Colchicine tablets (Lot number: 220204) were purchased from Banna Pharmaceutical Co., Ltd. Alanine aminotransferase (ALT) kit (Lot number: 20231024), aspartate aminotransferase (AST) kit (Lot number: 20231024), type IV collagen (COL-Ⅳ) kit (Lot number: 20231123), type III procollagen (PC-III) kit (Lot number: 20231123), laminin (LN) kit (Lot number: 20231123), and hyaluronic acid (HA) kit (Lot number: 20231123) were provided by Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Vitexin (Lot number: RFS-M02302002022, purity >98%), isovitexin (Lot number: RFS-Y11602107028, purity >98%), protocatechuic acid (Lot number: RFS-Y03111812016, purity >98%), rosmarinic acid (Lot number: AF22022152, purity >98%), schaftoside (Lot number: AF21061501, purity >98%), isoschaftoside (Lot number: AF21062602, purity >98%), and vicenin-2 (Lot number: AF20112508, purity >98%) were obtained from Chengdu Aifa Biotechnology Co., Ltd. (Chengdu, China). Electrophoresis buffer (Lot number: 20,220,810), electrotransfer buffer (Lot number: 240005008), TBST solution (Lot number: 20230214), Tumor necrosis factor-α (TNF-α) enzyme-linked immunosorbent assay (Elisa) kit (Lot number: CSB-E04741m), interleukin-1 beta (IL-1β) Elisa kit (Lot number: CSB-E8054m), interleukin 6 (IL-6) Elisa kit (Lot number: CSB-E04639m), and transforming growth factor beta 1 (TGF-β1) Elisa kit (Lot number: CSB-E04726m) were provided by Cusabio (Wuhan, China). Primary antibodies against toll-like receptor 4 (TLR4) (Lot number: 19811-1-AP), myeloid differentiation primary response protein 88 (MyD88) (lot number: 23230-1-AP), nuclear factor kappa-B (NF-κB) p65 (Lot number: 10745-1-AP), and Phospho-NF-κB p65 (Lot number: 82335-1-RR) were obtained from Proteintech (Chicago, United States). BCA protein assay kit (Lot number: 040224240722), Primary antibody diluent (Lot number: 011422220520), secondary antibody diluent (Lot number: 112522221205) and ECL chemiluminescence kit (Lot number: 091523231011) were supplied by Biyotime Biotechnology (Shanghai, China).

The plant material was sourced from Guangxi Xianzhu Traditional Chinese Medicine Technology Co., Ltd. (Nanning, China). It was authenticated by Professor Yong Tan of Guangxi University of Chinese Medicine as the dried aerial portions of Isodon lophanthoides (Buch.-Ham. ex D. Don) H. Hara, belonging to the Lamiaceae family (Supplementary Table S1). A reference specimen (voucher no. GUCM2021013) has been archived in the Department of Drug Analysis at Guangxi University of Chinese Medicine. The collection and use of plant materials complied fully with the Nagoya Protocol on Access and Benefit Sharing, the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), and all relevant phytosanitary and biodiversity regulations. The plant species used (Isodon lophanthoides (Buch.-Ham. ex D. Don) H. Hara) is not listed under CITES and was obtained from a legal commercial source in China. No international transport of plant materials was involved.

The plant material used in this study was Isodon lophanthoides (Buch.-Ham. ex D. Don) H. Hara [Lamiaceae; Herba Isodonis Lophanthoidis], traditionally used in Yao medicine. A total of 10 kg of dried aerial parts were extracted with 80 L of purified water. The extraction process involved two consecutive decoctions, each lasting 1.5 h. The supernatant from each decoction was filtered through 200-mesh gauze, combined, and concentrated under reduced pressure, yielding 1.2 kg of ILW. This product (Batch No. 202209013; Production date: 12 September 2022; Best before: 11 September 2025) is a research-grade aqueous extract developed for pharmacological investigation. It is not marketed as a finished pharmaceutical product and is not subject to national drug regulatory approval. Extraction and sample preparation were conducted by the Preparation R&D Center, Guangxi Hospital of Traditional Chinese Medicine (Nanning, China). No traditional processing methods (e.g., steaming, frying, roasting) were applied prior to extraction.

Quantification of key metabolites in ILW, including caffeic acid, rosmarinic acid, schaftoside, and isoschaftoside, was carried out by an Agilent 1290 UPLC system. Calibration curves were established for each analyte to ensure accurate measurement. Specific protocols for sample processing, instrumental settings and selection of metabolites for quantification are detailed in the Supplementary Material. To assess batch-to-batch consistency, chromatographic fingerprinting was performed on 12 independent ILW batches using UPLC method. Data analysis was carried out with the Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine (Version 2012.130723, National Pharmacopoeia Commission of China), as detailed in the Supplementary Material.

Liver tissues from control, model, and ILW-H groups (n = 3) were retrieved from −80°C. After total RNA was extracted, the concentration and integrity were assessed by a Qubit 4.0 fluorometer and Qsep400 bioanalyzer, respectively. Poly(A) mRNA was enriched using AMPure XP beads, fragmented, and reverse-transcribed into cDNA. After end-repair, adapter ligation, and 15 PCR cycles, libraries (∼300 bp) were constructed and validated using an Agilent 2100 Bioanalyzer, then sequenced using Illumina platform. Then, raw reads were filtered, aligned to the reference genome, and quantified. DESeq2 (v1.38.3) was used to identify differentially expressed genes (DEGs) with |log₂Fold Change (FC)| > 0.58 and FDR <0.05. Up- and downregulated genes were defined by FC > 1.5 and <0.67, respectively. Venny 2.1 (https://bioinfogp.cnb.csic.es/tools/venny/) was employed to visualize DEG overlaps, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment was conducted by clusterProfiler in R (v4.3.1). Data visualization was carried out using the Bioinformatics Platform (https://www.bioinformatics.com.cn/). The transcriptomic and proteomic datasets have been deposited in the Zenodo platform. Accession number is 15460165.

Liver tissues (n = 3 per group from control, model, and ILW-H cohorts) were pulverized in liquid nitrogen. The homogenized powder was lysed in buffer (8 M urea, 1 mM PMSF, 2 mM EDTA), followed by sonication and centrifugation (15,000 rpm, 10 min, 4°C). Then, protein concentrations quantified using a BCA assay. Proteins were digested overnight at 37°C using trypsin in 25 mM ammonium bicarbonate. Peptides were acidified (pH 2–3) with 20% TFA, purified via C18 cartridges, and quantified using the Pierce™ peptide detection kit. Peptide mixtures were resolved on a Vanquish Neo UHPLC system using a dual-column configuration (PepMap Neo trap and analytical columns). Mass spectra were acquired on an Orbitrap Astral MS in positive ion mode. Survey scans covered an m/z range of 380–980 at a resolution of 240,000. DIA acquisition used 299 variable isolation windows (2 Th), HCD at 25%, and dynamic accumulation (AGC target 500%, maximum IT 3–5 ms).

Raw spectral data were analyzed using Proteome Discoverer 2.5. Protein identification was performed using DIA-NN (v1.8.1). A 1% FDR cutoff was applied for confident protein assignment. Differentially expressed proteins (DEPs) analysis used a threshold of p < 0.05 and FC > 1.2 or <0.83. Proteins exceeding a FC of 1.2 were deemed upregulated, while those below 0.83 were downregulated. Overlapping DEPs across group comparisons (model vs. control, ILW-H vs. model) were visualized using Venny 2.1. KEGG pathway enrichment was carried out vis KOBAS (http://kobas.cbi.pku.edu.cn/), and results were graphically rendered via the Bioinformatics online platform.

Liver tissues (n = 3 per group from control, model, and ILW-H cohorts) were subjected to metabolomic profiling. Approximately 50 mg tissue was extracted with 500 μL of prechilled 70% methanol (−20°C), and centrifuged at 12,000 rpm for 10 min. A 300 μL aliquot of the supernatant was kept at −20°C for 30 min, then centrifuged again under the same conditions. The final supernatant (200 μL) was passed through a protein precipitation plate prior to analysis using a SCIEX QTRAP 6500+ system. To fully investigate the metabolites, two chromatography modes, T3 column and Amide column, were used for sample separation, and the detail parameters are presented in the Supplementary Material. Mass spectrometry was carried out using ESI in positive and negative modes. The source temperature was 550°C, with ion voltages set at 5.5 kV (+) and 4.5 kV (−). Curtain gas was maintained at 35 psi. Data acquisition used dynamic multiple reaction monitoring (MRM) mode.

Raw data were handled using Analyst 1.6.3 and MultiQuant 3.03 for peak detection, alignment, and integration. Principal component analysis (PCA) and pathway enrichment were conducted using MetaboAnalyst 6.0. Metabolites with p < 0.05 and FC > 1.5 or <0.67 were considered significantly altered. Shared differential metabolites (DMs) across comparisons were identified using Venny 2.1 and further analyzed via KEGG pathway enrichment. Pearson correlation analysis of QC samples confirmed data quality (Supplementary Figure S1).

Proteins were isolated from liver tissues and measured using BCA method. After normalization of protein levels with deionized water and the addition of 5× sample buffer, the mixtures were heat-treated at 95°C for 10 min to achieve denaturation. Equivalent protein quantities were loaded onto 10% SDS-PAGE gels for separation, followed by electrotransfer onto PVDF membranes. Membranes were first treated with primary antibodies before being probed with horseradish peroxidase-labeled anti-rabbit secondary antibodies. Protein signals were detected and captured with a Bio-Rad imaging platform (Bio-Rad, United States). Band intensity was quantitatively analyzed using ImageJ software.

Quantitative analysis of inosine monophosphate (IMP), adenine, hypoxanthine, adenosine, N⁶-methyladenosine (m6A), and adenosine monophosphate (AMP) in mouse liver tissues was performed by the external standard calibration method. The chromatographic and MS parameters agreed with those utilized in the metabolomics analysis. Details of the standard curves and scanning parameters are provided in Supplementary Tables S2-S3.

Liver concentrations of TNF-α, IL-1β, IL-6, and TGF-β1 were determined using ELISA kits, strictly following the manufacturers’ recommended protocols.

Data were presented as mean ± SEM. Statistical comparisons across groups were made using one-way ANOVA, followed by Tukey’s post hoc test. Analyses were conducted using GraphPad Prism 8.0, and p-values less than 0.05 were deemed statistically significant.

A comprehensive chemical analysis of ILW was conducted using UPLC-Q-TOF/MS in positive and negative modes (Figure 2). A total of 32 metabolites were identified, including 9 flavonoids, 6 terpenoids, 14 organic acids, and 3 other metabolites. Among these, 7 metabolites were unequivocally confirmed through comparison with authentic reference standards. An additional 8 metabolites were annotated based on database searches, including PubChem (https://pubchem.ncbi.nlm.nih.gov/), MassBank (https://massbank.eu/MassBank/), and HMDB (https://www.hmdb.ca/). The remaining 17 metabolites were tentatively identified by comparison with previously published literature. Detailed information on all identified metabolites in ILW is summarized in Supplementary Table S4. For instance, metabolite F1 showed a [M + H]⁺ ion at m/z 593.150 8, consistent with the molecular formula C₂₇H₃₀O₁₅. Sequential neutral losses of C₃H₆O₃ and C₄H₈O₄ yielded fragment ions at m/z 503.120 5 and 473.107 4, respectively. Subsequent loss of either C₃H₆O₃ or C₄H₈O₄ from m/z 473.107 4 produced ions at m/z 383.074 0 and 353.065 9. This fragmentation behavior, involving characteristic cross-ring cleavages of sugar moieties, is typical of flavone C-glycosides (Guo et al., 2022). Based on MS/MS data, literature comparisons (Picariello et al., 2025), and a reference standard, F1 was identified as vicenin-2. The proposed fragmentation pathway is illustrated in Supplementary Figure S2. In addition, the UPLC chromatograms of the extracts and reference standard of the powdered plant material at three wavelengths are listed in Supplementary Figure S3.

The chromatographic profiles of ILW and corresponding reference standards are presented in Supplementary Figure S4, with calibration curves provided in Supplementary Table S5. The amounts of the following main metabolites in ILW were ascertained by quantitative analysis: caffeic acid (1.10 mg g^-1), schaftoside (0.12 mg g^-1), isoschaftoside (0.18 mg g^-1), and rosmarinic acid (5.13 mg g^-1). Furthermore, reproducibility was assessed across 12 batches of ILW samples (Supplementary Figure S5). The high similarity values (>0.95), as detailed in Supplementary Tables S6 and S7, demonstrate good batch-to-batch consistency and support the reliability and robustness of the ILW preparation process.

ALT and AST are widely recognized indicators of liver damage. As shown in Figures 3A,B, serum levels of both ALT and AST were significantly elevated in the model group compared with the control group (p < 0.01 or p < 0.05), reflecting substantial liver damage induced by fibrosis. Remarkably, ILW administration resulted in a significant reduction in these enzyme levels, with the most pronounced effects observed in the high-dose group (p < 0.001), suggesting its potential to alleviate liver injury and restore hepatic function.

Key indicators of hepatic fibrosis, including serum levels of PC-III, COL-IV, LN, and HA, were markedly higher in the model group than in the control (p < 0.001 or p < 0.01), as illustrated in Figures 3C–F. ILW treatment significantly reduced their levels across all dosage groups (ILW-H, ILW-M, and ILW-L) (p < 0.01 or p < 0.05). Notably, ILW-H group exhibited the most pronounced reductions, particularly in PC-III, LN, and HA, surpassing the efficacy of colchicine. These findings highlight ILW’s potential antifibrotic activity and its capacity to mitigate liver fibrosis progression.

Histological examination through HE staining revealed extensive hepatocellular disorganization, inflammatory cell infiltration, fibrous tissue proliferation, and architectural disruption of hepatic lobules in the model group (Figure 3G). Colchicine treatment partially restored hepatocyte morphology, improving cellular arrangement and reducing inflammation. Similarly, ILW treatment (high, medium, and low doses) alleviated hepatic injury to varying degrees, progressively restoring normal hepatocyte structure (Figure 3G). Masson staining further confirmed extensive collagen fiber accumulation and fibrous septa formation in the model group, indicating severe hepatic fibrosis. In contrast, colchicine treatment preserved liver architecture with reduced collagen deposition. ILW administration markedly diminished fibrosis, as evidenced by decreased collagen accumulation and improved hepatic lobule integrity across all treatment groups (Figure 3G). These findings highlight ILW’s potential in protecting hepatocytes and mitigating liver fibrosis.

The liver, as the primary target of liver fibrosis, undergoes extensive gene expression alterations during disease progression. To explore ILW’s regulatory effects, transcriptome sequencing was conducted on liver tissues. PCA revealed distinct clustering among the groups, with ILW-H partially reversing CCl₄-induced gene expression changes along the Y-axis (Figure 4A). The two principal components accounted for 26.60% and 22.29% of the variance. A total of 2,341 differentially expressed genes (DEGs) were screened between the Model and Control groups, together with between the ILW-H and Model groups, based on volcano plot analysis using thresholds of FDR <0.05 and FC > 1.5 or <0.67 (Figures 4B,C; Supplementary Figure S6). Specifically, 343 DEGs (123 upregulated, 220 downregulated) distinguished the model from the control group, while ILW-H treatment resulted in 2,169 DEGs (978 upregulated, 1,191 downregulated) in the ILW-H vs. Model comparison (Figures 4D,E). Further analysis identified 171 overlapping DEGs between the two comparisons (Supplementary Table S8). KEGG pathway enrichment analysis highlighted several pathways implicated in disease progression and ILW-mediated therapeutic effects. Seven pathways were enriched in both comparisons (Figure 4F), including those related to inflammation and immunity (Rap1 signaling), metabolism (purine metabolism, glycerolipid metabolism, HIF-1 signaling), and cell proliferation and apoptosis (PI3K-Akt, calcium, and phosphatidylinositol signaling pathways).

To investigate protein expression changes associated with liver fibrosis and ILW intervention, label-free proteomic profiling was performed on liver tissues utilizing UHPLC-Orbitrap Astral MS. PCA analysis (Figure 5A) revealed distinct clustering among groups, indicating significant proteomic differences, while intra-group samples remained tightly clustered. DEPs were identified based on a threshold of p < 0.05 and FC > 1.2 or <0.83. A total of 1,305 DEPs were detected across comparisons involving the Model group versus the Control and ILW-H groups (Figures 5B,C; Supplementary Figure S7). Specifically, 853 DEPs (504 upregulated, 349 downregulated) were detected in the Model vs. Control group, while ILW-H treatment resulted in 673 DEPs (276 upregulated, 397 downregulated) in the ILW-H vs. Model group (Figures 5D,E). Further analysis identified 221 overlapping DEPs between the two comparisons (Supplementary Table S9). KEGG pathway enrichment analysis highlighted 14 common pathways (Figure 5F), encompassing inflammation and immunity (Rap1, MAPK, NF-κB, and JAK-STAT pathways), metabolism (purine metabolism, sphingolipid signaling, AMPK, HIF-1 pathways), and cell proliferation and apoptosis (PI3K-Akt, calcium, Ras, FoxO, ErbB, and phospholipase D pathways).

Comprehensive metabolic profiling of hepatic tissues from the control, model, and ILW-H-treated groups was performed using UPLC-MS/MS, resulting in the detection of 244 metabolites. PCA analysis (Figure 6A) revealed clear separation of the samples into three distinct clusters: the control, model, and ILW-H groups. This indicated significant differences between CCl₄-induced liver fibrosis and the control, while the ILW-H group showed a partial restorative effect. DMs were identified based on volcano plot criteria (p < 0.05, FC > 1.5 or <0.67). A total of 121 DMs were detected between the model and control groups, including 49 upregulated and 72 downregulated metabolites. In comparison, ILW-H treatment resulted in 62 DMs relative to the model group, with 20 upregulated and 42 downregulated (Figures 6B,C). Venn diagram analysis revealed 35 overlapping DMs between the two comparisons (Figure 6D; Table 1), and their relative abundance is visualized in a heatmap (Figure 6E). Pathway enrichment analysis using the MetaboAnalyst platform identified five significant metabolic pathways (p < 0.05), including purine metabolism, cysteine and methionine metabolism, glycerolipid metabolism, pentose phosphate pathway, and glycolysis/gluconeogenesis (Figure 6F).

To uncover central molecular mechanisms underlying ILW’s antifibrotic action, an integrated analysis of transcriptomic, proteomic, and metabolomic datasets was performed. Cross-omics comparison identified 12 enriched pathways shared between transcriptomics and proteomics, 3 overlapping between transcriptomics and metabolomics, and another 3 common to proteomics and metabolomics (Figure 7A). Notably, purine metabolism and glycerolipid metabolism were consistently enriched across all three omics layers (Figure 7B). Purine metabolism plays an essential role in maintaining intracellular energy homeostasis, regulating inflammatory signaling cascades, and modulating the activation of hepatic stellate cells (HSCs). In parallel, glycerolipid metabolism is integral to lipid equilibrium and membrane remodeling. In the context of CCl₄-induced liver fibrosis, purine metabolism emerges as particularly relevant, as its intermediates—most notably adenosine—function as key bioactive molecules capable of modulating the NF-κB signaling pathway. This pathway, consistently identified in both transcriptomic and proteomic analyses, serves as a central node in the pathogenesis of liver fibrosis, orchestrating inflammatory responses, immune cell recruitment, HSC activation, and apoptotic processes. These findings collectively underscore the purine metabolism and NF-κB signaling as critical pathways for further in-depth mechanistic investigation.

To further elucidate ILW’s influence on purine metabolism, DEGs, DEPs, and DMs were mapped to this pathway. As shown in Figure 8A, ILW-H modulated a total of 9 DEGs, 5 DEPs, and 6 DMs. Specifically, 5 DEGs (Pde4a, Pde4c, Adcy4, Adcy6, and Adcy7) were upregulated, while 4 (Adcy1, Enpp4, Ak7, and Itpa) were downregulated. Additionally, ILW-H treatment increased the expression of 3 DEPs (Nt5e, Ak2, and Ak3) while reducing 2 (Nme6 and Hprt1). Metabolomic profiling further identified 3 upregulated DMs (adenosine, hypoxanthine, and m6A) and 3 downregulated DMs (IMP, AMP, and adenine). Functionally, adenylate cyclase types 1, 4, 6, and 7 (ADCY1, ADCY4, ADCY6, ADCY7) convert adenosine triphosphate (ATP) to cyclic adenosine monophosphate (cAMP), whereas phosphodiesterase 4A and 4C (PDE4A and PDE4C) hydrolyze cAMP into adenosine monophosphate (AMP). Meanwhile, adenylate kinases (AK7, AK2, and AK3) facilitate the exchange of phosphate groups between ATP and AMP, maintaining the balance of ATP, AMP and adenosine diphosphate (ADP). Besides, Hprt1 and Itpa are involved in the synthesis of inosine monophosphate (IMP).

Based on metabolomics results, we further analyzed seven purine metabolites, including IMP, AMP, adenine, adenosine, hypoxanthine, m6A, and inosine. These are key intermediates in both degradation and salvage branches of purine metabolism. Inosine was included due to its close connection—it is derived from AMP, IMP, or adenosine, and is a precursor of hypoxanthine. Targeted quantification (Supplementary Tables S2–S3) showed that ILW-H significantly decreased hepatic IMP, AMP, and adenine levels (p < 0.05), while increasing adenosine, hypoxanthine, and m6A (p < 0.01) (Figure 8B). Inosine showed a rising trend after ILW-H intervention, though the change was not statistically significant (Supplementary Figure S8). Notably, Nt5e protein, which catalyzes AMP dephosphorylation to adenosine, was significantly upregulated in ILW-H group (p < 0.05, Supplementary Figure S9). This aligns with the observed elevation in adenosine. Together, these results suggest that ILW exerts a multi-level regulatory effect on purine metabolism in fibrotic liver tissue.

Similarly, we mapped the data from transcriptomics, proteomics, and metabolomics to the NF-κB pathway (Figure 9A). Relative to the model group, ILW-H treatment led to reduced expression of seven critical genes (LBP, CD14, TLR4, TAB, TNF-α, TNF-R1, and IL-1R) as well as two key proteins, CD14 and TLR4, suggesting a suppressive effect on the NF-κB pathway. Considering the pivotal role of TLR4, MyD88, and NF-κB in mediating inflammatory responses, their protein expression levels were further evaluated by Western blot analysis. As shown in Figure 9B, the model group exhibited significantly increased expression of TLR4, MyD88, phosphorylated NF-κB p65 (p-NF-κB), and the p-NF-κB/NF-κB ratio compared with the control group (p < 0.05, p < 0.01, or p < 0.001). These increases were attenuated by ILW treatment, particularly at medium (ILW-M) and high doses (ILW-H) (p < 0.05, p < 0.01, or p < 0.001), indicating effective suppression of NF-κB activation. Moreover, downstream cytokines regulated by NF-κB signaling, including TNF-α, IL-1β, TGF-β1, and IL-6, were quantitatively assessed in liver tissues via ELISA (Figure 9C). The model group showed significantly elevated levels of these cytokines relative to the control group (p < 0.01 or p < 0.001). ILW treatment significantly reduced hepatic concentrations of TNF-α (p < 0.05 or p < 0.01), IL-6 (p < 0.05 or p < 0.01), IL-1β (p < 0.05 or p < 0.01), and TGF-β1 (p < 0.05 or p < 0.01), especially at medium and high doses. Collectively, these results indicate that ILW mitigates the activation of the TLR4/MyD88/NF-κB signaling pathway and suppresses the production of pro-inflammatory cytokines, thereby exerting protective effects against CCl₄-induced hepatic inflammation and fibrosis.