Gandouling Inhibits the Sinusoid Capillarization Associated with Liver Fibrosis in Wilson’s Disease by Blocking the Communication Between Hepatic Stellate Cells and Liver Sinusoidal Endothelial Cells

TL;DR

Gandouling (GDL) reduces liver fibrosis in Wilson’s disease by blocking communication between liver cells, potentially offering a new treatment approach.

Contribution

The study reveals a novel mechanism by which GDL inhibits sinusoid capillarization in liver fibrosis through PDGFRβ/ERK/VEGFA signaling.

Findings

GDL alleviated liver fibrosis and inhibited sinusoid capillarization in Wilson’s disease mouse models.

GDL blocked VEGFA-mediated communication between hepatic stellate cells and liver sinusoidal endothelial cells.

GDL reduced PDGFRβ/ERK signaling and VEGFA expression, improving Wilson’s disease outcomes.

Abstract

Background: Gandouling (GDL) is a compound prepared in Chinese medicine and demonstrates favorable clinical efficacy. Studies have shown that sinusoid capillarization promoted hepatic fibrosis and was a potential target for preventing and treating liver fibrosis in Wilson’s disease (WD). This study aimed to explore whether GDL inhibited the sinusoid capillarization in WD by blocking the communication between hepatic stellate cells (HSCs) and liver sinusoidal endothelial cells (LSECs). Methods: In this study, Atp7b-H1071Q (TX) mice were used as the WD model mice, and CuSO4⋅5H2O treated LX-2 cells were used as the HSC activation model. We used scanning electron microscopy, vascular tube formation assay, Western blot, cell transfection, and co-culture system to study how GDL blocked the communication between HSCs and LSECs, as well as its inhibitory effect on the sinusoid capillarization. Results: We found that GDL alleviated liver fibrosis in TX mice, inhibited HSC activation, and sinusoid capillarization in TX mice. Excessive secreted VEGFA by LX-2 cells promoted the sinusoid capillarization, played the role of a messenger molecule, and GDL blocked the VEGFA-mediated HSCs-LSECs communication. Furthermore, bioinformatics analysis, molecular docking, and molecular dynamics suggested that GDL may exert its effect by modulating the PDGFRβ/ERK/VEGFA signaling axis. We validated the above observation through experiments, that GDL reduced PDGFRβ/ERK signal pathway in LX-2 cells, inhibited the expression of messenger molecule VEGFA, blocked HSCs-LSECs communication, inhibited sinusoid capillarization, and improved WD. Conclusions: GDL blocked the communication between HSCs and LSECs and inhibited the sinusoid capillarization associated with liver fibrosis in WD by the PDGFRβ/ERK/VEGFA signaling axis.

Figures

Full text

1. Introduction

Wilson’s disease (WD) is caused by the ATP7B gene mutation, which leads to Cu^2+^ metabolism disorder, and is an autosomal recessive genetic disorder [1]. In WD, the Cu^2+^ secretion pathway is restricted to varying degrees, resulting in copper accumulation that can cause body pathologies [2]. Early symptoms of WD include liver lesions, such as hepatic steatosis, cholestasis, and liver fibrosis [3]. The incidence rate of WD ranges from 1/10,000 to 1/30,000 worldwide, with a prevalence of 5.87/100,000 in China [4].

Liver fibrosis occurs in the early stage of liver involvement in patients with WD; early prevention and treatment of liver fibrosis are crucial for the treatment of WD [5]. Previous studies have shown that hepatic stellate cell (HSC) activation and sinusoid capillarization are the factors that trigger the pathological process of liver fibrosis [6]. Upon activation of HSCs, they stimulate the synthesis of numerous extracellular matrix components, including collagen I, III, and IV, as well as α-SMA, causing liver fibrosis [7,8]. Research demonstrates that compared with portal angiogenesis, sinusoid capillarization promotes liver fibrosis and is a probable treatment for improving liver fibrosis [9]. Therefore, we may be able to improve WD liver fibrosis by inhibiting the sinusoidal capillarization.

Liver sinusoidal endothelial cells (LSECs) have a unique fenestrated structure on their surface, which promotes the bidirectional exchange of cells and blood in the hepatic sinusoids and prevents HSC activation. However, in the pathological state, LSECs are subjected to external stimuli to undergo morphological transformation from a basement membrane-free state to the organized basement membrane, and fenestrations are decreased. This process is called sinusoid capillarization (also known as LSEC capillarization) [10]. After sinusoid capillarization, LSECs promote pathological angiogenesis and secrete vascular cell adhesion molecule 1 (VCAM1) to promote liver fibrosis [11,12]. Activated HSCs promote the sinusoid capillarization, and LSECs secrete fibronectin, TGF-β1, and endothelin after capillarization to accelerate the HSC activation, thereby exacerbating liver fibrosis [13]. However, the mechanism of interaction between HSCs and LSECs has not been fully elucidated.

Platelet-derived growth factor receptor β (PDGFRβ) is a target to regulate liver fibrosis and is related to HSC activation. PDGFRβ is the target for Roseotoxin B, salvianolic acid B, and Gomisin D in improving liver fibrosis [14,15,16]. PDGFRβ activates the extracellular signal-regulated kina (ERK) pathway to activate HSCs and participates in the progression of liver fibrosis. In addition, the PDGFRβ signaling pathway is involved in blood flow regulation and angiogenesis; it is a molecular target for regulating pathologic vasculature [17,18,19]. PDGFRβ overexpression promotes angiogenesis in osteoporosis and vascular endothelial cells (VECs) [20,21]. In nephroblastoma, PDGFRβ affects the process of angiogenesis by regulating the secretion of vascular endothelial growth factor A (VEGFA) [22]. The use of PDGFRβ and ERK inhibitors can inhibit the expression of VEGFA [23,24,25].

GDL is a commonly used traditional Chinese medicine for the clinical treatment of WD and demonstrates excellent clinical efficacy. It has been confirmed by Anhui Provincial Food and Drug Administration (approval number: Z20050071). Evidence suggests that GDL exerts anti-WD effects by anti-inflammatory, regulating the mitochondria autophagy, and ferroptosis [26,27,28]. GDL is composed of Salvia miltiorrhiza Bge., Spatholobus suberectus Dunn, Rheum palmatum L., Curcuma aeruginosa Roxb, Coptis chinensis Franch., and Curcuma longa L. Curcumin, a major component of Curcuma longa L., inhibits liver fibrosis by inducing the HSC senescence and anti-inflammatory activity [29]. The components of Rheum palmatum L., emodin and rhein, exhibit antioxidant and hepatoprotective activities [30]. Phytoconstituents of Salvia miltiorrhiza Bunge exhibit significant anti-liver fibrosis effects [31]. Spatholobus suberectus Dunn and Coptis chinensis Franch have anti-inflammatory and antioxidant properties [32,33]. The research group previously found that GDL alleviated WD fibrosis by inhibiting the activation of HSCs, but the specific mechanism was not clear.

Our preliminary findings showed that GDL inhibited the sinusoid capillarization, HSC activation, and liver fibrosis in TX mice. Therefore, we used co-culture, immunofluorescence, immunohistochemical staining, and vascular tube formation assay to investigate the relationship between HSC activation and sinusoid capillarization, as well as how GDL regulated the communication between HSCs and LSECs to inhibit liver fibrosis. Our research revealed a new molecular mechanism of GDL against WD liver fibrosis, provided new scientific evidence for its application in WD treatment, and promoted the clinical application of GDL.

2. Results

2.1. GDL Alleviated Liver Fibrosis in TX Mice

Through HE staining, it was observed that the hepatocytes of normal individuals had regular shapes, the hepatic cords were arranged neatly, and the liver lobule morphology and structure were normal. In TX mice, there was extensive edema, partial necrosis, hepatic cord disorganization, and narrowing of hepatic sinusoidal structures in the central venous confluence area. After GDL treatment, the liver structural damage in TX mice was markedly improved, hepatocyte edema and necrosis decreased, especially in the GDL-H group (Figure 1A). We found that the TX mice exhibited significant fibroplasia and collagen deposition around the central venous confluence area by Masson staining. Whereas, after GDL treatment, collagen deposition and fibroplasia were reduced (Figure 1B). The levels of ALT, AST in the serum and liver copper of TX mice were significantly increased, and GDL inhibited their levels, and GDL restored the serum copper level in TX mice (Figure 1C–F). These data demonstrated that GDL alleviated liver fibrosis in TX mice.

2.2. GDL Inhibited the Activation and Migration of HSCs

During the activation process of HSCs, the levels of collagen I and α-SMA increased. We used immunohistochemistry to detect their expression, and the results suggested that their levels in TX mice were markedly higher; different doses of GDL inhibited their expression to varying degrees (Figure 2A–B1). We treated LX-2 cells with 100 nM Cu^2+^ to establish the HSC activation model. The results demonstrated that the levels of α-SMA and collagen I increased in the model group, indicating the successful establishment of the HSC activation model, whereas serum containing GDL downregulated their expression (Figure 2C–E1). Further research revealed that after treating LX-2 cells with Cu^2+^, their migration markedly increased, and serum containing GDL inhibited their migration (Figure 2F,F1). We confirmed that GDL inhibited the activation of HSCs through in vitro and in vivo experiments.

2.3. GDL Inhibited the Sinusoid Capillarization in TX Mice

The presence of fenestration is the gold standard for confirming the identity of LSECs, and the decrease in fenestration represents capillarization. Scanning electron microscopy showed that a large number of fenestrations were visible in LSECs in the normal group, which were relatively uniform in size and aggregated to form sieve plates. The fenestration quantity and fenestration area on the surface of LSECs were significantly reduced in TX mice. However, the GDL treatment restored the number of fenestrations (Figure 3A–A2). Immunofluorescence showed that the levels of capillarization markers CD31, CD34, and vWF proteins in TX mice liver tissue were significantly upregulated, and GDL downregulated their expression. The GDL high-dose group showed the most obvious effect (Figure 3B–D1).

2.4. Excessive Secreted VEGFA by LX-2 Cells Promoted the Sinusoid Capillarization

We used ELISA to detect LX-2 cells’ supernatants and found that VEGFA secretion increased in the model group (Figure 4A). We used the VEGFR2 interference to knockdown the VEGFR2 in LSECs, then different groups of LX-2 cells were co-cultured with LSECs (Figure 4B). The vascular tube formation assay revealed an increase in the number of M-LX-2→LSEC group constituent tubes and a decrease in the number of M-LX-2→LSEC+si-VEGFR2 constituent tubes (Figure 4C,C1). Subsequently, we detected the expression of capillarization markers CD31, CD34, and vWF in LSECs. Immunofluorescence revealed that their levels were upregulated in the M-LX-2→LSEC group. However, knockdown of VEGFR2 significantly reduced their expression (Figure 4D–F1). Our results demonstrated that after HSCs were activated, they secreted a lot of VEGFA, which bound to the VEGFR2 on the surface of LSECs, thereby promoting the sinusoid capillarization.

2.5. GDL Inhibited the Sinusoid Capillarization by Blocking the HSCs-LSECs Communication

Previously, we found that GDL inhibited the expression of VEGFA in LX-2 cells. LX-2 cells in the normal group/model group/GDL group/NC group were co-cultured with LSECs, respectively. The vascular tube formation assay revealed that the constituent tube number of the GDL-LX-2→LSEC group was significantly reduced compared with the M-LX-2→LSEC group (Figure 5A,A1). Furthermore, immunofluorescence detection of LSECs suggested that levels of CD31, CD34, and vWF were downregulated in the GDL-LX-2→LSEC group (Figure 5B–D1). GDL inhibited the expression of messenger molecule VEGFA, thereby inhibiting the sinusoid capillarization.

2.6. Bioinformatics Analysis of GDL Treatment Mechanisms

543 chemical components and 749 corresponding targets in GDL were collected from the TCMSP website and visualized using Cytoscape 3.10.3 (Figure S1A,B). Obtained 5810 disease genes associated with WD liver fibrosis by OMIM and GeneCards website, and 466 intersecting targets of GDL and WD liver fibrosis were obtained using the Venny website (Figure S1C). We introduced overlapping genes in a string to obtain PPI networks (Figure S1D). The top twenty targets were derived using the MCC algorithm of Cytoscape 3.10.3 (Figure S1E). GO and KEGG analyses showed a high enrichment of multiple signaling pathways, including the Ras, MAPK, VEGFA, and HIF-1 signaling pathways (Figure S1F,G). Through preliminary experiments and literature review, we discovered that PDGFRβ was a key target for activating the ERK signaling pathway. Therefore, we predicted that the PDGFRβ/ERK/VEGFA signal axis may be a crucial pathway for GDL to exert its function.

Based on bioinformatics analysis and previous research findings, we analyzed the correlation between GDL and PDGFRβ protein by molecular docking and molecular dynamics. Molecular docking revealed that eight signature monomer molecules of GDL exhibited good binding ability to PDGFRβ protein (Supplementary Table S3). Visual results of eight signature monomer molecules in GDL binding to PDGFRβ protein were obtained by pymol (Figure 6A–H). We took the four monomers with higher binding energies for molecular dynamics simulations. Results revealed that Coptisine was most stable within the binding site; Coptisine and Tanshinone IIA exhibited the optimal binding stability. Catechin has the highest free binding energy and the highest stability (Supplementary Table S4). All four compounds significantly stabilized key residues of PDGFRβ (Figure 6I–N). Therefore, GDL may exert its effect by targeting the PDGFRβ. This prediction needed to be verified through in vivo and in vitro experiments.

2.7. GDL Blocked the PDGFRβ/ERK/VEGFA Signal Axis in LX-2 Cells

We detected the expression of members of the PDGFRβ/ERK/VEGFA signal axis. In the model group, the levels of p-PDGFRβ, p-ERK, and VEGFA were increased, and serum containing GDL significantly downregulated their expression (Figure 7A–D). To verify the regulation of PDGFRβ on the ERK and VEGFA, we performed the PDGFRβ overexpression in LX-2 cells (Figure 7E). After the PDGFRβ overexpression, the mRNA levels of α-SMA, collagen I, and VEGFA were increased (Figure 7F–H), and the protein expression of p-ERK, α-SMA, and VEGFA showed the same trend (Figure 7I–K1). Our results demonstrated that GDL exerted its effect by blocking the PDGFRβ/ERK/VEGFA signal axis in LX-2 cells.

2.8. GDL Inhibited the PDGFRβ/ERK/VEGFA Signal Axis in TX Mice

To further study the regulation of the PDGFRβ/ERK/VEGFA signaling axis by GDL, we detected its expression in vivo experiments. Immunohistochemistry revealed significantly elevated expression of p-PDGFRβ, p-ERK, and VEGFA in the livers of TX mice compared to normal mice. Following GDL treatment, p-PDGFRβ, p-ERK, and VEGFA levels were reduced, with the highest dose GDL group exhibiting the most pronounced effect (Figure 8A–C1). In addition, WB analysis demonstrated that GDL decreased the levels of p-PDGFRβ, p-ERK, and VEGFA (Figure 8D–F1).

3. Discussion

WD is characterized by abnormal accumulation of Cu^2+^ [34]. The pathogenic gene ATP7B is primarily expressed in the liver. Therefore, patients with WD are prone to liver damage in the early stages, and various neurological or psychiatric symptoms occur with age. WD is treated using the Cu^2+^ repellent Drugs, traditional Chinese medicine, and surgery [35]. Current treatments improve the WD symptoms. However, WD is hard to treat completely. Increased hepatic angiogenesis occurs during liver fibrosis, including portal vein, hepatic sinusoids, and central vessels, and targeted vascular therapy can improve liver fibrosis [36]. Carthamiflos prevents liver fibrosis progression by repairing the fenestrations of LSECs and modulating pathologic angiogenesis [37]. Therefore, the regulation of sinusoid capillarization may represent a potential mechanism for improving liver fibrosis.

PCA is the medicine of choice for the therapy of WD. However, it is associated with numerous adverse effects [38]. GDL exhibits significant regulation of mitochondrial autophagy, ferroptosis, and anti-inflammatory [39]. In this study, Atp7b-H1071Q mice were used as the WD model mice, and they were treated with GDL for 6 weeks. We observed that GDL inhibited the sinusoid capillarization and liver fibrosis in TX mice by HE, Masson staining, scanning electron microscopy, and IF assay. These results suggested GDL can improve hepatic microcirculation, restore the fenestrated structure of LSECs and the function of hepatic sinusoids, and promote the nutrient exchange of hepatocytes. If GDL is used in combination with a chelating agent, it may further repair the structure of hepatic sinusoids on the basis of copper excretion, achieving multi-link intervention in liver injury of WD. Our previous work showed that GDL improved liver fibrosis in WD by inhibiting the activation of HSCs. However, the mechanism by which GDL regulated the interaction between HSCs and LSECs was not clear.

VEGFA and its two receptors, VEGFR1 and VEGFR2, mediate the key signaling pathways in angiogenesis [40]. VEGFA and VEGFR2 binding promote angiogenesis [41]. VEGFA is a key regulator and has a “dual-way” function in liver fibrosis. Normally, HSCs and hepatocytes express VEGFA to maintain the fenestration structure of LSECs. However, in the pathological state, VEGFA overexpression induces angiogenesis, resulting in sinusoid capillarization and reducing fenestration [42,43]. Neuropilin-1 (NRP-1) promotes the sinusoid capillarization through the VEGFR2-dependent PI3K/Akt pathway [44]. We used the co-culture system of LX-2 cells with LSECs to investigate the regulatory mechanism of HSC activation on the sinusoid capillarization. We found that activated HSCs excessively secreted VEGFA to promote the sinusoid capillarization through vascular tube formation and Immunofluorescence staining. GDL blocked the VEGFA-mediated HSCs-LSECs communication and inhibited the sinusoid capillarization. The limitation was that the vascular tube formation assay in vitro cannot fully replace the functions of hepatic sinusoids.

It is worth noting that the abnormal secretion of VEGFA in HSCs is related to the PDGFRβ signaling pathway. PDGFRβ is expressed on HSC membranes, and its phosphorylation activates the downstream MAPK pathway [45]. ERK is a member of the MAPK family and plays a crucial role in HSC activation during liver fibrosis [46]. Blocking PDGFRβ in HSCs may attenuate the progression of liver fibrosis. Evidence suggests that knockdown of ERK in HSCs inhibits HSC proliferation and prevents the PDGFRβ-induced proliferative effect on HSCs [47]. Moreover, the PDGF signaling pathway is crucial for angiogenesis and promotes the expression of VEGFA in HSCs. The Binding of ligands to PDGFRβ exacerbates sinusoid capillarization [13]. Based on bioinformatics analysis, we speculated that GDL inhibited the VEGFA secretion by suppressing the PDGFRβ/ERK pathway. Molecular docking and dynamics indicated that PDGFRβ was a target of GDL action. In vivo and in vitro experiments demonstrated that GDL inhibited the PDGFRβ/ERK/VEGFA signaling axis, thereby blocking the communication between HSCs and LSECs.

This study focused on the unidirectional regulation of sinusoid capillarization by HSC activation and clarified the mechanism by which activated HSCs oversecreted VEGFA to drive abnormal LSEC phenotypes. However, it failed to further explore the reverse effect of sinusoid capillarization on HSC activation. In addition, GDL was a multi-herbal formulation; this study did not rule out the possibility that GDL regulated other fibrosis-related pathways. In the future, we will further investigate the interaction between HSCs and LSECs, search for other pathways through which GDL regulates liver fibrosis, increase the sample size, and collect clinical samples to enhance the persuasiveness of our research.

4. Materials and Methods

4.1. Cell Culture

The human HSC line LX-2 (CL-0560) was purchased from Pricella Company (Wuhan, China), cultured in DMEM (CM-0560) containing 10% fetal bovine serum (FBS), 1% penicillin, and streptomycin. The human LSEC line (iCell-0019a) and complete primary endothelial cell medium (iCell-0019a-001b) were provided by iCell Bioscience (Shanghai, China). The cells were incubated at 37 °C with 5% CO_2_. LX-2 cells were treated with 100 nM CuSO_4_⋅5H_2_O (Sigma Aldrich, Mo, USA) to establish the HSC activation model [28].

4.2. Model Animals and Grouping

We used Atp7b-H1071Q (TX) mice as the WD model; they were housed at the Animal Center of Anhui University of Chinese Medicine, and divided into model, low dose, medium dose, high dose GDL, and positive drug control (penicillamine, PCA) groups. The low dose was 0.29 g/kg/d, the medium dose was 0.58 g/kg/d, the high dose was 1.16 g/kg/d, and the PCA group received 0.09 g/kg/d. In addition, C57 mice were selected as the normal group and were gavaged with normal saline [48]. After six weeks of GDL treatment, the serum and tissues were collected for subsequent experiments. All animal experimental protocols were approved by the Animal Ethics Committee of Anhui University of Chinese Medicine (Animal number: AHUCM-mouse-2022111).

4.3. Preparation of GDL-Containing Serum

Sprague-Dawley rats were randomly divided into normal and GDL groups after one week of acclimatization. The GDL group was gavaged at a dose of 0.4 g/kg/d, and the normal group was administered an equal amount of saline twice daily for 7 days. After the last administration, the rats were anesthetized using 20% urethane, the rats’ blood samples were collected via the abdominal aorta, then centrifuged and inactivated in a 56 °C water bath. Serum containing GDL was stored at −80 °C [48].

4.4. Cell Transfection

PDGFRβ overexpression plasmid and VEGFR2 knockdown plasmid were provided by Sangon Biotech (Shanghai, China). We cultivated LX-2 cells and LSECs in six-well plates at 37 °C and 5% CO_2_ until they covered 90% of the bottom of six well plate. For LX-2 cells, dissolved 4 μg of PDGFRβ plasmids in 250 μL opti medium, mixed well, and waited for 5 min. Dissolved 10 μL LipoHigh in 250 opti medium, mixed well, and waited for 5 min. Mixed the plasmid medium and LipoHigh medium well, added to LX-2 cells and gently mixed, cultured them for 4–6 h at 37 °C and 5% CO_2_. Replaced with non-double antibody medium, continued to culture for 24–48 h. For LSECs, dissolved 100 pmol of VEGFR2 plasmids in 250 μL opti medium to make plasmid medium. Dissolved 10 μL LipoHigh in 250 opti medium to make LipoHigh medium. The VEGFR2 knockdown plasmid sequences are shown in Supplementary Table S1.

4.5. Co-Culture of HSC Cell Line LX-2 Cells with LSECs

We used the transwell to establish the co-culture system. The LSECs were cultivated in the lower chamber of the transwell, and the LX-2 cells in each group were cultivated in the upper chamber of the transwell for 24 h at 37 °C and 5% CO_2_. The experiments were divided into 6 groups as follows: (i) The normal LX-2 cells (N-LX-2)→LSEC group (N-LX-2 cells in the upper chamber, and LSECs in the lower chamber); (ii) The model LX-2 cells (M-LX-2)→LSEC group; (iii) The GDL-LX-2→LSEC group; (iv) The NC-LX-2→LSEC group; (v) The M-LX-2→LSEC+si-VEGFR2 group; (vi) The M-LX-2→LSEC+si-NC group.

4.6. Bioinformatics Analysis

We used the TCMSP (https://www.tcmsp-e.com/load_intro.php?id=43 accessed on 30 November 2025) database to search bioactive components and corresponding targets of six GDL herbs, processed bioactive components and targets of GDL with Cytoscape 3.10.3. WD liver fibrosis targets were searched by GeneCards (https://www.genecards.org/), used Venny website (https://bioinfogp.cnb.csic.es/tools/venny/ accessed on 30 November 2025) was used to obtain intersecting targets of diseases and drugs. Then used the STRING website (https://cn.string-db.org/) to perform PPI analysis on the overlapping targets. Finally, GO and KEGG analyses were performed to obtain the potential targets and pathways of GDL [28].

4.7. Molecular Dynamics and Molecular Docking

Download the PDB formats of PDGFRβ protein and GDL monomer components (including Berberine, Coptisine, Aloe-eModin, Catechin, Chrysophanic acid, Emodin-3-methyl ether, Emodin, and Tanshinone IIA) from the RCSB (https://www.rcsb.org/) and TCMSP websites, respectively. The binding energy was obtained by docking the PDGFRβ protein and GDL monomer components using the AutoDock Vina v1.2.3. The binding sites between the GDL monomer components and the PDGFRβ protein were analyzed using PyMol v2.6. Then, we completed molecular dynamics (MD) of the PDGFRβ protein with berberine, coptisine, catechin, and tanshinone IIA using the Gromacs v2022.03. On the basis of the MD results, we analyzed the SASA, RMSF, RMSD, Rg, and H-bonds of these complexes. Finally, we computed the small molecule-protein complexes’ free binding energy [49].

4.8. Transwell Migration of LX-2 Cells

Added 500 μL of medium containing 10% FBS to the lower chamber and 150 μL of medium containing 5% FBS to the upper chamber of the transwell chamber. Then, added 4 × 10^4^ LX-2 cells to the upper chamber and cultured for 24 h. After fixation, stained with crystal violet. Then, took out the chamber and gently washed the upper surface of the chamber with phosphate-buffered saline (PBS), and photographed using an inverted microscope (AE2000, Motic, Xiamen, China).

4.9. Immunofluorescence

The LSECs were fixed with 4% paraformaldehyde at room temperature for 30 min. Washed twice with PBS, then permeabilized with 0.5% Triton X-100 (Solarbio, Beijing, China) for 15 min. Sealed cells with 5% bovine serum albumin (BSA), treated with the primary antibody and secondary antibody. Treated with DAPI solution for 10 min and photographed images using the Leica microscope (Leica DMi8, Wetzlar, Germany), and the antibodies used in this study included the CD31 (Proteintech, 28083-1, 1:4000, Wuhan, China), CD34 (Proteintech, 14486-1, 1:200, Wuhan, China), and vWF (Bioss, bs-0586R, 1:200, Bejing, China).

4.10. Immunohistochemistry

The slices were baked in a drying oven at 65 °C for 20 min, deparaffinized by soaking three times in xylene, and then treated with an anhydrous ethanol gradient. The washed slices were treated with a citrate repair solution, incubated at room temperature in 3% H_2_O_2_, and rinsed thrice with PBS. The slices were treated with the primary antibodies and secondary antibodies. Then added DAB color developer and hematoxylin in slices. The images were gained by a slide scanner [50]. The antibodies used in the study included the α-SMA (Abcam, ab124964, Cambridge, UK), Collagen I (Abcam, ab316222), p-PDGFRβ (Biodragon, BD-PP1123, Suzhou, China), p-ERK1/2 (Abcam, ab76299), and VEGFA (immunoway, YN5444, Plano, TX, USA).

4.11. Scanning Electron Microscopy

TX mice were anesthetized using 20% urethane; dissected the mice and performed perfusion fixation of the liver. Tied the portal vein with a suture and inject 2.5% glutaraldehyde through the inferior vena cava for perfusion until the liver hardened and turned white. Cut the liver into 3 mm^2^ slices and fixed them, transferred tissue blocks into 1% OsO_4_ in 0.1 M PB (pH 7.4) for 1–2 h at room temperature. Then washed tissue blocks in 0.1 M PB (pH 7.4) for 3 times, 15 min each. Dehydrated with ethanol of increasing concentrations, dried samples with a Critical Point Dryer. Used a microscope (SU8100, Hitachi, Tokyo, Japan) to locate the field of view containing LSEC fenestration, then observed and captured images. Calculations: Fenstration frequency (number/μm^2^) = Fenestration quantity/(Total area analyzed—Area of gaps), Porosity (%) = Fenestration area/(Total area analyzed—Area of gaps) [51].

4.12. Vascular Tube Formation Assay

LSECs were co-cultured with the LX-2 cells until cell growth reached 90%. Laid Matrigel (Absin, abs9414, Shanghai, China) on a 96-well plate at 50 μL/well. The LSECs were added in Matrigel at 1 × 10^4^ cells/well and cultured for 3–12 h at 37 °C and 5% CO_2_. Images were gained by an inverted microscope (AE2000, Motic, Xiamen, China) and analyzed using the ImageJ (1.54k version) software to quantify tube formation.

4.13. RT-qPCR

We used TRIzol (Invitrogen, CA, USA) to extract total RNA from cells and tissue. After quantifying the total RNA, the cDNAs were synthesized using the Reverse Transcription Kit (BL696A, Biosharp, Beijing, China), and they were stored in −80 °C refrigerator. Used the PCR LightCycler^®^ 96 instrument (Roche, Basel, Switzerland) and SYBR kit (BL697A, Biosharp, China) to perform PCR amplification [52]. The gene sequences were in Supplementary Table S2.

4.14. Western Blot (WB)

Treated cells and tissues with PIPA lysis buffer and PMSF to obtain proteins. After all proteins were extracted, they were quantified and then subjected to electrophoresis and transferred to PVDF membranes (IPVH00010, Immobilon, Mo, USA). Washed and sealed using 5% skimmed milk solution. After washing PVDF membranes thrice with TBST, incubated them with primary antibodies, including β-actin (Bioworld, BS6007M, 1:10,000, Beijing, China), PDGFRβ (ab313777, 1:1000), p-PDGFRβ (ab218534, 1:1000), ERK1/2 (ab184699, 1:10,000), p-ERK1/2 (ab76299, 1:10,000), α-SMA (ab124964, 1:10,000), and VEGFA (ab214424, 1:1000) at 4 °C for 12 h, purchased from Abcam. Then treated with secondary antibody (ZSGB-BIO, ZB-2301, 1:10,000, Beijing, China) for 2 h. Detected membranes using the BeyoECL Plus kit in Tanon 5200, and analyzed the gray value analysis by the ImageJ.

4.15. Statistical Analysis

The data were expressed as mean ± standard deviation (SD). For statistical analysis, the t-test was used for comparing two groups, while one-way ANOVA was employed for multiple groups. In this study, the sample size n = 3 for all experiments represented three independent biological replicates. p ≤ 0.05 indicated a statistically significant difference in the mean between the two groups.

5. Conclusions

In this study, we identified that VEGFA was the messenger molecule regulating the relationship between HSC activation and sinusoid capillarization. GDL modulated the PDGFRβ/ERK/VEGFA signaling axis, suppressed the VEGFA expression, blocked the HSCs-LSECs communication, and inhibited the sinusoid capillarization. This research provided novel insights for the prevention and treatment of WD liver fibrosis and offered theoretical support for the clinical application of GDL.