LM3 (cat. no. CL-0278), Huh7 (CL-0120), MHCC-97H (CL-0497) and Hep3B (cat. no. CL-0102) cells were purchased from Procell Life Science & Technology Co., Ltd. Cells were cultured in DMEM high glucose supplemented with 10% FBS, 100 mg/ml penicillin G, and 50 mg/ml streptomycin in a humidified incubator at 37°C supplied with 5% CO₂. MHCC-97H (CL-0497) cells were used for western blotting.

TGF-β was purchased from PeproTech, Inc. FXR1 (cat. no. 67813-1-Ig, 1:10,00 or 1:100), Ki67 (27309-1-AP; 1:100) and β-actin (cat. no. 66009-1-Ig, 1:10,00) antibodies were purchased from ProteinTech Group, Inc. The FXR1 antibody was diluted 1:1,000 in the western blotting assay and 1:100 in immunohistochemistry assays. SMAD2/3 (cat. no. # 8685S, 1:1,000), N-Cadherin (cat. no. # 4061, 1:1,000), and slug (cat. no. 9585S, 1:1,000) antibodies were purchased from Cell Signaling Technology, Inc. The secondary antibodies used were HRP-conjugated Affinipure Goat Anti-Rabbit (1:2,000 or 1:50; cat. no. SA00001-2) and HRP-conjugated Affinipure Goat Anti-Mouse (1:2,000; cat. no. SA00001-1) (both from ProteinTech Group, Inc.). HRP-conjugated Affinipure Goat Anti-Rabbit was diluted 1:2,000 in the western blotting assay and 1:50 in immunohistochemistry assays. SMAD2/3 siRNA was purchased from Santa Cruz Biotechnology, Inc. (cat. no. sc-37238). shRNA was purchased from Public Protein/Plasmid Library. The sequences of the FXR1 knockdown shRNAs were: shFXR1#1, 5′-GCTAGAGGTTTCTTGGAATTT-3′; shFXR1#2, 5′-CGCCAGGTTCCATTTAATGAA-3′; and negative control (shNC), 5′-GTTCTCCGAACGTGTCACGTT-3′. The FXR1 expression plasmid with a 3′FLAG tag was purchased from Public Protein/Plasmid Library.

TRIzol^® reagent was used to extract total RNA from tissues or cells, and 1 µg RNA was reverse transcribed to cDNA using a Vazyme reverse transcription kit according to the manufacturer's protocol (Vazyme Biotech Co., Ltd.; cat. no. R323-01). Amplification was performed using amplification reagent according to the manufacturer's protocol (Vazyme Biotech Co., Ltd.; cat. no. Q711-02). The thermocycling conditions for amplification were as follows: 95°C for 5 min, followed by 40 cycles at 95°C for 10 sec, 60°C for 30 sec, and a final step at 95°C for 15 sec, 60°C for 60 sec and 95°C for 15 sec. β-actin was used as the housekeeping gene. The sequences of the primers used for amplification were: β-actin forward, 5′-CATGTACGTTGCTATCCAGGC-3′ and reverse 5′-CTCCTTAATGTCACGCACGAT-3′; SMAD2 forward, 5′-GATCCTAACAGAACTTCCGCC-3′ and reverse, 5′-CACTTGTTTCTCCATCTTCACTG-3′; and SMAD3 forward, 5′-TCCATCCCCGAAAACACTAAC-3′ and reverse, 5′-CATCTTCACTCAGGTAGCCAG-3′ (23). The relative expression level of the target genes was calculated using the 2^−ΔΔCq method (23).

For transient transfection of plasmids, LM3 cells were transfected using Lipofectamine^® 3000 (Thermo Fisher Scientific, Inc.). For transfection of shRNAs (final concentration, 30 µM), proliferating cells in a 6-well plate were incubated in serum-free DMEM containing Lipofectamine^® 3000. After 4–6 h, the cells were incubated in supplemented DMEM for 48 h, and then treated with the newly configured puromycin (2 µg/ml) screening medium every day until no cell death was observed. The maintenance medium contained puromycin (2 µg/ml). To transfect siRNAs (final concentration, 20 µM), proliferating cells in a 6-well plate were incubated with serum-free DMEM supplemented with Lipofectamine^® 3000. After 4–6 h, the cells were incubated in complete DMEM for 48 h and then selected with puromycin (2 µg/ml) until no cellular death was observed. The maintenance medium contained puromycin (2 µg/ml).

Total protein was extracted using RIPA lysis buffer supplemented with a mixture of PMSF, aprotinin, and phosphatase inhibitors. LM3 (cat. no. CL-0278), Huh7 (CL-0120), MHCC-97H (CL-0497) and Hep3B (cat. no. CL-0102) cells were used for western blotting. The protein concentration was measured using a Bicinchoninic protein assay (Thermo Fisher Scientific, Inc.). Proteins (30 µg per lane) were separated using SDS-PAGE on a 10% SDS gel, transferred to PVDF membranes, and incubated with the primary antibody, followed by incubation with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody at room temperature for 1 h. BeyoECL Plus (Beyotime Institute of Biotechnology) was used for signal visualization, and images were obtained using a Fushon Fx (Vilber Lourmat) imaging system.

A Beyotime Institute of Biotechnology apoptosis kit was used for testing (cat. no. C1069L). The cells were washed with PBS once and detached using pancreatic enzyme cell digestion solution, at room temperature until gentle aspiration with a pipette was sufficient to lift cells (ensuring over-digestion was avoided), after which the pancreatic enzyme cell digestion solution was removed. The cells were collected, mixed gently, and centrifuged at 1,000 × g at 4°C for 5 min, after which the cells were collected and gently resuspend in PBS for counting. A total of 50–100 µl of the resuspended cell solution was taken, centrifuged at 1,000 × g at 4°C for 5 min, the supernatant was discarded, 195 µl annexin V-mCherry Binding Buffer was added, and cells were gently resuspended. A total of 5 µl annexin V-mCherry was added and mixed gently, incubated at room temperature (20–25°C) for 10–20 min, avoiding light, followed by incubation in an ice bath. Aluminum foil was used to protect against light. Cells were resuspended 2–3 times during incubation to ensure appropriate and equal staining. Subsequently, cells were analyzed using flow cytometry (BD FACSymphony™ A3; BD Biosciences) for Annexin V-mCherry red fluorescence (Beyotime Institute of Biotechnology apoptosis kit; cat. no. C1069L). FlowJo™10 (BD Biosciences) was used for analysis.

After digestion of cells using cellular pancreatic enzyme during the logarithmic growth phase, cells were resuspended in supplemented media and counted. A total of 400–1,000 cells/well (generally 700 cells/well) were seeded in each experimental group in a 6-well culture plate and cultured for 14 days, changing the media every 3 days. After 14 days, 1 ml crystal violet dye solution was added to each well and incubated at room temperature for 10–20 min, after which cells were washed several times, dried, and imaged with a digital camera attached to a bright field microscope at a magnification of ×200. Colonies consisting of ≥50 cells were counted.

Cell migration was determined using a Transwell assay. Transfected Hep3B or LM3 cells (5×10⁴ per well) or cells treated with 20 ng/ml TGF-β were resuspended in 500 µl serum-free medium containing 2 µg/ml mitomycin C and added to the upper chamber to inhibit cell proliferation, and 500 µl supplemented media was added to the lower chamber. After incubation at 37°C for 48 h, the cells on the upper surface of the filter membrane were removed using a swab and the filter membrane was incubated in 100% methanol at room temperature for 2 min. The cells that had migrated to the underside of the filter were stained with 0.5% crystal violet at room temperature for 20 min, and images were obtained using a bright field microscope. The number of cells that had migrated was counted using Image Pro Plus 6.0 (Media Cybernetics, Inc.). Cell invasion was determined using the same method as that for cell migration except cells were seeded in a Transwell chamber covered with Matrigel (BD Biosciences).

A total of 20 BALB/c nude mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. Mice were maintained with a 12-h light/dark cycle under specific pathogen-free conditions. All animal experiments were approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University (approval no. 2019-KY-21) and in accordance with the guidelines of the Office of Laboratory Animal Welfare. Mice were monitored daily and euthanized according to NIH-approved criteria for institutional humane endpoints if they failed to demonstrate a corrective response. For tumor growth experiments, stably transfected LM3 cells (1×10⁵) were injected into mice by subcutaneous injection in the abdomen. The volume of the tumor was measured once every 2 days as follows: Volume=0.5× length × width × height; the total volume of each tumor did not exceed 4,400 mm³. When the tumor size reached 12–20 mm the mice were euthanized via cervical dislocation; death was confirmed by lack of breathing and heartbeat.

The study included 88 patients with HCC (74 male and 14 female patients) treated at the First Affiliated Hospital of Zhengzhou University. HCC (2–3 cm) and normal tissue samples were obtained after surgery between January 2020 and December 2021. Patients were aged 30–76 years (median age, 56 years). Complete clinical information was available for included patients, and none of the patients had other malignancies or a history of chemotherapy or radiation therapy.

A total of 88 pairs of HCC and normal tissues (74 males and 14 female patients) were obtained from the First Affiliated Hospital of Zhengzhou University (Zhengzhou, China). The tumors of nude mice were tested for Ki67 after removal.

For immunohistochemistry, tissue sections were mounted on a glass slide, deparaffinized, and rehydrated. Samples were soaked in 100% alcohol, 100% alcohol, 95% alcohol, 75% alcohol and water for 5 min each for hydration. Heat-induced antigen retrieval was subsequently performed using 10 nmol/l citrate buffer (pH=6.0) for 10 min in a microwave oven (high heat). The slides were incubated with FXR1 (cat. no. 67813-1-Ig; 1:100) or Ki67 (27309-1-AP; 1:100) antibody overnight at 4°C, followed by incubation with a HRP labeled secondary antibody (cat. no. SA00001-2; 1:50) at room temperature for 15 min. The slides were further stained with DAB developer and counterstained with 5% hematoxylin at room temperature for 3 min. Positive cells were stained brown.

The expression levels were scored based on the percentage of positive cells and the intensity of staining. The score was calculated as follows: 0, <5% positively stained cells; 1, 5–25% positively stained cells; 2, 25–50% positively stained cells; 3, 50–75% positively stained cells; and 4, >75% positively stained cells. The intensity was scored as follows: 0, negative staining; 1, light yellow staining; 2, brownish yellow staining; and 3, chocolate-brown staining. The final score was calculated by multiplying the percentage score by the intensity score. A score <8 was considered low, and a score ≥8, was considered high.

The Oncomine online database (http://www.oncomine.org) was used to analyze FXR1 transcripts in HCC from the Rossher liver and Rossher liver 2 datasets. UALCAN (http://ualcan.path.uab.edu) was used to analyze the expression of FXR1 in HCC data from TCGA. GEPIA (http://gepia.cancer-pku.cn/) was used to analyze the correlations between i) FXR1 and SMAD2/3/4 mRNA levels and ii) FXR1 expression in TCGA. The threshold for mRNA levels was customized to allow for the stratification of subjects with low or high FXR1 mRNA levels. TCGA raw data, such as gene expression and clinicopathological data, were included in the UALCAN and cBioPortal analyses (http://www.cbioportal.org). Gene Set Enrichment Analysis (GSEA) was performed to explore potential biological pathways based on Kyoto Encyclopedia of Genes and Genomes (KEGG) gene sets using R software (version 4.1.2; http://www.r-project.org/). For KEGG enrichment analysis, the ‘clusterProfiler’ R package was used to decode potential targets downstream of FXR1, the gseKEGG function was used to explore the underlying biological significance. Based on gene count thresholds generated from the median value of FXR1 expression (TPM value=4.16), patients were classified into low and high groups.

SPSS software (version 22.0; IBM Corp.) was used for statistical analyses. A one-way ANOVA followed by a post hoc Bonferroni correction was used for the statistical analysis of mRNA expression, cell migration, and invasion. A Student's t-test was used to compare the differences between two groups. A Pearson's correlation test was used to assess correlations. Paired samples were tested using a paired t-test. P<0.05 was considered to indicate a statistically significant difference.

To investigate FXR1 expression in HCC, data from the HCC cohort in TCGA was used; FXR1 was upregulated in HCC (Fig. 1A). Validation of FXR1 expression was also performed using HCC and normal tissues, and the results showed that FXR1 was highly expressed in HCC at the protein level (Fig. 1B and C). Next, FXR1 expression was assessed in HCC and normal tissues using immunohistochemical microarrays, and similar results were observed; FXR1 expression was higher in HCC tissues than in normal tissues (Fig. 1D and E).

To explore potential targets downstream of FXR1, GSEA was used to identify the latent biological pathways based on KEGG gene sets; the top 10 corresponding genes are shown in Fig. 2A. Bioinformatics analysis was used to assess the correlation between FXR1 and genes from TCGA. The mRNA expression levels of FXR1 were positively correlated with the mRNA expression levels of SMAD2/3 (Fig. 2B).

We further investigated the effects of FXR1 on the migratory and invasive effects of HCC and the involvement of SMAD2/3. The protein expression levels of FXR1 in five different HCC cell lines were examined, and the results showed that FXR1 was relatively highly expressed in LM3 and Hep3B cells (Fig. 3A). shRNAs were used to knock down FXR1 expression in LM3 and Hep3B cells, and knockdown was confirmed using western blotting (Fig. 3B). Western blotting also showed a decrease in the expression of N-cadherin and Slug in FXR1 knockdown cells (Fig. 3B). These results suggest that FXR1 affects the expression of EMT-related proteins. In addition, the expression levels of SMAD2/3 decreased upon FXR1 knockdown (Fig. 3B).

TGF-β promotes tumor migration and invasion through EMT (24,25). The effect of FXR1 on the TGF-β-induced migration and invasion in HCC cells was thus next assessed. FXR1 knockdown cells transfected with the two different shRNAs consistently inhibited the migration and invasion of LM3 and Hep3B cells (Fig. 3C and D). Additionally, shRNA-induced FXR1 knockdown inhibited the TGF-β-induced increase in HCC cell migration and invasion when compared with cells treated with TGF-β and transfected with shNC (Fig. 3C and D).

To validate the effect of FXR1 on apoptosis in HCC cells, two different shRNAs were used to knockdown FXR1 in Hep3B and LM3 cells, and apoptosis was assessed using flow cytometry. The results showed that FXR1 knockdown enhanced early apoptosis in HCC cells (Fig. 4A and B). Additionally, the knockdown of FXR1 inhibited the proliferation of HCC cells in the colony formation assay. These results were validated in both Hep3B and LM3 cells (Fig. 4C and D). These results suggest that knockdown of FXR1 inhibited the proliferation of HCC cells and promoted early apoptosis of HCC cells.

FXR1 overexpressing LM3 cell lines with or without SMAD2/3 knockdown were established. RT-qPCR was used to verify the knockdown efficiency of siSMAD2/3 (Fig. S1). The results showed that FXR1 overexpression enhanced LM3 cell invasion. Knockdown of SMAD2/3 inhibited the FXR1 overexpression-induced increase in the invasion of LM3 cells (Fig. 5B). The transfection efficiency of FXR1 and SMAD2/3 was verified using western blotting (Fig. 5A). The results of this experiment validated the role of SMAD2/3 in the FXR1-induced increase in HCC invasion.

To verify the role of FXR1 in vivo, mice were injected with shNC-transfected LM3 cells or shFXR1-transfected LM3 cells. By measuring the volume of the tumors every 2 days, the results showed that tumors grew slower in FXR1 knockdown cells (Fig. 6A and B). The average diameter of the NC tumors was 14.5 mm (range, 12.66-16.84 mm) and the mean diameter of the shFXR1 tumors was 11.9 mm (range, 9.15-14.70 mm); the difference was statistically significant. The expression of Ki67 in the tumors was assessed using immunohistochemistry, which showed a significant decrease in Ki67 expression in the tumors after FXR1 knockdown (Fig. 6C and D). These data indicate that FXR1 knockdown inhibits HCC proliferation in vivo.