Recent findings verify hepatocyte-derived oxidative stress may drive fibrosis: in hepatocyte-HSC co-cultures, knockout of hepatocyte Nrf2 significantly upregulated α-smooth muscle actin (α-SMA)—an indicator of HSC activation [12]. However, the specific mechanisms underlying its regulation of HSC activation remain elusive. The Shentao Ruangan formula (STR)—a ginseng-containing formula developed by prominent TCM expert Professor Zhou Daihan, via modifying three classic prescriptions (Ginseng Danggui, Xia Yuxue, Yinchen) [13]. Guided by the TCM principle of “strengthening healthy qi and eliminating pathogenic factors (Fu Zheng Qu Xie, FZQX)", this herbal formula has long been used clinically for treating chronic liver diseases, including hepatic fibrosis and HCC [13,14]. A prior investigation into TCM-centered hepatic fibrosis therapies identified 5 of the 18 most common herbs as key STR components [15], and 20-year CNKI data revealed 6 of the 20 most widely used anti-hepatic fibrosis herbs are STR constituents [16]. Clinical investigations also verify STR ameliorates hepatic fibrosis biomarkers and retards HCC progression [17]. This observation aligns with TCM theory, which holds that the pathogenesis of hepatic fibrosis is rooted in “deficiency leading to accumulation (Xu Sun Sheng Ji, XSSJ)" and “healthy qi deficiency with blood stasis (Zheng Xu Xue Yu)", with treatment guided by the principle of “FZQX” [18]. However, critical questions remain unaddressed: the cellular and molecular mechanisms governing the TCM pathogenesis of “XSSJ” in hepatic fibrosis, as well as the manner in which the ginseng-based compound STR achieves “FZQX” during hepatic fibrosis. In this study, single-cell RNA sequencing of liver tissues from patients with hepatic fibrosis was used to analyze the fibrotic microenvironment, revealing the pathogenesis of the TCM syndrome “XSSJ”. Integrating network pharmacology, RNA sequencing, and in vitro/in vivo experimental verification, this research confirms STR efficiently suppresses reactive oxygen species (ROS)-driven Transforming Growth Factor-β (TGF-β) release and the downstream fibrotic cascade via activating hepatocellular Nrf2 signaling. Furthermore, molecular docking, Nrf2 inhibitors and Nrf2 knockout (Nrf2−/−) mouse model studies reveal Nrf2 is the core target mediating STR's anti-fibrotic efficacy. In summary, this study provides initial insights into the cellular and molecular mechanisms that may underlie the TCM pathogenesis of “XSSJ” in hepatic fibrosis, while also exploring the potential antifibrotic mechanism of ginseng-containing formulas represented by STR that align with the “FZQX” principle. Additionally, our findings suggest that the hepatocellular Nrf2-ROS-TGF-β axis may serve as a potential novel target for antifibrotic drug development. Materials and methods Data collection and processing The human hepatic fibrosis dataset GSE136103 (platform: GPL20301) [19] utilized herein was retrieved from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) [20]. This dataset comprises liver tissue RNA sequencing data from 5 hepatic fibrosis patients and 5 controls, with all data undergoing strict quality control and standardization. The Harmony algorithm was used to correct batch effects between samples, and a neighborhood graph was constructed based on the corrected space for cell clustering. Uniform Manifold Approximation and Projection dimensionality reduction was applied for visualization, and manual cell type annotation was performed by combining known lineage marker genes. Cell communication analysis Gene expression matrices and metadata were extracted from disease group samples, and CellChat [21] objects were created with grouping by “celltype”. Using the human secretory signaling database, overexpressed genes were identified, ligand-receptor pairs characterized, and these mapped to protein-protein interaction (PPI) networks. Intercellular communication probabilities were calculated, interactions with <10 cells filtered out, interaction counts and weights visualized via netVisual_circle, and fibrotic tissue intercellular communication analyzed using CommPath. Cell type-specific ligand-receptor markers were detected via the Wilcoxon test significant LR pairs were filtered with logFC >0.3 and p-value <0.01. These were associated with Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, and pathway activities were scored using the gene set variation analysis (GSVA). For differential analysis, only significantly upregulated pathways were retained (only.posi = TRUE, only.sig = TRUE), and ligand-receptor pairs and pathways were filtered based on these results. All analyses were performed using R software and the CellChat package. Functional and gene enrichment analysis To explore biological functions of identified cell subpopulations, gene symbols were converted to Entrez IDs using org.Hs.eg.db for subsequent analyses. Genes were subjected to KEGG and Gene Ontology (GO) analyses with significantly enriched terms filtered at adjusted p-value <0.05, the same threshold filtered significant pathways for enrichment analysis. Gene set variation analysis of ROS pathways in hepatic fibrosis By integrating hepatic fibrosis-related GEO datasets (GSE25097, GSE89377), GSVA was performed to assess the enrichment level of the ROS gene set. Differences between groups were analyzed using the Wilcoxon rank-sum test, and box plots were generated for visualization using the ggpubr package. Preparation of STR STR, which was provided by the First Affiliated Hospital of Guangzhou University of Chinese Medicine (GUCM), consists of Panax quinquefolius 6g, Prunus persica (L.) Batsch 6g, Angelica sinensis (Oliv.) Diels 6g, Eupolyphaga sinensis Walker 6g, Rheum palmatum L. 6g, Agrimonia pilosa Ledeb. 15g, Artemisia capillaris Thunb. 6g and Salvia miltiorrhiza Bunge 6g. STR was prepared using the method described in previous studies [13,22]. The specific ingredients of STR [13,22] and its major chemical compounds identified by high-performance liquid chromatography [22] have been described previously and are provided in the Supplemental Material. Chemical and reagents Methionine- and choline-deficient (MCD) chow was purchased from Nantong Trophic (TP3006; Jiangsu, China). Alanine aminotransferase (ALT; C009-2-1), aspartate aminotransferase (AST; C010-2-1), triglyceride (TG; A110-1-1), total cholesterol (TC; A111-1-1) kits, commercial malondialdehyde (MDA; A003-1-2), catalase (CAT; A007-1-1), superoxide dismutase (SOD; A001-3-2), and glutathione (GSH; A006-2-1) assay kits were provided by Jiancheng Institute of Biotechnology (Jiangsu, China). A hydrogen peroxide assay kit (H2O2; S0038) was provided by Beyotime Biotechnology (Shanghai, China). Hematoxylin and eosin (H&E) dye solution was obtained from Regen Biotech (Beijing, China), Red Oil-O dye solution was obtained from Sigma (O0625-25g), Masson and Sirius Red dye solutions were provided by LEAGENE Biotech (Beijing, China), and all enzyme-linked immunosorbent assay (ELISA) kits were purchased from Ruixin Biotechnology Company (Guangzhou, China). DCFH-DA Reactive Oxygen (ROS) Fluorescent Probe (Red, M176010) was purchased from Beijing MREDA Technology Co., Ltd. (Beijing, China). TRIzol reagent and RNA Storage Solution were provided by Invitrogen (10296028, AM7000, USA). ABScript Ⅲ RT Master Mix for qPCR with gDNA Remover (RK20433) and Universal SYBR Green Fast qPCR Mix (RK21203) were provided by ABclonal (Hubei, China). The primary antibodies against collagen, type 1, alpha 1 (Col1a1; AF6524), Nrf2 (AF7623), and TGF-β (AF0297) were obtained from Beyotime Biotechnology. α-SMA (14395-1-AP), HO-1 (10701-1-AP), and β-actin (81115-1-RR) antibodies were purchased from Proteintech (Hubei, China). Anti-rabbit horseradish peroxidase (HRP; CSB-PA564648) was obtained from CUSABIO (Hubei, China). ColorMixed Protein Marker (RM19001) was obtained from ABclonal. Animals and treatments Forty 6-week-old male C57BL/6J mice (20-22 g) were obtained from the Animal Model Research Center of GUCM (SCXK 2022-0002; Guangzhou, China). An additional 12 Nrf2 knockout (Nrf2−/−) mice were provided by Professor Yong Gao (GUCM, Guangzhou). The mice were fed and watered freely under standard humidity and temperature conditions (12-h light/dark cycle). After acclimatization period, mice were randomly assigned to five groups (n = 8 per group): normal (control), MCD diet model (model), MCD + STR 0.62 g/kg/d (STR-L), MCD + STR 1.24 g/kg/d (STR-M), and MCD + STR 2.48 g/kg/d (STR-H). The doses were determined based on previous studies [23], which was converted from the recommended clinical dosage of STR to the murine dose using the standard body surface area conversion method. STR was administered via gavage at a volume of 0.01 mL/g once daily. For 8 weeks, the control group was fed a regular chow diet, while all other groups were fed an MCD diet. During the last 2 weeks, the STR-L, STR-M, and STR-H groups were gavaged with STR at respective doses, whereas the control and model groups were gavaged with an equal volume of saline. Similarly, 18 C57BL/6J mice and 12 Nrf2−/− mice were divided into five groups (n = 6 per group): control, model, STR-H, Nrf2−/− model, and Nrf2−/− STR-H, and treated following the same protocol. Serum and tissue samples were collected on the final day after anesthesia. The animal studies were approved by the Animal Ethics Committee of the GUCM (Ethics number: No. 20240302005). Biochemical analysis Blood collected from mouse eyes was centrifuged at 4000 rpm for 10 min, with serum AST and ALT levels measured following the manufacturer's protocol. The TG and TC levels were measured as described above using TG or TC kits. Serum Hyaluronic acid (HA), procollagen Ⅲ (PC Ⅲ), laminin (LN) and Collagen type Ⅳ (Col Ⅳ) were tested using ELISA kits following the manufacturer's instructions. H2O2, MDA, CAT, GSH, and SOD levels were measured according to the protocols of the assay kits described above. Histopathology and immunohistochemistry Hepatic specimens were fixed in paraformaldehyde, embedded in paraffin, sectioned, and stained with H&E, Sirius Red, and Masson. Some hepatic tissues were frozen at optimal cutting temperature, sectioned, and stained with Red Oil-O for hepatic lipid content analysis. For Immunohistochemistry (IHC), sections were blocked with 10% bovine serum albumin prior to incubation with α-SMA primary antibody (1:2000). After three washes with phosphate-buffered saline, sections were incubated with HRP, using 3,3′-diaminobenzidine hydrochloride as the chromogen. Digital images were obtained using a KFBIO (KF-PRO-005-EX) Digital Pathology Scanner (Zhejiang, China). Network pharmacology study STR constituents were screened using the ADME system in Traditional Chinese Medicine Systems Pharmacology (TCMSP, http://tcmspw.com/) database [24], retaining those with oral bioavailability ≥30% and drug likeness ≥0.18. Potential component targets were retrieved from TCMSP, HERB (http://herb.ac.cn/), and GeneCards (https://www.genecards.org/). Hepatic fibrosis targets from GeneCards [25] using “hepatic fibrosis” and “liver fibrosis” and component targets from GeneCards requiring relevance score ≥20. TCMSP component targets were mapped to standard human gene symbols via UniProt, their intersection with disease targets shown in a Venn diagram. Metascape (https://metascape.org/) conducted KEGG and GO analyses of matched STR-hepatic fibrosis genes, action targets were imported into STRING (V 12.0), and key ingredients screened via Cytoscape 3.10.0 topological analysis to build a PPI network ranked by betweenness centrality. Molecular docking Molecular docking was performed as described previously [26]. The 3D structures of the possible active ingredients (Daturilin, picroside iii, 20-(S)-Ginsenoside-Rg3_qt, ginsenoside Rb3_qt, ginsenoside rh2) obtained from the network pharmacology study in SDF format were downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov/) [27], whereas that of Keap1 (1ZGK) was obtained from the PDB database using PyMOL 2.3.0. ROS analysis ROS accumulation in the liver was detected as described previously [28]. Fresh liver tissues were taken, and unfixed 10-μm-thick fresh frozen sections were prepared within 2h ex vivo and stored at −20 °C away from light according to the ROS Assay Kit instructions, including cleaning the working solution clearing slices, staining the working solution, incubating away from light at 37 °C, removing the staining solution, washing with phosphate-buffered saline, sealing with glycerol. Quantitative real-time polymerase chain reaction Total mRNA was extracted from hepatic tissues or cells using TRIzol reagent. High-capacity cDNA reverse transcription kits were used to perform the reverse transcription. Quantitative real-time polymerase chain (qPCR) was performed using a SYBR Green Master Mix kit. All primers used are listed in Supplemental Table 1. We selected β-actin as the internal control to normalize relative gene expression levels, and the relative mRNA expression levels were calculated using the 2−ΔΔCT method. Western blotting analysis Total protein was collected from hepatic tissues lysed with RIPA buffer, and 20-μg protein samples were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to 0.45-μm polyvinylidene fluoride (PVDF) membranes. The PVDF membranes were then blocked by 5% nonfat milk and incubated with the primary antibodies for α-SMA (1:2000), Nrf2 (1:1000), HO-1 (1:3000), and β-actin (1:5000) overnight at 4 °C. Finally, the PVDF membranes were incubated with HRP (1:5000) and visualized using Clarity Western ECL Substrate. A densitometric analysis of the bands was performed using ImageJ software to determine the relative expression levels of each protein. Cell culture and cell viability assay Mouse primary hepatocytes (MPHs) were isolated and cultured as previously described [28]. HepG2 cells (CL-0103) and LX-2 cells (CL-0560) were procured from Pricella Biotechnology Co., Ltd. and used between passages 5 and 10. Those cells cultivated in Dulbecco's modified Eagle's medium with 10 % fetal bovine serum, 100 unit/mL penicillin, and 100 μg/mL streptomycin (5 % CO2, 37 °C). Cell viability was assessed using a Cell Counting Kit-8 (CCK-8) assay. Finally, cell viability was analyzed at 540 nm. For oxidative stress environment simulation, MPHs and HepG2 cells were incubated with hydrogen peroxide for 24 h [29] to support the study. To confirm whether STR alleviates oxidative stress in hepatocytes and reduces TGF-β secretion, thereby inhibiting HSC activation, LX-2 cells were incubated for 24 h with conditioned medium (CM) from HepG2 cells subjected to various treatments. Immunofluorescence staining Immunofluorescence (IF) staining of MPHs and HepG2 cells were performed according to general procedures. After being fixed with paraformaldehyde, the cells were permeabilizated with Triton X-100, blocked with 2% bovine serum albumin, treated with primary antibody against Nrf2 (1:200), HO-1 (1:400), TGF-β (1:200), α-SMA (1:1000), or Col1a1 (1:200) overnight at 4 °C, and incubated with secondary antibodies and DAPI staining. The stained cells were observed under a fluorescence microscope. RNA-sequencing and bioinformatics analysis An RNA-sequencing analysis was performed as previously described [30]. Genes with an adjusted value of p < 0.05 and fold-change ≥2 were identified as differentially expressed genes (DEGs). KEGG pathway and GO enrichment analyses of the genes were performed of the matched genes between STR and hepatic fibrosis using the R package ClusterProfiler. The GO and KEGG entries were then ranked from smallest to largest according to their p-values, and the top 10 or 20 significant entries in each category were displayed. Statistical analysis Each experiment was performed in triplicate, and the data are expressed as mean ± standard error of the mean (SEM). The statistical analyses were performed using GraphPad Prism (V 9.5.0; GraphPad, San Diego, CA, USA). Intergroup differences were analyzed using Student's t-test. One-way analysis of variance with Tukey's post-hoc test was used for multiple-group comparisons. Statistical significance was set at values of p < 0.05. Results Hepatocellular oxidative stress-mediated Hepatocyte-HSC crosstalk as the core of the fibrotic microenvironment Visualization (Fig. 1A) and intercellular communication analyses of single-cell sequencing showed that hepatocytes exerted the most notable impact on HSCs (Fig. 1B and C). Meanwhile, analysis of intercellular communication between different cell types revealed that the interactions between hepatocytes and HSCs were the most frequent and weighted most heavily in liver tissues (Fig. 1D and E), which indicate that the crosstalk between hepatocytes and HSCs is the core of the hepatic fibrotic microenvironment. Focusing on hepatocytes and HSCs, the results revealed that Chemical carcinogenesis - reactive oxygen species is the primary pathway shared by HSCs and hepatocytes in hepatic fibrosis (Fig. 1F). Moreover, enrichment analysis of pathways associated with DEGs in each cell type indicated that oxidative stress was most conspicuous in hepatocytes among all cell types (Fig. 1G). Subsequent KEGG and GO enrichment analyses of differentially expressed genes in hepatocytes revealed the critical role of ROS in hepatocytes of patients with hepatic fibrosis (Fig. 1H and I). Furthermore, datasets of patients with hepatic fibrosis collected from the GEO database reaffirmed the important role of ROS in hepatic fibrosis (Fig. 1J and K). These findings indicate that oxidative stress is a key mechanism in hepatic fibrosis, with hepatocellular oxidative stress being critical within its microenvironment. Single-cell sequencing analysis. (A) Cellular distribution profile in liver tissue. (B-C) Analysis of intercellular communication. (D-F) Visualization of the number and weight of intercellular communication interactions. (G) Heatmap of pathway enrichment across various cell types. (H-I) KEGG and GO enrichment analysis. (J-K) Gene set variation analysis of GEO datasets. STR alleviated hepatic fibrosis in vivo MCD-treated mice significantly lost body weight, and this trend was alleviated by high-dose STR (Fig. S1A). However, there was no significant effect on the liver weight (Fig. S1B). In a dose-dependent fashion, STR significantly reduced serum AST/ALT levels (Fig. 2A and B), and hepatic and serum TG/TC accumulation (Fig. 2C and D; Fig. S1C–D). Additionally, H&E and Red Oil-O revealed that STR notably improved liver injury and lipid accumulation (Fig. 2E and F). MCD elevated hepatic fibrosis markers which were mitigated by STR treatment (Fig. 2G and H; Fig. S1E–F). Masson, Sirius Red staining (SR) and SR positive area showed similar results (Fig. 2I–K). STR significantly reversed MCD-induced elevation of TGF-β, α-SMA, and Col1a1 mRNA expression (Fig. 2L). Western blotting and IHC staining demonstrated STR strongly downregulated hepatic α-SMA protein in MCD-fed mice relative to the model group (Fig. 2M and N). These results verify STR mitigates hepatic fibrosis progression in vivo. STR alleviated hepatic fibrosis. (A-D) ALT, AST, TG, and TC levels (n = 8). (E-F) H&E and Red Oil-O staining (400 × ). (G-H) Serum contents of fibrosis markers (n = 7). (I-K) Masson, SR staining (400 × ), and SR positive area (n = 3). (L) Relative mRNA expression levels of hepatic fibrosis (n = 8). (M) Effects of STR on the expression of α-SMA proteins by western blotting. (N) IHC staining images of hepatic α-SMA (400 × ). Scale bars indicate 50 μm. The data are shown as mean ± SEM. #p < 0.05, ##p < 0.01, ####p < 0.0001 vs. Control group, ∗p <0.05; ∗∗p <0.01; ∗∗∗p <0.001; ∗∗∗∗p < 0.0001; vs. Model group. Network pharmacology study and RNA-sequencing analysis The STR component database comprised 143 effective monomer components from herbs, with compound and target counts per herbal medicine in Supplemental Table 2. Fig. 3A shows 41 STR targets associated with hepatic fibrosis therapy. The top 20 KEGG pathways showed primary enrichment in lipid metabolism, inflammation, and ROS-related pathways (Fig. 3B). Fig. 3C further depicts the top 10 GO terms. The PPI network identified 5 core targets (TGF-β, IL-6, TNF-α, Akt, HO-1) (Fig. 3D) and additional targets (NRF2, IL-1β, TP53) primarily related to inflammation and oxidative stress. Network pharmacology study and RNA-sequencing analysis. (A) Venn diagram. (B) KEGG enrichment analysis. (C) GO enrichment analysis. (D) Core targets of PPI network. (E) KEGG enrichment analysis from the sequencing data. (F) Relative expression level of heatmap. Subsequent RNA-sequencing analysis was conducted to further explore the anti-hepatic fibrosis mechanism of STR. Fig. S1G shows 15 upregulated and 138 downregulated hepatic DEGs. KEGG and GO analyses were therefore conducted on genes with p < 0.05 (Fig. 3E, Fig. S1H). ROS appeared in the top 20 KEGG pathways, which aligned with the findings from network pharmacology. A heatmap (Fig. 3F) of oxidative stress/fibrosis-associated gene transcripts points to mitochondria- and Nrf2 pathway-related mechanisms. STR alleviated hepatic oxidative stress in MCD diet-induced hepatic fibrosis and targeted Nrf2 The MCD diet notably elevated oxidative stress markers (H2O2, MDA), and STR reduced these dose-dependently (Fig. 4A and B). Conversely, STR-treated groups exhibited robust restoration of hepatic antioxidants (Fig. 4C–E). Hepatic ROS was strongly induced by MCD but greatly reduced by STR (Fig. 4F). Furthermore, STR treatment also upregulated hepatic antioxidant gene expression compared to the model group (Fig. 4G). Western blotting revealed MCD diminished hepatic Nrf2 and HO-1, which STR reversed (Fig. 4H and I), demonstrating STR-activated Nrf2/HO-1 signaling. STR alleviated oxidative stress and activated the Nrf2/HO-1 pathway in the MCD-fed wild-type mouse model. (A-E) H2O2, MDA, CAT, GSH, and SOD level in hepatic tissues (n = 6). (F) Liver ROS levels. (G) Relative mRNA expression levels of oxidative stress (n = 8). (H-I) Effects of STR on Nrf2 and HO-1 protein expressions. The data are shown as mean ± SEM. ##p < 0.01, ###p < 0.001, ####p < 0.0001 vs. Control group, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; vs. Model group. STR alleviated oxidative stress via Nrf2 pathway activation in hepatocytes Hepatocyte-HSC crosstalk is central to the fibrotic microenvironment, with hepatocellular oxidative stress as a key driver. We proposed STR mitigates hepatic fibrosis by reducing oxidative stress-induced hepatocyte damage through the Nrf2 pathway, inhibiting HSC activation. STR upregulated antioxidant genes in MPHs (Fig. 5A) and IF verified enhanced Nrf2/HO-1 in H2O2-stimulated MPHs (Fig. 5C, Fig. S3D). Subsequent qPCR and IF in HepG2 cells replicated it (Fig. 5B and D). STR alleviated oxidative stress via Nrf2 pathway activation in hepatocytes. (A) Relative mRNA expression levels of oxidative stress in MPHs (n = 4). (B) Relative mRNA expression levels of oxidative stress in HepG2 cells (n = 4). (C) IF staining for Nrf2 expression in MPHs (1000 × ). (D) IF staining for Nrf2 expression in HepG2 cells (400 × ). (E-I) Molecular docking conformation of Daturilin, picroside iii, 20-(S)-Ginsenoside-Rg3_qt, ginsenoside Rb3_qt, and ginsenoside rh2 with Keap1. The data are shown as mean ± SEM. #P < 0.05, ##P < 0.01, ####P < 0.0001 vs. Control group, ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001; vs. H2O2 group. To investigate the potential role of bioactive components from Panax quinquefolius in STR, molecular docking was performed against Keap1. This approach aligns more closely with the physiological mechanism of Nrf2 activation, which involves disrupting the Keap1-Nrf2 interaction. Through network pharmacology analysis, 12 active ingredients from Panax quinquefolius, including ginsenoside Rh2, were identified. In addition, given the pharmacological importance of ginsenosides, representative ginsenoside components with oral bioavailability ≥29% and drug likeness ≥0.18 were included in the computational analysis (Supplemental Table 3). The five active components with the lowest binding energies were visualized: Daturilin (−10.99, MOL008397), picroside iii (−10.93, MOL002811), 20-(S)-Ginsenoside-Rg3_qt (−10.84, MOL005288), ginsenoside Rb3_qt (−10.51, MOL006754), and ginsenoside rh2 (−10.5, MOL005344) (Fig. 5E–I). STR inhibits LX-2 activation via hepatocyte Nrf2-mediated suppression of TGF-β secretion in vitro Integrated single-cell sequencing, network pharmacology and prior work [31] highlighted TGF-β′s central role in STR-mediated hepatic fibrosis suppression, we proposed a hepatocyte-HSC crosstalk regulatory axis: hepatocellular Nrf2-ROS-TGF-β-HSC activation. We then validated STR's effects on TGF-β in MPHs and HepG2 cells. CCK-8 assay established safe STR doses in MPHs (<2 mg/mL; Fig. S2A), and STR doses (0.25, 0.5, 1 mg/mL) were selected. H2O2 reduced MPH Nrf2/HO-1, while STR upregulated both dose-dependently (Fig. S2B–C). ELISA analysis demonstrated STR decreased H2O2-induced TGF-β secretion in MPH media (Fig. 6A). qPCR showed H2O2 elevated TGF-β in MPHs and HepG2 cells while STR reversed this dose-dependently (Fig. 6B and C), with IF confirming reduced TGF-β in H2O2-stimulated cells (Fig. 6D and E). STR inhibits LX-2 activation via hepatocyte Nrf2-mediated suppression of TGF-β secretion in vitro. (A) TGF-β in the culture medium of MPHs by ELISA analysis. (B-C) Relative mRNA expression levels of TGF-β in MPHs or HepG2 cells (n = 4). (D-E) IF staining for TGF-β expression in MPHs (1000 × ) and HepG2 cells (400 × ). (F) Design of the in vitro experiment. (G) Effects of different CM on the mRNA levels of α-SMA and Col1a1 by qPCR in LX-2 cells (n = 4). (H-I) IF staining for α-SMA and Col1a1 expression in LX-2 cells (400 × ). (J-M) Relative mRNA expression levels of oxidative stress in HepG2 cells (n = 4). (N) Effects of different CM on the mRNA levels of α-SMA and Col1a1 by qPCR in LX-2 cells (n = 4). (O) IF staining for α-SMA expression in LX-2 cells (400 × ). The data are shown as mean ± SEM. #P < 0.05, ##P < 0.01, ###P < 0.001, ####P < 0.0001; ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. Considering the fact that oxidative stress in hepatocytes and the TGF-β excreted by them are vital to HSC activation, CM was collected from HepG2 cells and used to stimulate LX-2 cells to investigate the underlying mechanism (Fig. 6F). CM from H2O2-treated HepG2 cells significantly increased the mRNA expressions of α-SMA and Col1a1 in LX-2 cells, which was lower in H2O2/STR-H-treated CM (Fig. 6G). IF staining of α-SMA and Col1a1 showed that CM from H2O2-exposed HepG2 cells activated HSCs significantly less than that from H2O2/STR-H-treated cells (Fig. 6H and I). To determine whether this protective effect is Nrf2-dependent, HepG2 cells were treated with the Nrf2 inhibitor ML385, and the mRNA expression of oxidative stress–related genes was measured by qPCR (Fig. 6J–M). Subsequently, conditioned medium co-culture experiments were repeated under ML385 intervention. As shown in Fig. 6N and O, qPCR and IF results revealed that the ability of STR to reduce TGF-β secretion and inhibit LX-2 activation is dependent on hepatocyte Nrf2. STR failed to alleviate hepatic fibrosis in Nrf2−/− mice Nrf2−/− mouse identification is shown in Fig. S3A–B, and body and liver weight results are displayed in Fig. S3C–D. Nrf2−/− model mice had higher ALT/AST (Fig. 7A and B) and TG/TC (Fig. 7C–D, S3E–F) levels than WT mice, H&E and Oil Red-O staining further supported this (Fig. 7E and F). Notably, STR exerted less potent beneficial effects on liver injury and lipid deposition in Nrf2−/− mice. Further experiments validated Nrf2's role in hepatic fibrosis. Compared to WT mice, Nrf2−/− model mice had markedly higher fibrosis markers (Fig. 7G and H), while Masson's trichrome staining, Sirius Red staining, and Sirius Red-positive area analyses demonstrated Nrf2's protective effect (Fig. 7I–K). Critically, STR showed limited efficacy in Nrf2−/− mice. In addition, hepatic TGF-β, αSMA, and Col1a1 mRNA levels were barely improved by STR (Fig. 7L), and IHC indicated STR's ability to inhibit hepatic α-SMA was diminished in Nrf2−/− mice (Fig. 7M). STR failed to alleviate hepatic fibrosis in Nrf2−/− mice. (A-D) ALT, AST, TG, and TC levels (n = 5). (E-F) Representative images of H&E and Red Oil-O (400 × ). (G-H) Serum contents of fibrosis markers were measured (n = 5). (I-K) Masson, SR staining (400 × ) and SR positive area (n = 3). (L) Relative mRNA expression levels of hepatic fibrosis (n = 5). (M) Representative images of IHC of hepatic α-SMA (400 × ). (N-R) H2O2, MDA, CAT, GSH, and SOD levels in hepatic tissues (n = 5). (S) Liver ROS levels. (T) Relative mRNA expression levels of antioxidant genes (n = 5). The data are shown as mean ± SEM. #p < 0.05, ##p < 0.01, ###p < 0.001, ####p< 0.0001 vs. control group; ∗p< 0.05, ∗∗p< 0.01, ∗∗∗p< 0.001, ∗∗∗∗p < 0.0001. We further evaluated the impact of Nrf2 ablation on STR's antioxidant capacity in hepatic fibrosis. MCD-fed Nrf2−/− mice had elevated H2O2 and MDA (Fig. 7N and O) with reduced CAT, GSH, and SOD (Fig. 7P–R), and STR showing minimal amelioration. Hepatic ROS levels were significantly higher in Nrf2−/− mice, and STR failed to reduce ROS in Nrf2−/− mice (Fig. 7S). Moreover, STR's enhancement of antioxidant genes was attenuated in Nrf2−/− mice (Fig. 7T). Discussion As the primary effector cells in fibrosis, HSCs function as a central hub in fibrotic progression, yet the dynamic interplay between HSCs and other cells in the hepatic fibrotic niche is poorly defined [31]. In TCM, the pathogenesis of hepatic fibrosis is characterized by “XSSJ” [18,32], though the corresponding modern biological underpinnings await further clarification. In this study, single-cell sequencing analysis of liver tissues from patients with hepatic fibrosis identified abnormal changes in the hepatocyte-to-HSC ratio and revealed that their dynamic interplay is a defining feature of the fibrotic microenvironment. Differential gene expression analysis showed that hepatocytes and HSCs are co-enriched in pathways involving chemical carcinogenesis-reactive oxygen species. Among these, the reactive oxygen species pathway was most highly expressed in hepatocytes, suggesting that hepatocellular oxidative stress is a key mediator of intercellular crosstalk. While previous studies have primarily focused on direct effects of ROS on HSCs, our research is the first to demonstrate that hepatocellular oxidative stress activates HSCs via the ROS-TGF-β axis, leading to ECM deposition. This finding offers a potential molecular interpretation consistent with the TCM concept of “XSSJ”, in which hepatocellular injury induced by oxidative stress may correspond to “Xu Sun”, while abnormal ECM deposition may correspond to “Sheng Ji”. STR, developed by Professor Zhou Daihan based on TCM principles such as “FZQX” and derivative theories like “Yun Pi Hua Ji”. In this study, STR dose-dependently improved collagen deposition and reduced the levels of fibrotic markers in a mouse model of MCD diet-induced hepatic fibrosis. Integration of network pharmacology and RNA sequencing results revealed that the mechanism of STR action is associated with oxidative stress. Mechanistic studies further demonstrated that STR activates the Nrf2 pathway in hepatocytes, upregulating antioxidant genes such as HO-1, NQO1, GCLC and GCLM, thereby reducing ROS levels and TGF-β secretion, and inhibiting LX-2 cell activation and ECM synthesis. Subsequent conditioned medium co-culture experiments with or without the Nrf2 inhibitor further demonstrated that STR attenuates LX-2 activation via hepatocyte Nrf2-mediated suppression of TGF-β secretion. Experiments using Nrf2−/− mouse models suggest that the antifibrotic effects of STR may depend on Nrf2-mediated regulation of the ROS-TGF-β axis. However, these findings warrant further validation using hepatocyte-specific Nrf2 knockout mouse models. By innovatively integrating single-cell sequencing, network pharmacology, and gene editing technologies, this study is the first to systematically delineate the “hepatocyte Nrf2-ROS-TGF-β-HSC” intercellular regulatory model, providing a potential modern biological and molecular framework for the TCM pathogenesis of “XSSJ”. STR blocks the pathological cascade of “hepatocyte injury → HSC activation → ECM deposition” by enhancing hepatocellular antioxidant capacity (reflecting “Fu Zheng”) and suppressing fibrosis drivers (reflecting “Qu Xie”), offering a novel therapeutic strategy for hepatic fibrosis (Fig. S4). This work not only clarifies the molecular mechanism of STR's antifibrotic effects but also provides a potential modern biological interpretation of the TCM principle of “FZQX". Several limitations of this study should be acknowledged. We did not directly isolate distinct liver cell populations (including hepatocytes, Kupffer cells, and HSCs) from clinical or animal liver tissues to identify the primary cellular sources of ROS and TGF-β. In addition, we did not perform cross-validation using alternative models of hepatic fibrosis, such as those induced by alcohol, high-fat diet, or carbon tetrachloride. While we revealed the “hepatocyte Nrf2-ROS-TGF-β-HSC” regulatory axis in the fibrotic microenvironment, multiple other liver cell types also play important roles in hepatic fibrosis progression [11,33,34], and their specific regulatory mechanisms require further exploration. These limitations will be addressed in subsequent studies. Conclusion In conclusion, addressing the unmet clinical need for effective hepatic fibrosis interventions and anchored in TCM principles, this study integrates a diverse array of technical strategies to advance both our understanding of disease pathogenesis and the development of targeted therapeutics—with a focus on the ginseng-containing STR formula. Our findings identify aberrant hepatocyte-HSC crosstalk as a core feature of the fibrotic liver microenvironment. Critically, the mechanistic link we established —whereby hepatocellular oxidative stress drives HSC activation via the ROS-TGF-β axis—forges a direct connection between the TCM pathogenesis of “XSSJ” and modern biology: hepatocyte injury (a manifestation of “Xu Sun”) propagates ECM deposition (the “Ji” of fibrosis) through this pro-fibrotic signaling cascade, thereby endowing this ancient TCM concept with precise molecular underpinnings. For the ginseng-containing STR formula, our data demonstrate that its ability to inhibit hepatocellular oxidative stress and disrupt fibrotic cascades via Nrf2 activation embodies the TCM doctrine of “FZQX” in a mechanistically defined manner. Specifically, STR's activation of hepatocellular Nrf2 to enhance antioxidant capacity constitutes “Fu Zheng”, while its subsequent suppression of ROS-TGF-β -mediated HSC activation and ECM deposition represents “Qu Xie”. This dual action not only positions STR as a promising candidate for targeted hepatic fibrosis therapy, but also provides a paradigm for integrating TCM theory with rigorous modern biomedical research. CRediT authorship contribution statement Shangyi Huang: Investigation, Validation, Formal analysis, Writing–original draft. Ruisheng Zhou: Conceptualization, Methodology, Writing–original draft. Jing Li: Validation, Writing–original draft. Yi Zheng: Validation. Lunwei Yang: Validation. Tianqi Gao: Formal analysis. Yi Yuan: Formal analysis. Dailin Wu: Formal analysis. Yafei Shi: Conceptualization. Dayong Zheng: Supervision, Funding acquisition. Daihan Zhou: Conceptualization, Methodology. Ying Tang: Project administration, Supervision, Funding acquisition, Methodology. Declaration of competing interest The authors declare that the study was performed without any commercial conflict. Acknowledgements This work was supported by the State Key Laboratory of Traditional Chinese Medicine Syndrome Projects (Grant No. SKLKY2024B0013) and Science and Technology Department of Xinjiang Uygur Autonomous Region (Grant No. 2021D01C016). We are grateful to Professor Yong Gao (Guangzhou University of Chinese Medicine, Guangzhou) for providing the Nrf2−/− mice. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.jgr.2026.101053. Contributor Information Dayong Zheng, Email: zhengdy@smu.edu.cn. Daihan Zhou, Email: 13902206571@163.com. Ying Tang, Email: tangying@gzucm.edu.cn. Supplementary data The following are the Supplementary data to this article. Multimedia component 1 Multimedia component 2 Multimedia component 3 Multimedia component 4 Multimedia component 5 Multimedia component 6 Multimedia component 7 Multimedia component 8 Multimedia component 9 Multimedia component 10 Multimedia component 11 Multimedia component 12 Multimedia component 13 Multimedia component 14 Multimedia component 15 Multimedia component 16 Multimedia component 17 Multimedia component 18 References Associated Data Supplementary Materials Multimedia component 1 Multimedia component 2 Multimedia component 3 Multimedia component 4 Multimedia component 5 Multimedia component 6 Multimedia component 7 Multimedia component 8 Multimedia component 9 Multimedia component 10 Multimedia component 11 Multimedia component 12 Multimedia component 13 Multimedia component 14 Multimedia component 15 Multimedia component 16 Multimedia component 17 Multimedia component 18