Ganlong capsules were purchased from Saino Pharmaceutical Co., Ltd., colchicine tablets from Yunnan Phytology Co., LTD. and aminotransferase (AST), alanine aminotransferase (ALT) assay kits from Nanjing Jiancheng Biological Engineering Institute.

Laminin (LN), type III procollagen (PC III), hyaluronic acid (HA) and type IV collagen (Col-IV) were assayed in liver tissue by ELISA kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). CCl₄ was purchased from Tianjin Fuchen Chemical Reagent Factory (Lot No: 2021030), Voila extra virgin olive oil from yihai grain and oil industry co., LTD. (Lot No: 20201205), ammonium acetate (Lot No: 631-61-8) from SIGMA and methanol (Lot No: 67-56-1) and acetonitrile (Lot No: 75-05-8) from CNW.

Six-week-old Sprague-Dawley (SD) male rats (160–200 g) were purchased from Hunan Jingda Experimental Animal Co., Ltd. (No. SXCK 2019–0004) and acclimated to a temperature of 25°C ± 2°C with a 12 h light/12 h dark cycle for 3 days prior to experiments.

Ethical approval was granted by the Experimental Animal Ethics Committee of Dali University.

CCL₄ diluted with olive oil to give a 40% solution was given at a dose of 1 mL/kg body weight by intraperitoneal injection twice a week for 8 weeks.

Thirty-six rats were randomly divided into two groups: normal control (n = 6) and model (n = 30). Model rats received CCl₄ and control rats an equal volume of solvent by intraperitoneal injection twice a week for 9 weeks. Liver fibrosis was ascertained by hematoxylin-eosin (HE) staining of liver samples.

After confirmation of successful modelling, model rats were randomly divided into five groups (n = 6): Ganlong capsules high-dose (250 mg/kg), medium-dose (125 mg/kg) and low-dose (62.5 mg/kg), colchicine (0.2 mg/kg) and model (untreated).

Colchicine and Ganlong capsules were given at a dose of 0.5 mL/100 g by intragastric administration while model and control rats received an equal volume of normal saline by intragastric administration once a day for 4 weeks. Rat body weights were recorded at 2 days intervals for 13 weeks. Urine was collected from fasting rats during 12 h of darkness following the final treatment administration and blood collected from the abdominal aorta after anesthetization. Livers and spleens were extracted and weighed for calculation of the viscera index (%) by the following formula: viscera weight (g)/body weight (g) × 100%.

Rats were euthanized by intraperitoneal injection of 300 mg per kg pentobarbital sodium, according to the animal euthanasia guidelines of the Experimental Animal Ethics Committee of Dali University.

Blood was left to stand at room temperature for 2 h and centrifuged at 3, 000 rpm for 10 min to obtain serum. Hepatic function was assessed by measurement of serum ALT and AST by kit, according to the manufacturer’s instructions.

Hyaluronic acid (HA), laminin (LN), type III procollagen (PC III) and type IV collagen (Col-IV) in liver tissue were detected by ELISA kit, according to the manufacturer’s instructions.

One mm³ of rat liver was fixed with 4% tissue cell fixative, embedded in paraffin and sectioned. HE and Masson staining were performed and observed under an inverted microscope at ×200 magnification to score the degree of liver fibrosis.

Differential metabolites in serum, urine and liver tissue from control, model and Ganlong capsules treated groups were screened by liquid chromatography-mass spectrometry (LC-MS).

100 μl serum or urine was transferred to EP tubes, 400 μl extraction solution (equal volumes of methanol and acetonitrile) added with vortex mixing for 30 s. Samples were sonicated in ice water for 10 min, left to stand at −40°C for 60 min and centrifuged at 12,000 rpm at 4°C for 20 min.

25 mg liver tissue was added to 500 μl extraction solution (methanol: acetonitrile: water at 2:2:1 (V/V), containing isotope-labeled internal standard mixture), ground for 4 min (35 Hz) and sonicated for 5 min in an ice water bath. The above steps were repeated 3 times. Samples were incubated in a water bath at 40°C for 1 h and centrifuged at 12,000 rpm at 4°C for 20 min. Supernatants were used for analysis and quality control samples prepared by mixing equal amounts of supernatant from all samples.

Target compounds were separated by Vanquish UPLC with an ACQUITY UPLC HSS T3 1.8 μm 2.1 × 100 mm liquid chromatography column. The liquid chromatographic A phase contained 25 mmoL/L ammonia and 25 mmoL/ammonium acetate and the B phase acetonitrile. LC-MS was performed, according to the instrument operation manual.

UHPLC-Q Exactive mass spectrometry information was visualized using R language, peak area data retained for all groups with no more than 50% null values, normalized (using internal standard) and missing values simulated (minimum value of one-half).

Multivariate analysis of pre-processed data was performed by SIMCA-P 16.0 and unsupervised principal component analysis (PCA) applied to asses system stability and repeatability and distribution among samples, followed by supervised mode orthogonal partial least squares analysis (OPLA). Sample metabolic profiles and differential metabolites were evaluated by partial least squares-discrimination analysis (OPLS-DA), verified by response permutation testing (RPT) to identify “overfitting”. Fold change (FC), variable important in projection (VIP) and p-values were visualized by the OPLS-DA model with FC > 1.2 or FC < 0.8 and VIP >1. Values of FC < 0.8, VIP >1 and p < 0.05 were used to screen differential metabolites considered as potential biomarkers.

Differential metabolites were also identified using the human metabolome database HMDB (http://www.hmdb.ca/) and the KEGG database (http://www.kegg.jp/kegg/pathway.html).The MetaboAnalyst 5.0 website (http://www.metaboanalyst.ca/) was used to perform metabolic pathway analysis and clustering heat mapping on the basis of Raw p < 0.05.

All data are expressed as mean ± SD for at least three independent experiments. Prism 9.0 software (GraphPad Software, United States) was used for statistical analysis and comparisons between groups assessed by one-way ANOVA. A value of p < 0.05 was regarded as statistically significant.

Liver index, spleen index, body weight, levels of liver fibrotic markers, PC III, LN, Hyp, Col-IV, HA, activities of AST, and ALT and pathological changes in the liver were all measured 13 weeks after administration of Ganlong capsules.

Control rats gained weight at non-statistically significantly higher rate than rats in other groups (Figure 1A) and CCl₄ treatment resulted in significantly higher liver and spleen indexes, although values were reduced by Ganlong capsules (p < 0.01, Figure 1B).

ALT and AST activities were higher in fibrosis model rats (p < 0.01) and administration of Ganlong capsules reduced the degree of liver fibrosis (p < 0.01, Figure 1C) and downregulated markers of hepatic fibrosis, including Hyp, HA, LN, PC III, and Col-IV (Figures 1D, E).

General liver morphology and Masson and HE staining indicated that liver cells of control rats were closely and radially arranged with few inflammatory cells but enlarged confluent area, obvious inflammatory cell infiltration, widely spaced collagen fibres and disorganized lobular structure and hepatocyte arrangement were all present in fibrosis model rats (Figure 2A).

Treatment with either colchicine and Ganlong capsules ameliorated liver damage (Figure 2B) and reduced total collagen deposition, decreased fibrotic lesions and normalized lobular structures (Figure 2C).

Endogenous metabolites were detected in positive and negative ionization modes by UHPLC-Q Exactiv and the Base Peak Chromatogram (BPC) obtained for each group (Figure 3).

PCA showed tight aggregation in both positive and negative ion modes for quality control samples (QC), indicating good system stability and reproducibility (Figure 4A). R2X values were all greater than 0.5. Significant separation between the fibrosis model and control groups was seen in both positive and negative ion patterns, indicating significant differences in metabolic profile. Values for the group treated with ganlong capsules showed a trend away from the model group tending towards the normal control group. Two-by-two analysis revealed that the control and model groups were completely separated in the positive and negative ion mode (Figure 4B) and that the ganlong capsules administration group was separated from the model group (Figure 4C).

OPLS-DA analysis of metabolic profiles showed the separations of both the normal control and ganlong capsules group from the fibrosis model group in positive and negative ion mode (Figures 4D, G). OPLS-DA parameters gave R2Y values greater than 0.9 and Q2 values greater than 0.5 with positive slopes for R2Y and Q2 and negative intercepts for Q2, indicating no “overfitting” and good validity and reliability (Figures 4E, H). VIP, P, and FC values were visualized by volcano plot to enable screening for differential metabolites (Figures 4F, I).

Intraperitoneal injection of CCl₄ led to significant differences in liver metabolic profiles between the model and control groups, indicating the successful generation of the liver fibrosis model at the metabolic level. Separation of the ganlong capsules and model groups indicated alterations of liver metabolic profiles resulting from treatment.

A total of 43 differential metabolites were identified with a VIP >1 and p < 0.05 (fold change (FC) > 1.2 or <0.8) between model vs. control groups or Ganlong capsules vs. model groups (Table 1). Clustering analysis was performed via the MetaboAnalyst 5.0 website (Figure 5).

Levels of phosphatidylethanolamine [PE (16:0/18:3 (9Z, 12Z, 15Z], phosphatidylcholine [PC (20:4 (8Z, 11Z, 14Z, 17Z)/P-18:0)], ethanolamine phosphate (O-Phosphoethanolamine) (R) -3-hydroxybutyric acid (R) -3-Hydroxybutyric acid), gamma-Aminobutyric acid (gamma-Aminobutyric acid) and 20 other metabolites were significantly elevated in fibrotic liver tissue compared with controls (p < 0.05) and 23 metabolites, including 2-ketobutyric acid (2-Ketobutyric acid), were significantly decreased (p < 0.05). Ganlong capsule administration produced a significant trend of regression towards control values (p < 0.05).

The pathways of glycerophospholipid metabolism (p = 0.042, impact = 0.223) and butanoate metabolism (p = 0.041, impact = 0.032) were analyzed on the basis of values of p < 0.05 (Table 2).

Metabolic characteristics of endogenous serum metabolites were analyzed by UHPLC-Q Exactive system and serum BPC plots obtained in positive and negative ion mode (Figure 6).

Unsupervised PCA analysis showed aggregation of QC samples in positive and negative ion modes, indicating good stability and reliability. Some separation of the control, model and Ganlong capsule groups was shown but was incomplete, perhaps due to noise interference (Figure 7A). PCA maps also showed incomplete separation of the model versus control groups and the Ganlong capsule versus model groups in the positive and negative ionization mode (Figures 7B, C).

Supervised OPLS-DA modeling analysis was implemented to eliminate noise and showed the control group to be completely separated from the model group and the Ganlong capsule group to be separated from the model group under the positive and negativeion model (Figures 7D, G). R2Y values were greater than 0.9 but Q2 values mostly less than 0.5, indicating excellent explanatory but poor predictive ability of the models. Response ranking tests gave positive R2Y and Q2 slopes and a negative Q2 intercept, excluding “overfitting” and confirming validity and reliability (Figures 7E, H). Red and blue dots represent differential metabolites meeting the conditions of VIP >1, P < 0.05 and FC > 1.2 or FC < 0.8 in Figures 7F, I.

A total of 21 potential biomarkers were identified in rat serum as being either up- or downregulated (p < 0.05) by Ganlong capsule treatment. Serum PC (PC (20:2 (11Z, 14Z)/20:4 (5Z, 8Z, 11Z, 14Z)), PC (20:4 (8Z, 11Z, 14Z, 17Z)/P-18:0)), PE (PE (O-16:1 (1Z)/22:6 (4Z, 7Z, 10Z 13Z, 16Z, 19Z)) and thymidine (thymine) showed 17 significantly elevated metabolites (p < 0.05) and malonic acid four significantly lower metabolites in the model compared with control groups and restoration of values was shown for all in the Ganlong capsule group (p < 0.05; Table 3). Heat maps were drawn using the Metaboanalyst 5.0 website to visualize clustering of potential serum biomarkers.

Pathways of glycerophospholipid metabolism (p = 0.021, impact = 0.199), pyrimidine metabolism (p = 0.024, impact = 0.097) and linoleic acid metabolism (p = 0.032, impact = 0.00) were analyzed since metabolites had values of p < 0.05 (Figure 8 and Table 4).

Metabolic characteristics of endogenous urine metabolites were analyzed by UHPLC-Q Exactive system and serum BPC plots obtained in positive and negative ion mode (Figure 9).

Unsupervised PCA of QC samples produced aggregated results showing good reliability. However, the separation of samples from the three groups was not clear in positive and negative ion modes (Figure 10A). PCA plots of model versus control (14 B) and Ganlong capsule versus model (14C) were separate in positive and negative ion mode with no outlier points.

OPLS-DA analysis was performed and validated by 200 times response ranking test. A volcano plot was constructed to identify potential biomarkers.

OPLS-DA results showed separation of controls from model and of Ganlong capsules from model samples in positive and negative ion mode. R2Y values were all greater than 0.9 and Q2 values all greater than 0.5 (Figures 10D, G). The 200 response ranking tests showed R2Y and Q2 values to be consistent with the model test parameters and “overfitting” to be absent (Figures 10F, H). Dot size corresponds to VIP value in the volcano plot with metabolites arranged in a V-shape, higher p-values at the bottom and smaller p-values at the top.

A total of 20 potential biomarkers which had been up- or downregulated by Ganlong capsules (p < 0.05) were identified in the urine of rats with liver fibrosis. Ginkgolide B levels were higher in urine from the model group (p < 0.05) and decreased after Ganlong capsule treatment (p < 0.05). Levels of 19 metabolites of choline (Choline), betaine aldehyde (Betaine), and L-leucine (L-Leucine) were lower in fibrotic urine (p < 0.05) and increased after Ganlong capsule treatment (p < 0.05) (Table 5). Heat maps of potential biomarker clusters were drawn via the Metaboanalyst 5.0 website.

The identification of differential metabolites allowed the selection of two pathways of interest: glycine, serine and threonine metabolism (p = 0.014, impact = 0.026) and valine, leucine and isoleucine biosynthesis (p = 0.046, impact = 0.000; Figure 11 and Table 6).

Glycerophospholipid metabolism in liver (p = 0.042, impact = 0.223) and serum samples (p = 0.021, impact = 0.199) and linoleic acid metabolism (p = 0.032, impact = 0.000) were aspects of lipid metabolism found to be affected in CCl₄-induced liver fibrosis. Glycerophospholipids, such as PE and PC, are key components of biological membranes and their metabolism affects membrane fusion, endocytosis and transport. Disorders of glycerophospholipid metabolism are thought to adversely affect the liver (Xue et al., 2020). The essential omega-6 polyunsaturated fatty acid (PUFA), linoleic acid, is essential for growth and development, cell function, signal transduction and the immune response (Lv et al., 2020). CCl₄ has been shown to damage hepatocyte membranes and its metabolism by hepatocyte cytochrome P450 enzymes may generate free radicals. The consequent lipid peroxidation damages cell and organelle membranes, allowing accumulation of PC and PE, leading to hepatic steatosis and aggravating liver fibrosis (Weber et al., 2003; Yumeng et al., 2021). Raised PC and PE levels are considered to be markers of liver fibrosis in the clinic (Huang et al., 2011). The identification of five differential metabolites of PE, PC and ethanolamine phosphate with elevated levels during the present work indicates the involvement of glycerophospholipid and linoleic acid metabolism in fibrosis. Ganlong capsule treatment reduced the levels of these metabolites. Thus, Ganlong capsules ameliorated CCl₄-disrupted glycerophospholipid metabolism, in agreement with previous reports (Yongcan, 2022).

Altered butyric acid metabolism is central to CCl₄-induced liver fibrosis with elevated levels of gamma-aminobutyric acid (GABA), an inhibitory neuro-signalling transmitter, being involved (Zhu et al., 2018). Liver fibrosis and cirrhosis are thought to be associated with an imbalance in amino acid metabolism that raises blood ammonia levels and triggers hepatic encephalopathy (Butterworth, 2019). The present study found two metabolites of butyric acid (R) -3-hydroxybutyric acid and γ-aminobutyric acid, to be significantly upregulated in the fibrosis model, a change which was reversed by Ganlong capsule treatment.

Liver tissue has a regenerative capacity which depends on active pyrimidine anabolism. Furthermore, inhibition of pyrimidine synthesis prevents excessive cell proliferation and protects against liver fibrosis (Cunshuan, 2010). Thymidine levels were greatly enhanced in the liver fibrosis model of the current study and decreased following Ganlong capsule administration. CCl₄-damaged liver tissue may have activated the self-protective mechanism of secreting ECM, increasing formation of collagen fibres. In contrast, Ganlong capsules reduced pyrimidine levels while alleviating abnormal proliferation of liver cells, reversing liver fibrosis. The levels of deoxycytidine and uracil were much higher in model rats than in controls, consistent with previous reports (Liu et al., 2013). Ganlong capsules reduced deoxycytidine and uracil levels to approximate control values. Thus, Ganlong capsules may exert anti-liver fibrosis effects by lowering pyrimidine levels and reducing excessive tissue proliferation.

The present study showed glycine, serine and threonine metabolism and valine, leucine and isoleucine biosynthesis to be involved in CCl₄-induced liver fibrosis. The liver is a key site for amino acid metabolism and damage to this organ is known to impact these metabolic pathways (Trefts et al., 2017).

Choline and betaine aldehyde, involved in glycine, serine and threonine metabolism and L-leucine, involved in valine, leucine and isoleucine biosynthesis, were metabolites identified during the present work as being central to fibrosis. Choline and betaine aldehyde were lower in fibrosis but restored by Ganlong capsules. This suggests that CCl₄-induced liver fibrosis is associated with glycine, serine and threonine metabolism and Ganlong capsules exert hepatoprotective effects through an impact on these pathways. The branched-chain amino acid, L-leucine, was lower in fibrosis and restored by Ganlong capsules. Thus, Ganlong capsules may exert anti-fibrotic effects via influencing branched chain amino acid metabolism.

Liver damage, resulting from fibrosis or hepatitis, is characterized by an inflammatory state. Arachidonic acid is stored as a membrane phospholipid and converted to inflammation-inducing prostaglandins by cyclooxygenase to exacerbate inflammation (Stringer, et al., 2016). The current investigation found arachidonic acid metabolism to be disturbed by CCl₄, producing high levels of prostaglandin E2 which was reduced by Ganlong capsules. This may be another route by which Ganlong capsules exert a hepatoprotective effect.