This retrospective study utilized archived paraffin-embedded tissue specimens and clinical records from 60 HCC patients who underwent surgical resection at the General Hospital of Ningxia Medical University between 2019 and 2022. Patient inclusion required: (1) postoperative pathological confirmation of HCC, (2) no history of prior anticancer therapy, (3) availability of complete clinical data, and (4) procurement of both tumor and adjacent normal tissue samples (located $>$2 cm from the tumor margin). Exclusion criteria comprised a diagnosis of other concurrent malignancies or significant cardiovascular/respiratory comorbidities. The research was approved by the Medical Research Ethics Review Committee of the General Hospital of Ningxia Medical University (No. KYLL-2019-080). A written consent was signed by the patients or their families/legal guardians.

Antibodies included CCN2 (ab6992, Abcam, Cambridge, Cambridgeshire, UK); $\beta$-actin (ab8226, abcam, Cambridge, Cambridgeshire, UK); c-Met (8198T, Cell Signaling Technology, Danvers, MA, USA); p-Met (3077T, Cell Signaling Technology, Danvers, MA, USA); Integrin$\alpha$V (3019, Cell Signaling Technology, Danvers, MA, USA); FAK (3285, Cell Signaling Technology, Danvers, MA, USA); p-FAK (3283, Cell Signaling Technology, Danvers, MA, USA); YAP (14074, Cell Signaling Technology, Danvers, MA, USA); and p-YAP (13008, Cell Signaling Technology, Danvers, MA, USA).

Paraffin-embedded tissue sections underwent sequential deparaffinization using xylene gradients and rehydration through graded ethanol series. Antigen retrieval was achieved through microwave irradiation in sodium citrate buffer (pH 6.0), followed by natural cooling to room temperature. The staining protocol included three cycles of PBS washing (5 min each), endogenous peroxidase inactivation with 3% H₂O₂ (10 min, room temperature), and non-specific binding blockage using 5% BSA (30 min, RT). Primary antibody incubation was performed at 4 °C overnight, followed by HRP-conjugated secondary antibody (1:200) treatment for 1 h at room temperature. Signal visualization employed 3,3^′-Diaminobenzidine (DAB) chromogenic substrate with hematoxylin counterstaining, culminating in ethanol dehydration, xylene clearing, and neutral resin mounting. Two blinded pathologists independently evaluated staining outcomes using a semi-quantitative scoring system: (1) Positive cell proportion: 1 ($\le$25%), 2 (26–50%), 3 (51–75%), 4 ($>$75%); (2) Staining intensity: 0 (negative), 1 (weak), 2 (moderate), 3 (strong). Final scores (range: 0–12) were calculated by multiplying proportion and intensity scores. Specimens were stratified into low-expression (median score) and high-expression ($\ge$median score) cohorts.

Cell lines including MHCC97L, MHCC97H, HCCLM3, Hep3B, Huh-7, and LO2 were all kindly provided by the Liver Disease Research Institute affiliated with Fudan University in Shanghai. These cells were cultured in Dulbecco’s modified Eagle medium (DMEM, 2329225, VivaCell, Shanghai, China) supplemented with 10% fetal bovine serum (FSP500, ExCell Bio, Shanghai, China) and 1% penicillin-streptomycin (P1400, Solarbio, Beijing, China). All cells were maintained at 37 °C with 5% CO₂. All cell lines used in this study were authenticated by short tandem repeat (STR) profiling and confirmed to be mycoplasma-negative. Lentiviruses for CCN2 overexpression and knockdown were purchased from Beijing Tsingke Biotech Co., Ltd. (China). Lentiviral transfections were conducted according to the manufacturer’s instructions, with Lipo3000 (L3000015, Invitrogen, Carlsbad, CA, USA) used as recommended. Screening was performed using puromycin (P8230, Solarbio, Beijing, China).

Total RNA was extracted from cells using the MolPure® Cell/Tissue Total RNA Kit (19221ES50, Yeasen Biotechnology, Shanghai, China), and cDNA was synthesized using the PrimeScript™ RT Master Mix (Perfect Real Time) (RR036A, Takara Bio, Kyoto, Japan). Subsequently, RT-qPCR reactions were performed using the TB Green® Premix Ex Taq™ RNaseH Plus kit (RR820A, Takara Bio, Kyoto, Japan). The primers (Table 1) used in this study were designed and synthesized by Tsingke Biotech Co., Ltd. (Beijing, China).

Protein extraction from hepatic carcinoma specimens and cultured cells was performed using radioimmunoprecipitation (RIPA) lysis buffer (P0013B, Beyotime Biotechnology, Shanghai, China) supplemented with protease/phosphatase inhibitors. After 20-min ice incubation, lysates were clarified by centrifugation (12,000 $\times$g, 10 min, 4 °C) to remove insoluble components. Protein quantification was conducted via bicinchoninic acid (BCA) assay (KGB2101-100, KeyGEN BioTECH, Nanjing, Jiangsu, China) with bovine serum albumin as standard. Equal amounts of protein (30 µg/lane) were resolved on 10% SDS-polyacrylamide gels and electrophoretically transferred to polyvinylidene fluoride (PVDF) membranes. Membranes were blocked with 5% BSA or non-fat milk (1 h, RT) prior to overnight incubation with primary antibodies (4 °C). Following three TBST (Tris-buffered saline, 0.1% Tween 20) washes, membranes were probed with species-matched HRP-conjugated secondary antibodies (1:5000, 1 h, RT). Subsequent to additional TBST rinses, chemiluminescent signals were developed using ECL Plus substrate (PE0010, Solarbio, Beijing, China) and quantified with the Tanon-5200 automatic chemiluminescence imaging analysis system (Tanon, Shanghai, China).

In the lower chamber of a Transwell insert (16122062, Corning, New York, NY, USA), 500 µL of DMEM containing 10% fetal bovine serum was added, while 200 µL of DMEM containing 8 $\times$ 10⁴ cells was added to the upper chamber. Following 48 h of incubation, the inserts were removed, fixed with 4% paraformaldehyde for 15 min, and stained with 0.1% crystal violet for 30 min. Images were captured using an inverted microscope (DMi8, Leica, Wetzlar, Hesse, Germany), and the number was counted using ImageJ software (version 1.8.0, National Institutes of Health, Bethesda, MD, USA).

Hepatocellular carcinoma cells were plated in six-well plates and cultured to confluence, while wounds were inflicted using a 200 µL pipette tip. After removing cell debris with PBS, cells were cultured in serum-free medium. Wound closure was monitored using an inverted microscope (DMi8, Leica, Wetzlar, Hesse, Germany) every 0 and 24 h.

Cell proliferation was assessed using the Cell Counting Kit-8 (B34304, Biomake, Beijing, China). HCC cells were seeded in 96-well plates at a density of 2000 cells per well and cultured at 37 °C for 12, 24, 48, 72, and 96 h, respectively. Each well was subsequently treated with 10 µL of CCK-8 solution and 100 µL of DMEM. Upon incubation for 2 h, absorbance was measured at 450 nm using a microplate reader (Multiskan SkyHigh, Thermo Scientific, Waltham, MA, USA).

Hepatocellular carcinoma cells were seeded in 6-well plates at a density of 500 cells per well. After 14 d of culture, the cells were fixed with 4% paraformaldehyde for 15 min, washed three times with PBS, and stained with 0.1% crystal violet for 30 min. Afterwards, the number of colonies was counted using ImageJ software (version 1.8.0, National Institutes of Health, Bethesda, MD, USA).

Cells were seeded into 96-well plates at a density of 5000 cells per well and incubated. Once the cells adhered, varying concentrations of Sorafenib (0 µM, 1.25 µM, 2.5 µM, 5 µM, 10 µM, 20 µM, 40 µM, and 80 µM) were added according to the experimental groups. Upon further incubation for 24 h, CCK-8 solution was added, the absorbance at 450 nm was measured, and cell viability was calculated.

RNA-seq data for HCC were obtained from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). Based on the median expression level of CCN2, HCC patients were categorized into high-expression (n = 187) and low-expression (n = 187) groups, and differentially expressed genes between these two groups were identified. The resulting CCN2-associated differentially expressed genes were subsequently analyzed and functionally enriched using the GSEA platform (https://www.gsea-msigdb.org/gsea/index.jsp). Transcription factors potentially regulating the c-Met gene were predicted via the humanTFDB database. Finally, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was conducted on these c-Met-related transcription factors using The Database for Annotation, Visualization and Integrated Discovery (DAVID) database (https://davidbioinformatics.nih.gov/summary.jsp). The correlation between Met and Yap expression was analyzed using the Gene Expression Profiling Interactive Analysis (GEPIA) database (http://gepia.cancer-pku.cn/).

Protein complexes were isolated using Protein A/G Magnetic Beads (88802, Thermo Scientific, Carlsbad, CA, USA) according to manufacturer specifications. For antibody conjugation, 20 µL bead suspension was incubated with either control IgG (2729, Cell Signaling Technology, Danvers, MA, USA) or CCN2-specific antibody (50 µg/mL final concentration, ab6992, Abcam, Cambridge, Cambridgeshire, UK) under rotary agitation (25 °C, 1 h). Following magnetic separation (800 $\times$g, 30 sec) and three wash cycles with PBS-0.1% Tween, pre-cleared cellular/tissue lysates were introduced for target protein capture (4 °C, overnight rotation). Sequential purification steps included: ① Four cycles of binding/wash buffer treatment with magnetic separation. ② Two washes with 20 bead-bed volumes of Tris-buffered saline (TBS, pH 7.4). ③ Acidic elution using 0.1 M glycine-HCl (pH 3.5, $\le$20 min incubation).

Eluted proteins were denatured in 5$\times$ Laemmli buffer (95 °C, 10 min) prior to immunoblotting analysis. This optimized protocol maintained protein complex integrity while effectively reducing non-specific binding, achieving $>$85% capture efficiency as validated by tandem mass spectrometry.

Cells were plated in 24-well plates with cover slips and cultured to 50%–70% confluence. The medium was removed, and cells were washed three times with PBS, fixed with 4% paraformaldehyde at room temperature for 20 min, and washed again three times with PBS. Cells were then permeabilized with 0.3% Triton X-100 for 20 min, blocked with 5% BSA for 1 h, and incubated with primary antibodies overnight at 4 °C. Following this, cells were incubated with fluorescent secondary antibodies for 1 h at room temperature, stained with DAPI for nuclear counterstaining, and mounted. Images were captured using a fluorescence inverted microscope (DMi8, Leica, Wetzlar, Hesse, Germany) .

Female BALB/c-nu nude mice (4–5 weeks old) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. and housed under specific pathogen-free (SPF) conditions. Subcutaneous xenograft models were established by injecting 150 µL of a cell suspension containing 5 $\times$ 10⁶ cells into the left axilla of each mouse. On day 30 post-inoculation, mice were humanely euthanized through a two-stage procedure: initial anesthesia was induced with isoflurane inhalation (4% in oxygen), followed by an intraperitoneal injection of sodium pentobarbital (100 mg/kg) to ensure cardiac arrest. Death was confirmed by the cessation of circulatory function via venous puncture. The tumors were then excised, and their volume (calculated as 0.5 $\times$ length $\times$ width²), weight, and macroscopic appearance were recorded. All experimental procedures involving animals were performed in strict accordance with the National Institutes of Health guidelines for the care and use of laboratory animals and were approved by the Medical Research Ethics Review Committee of the General Hospital of Ningxia Medical University (No. KLYY-2019-080).

Statistical analysis and graphical presentation were conducted using SPSS (version 26.0, IBM, Armonk, NY, USA) and GraphPad Prism (version 8.0, San Diego, Boston, MA, USA). For quantitative data following a normal distribution, comparisons between two groups were made using t-tests, and comparisons among multiple groups were performed using analysis of variance (ANOVA). The association between CCN2 and clinical pathological characteristics was further assessed using t-tests, Wilcoxon rank-sum tests, chi-square tests, and Fisher’s exact tests, with p 0.05 indicating statistical significance.

Clinical data and specimens from 60 HCC patients treated at the General Hospital of Ningxia Medical University between 2019 and 2022 were analyzed in this study. Immunohistochemical analysis localized CCN2 predominantly to the cytoplasm of HCC cells (Fig. 1A). This study investigated the correlation between CCN2 expression levels and the clinicopathological characteristics of HCC patients in a real-world study (Table 2). The results revealed that CCN2 expression was significantly associated with a history of hepatitis B, aspartate aminotransferase (AST) levels, the severity of liver cirrhosis, and tumor size (p 0.05). Consistent with these clinical findings, Western blot analysis revealed differential CCN2 protein expression across hepatocellular carcinoma cell lines. Specifically, CCN2 expression was markedly elevated in Hep3B and Huh-7 cell lines compared to the normal hepatic LO2 cell line (Fig. 1B). To investigate the functional role of CCN2, we generated stable CCN2-overexpressing LM3 cells and CCN2-knockdown Hep3B/Huh-7 cells using lentiviral transduction. The efficiency of genetic manipulation was confirmed at both the protein and transcript levels by Western blot (Fig. 1C) and RT-qPCR (Fig. 1D), respectively. CCK-8 and colony formation assays demonstrated that CCN2 knockdown significantly suppressed the proliferation of hepatocellular carcinoma cells, whereas CCN2 overexpression enhanced it (Fig. 1E,F). Similarly, Transwell migration assays revealed that CCN2 depletion inhibited cell migration (Fig. 1G), and wound healing assays confirmed that CCN2 overexpression promoted migratory capacity (Fig. 1H).

Fig. 1.

Expression and biological functions of CCN2 in HCC. (A) Immunohistochemical assessment of CCN2 expression in hepatocellular carcinoma tissues. Scale bar: 200 µm (upper) and 50 µm (lower). (B) Western blot analysis of CCN2 expression levels in hepatocellular carcinoma cell lines. (C,D) Establishment of stable hepatocellular carcinoma cell lines with CCN2 overexpression and knockdown. (E,F) CCK-8 and colony formation assays to evaluate the effects of CCN2 on hepatocellular carcinoma cell proliferation and clonogenic potential. (G,H) Transwell and wound-healing assay to assess the effects of CCN2 on hepatocellular carcinoma cell migration. Scale bar: 300 µm (G) and 500 µm (H). Per group n = 3, ***p 0.001.

The in vitro sensitivity of various hepatocellular carcinoma cell lines (MHCC97L, MHCC97H, HCCLM3, Huh-7, Hep3B) and the normal LO2 cell line to sorafenib was assessed, yielding IC₅₀ values of 16.92, 21.76, 17.42, 11.66, 10.48, and 14.84 µM, respectively (Fig. 2A). Subsequent experiments demonstrated that CCN2 overexpression in HCCLM3 cells increased the sorafenib IC₅₀ (CON vs. OV: 16.95 vs. 23.29 µM, p 0.001), whereas CCN2 knockdown in Hep3B cells decreased it (shCON vs. sh#3: 10.91 vs. 7.293 µM, p 0.001) (Fig. 2B), indicating that CCN2 downregulation enhances sorafenib sensitivity. This conclusion was further supported by CCK-8 assays, in which combined CCN2 knockdown and sorafenib treatment significantly suppressed HCC cell proliferation (p 0.001; Fig. 2C).

Fig. 2.

Impact of CCN2 expression changes on sorafenib resistance in hepatocellular carcinoma cells in vitro. (A) Determination of sorafenib IC₅₀ values in hepatocellular carcinoma cell lines MHCC97L, MHCC97H, HCCLM3, Huh-7, Hep3B, and the normal liver cell line LO2. (B) Assessment of sorafenib IC₅₀ values in hepatocellular carcinoma cell lines with CCN2 knockdown or overexpression. (C) CCK-8 assay evaluating the effect of CCN2 knockdown combined with sorafenib on the proliferation of hepatocellular carcinoma cells. ***p 0.001.

Sorafenib, the first FDA-approved multi-targeted tyrosine kinase inhibitor for advanced hepatocellular carcinoma, inhibits tumor growth primarily by suppressing angiogenesis. To identify potential mediators of resistance, a tyrosine kinase PCR array profiling 90 tyrosine kinases (60 receptor and 30 non-receptor types) was used in CCN2-knockdown Hep3B cells. This screen revealed a broad downregulation of tyrosine kinase mRNAs, most notably c-Met (Fig. 3A). Subsequent validation by RT-qPCR and Western blot confirmed that CCN2 knockdown specifically reduces c-Met expression at both the mRNA and protein levels (Fig. 3B,C). These findings indicate that CCN2 promotes sorafenib resistance in HCC cells, at least in part, through the activation of c-Met.

Fig. 3.

CCN2 enhancing sorafenib resistance in hepatocellular carcinoma cells through activation of the c-Met signaling pathway. (A) Tyrosine kinase PCR array analysis. (B) RT-qPCR detection of changes in CCN2 and c-Met expression levels upon CCN2 knockdown. (C) Western blot analysis of changes in c-Met protein expression levels upon CCN2 knockdown or overexpression. ***p 0.001.

To further elucidate the specific mechanisms of CCN2 regulating c-Met transcription, the HumanTFDB was hereby used to predict transcription factors of the c-Met gene, and the DAVID database was employed for pathway enrichment analysis. The results indicated that the Hippo Signaling Pathway was involved in the transcriptional regulation of c-Met by CCN2 (Fig. 4A). Additionally, GSEA enrichment analysis showed significant enrichment of Hippo Signaling Regulation Pathways in populations with high CCN2 expression (Fig. 4B). The Hippo kinase pathway was identified as a carcinogenic resistance mechanism following many traditional small-molecule therapies, with YAP, a key transcriptional coactivator, translocating to the nucleus upon activation to regulate the transcription of downstream genes. The GEPIA database also demonstrated a positive correlation between YAP and c-Met expression (Fig. 4C). Additionally, subsequent immunofluorescence revealed that overexpression of CCN2 promoted YAP nuclear translocation, while CCN2 knockdown reduced YAP nuclear aggregation (Fig. 4D), demonstrating CCN2-mediated YAP nuclear translocation as a critical mechanism in regulating c-Met expression.

Fig. 4.

CCN2 upregulating c-Met expression in hepatocellular carcinoma cells via integrin-mediated FAK/YAP activation. (A) Enrichment of c-Met transcription factor signaling pathways. (B) GSEA enrichment analysis indicating an association between CCN2 and Hippo Signaling Regulation Pathways. (C) Positive correlation between YAP and c-Met. (D) Immunofluorescence detection of YAP changes following CCN2 overexpression or knockdown. Scale bar: 10 µm. (E) Co-immunoprecipitation (Co-IP) analysis of the interaction between CCN2 and integrin $\alpha$V. (F) Western blot analysis of FAK, p-FAK, and integrin $\alpha$V expression changes following CCN2 overexpression or knockdown. (G) Treatment of LM3 cells overexpressing CCN2 with the integrin inhibitor Cyclo(-RGDfK) resulting in corresponding inhibition of FAK phosphorylation levels. (H,I) CCN2 upregulating c-Met expression via integrin-mediated FAK/YAP activation. FAK/YAP, focal adhesion kinase/yes-associated protein.

Following that, the mechanisms of CCN2 regulating YAP were investigated. Integrin subunit $\alpha$V (Integrin $\alpha$V), a component of integrins and a key target of CCN2, transmits signals from integrins to downstream targets via FAK, thereby promoting tumor progression. FAK also serves as a critical molecule mediating YAP nuclear translocation and transcriptional regulatory effects. In this study, to clarify the relationship between CCN2 and Integrin $\alpha$V, co-immunoprecipitation (Co-IP) results confirmed that CCN2 bound to and interacted with Integrin $\alpha$V (Fig. 4E). Western blot analysis showed that the phosphorylation levels of FAK decreased or increased correspondingly upon CCN2 knockdown or overexpression, while no significant changes were observed in Integrin $\alpha$V expression (Fig. 4F). This suggested that Integrin $\alpha$V might be activated upon binding to CCN2, thereby promoting the phosphorylation of downstream FAK. To test this hypothesis, the integrin inhibitor Cyclo(-RGDfK) was utilized to interfere with the activity of Integrin $\alpha$V, and the results showed that Cyclo(-RGDfK) effectively reversed the activation of FAK by CCN2 (Fig. 4G).

To investigate the dependence of CCN2’s regulation of c-Met on the Integrin $\alpha$V/FAK/YAP axis, the activity of Integrin $\alpha$V, FAK, and YAP was sequentially inhibited using the integrin inhibitor Cyclo(-RGDfK), the FAK inhibitor Defactinib, and the YAP inhibitor Verteporfin. Western blot analysis demonstrated that silencing CCN2 decreased the activity of FAK and YAP, as well as the expression of c-Met. In contrast, the addition of Cyclo(-RGDfK), Defactinib, and Verteporfin progressively suppressed the activity or expression of corresponding downstream molecules. Overexpression of CCN2 increased the activity of FAK and YAP, and the expression of c-Met, an effect effectively reversed by inhibiting the activity of Integrin $\alpha$V, FAK, and YAP (Fig. 4H,I). These results suggested that the secreted protein CCN2 bound to membrane receptors such as integrins to activate downstream FAK, which subsequently mediated the nuclear translocation of the transcriptional coactivator YAP, thereby enhancing the transcriptional activity of c-Met in hepatocellular carcinoma cells.

To investigate the role of CCN2 in promoting tumor growth in vivo and conferring resistance to sorafenib, Huh-7 cells with either control or CCN2 knockdown (assigned to shCON, shCCN2, shCON + Sorafenib, and shCCN2 + Sorafenib groups), as well as HCCLM3 cells with control or CCN2 overexpression (designated as control and CCN2 groups), were subcutaneously injected into the left anterior axilla of female BALB/c nude mice to generate xenograft tumor models (n = 6 per group). Mice in the Sorafenib and shCCN2 + Sorafenib groups were administered 25 mg/kg sorafenib via intraperitoneal injection every three days, while control groups received an equivalent volume of saline. Tumor growth was monitored tri-weekly. Following a 30-day period, the mice were euthanized, and the tumors were excised for volume and weight measurement (Fig. 5A–C). The results showed that tumors in the CCN2-overexpressing group were significantly larger than those in the control group (806.87 mm³ vs. 502.49 mm³). Critically, the shCCN2 + Sorafenib combination group exhibited the smallest tumor volume (135.10 mm³), which was substantially reduced compared to the shCON (649.33 mm³), Sorafenib-alone (353.59 mm³), and shCCN2-alone (315.29 mm³) groups (Fig. 5A,B). Corroborating this functional data, immunohistochemical analysis confirmed the involvement of the c-Met pathway: c-Met expression was significantly lower in the shCCN2 group than in the shCON group, and higher in the CCN2-overexpressing group compared to its control (Fig. 5D). Collectively, these findings indicate that CCN2 interference alleviates c-Met-mediated sorafenib resistance, and that combining sorafenib with CCN2 targeting enhances its antitumor efficacy.

Fig. 5.

In vivo xenograft tumor study. (A) Representative images of xenograft tumors after excision from each group (n = 6). (B) Changes in tumor volume across groups. (C) Trends in body weight of nude mice across groups. (D) Immunohistochemical detection of CCN2 and c-Met expression levels. Scale bar: 200 µm (upper) and 50 µm (lower). ***p 0.001.