EGFR inhibitor: Gefitinib (No. #S1025); PI3K inhibitor: Buparlisib (No. #S2247); Akt inhibitor: Perifosine (No. #S1037); braf inhibitor: Sorafenib (BAY 43-9006) tosylate (No. #S1040); ERK inhibitor: Temuterkib (No. #S8534). All above inhibitors were bought from Selleck (USA). RAS inhibitor, RAS GTPase inhibitor 1 (No. # T12692), was purchased from Topscience (China).

Human normal hepatocyte LO2 cells were cultured in DMEM medium (Gibco, Carlsbad, USA) supplemented with 10% fetal bovine serum (FBS; Gibco) with 95% O₂ and 5% CO₂ at 37°C. The cell lines were gifted from the Center of Hepato-Pancreato-Biliary Surgery, the First Affiliated Hospital of Sun Yat-sen University.

pEGFP-C1-CsGRN recombinant plasmid and pEGFP-C1 vector (Promega) were transfected into human liver cell lines LO2 (CsGRN-LO2), respectively, as our previously described (Wang et al., 2017). The cells stably expressing EGFP-tagged CsGRN were selectively established by 400 μg/ml G418 and identified by fluorescence microscope and qRT-PCR analysis.

The methods of gene cloning and purification of recombinant CsGRN were used as our previous study (Wang et al., 2017). Briefly, the ORF of CsGRN (GenBank KY855531) has a genome length of 714 bp and was cloned into the pMAL-c2x vector (preserved in our lab). Following digestion with restriction enzymes, the plasmid DNA was attached to the pET-28a (+) expression vector (Novagen, Darmstadt, Germany), before being introduced into the E. coli BL21 (DE3) (Promega). Then, we induced selected clones at 20°C for 12–18 h with 1 mM IPTG (Sigma, Guangzhou, China) following growth. By using the Amylose Resin-Bind Purification Kit (BioLabs), we purified recombinant protein (MBP-GRN) from the supernatant.

Total RNA was obtained by using Trizol reagent (Invitrogen, Carlsbad, CA, USA), and the total cDNA was generated using ABM’s 5× All-In-One RT Master Mix (Transgen, Beijing, China). SYBR Premix Ex Taq (Takara, Dalian, China) was used for quantitative analysis of the expression of mRNA level of indicated gene. PCR quantification was performed using the CFX96 Real-Time PCR system (Bio-Rad). GAPDH was performed as the reference gene. Following are the formulae used to calculate relative mRNA expression: ΔCt = Ct (sample) − Ct (GAPDH), ΔΔCt (sample) = ΔCt (sample) −ΔCt (calibrator). The relative quantification of mRNA was used to calculate the fold change. cDNA was amplified using specific primers ( Supplementary Table ).

Cell proliferation was evaluated by Cell Counting Kit 8(CCK8) assay as well as Colony Formation assay, respectively. Briefly, cells were seeded onto 96-well plates with 8×10³/100 µl, after 24 h, replaced with serum-free medium 100 µl added with 10 µl CCK8 in each well, and then placed in the incubator for 30 min to 3 h, and the absorbance value of 450 nm wavelength was measured by microplate reader. The long-term effects of CsGRN on LO2 cell proliferation were analyzed with a colony formation assay. Cells were seeded into six-well plates and cultured for 2–3 weeks. Medium was renovated every 3 days. After washing with phosphate buffer saline (PBS) for three times, the cells were then fixed with 4% paraformaldehyde, stained with crystal violet solution, and washed with ddH₂O to remove the excess dye. The colony number was analyzed with Image J software.

Cell cycle was evaluated by flow cytofluorometry. Briefly, cells were seeded into six-well plates at a density of 10⁶ cells per well in completed medium. After 24 h, the cells were stained with Propidium iodide [MultiSciences (Lianke) Biotech Co, Ltd.] and analyzed by flow cytometry (Coulter Beckman Gallios) and analyzed with Flowjo software.

The invasion ability of CsGRN-LO2 cells was evaluated though Transwell invasion chambers (Costar, NY, USA). Transfected CsGRN-LO2 cells cultured in 24-well plates (1×10⁶cells/ml) with non-serum medium and Matrigel (BD Biosciences, Heidelberg, Germany) mixed on the upper chambers (1:8), while medium containing 10% FBS was added into the lower chambers for 24 h at 37°C. The invading cells were then stained with 1% crystal violet (Beyotime Biotechnology, Guangzhou, China) for 15 min, followed by imaging with light microscope (Leica DMI3000B, Wetzlar, Germany) and counting the numbers of migrated cells in five random fields. The cell motility points were measured by using Image J software.

The effects of CsGRN on LO2 cells migration were analyzed with a wound-healing assay. Transfected LO2 cells were moved into six-well plates and hatched for 24 h. When cells were grown to 80%, the bottom of plate was scratched with a 200 μl pipette tip, followed by observing and imaging with light microscope (Leica DMI3000B, Wetzlar, Germany) every 24 h for 72 h.

Cells were dissolved by RIPA buffer (Beyotime, Shanghai, China) and quantified by BCA Protein Assay (Trans, Beijing, China). Thirty μg total proteins were separated by SDS-PAGE, followed by transferring onto PVDF membranes (Whatman, Maidstone, UK) and then sealed with 5% skimmed milk in TBST for 2 h at room temperature. After incubation in primary and secondary antibodies, the blots were detected by ECL (Millipore, Billerica, USA). The following antibodies were used: anti-MBP tag monoclonal antibody (1:2,000, Novagen), rat anti-CsGRN serum (1:200), anti-E-cadherin (1:2,000), anti-vimentin (1:2,000), anti-N-cadherin (1:2,000), anti-ZO-1 (1:2,000), anti-β-catenin (1:2,000), anti-p18 (1:2,000), anti-p21 (1:2,000), anti-p27 (1:2,000), anti-CDK2 (1:2,000), anti-CDK4 (1:2,000), anti-CDK6 (1:2,000), anti-CyclinD1 (1:2,000), anti-CyclinD3 (1:2,000), p-ERK (1:2,000), ERK (1:2,000), p-AKT (1:2,000), p-EGFR (1:2,000), RAS (1:2,000), p-braf (1:2,000), and anti-GAPDH (1:2,000). All antibodies were products from Cell Signaling Technology (CST, Boston, USA). Secondary antibodies were purchased form for Proteintech Group: goat anti-mouse IgG-HRP (1:5,000) and goat anti-rabbit IgG-HRP (1:5,000). For quantification of western blot images, ImageJ (https://imagej.nih.gov/ij/index.html) was used.

Statistical analysis was performed using SPSS26.0 statistical software. Student’s t-test and ANOVA were used to analyze statistical differences. Data were shown as means ± S.D., and significance was described as P < 0.05.

To determine the potential role of CsGRN overexpression on inducing the malignant transformation of human hepatocytes, an LO2 cell line with stable expression of EGFP-Tagged CsGRN fusion protein was constructed, while empty vector transfected LO2 cells were used as negative controls. Under the inverted microscope, the green fluorescence was emitted by LO2 cells that were transformed with pEGFP-C1 and pEGFP-C1-CsGRN ( Figure 1A ). Both Western blot and qPCR revealed the higher expression of CsGRN in the stably transfected LO2 cells ( Figures 1B–D ). The eukaryotic expressing vector of pEGFP-C1 is successfully constructed (named as LO2-GFP), and the LO2 cell line that stably expresses pEGFP-C1-CsGRN (named as LO2-GRN) has provided solid experimental foundation for further studies.

Next, we explored the proliferation and cell cycle of LO2 cells. We used two cell models: CsGRN-overexpressed LO2 cells and recombinant CsGRN protein (called as MBP-GRN) treated-LO2 cells to simulate secreted granulin of Clonorchis sinensis in the liver. Compared with control cells, CsGRN-overexpressed and recombinant CsGRN protein treated-LO2 cells both grew quickly and formed more cell clones ( Figures 2A–D ). Since CsGRN induced hepatocyte cell growth, we further used flow cytometry analysis to determine whether the proliferation of CsGRN was due to cell cycle arrest. LO2-GRN cells and MBP-GRN cultured-LO2 cells showed an increased proliferation index with a significant amount of cell accumulation at both S and G2/M phase ( Figures 2E, F ), referring to that CsGRN overexpression induces cell proliferation by stimulating S phase entry and G2/M phase retardation.

To investigate the possible molecular mechanism of CsGRN-induced growth arrest in human hepatocyte, the mRNA and protein levels of various cell-cycle-related regulators were examined. Cell cycle regulation is a delicate biological process, and it forms a complex signaling molecular network system with the participation of multiple genes and proteins (Dynlacht, 1997). The significant increase of CyclinD1, CyclinD3, CDK2, CDK4, and CDK6, and obvious decrease of p18, p21, and p27, both in transcriptional and translational aspects, were detected in LO2-GRN cells and MBP-GRN treated-LO2 cells ( Figure 3 , Figure S1 ). Thereby, stimulating abnormal proliferation of LO2 cells through the regulation of cell-cycle-related genes resulted in malignant transformation of normal hepatocytes.

To identify the biological role of CsGRN in LO2 cells, we investigated the effect of CsGRN on the cell’s migration levels. The scratched wound healed quickly in LO2-GRN cells and MBP-GRN treated-LO2 cells ( Figures 4A, B ). Moreover, the effects on LO2 cells’ migration and invasion of CsGRN were examined by Transwell assay. The number of transmembrane cells, both with or without Matrigel, in the LO2-GRN group and MBP-GRN cultured group were significantly elevated ( Figures 4C, D ). Together, these data suggested that CsGRN promotes the migration and invasion in LO2 cells.

It seems that chronically injured liver cells do not undergo epithelial-to-mesenchymal transition (EMT) (Taura et al., 2010; Chu et al., 2011; Munker et al., 2017). In contrast, it has been reported that hepatocytes in damaged liver regeneration may undergo an EMT-like process (Oh et al., 2018). EMT is considered as a significant cellular event in the process of liver fibrosis (Rowe et al., 2011). Studies revealed that more than 80% of HCC occurs when liver fibrosis or cirrhosis develops, suggesting that liver fibrosis plays an important role in the premalignant environment of the liver (Affo et al., 2017). Therefore, we investigated the possibility that CsGRN affected hepatocytes’ migration and invasion by EMT. Western-blot results proved that there were significant differences between the LO2-GRN and LO2-GFP groups in the expression of vimentin, N-cadherin, β-catenin, and ZO-1 ( Figures 5A, B ). Additionally, the relative mRNA expression of matrix metalloproteinases (MMPs) in LO2-GRN cells exhibited that the MMP9 level was higher in LO2-GRN cells than that in LO2-GFP cells ( Figure 5C ). Thus, we considered that CsGRN enhanced LO2 cell migration through promotion of EMT.

The underlying mechanisms regulated by CsGRN-induced malignant transformation of normal hepatocytes were explored. As CsGRN is one member of the growth factor family, we investigated the potential downstream regulatory effects of EGFR-mediated signaling pathways. The mRNA levels of EGFR, PI3K, Akt, braf, RAS, and ERK were moderately increased both in LO2-GRN cells and MBP-GRN treated-LO2 cells ( Figures 6A, C ). While, the significant elevation of phosphorylated form of EGFR, PI3K, AKT, braf, and ERK proteins were more obvious than their non-phosphorylated form ( Figures 6B, D ). These results indicated that CsGRN activated the RAS/MAPK/ERK and PI3K/Akt signaling pathways through enhancing protein phosphorylation.

To identify the key role of EGFR in CsGRN-induced hepatocytes’ malignant transformation, we co-cultured recombinant CsGRN with EGFR inhibitors in LO2 cells. We found EGFR inhibitor could recover recombinant CsGRN-induced cell proliferation, migration, and invasion of LO2 cells ( Figures 7A–C ). In addition, treating LO2 cell with EGFR inhibitor, Akt inhibitor, RAS inhibitor, braf inhibitor, and ERK inhibitor, respectively, could block the activation of RAS/MAPK/ERK and PI3K/Akt pathways, which were enhanced by recombinant CsGRN ( Figure 7D , Figure S2 ). Therefore, we could draw the conclusion that secreted CsGRN protein promoted malignant transformation of hepatocytes by activation of RAS/MAPK/ERK and PI3K/Akt signaling pathways through EGFR ( Figure 8 ).