All animal-related protocols were approved by the Ethical Committee of Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China). Sprague Dawley (SD) rats (male, 6–8 weeks old) weighing 200–250 g were purchased from the Experimental Animal Center of School of Medicine, Shanghai Jiao Tong University (Shanghai, China). The rats were maintained in our specific pathogen-free facility under controlled conditions (22°C, 40–60% humidity, and 12-h light/dark cycle), and free access to tap water and standard rat food was given to the rats.

1-(4-trifluoromethoxyphenyl)-3-(1-propionylpiperidin-4-yl)-urea, TPPU, and 4-(5-phenyl-3-{3-[3-(4-trifluoromethylphenyl)-ureido]-propyl}-pyrazol-1-yl)-benzenesulfonamide, PTUPB, were synthesized according to the previous procedures (12). TPPU was generously provided by the laboratory of Dr. Bruce Hammock (UC Davis, USA) and stored at 20°C. TPPU and PTUPB were dissolved in PEG-400 to give a 10g/L clear solution. This solution was then added to warm drinking water with rapid stirring to give the 100 mg/L solution of TPPU/PTUPB in drinking water. Based on the estimation of daily water consumption, a concentration of 100 mg/L inhibitor in drinking water will result in a dose of approximately 10 mg/kg/day. The other rats received vehicle (PEG 400 diluted in water) as control. PTUPB and TPPU were administered for 4 weeks from the 12th week in the CCl₄ model group.

A total of 20 rats were used in this study. Rats were divided into four subgroups as follows: control group (OO-VEH) received pure olive oil injection with vehicle administration (n = 5); PHT group received carbon tetrachloride (CCl₄) (50% in olive oil, v/v, 1 ml/kg) by subcutaneous injection (s.c.) two times a week for 16 weeks with vehicle administration (CCl₄-VEH) (n = 5); PHT group with TPPU administration (CCl₄-TPPU) (n = 5); and PHT group with PTUPB administration (CCl₄-PTUPB). The detailed grouping strategy was shown in Figure 1.

After finishing modeling, rats were anesthetized with 40 mg/kg zolazepam and tiletamine (Zoletil 50, France) and 8 μg/kg dexmedetomidine hydrochloride (Dexdomitor^®, Pfizer Inc. USA) through intramuscular injections. A PE-50 catheter (Smiths Medical, UK) was inserted into the right femoral artery to determine the heart rate (HR) and mean arterial pressure (MAP). The catheter was then placed into the portal vein to determine the portal pressure (PP). A transducer linked the catheter to the monitor, and the values were acquired using a multichannel physiological signal collection system (ALC-MPA multichannel bioinformatics analysis system, Shanghai Alcott Biotechnology Co., Ltd., China). Following sacrifice of the rats, blood, liver, and superior mesenteric arteries (SMA) were taken for enzymatic analysis, histological, and molecular analysis.

Serum levels of interleukin (IL)-6 and hyaluronan were determined by IL-6 rat ELISA kit (Thermo Fisher, USA) and hyaluronan rat ELISA Kit (R&D Systems, USA), respectively, according to the manufacturer's instructions.

The liver sections from the right lobe and mesenteric tissues were fixed in 10% formalin buffer (pH 7.4) and embedded in a paraffin block. Hematoxylin–eosin (H&E), Masson, and Sirius Red staining were used on the liver sections, followed by random evaluation under a light microscope by an expert pathologist.

For IHC staining, liver sections were incubated with antimatrix metalloproteinase-2 (MMP-2) antibody (1:150, Servicebio, China), antimatrix metalloproteinase-9 (MMP-9) antibody (1:300, Servicebio, China), antivascular endothelial growth factor (VEGF) A antibody (1:250, Servicebio, China), antivon Willebrand factor (vWF) antibody (1:1,000, Servicebio, China), anticytokeratin 19 (CK-19) antibody (1:1,000, Servicebio, China), anti-sEH antibody (1:50, Absin, China), anti-CD31 antibody (1:300, Servicebio, China), and anti-CD68 antibody (1:150, Servicebio, China), overnight at 4°C with phosphate-buffered saline (PBS) as negative control. Subsequently, the sections were incubated with appropriate HRP-conjugated goat antirabbit secondary antibody for 60 min, followed by restaining with hematoxylin. The collagen deposition volume and stained area were calculated with IHC Profiler plugin in ImageJ (version 1.53, USA). The results were expressed as the proportions of the stained areas. The total area and average values were taken from five rats in each group.

Hepatic functions, including the total bilirubin (TBIL), direct bilirubin (DBIL), aspartate aminotransferase (AST), alanine transaminase (ALT), and γ-glutamyl transferase (GGT), were examined using kits from Changchun Huili Biotech (China) according to the manufacturer's instruction.

Samples from liver tissue were taken and kept at −80°C. To extract protein from the liver, the sample was crushed in liquid nitrogen and homogenized according to the manufacturer's instructions using RIPA buffer (Beyotime, China). The liver extracts were then centrifuged for 15 min at 10,000 g for 15 min at 4°C. The supernatant was immediately collected, and the BCA protein analysis kit was used to determine the total protein content (Beyotime, China). Equal amounts of proteins were electrophoresed on sodium dodecyl sulfate-polyacrylamide gels (SDS-PAGE) and then electrotransferred onto polyvinylidene difluoride (PVDF) membranes. Primary antibodies against α-smooth muscle actin (α-SMA) (1:1,000, Servicebio, China), COX-2 (1: 750, Servicebio, China), and sEH (1:1,000, Absin, China) were used to incubate the blots (1:1,000, Servicebio, China). The membranes were then treated with corresponding secondary antibodies. Immunoreactive bands were visualized using an electrochemiluminescence instrument (Vilber Lourmat, France) and quantified using digital image software (Kodak, USA).

The mRNA expressions of TGF-β, epoxide hydrolase 2 (EPHX2), COX2, α-SMA, collagen type I alpha 1 chain (COL1A1), MMP2, MMP9, VEGF, VWF, angiopoietin 1 (ANGPT1), endothelial nitric oxide synthase (eNOS), GTP cyclohydrolase 1 (GCH1), and adhesion G protein-coupled receptor E1 (Adgre1, F4-80) were determined. qRT-PCR was conducted on a Bio-Rad iCycleriQ real-time PCR detection system (Bio-Rad laboratories, Germany) using IQ SYBR Green supermix Kit (Bio-Rad). The reaction was carried out in triplicate. The data were analyzed using the iCycleriQ software system (Bio-Rad, Germany).

Potential target genes were obtained from Pharm Mapper (http://lilabecust.cn/pharmmapper/) and SwissTargetPrediction [https://swisstargetprediction.ch/; probability value ≥0.10 was selected (13)]. Afterward, we entered the keyword “liver fibrosis” into the GeneCards (https://www.genecards.org/) (14) to obtain target genes related to liver fibrosis. The differentially expressed genes related to PTUPB and liver fibrosis were intersected and depicted in Venn diagram by ggVennDiagram package.

Protein–protein interaction (PPI) network of common target genes was generated from STRING database [https://string-db.org/ (15)]. The PPI network was visualized using Cytoscape (version 3.6.1) (16). To determine the hub genes, we computed the centrality of each mRNA node using “MCC” method in CytoHubba, a Cytoscape plugin (17). The top ten genes were deemed hub mRNAs based on their degree of centrality.

The clusterProfiler package was used to perform gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis on common target genes (18). The GO terms describe gene functions in three aspects: biological processes, molecular functions, and cellular components. The KEGG analysis predicts the involvement of the common target genes in various biological pathways. The modified p < 0.05 was used as a cutoff value (19, 20).

All the statistical analyses were performed using R (version 4.1.0.). Continuous variables were expressed as means ± standard deviation (SD). N represents the number of rats. Statistical significance was calculated by Student's t-test or Mann–Whitney U test. Two-sided p < 0.05 was considered statistically significant.

The hemodynamic and general characteristics were measured including weight, liver weight, HR, MAP, and PP. After the modeling of PHT, CCl4-VEH group presented with a higher PP (17.50 ± 4.65 vs. 5.40 ± 1.13 mmHg), lower weight (465.75 ± 11.15 vs. 562.25 ± 65.38 g), and lower liver weights (21.63 ± 0.93 vs. 26.93 ± 1.77 g) compared with OO-VEH significantly (p < 0.05), although MAP and HR remained comparable. Following the treatment of PTUPB, the body and liver weights of PHT rats rebounded and PP decreased significantly from 17.50 ± 4.65 to 6.37 ± 1.40 mmHg (p < 0.001), whereas map and HR remained unchanged. Similar to PTUPB, TPPU treatment also increased weights (609.50 ± 25.12 vs. 465.75 ± 11.15 g) and liver weights (29.33 ± 2.76 vs. 21.63 ± 0.93 g) compared with CCl₄-VEH group and decreased PP (6.86 ± 1.44 vs. 17.50 ± 4.65 mmHg) significantly (p < 0.05). The results are shown in Table 1.

After 16 weeks of CCl4 treatment, rats in the CCl4-VEH group exhibited marked liver fibrosis. Liver in PHT rats showed increased vacuolar degeneration and lysis in H&E staining, more collagen deposition in Masson (24% vs. 6%) and Sirius Red (23% vs. 15%) staining (Figures 2A,E,F). The hepatic expression of α-SMA, COL1A1, and serum hyaluronan also increased significantly (Figures 2B–D). PTUPB treatment resulted in significant reduced collagen deposition in Masson (9% vs. 24%) and Sirius Red (13% vs. 23%) staining which could also been confirmed by macroscopy and H&E (Figures 2A,E,F). In addition, the hepatic expression of α-SMA, COL1A1, and serum hyaluronan also decreased significantly after PTUPB treatment (Figure 2B–D,G). The treatment of TPPU was relatively less effective. Only the Masson staining of CCl4-TPPU group (13%) was significantly less than CCl4-VEH. No significant difference was observed in other indicators between CCl₄-TPPU with CCl₄-VEH group including the Sirius Red staining, serum hyaluronan, and hepatic expression of α-SMA and COL1A1 (Figures 2B–F).

Liver fibrosis is often accompanied by an increase hepatic and systemic inflammation. In CCl4-VEH group, hepatic expression of F4/80 mRNA and CD68 staining area increased significantly, indicating the upregulated hepatic mononuclear or macrophage infiltration (Figures 3C,D). The serum IL-6 also elevated, suggesting the systemic inflammation in PHT models (Figure 3E). The PTUPB treatment groups showed significant lower levels of F4/80 mRNA and CD68 staining in liver and also serum IL-6 (Figures 3A–E). Similar but less-effective results were observed in CCl₄-TPPU group.

In addition, we found that the liver structure was destroyed and its functions were decreased severely in PHT models. CK-19, a classical marker of bile duct cell proliferation, was upregulated in CCl4-VEH group compared with OO-VEH group (5% vs. 2%). Liver function markers, such as ALT, AST and DBIL, were also elevated significantly in CCl4-VEH rats (Figures 3A,B,F–H). In TPPU and PTUPB treatment groups, CK-19, ALT, and AST were observed to decrease significantly (Figures 3B,F,G) whereas TBIL, DBIL, and GGT did not change significantly (Figures 3H–J).

Pathological angiogenesis and sinusoidal remodeling are remarkable pathological symptoms of cirrhosis which contribute to vascular resistance and PHT. Several mediators of angiogenesis and remodeling were significantly increased in PHT rats including MMP2, VEGF, vWF, and Angpt1 (Figures 4B–E,J), which were confirmed by immunohistochemical staining (Figure 4A). After PTUPB treatment, MMP2, VEGF, vWF, Angpt1, and CD31 were significantly reduced, whereas MMP9 remained unchanged (Figures 4B–E,J,K). However, in CCl₄-TPPU group, only vWF, Angpt1, and CD31 were reduced significantly. In addition to the sinusoid remodeling, sinusoidal dysfunction also contributes to vascular resistance. Our results indicated that the expression of GCH1 decreased significantly in PHT group. After PTUPB treatment, the hepatic GCH1 and eNOS mRNA increased remarkably (Figures 4L,M).

In CCl₄-induced PHT rats, the lumen wall of SMA became thinner while its vascular pattern became disrupted (Figures 5A,B). However, the lumen diameter stayed unchanged in CCl₄-VEH group (Figure 5C) despite the wall thickness or lumen diameter ratio still reduced remarkably (Figure 5D). Following the treatment of PTUPB and TPPU, SMA restored to normal state with thicker wall thickness and higher wall thickness or lumen diameter ratio (Figures 5A,B,D).

In addition, the expression of remodeling factors including MMP2 and VEGF in SMA also increased significantly in PHT group (Figures 6B,D). After PTUPB treatment, vWF, VEGF, and CD68 decreased significantly (Figures 6C–E). As a contrast, only vWF reduced significantly in TPPU group (Figure 6C).

A total of 4,679 genes related to liver fibrosis were predicted through the GeneCards database. Based on the pharmacophore model and the principle of structural similarity, a total of 366 PTUPB-related target genes were collected through the SwissTargetPrediction and the PharmMapper database. After intersecting and merging cirrhosis-related genes and PTUPB predicted targets, 230 overlapping targets were obtained as candidate genes, which were shown in the Venn diagram (Figure 7A). To clarify the interactions among common targets between liver fibrosis and PTUPB, a PPI network was constructed by String database and CytoHubba plugin of Cytoscape. Using MCC method, the hub genes with the highest DC values were identified based on three topological parameters (degree, betweenness, and closeness centrality). The top 10 hub genes included are MAPK1, CASP3, SRC, ALB, IGF1, EGFR, HSP90AA1, PTGS2, ESR1, and ANXA5 (Figure 7B).

Subsequently, enrichment analysis was conducted on common target genes. In GO analysis, the common targets mainly involved in regulation of inflammatory response, phosphatidylinositol 3-kinase signaling, response to reactive oxygen species, regulation of vasoconstriction, regulation of blood pressure, vasodilation, negative regulation of inflammatory response, tissue remodeling, nitric oxide metabolic process, regulation of tissue remodeling, nitric oxide synthase biosynthetic process, tube formation, VEGF receptor signaling pathway, macrophage chemotaxis, macrophage activation, regulation of fibroblast proliferation, fibroblast proliferation, collagen metabolic process, collagen catabolic process, and epoxygenase P450 pathway (Figure 7C).

Kyoto Encyclopedia of Genes and Genomes results suggested that the enrichment pathways included VEGF signaling pathway, tumor necrosis factor (TNF) signaling pathway, non-alcoholic fatty liver disease (NAFLD), lipid and atherosclerosis, hypoxia inducible factor (HIF)-1 signaling pathway, hepatitis B, drug metabolism—CYP450, chemokine signaling pathway, apoptosis, and ARA metabolism (Figure 7D).

As a dual COX-2/sEH inhibitor, we evaluated the inhibitory effect of PTUPB on COX-2/sEH pathways and proinflammatory cytokine TGF-β. In PHT rats, the hepatic expression of sEH, COX-2, and TGF-β increased significantly (Figures 8B–D). Besides, sEH staining in SMA also increased in PHT rats (A). Comparable results were found in western blot analysis (Figure 8F). After treatment with PTUPB, the expression of sEH, COX-2, and TGF-β in the liver decreased significantly (Figures 8B,C). Similar results were also seen in immunohistochemical analysis and western blot (Figures 8A,E,F). However, in the TPPU treatment group, only sEH and TGF-β decreased significantly, whereas COX-2 remained unchanged (Figures 8B–D,F).