EZH2 confers lenvatinib resistance in hepatocellular carcinoma by suppressing ACSL1-Mediated ferroptosis

Abstract

Background

Lenvatinib resistance significantly limits treatment efficacy in hepatocellular carcinoma (HCC), yet the underlying mechanisms remain poorly understood. This study investigates the role of EZH2 in mediating lenvatinib resistance through ferroptosis regulation, aiming to identify novel therapeutic targets for overcoming drug resistance in HCC.

Methods

EZH2 expression patterns were analyzed using TCGA datasets and validated in clinical HCC samples through RT-qPCR. Lenvatinib-resistant HCC cell lines were established to examine EZH2’s functional role. The impact of EZH2 on ferroptosis was evaluated by measuring cell proliferation, reactive oxygen species (ROS), malondialdehyde (MDA), and glutathione (GSH) levels. Mechanistic investigations were performed using EZH2 knockdown, ACSL1 expression analysis, and H3K27me3 modification assays. The therapeutic potential of EZH2 inhibition was further assessed in xenograft models.

Results

EZH2 was significantly overexpressed in HCC tissues and correlated with poor patient survival. Resistant cell models demonstrated EZH2-mediated suppression of ferroptosis through ACSL1 downregulation via H3K27me3, evidenced by altered ROS, MDA and GSH levels. Genetic inhibition of EZH2 restored lenvatinib sensitivity by upregulating ACSL1 and promoting ferroptosis. In vivo studies confirmed that EZH2 targeting enhanced lenvatinib’s antitumor effects in resistant HCC models.

Conclusions

Our findings establish EZH2 as a critical regulator of lenvatinib resistance in HCC through ACSL1-mediated ferroptosis suppression. The EZH2-H3K27me3-ACSL1 axis represents a promising therapeutic target for overcoming drug resistance, offering new strategies to improve HCC treatment outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-15086-9.

Background

Liver cancer stands as one of the most prevalent malignancies and a leading cause of cancer-related deaths worldwide, highlighting the pressing need to develop innovative and effective treatment approaches [1]. According to Global Cancer Statistics 2022, liver cancer ranks sixth in global incidence (4.3%) but third in mortality (7.8%), with approximately 860,000 new cases and over 760,000 deaths annually [2]. Despite advances in diagnostic and therapeutic strategies, the prognosis for HCC remains poor, with a 5-year survival rate of less than 20% due to late-stage diagnosis, high recurrence rates, and limited treatment efficacy [3]. In recent years, molecularly targeted therapies, particularly tyrosine kinase inhibitors (TKIs), have revolutionized the systemic treatment of advanced HCC [4]. Lenvatinib, a multi-targeted TKI that inhibits vascular endothelial growth factor receptors (VEGFRs), fibroblast growth factor receptors (FGFRs), and platelet-derived growth factor receptor α (PDGFRα), has emerged as a first-line therapy for unresectable HCC, demonstrating non-inferiority to sorafenib in overall survival with improved progression-free survival and objective response rates [5]. However, the clinical benefits of lenvatinib are frequently hampered by the development of drug resistance, which ultimately leads to disease progression in the majority of patients.

Recent studies have identified ferroptosis, a unique iron-dependent form of programmed cell death characterized by excessive lipid peroxidation, as a critical metabolic vulnerability in HCC [6]. Unlike apoptosis or necrosis, ferroptosis is specifically triggered by the accumulation of ROS in cellular membranes, making it particularly relevant to cancer cells with high metabolic demands [7, 8]. Growing evidence suggests that lenvatinib may exert their anti-tumor effects in part through ferroptosis induction [9, 10]. However, the emergence of lenvatinib resistance poses a significant clinical challenge, and the mechanisms by which HCC cells evade ferroptosis under therapeutic pressure remain poorly understood. Recent work has implicated various adaptive responses in ferroptosis resistance, including metabolic rewiring toward fatty acid desaturation, upregulation of ferroptosis suppressor proteins, and epigenetic reprogramming [11, 12]. Intriguingly, several epigenetic modifiers have been shown to regulate ferroptosis sensitivity in other cancer types [13–15], but whether similar mechanisms operate in HCC to confer lenvatinib resistance has not been investigated. This knowledge gap critically limits our ability to develop effective strategies to overcome treatment resistance.

As the core catalytic subunit of Polycomb repressive complex 2 (PRC2), EZH2 mediates gene silencing through catalyzing histone H3K27me3 modification and plays a pivotal role in tumor development [16, 17]. Studies have demonstrated that EZH2 is frequently overexpressed or harbors gain-of-function mutations in various malignancies (e.g., lymphoma, breast cancer, and hepatocellular carcinoma) [18–20]. Its overexpression promotes tumor cell proliferation, invasion, and metastasis by silencing tumor suppressor genes (e.g., CDKN2A, DAB2IP) [21, 22]. Additionally, EZH2 can activate oncogenic signaling pathways (e.g., Wnt, mTORC and NF-κB) in a PRC2-independent manner, further driving tumor progression [23, 24]. In recent years, the role of EZH2 in drug resistance has been increasingly uncovered. In multiple cancer models, EZH2 upregulation is closely associated with resistance to chemotherapy and targeted therapy [25]. In glioblastoma, EZH2 promotes therapeutic resistance by upregulating the expression of multidrug resistance genes such as ATP-binding cassette subfamily G member 2 (ABCG2) and other multidrug resistance proteins (MRPs), thereby facilitating drug efflux and reducing intracellular drug accumulation [26]. In small cell lung cancer, EZH2 contributes to resistance by mediating epigenetic silencing of the SLFN11 gene through H3K27 hypermethylation. SLFN11 is a key modulator of DNA damage response, and its repression impairs chemosensitivity to DNA-damaging agents [27]. Furthermore, in ovarian cancer, tumor-associated endothelial cells exhibit EZH2 upregulation in response to paracrine VEGF signaling. EZH2 silences anti-angiogenic genes such as vasohibin-1 via histone methylation, thereby promoting neovascularization and facilitating an acquired resistance phenotype. Notably, targeting EZH2 in tumor vasculature restores vasohibin-1 expression and impairs angiogenesis, highlighting its role in overcoming therapy resistance [28]. Preclinical studies have shown that combining EZH2 inhibitors with conventional chemotherapy or targeted agents significantly enhances antitumor efficacy [29, 30]. Notably, tazemetostat has been approved by the FDA for treating epithelioid sarcoma and follicular lymphoma [31, 32]. However, the therapeutic potential and resistance mechanisms of EZH2 inhibitors in solid tumors (particularly liver cancer) remain to be further explored. Importantly, although the role of EZH2 in drug resistance has been extensively studied, how it regulates ferroptosis to influence lenvatinib resistance in HCC cells remains unclear.

This study elucidates the molecular mechanism by which EZH2 confers resistance to lenvatinib-induced ferroptosis in HCC. EZH2 was found to be overexpressed in HCC tissues compared to healthy liver tissues, as evidenced by analysis of the The Cancer Genome Atlas (TCGA) dataset and RT-qPCR in patient samples. Higher EZH2 levels correlate with poor overall survival and disease-free survival, highlighting its role as a prognostic indicator. Furthermore, our findings suggest that elevated EZH2 is associated with resistance to lenvatinib, a common chemotherapeutic agent for HCC, as shown through analysis of the establishment of lenvatinib-resistant HCC cell lines. Mechanistically, EZH2 appears to inhibit ferroptosis in HCC cells by modulating acyl-CoA synthetase long-chain family member 1 (ACSL1) levels. Knockdown of EZH2 sensitizes resistant cell lines to lenvatinib and induces ferroptosis by increasing ACSL1 expression. Additionally, EZH2 regulation of ACSL1 is mediated through the H3K27me3. In vivo studies further validate that targeting EZH2 enhances the efficacy of lenvatinib in resistant HCC models by promoting ferroptosis. Collectively, these findings underscore the potential of EZH2 as a therapeutic target to overcome drug resistance in HCC, providing insights into its role in modulating treatment responses and patient prognosis.

Materials and methods

Cell culture

The human HCC cell lines Huh7 and Hep3B were obtained from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences (Shanghai, China). All cell lines were authenticated using short tandem repeat (STR) profiling to ensure genetic integrity. Cells were maintained under standard culture conditions at 37 °C in a humidified 5% CO2 atmosphere. The culture medium was supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin solution (Solarbio). Specifically: Hep3B cells were cultured in MEM medium (Biological Industries) Huh7 cells were grown in high-glucose DMEM medium Prior to experiments, all cell lines underwent routine PCR-based screening to confirm the absence of mycoplasma contamination.

Clinical specimens

A total of 12 HCC patients’ tissues, including paired samples of tumor tissues and adjacent normal tissues, were acquired from Affilllated Cancer Hospital and Institute of Guangzhou Medical University. Following the Declaration of Helsinki, this study acquired the approval by the Medical Ethics Committee of AffiliatedCancer Hospital and Institute of Guangzhou Medical University. Informed consent was acquired from all the patients.

Immunohistochemical staining

Analysis The immunohistochemical (IHC) staining procedure was performed following established protocols. Briefly, 4-µm-thick sections were prepared from formalin-fixed, paraffin-embedded tissue specimens. For immunostaining, tissue sections were incubated with a primary antibody targeting EZH2 (Cat# 21800-1-AP, Proteintech; working dilution 1:1000). Negative controls were established using rabbit IgG instead of primary antibody, while known IHC-positive sections served as positive controls. The immunoreactivity was evaluated through a semiquantitative scoring system: Staining extent was graded as: 0 (no detectable staining), 1 (≤ 10% positive cells), 2 (11%−50%), 3 (51%−75%), or 4 (> 75%). Staining intensity was classified as: 1+ (weak), 2+ (moderate), or 3+ (strong). The final immunoreactivity score (S) was calculated by multiplying the percentage of positive cells (P) by the staining intensity (I) (S = P × I), considering nuclear, cytoplasmic, or membrane expression patterns. Digital images were captured using a Leica DMi8 imaging system.

Real-time quantitative PCR

Total RNA was isolated from cells using TRIzol reagent (Thermo Fisher Scientific). Following the manufacturer’s protocol, 2 µg of RNA was reverse-transcribed into complementary DNA (cDNA) using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, USA). For gene expression analysis, equal amounts of cDNA were amplified and quantified using SYBR Green PCR Master Mix (Thermo Fisher Scientific) on an Applied Biosystems 7500 Real-Time PCR System. Semi-quantitative PCR and quantitative real-time PCR (RT-qPCR) were performed with the primer sets shown in Table. S1.

Western blot

Total protein was extracted using RIPA lysis buffer supplemented with a protease inhibitor cocktail. Protein concentrations were quantified and normalized using a BCA assay kit (Thermo Fisher Scientific, USA). Equal amounts of protein samples from each group were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore). The membranes were incubated overnight at 4 °C with the primary antibodies shown in Table. S2. GAPDH was selected as a loading control due to its stable expression under experimental conditions, as confirmed in preliminary studies. After primary antibody incubation, the membranes were probed with appropriate secondary antibodies for 1 h at 25 °C. Protein bands were visualized using an enhanced chemiluminescence (ECL) detection kit (Beyotime, China) following the manufacturer’s protocol. The intensity of each band was quantified by densitometry and normalized to β-actin as an internal control.

Establishment of lenvatinib resistance

The cells were cultured in complete medium supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomycin (Solarbio) under standard conditions (37 °C, 5% CO2). To evaluate drug sensitivity, we first determined the half-maximal inhibitory concentration (IC50) of lenvatinib in Huh7 and Hep3B cells by seeding them in 96-well plates and exposing them to increasing drug concentrations for 72 h, followed by cell viability assessment using CCK-8 assay. For developing lenvatinib-resistant cell lines, parental cells were continuously cultured for 3–6 months in medium containing progressively increasing concentrations of lenvatinib (Selleck Chemicals, starting at 1 µM), with medium replacement every 48 h and regular monitoring of resistance development. The established resistant sublines, designated as Huh7-LR and Hep3B-LR, were subsequently maintained in medium containing 5 µM lenvatinib to preserve their drug-resistant phenotype.

Establishment of EZH2 stable knockdown cell lines

To establish stable EZH2-knockdown cell lines in lenvatinib-resistant HCC models, Huh7-LR and Hep3B-LR cells were transduced with lentivirus carrying either of two validated shRNAs targeting human EZH2 (shEZH2-#1: TATGATGGTTAACGGTGATCA; shEZH2-#2: GCTAGGTTAATTGGGACCAAA), two validated shRNAs targeting human ACSL1 (shACSL1-#1: GCCCAGATGATACTTTGATAT; shACSL1-#2: CGCAGATAGATGACCTCTATT) or non-targeting control shRNA (shCtrl) using the pLKO.1-puro system. Briefly, HEK293T cells were co-transfected with shRNA plasmids along with psPAX2 and pMD2.G packaging plasmids using Lipofectamine 3000. Viral supernatants were collected at 48 and 72 h post-transfection, filtered (0.45 μm), and used to infect Huh7-LR and Hep3B-LR cells in the presence of 8 µg/mL polybrene. After 24 h incubation, stable pools were selected with 3 µg/mL puromycin for 7 days. Successful EZH2 knockdown was confirmed by qRT-PCR and Western blot.

Cell viability assay

Cell viability was quantitatively assessed using the CCK-8 assay kit (C0037, Beyotime Biotechnology) following the manufacturer’s instructions. In brief, cells were seeded in 96-well plates at an optimal density (typically 3–5 × 10³ cells/well) and cultured overnight to allow attachment. After experimental treatments, 10 µL of CCK-8 solution was added to each well containing 100 µL culture medium, followed by incubation at 37 °C for 1–4 h. The optical density was measured at 450 nm using a microplate reader (BioTek Instruments), with a reference wavelength of 650 nm to correct for nonspecific background. For each experimental condition, at least six replicate wells were included, and untreated cells served as negative controls. The percentage of cell viability was calculated as: (ODtreatment - ODblank)/(ODcontrol - ODblank) × 100%.

Malondialdehyde assay

The malondialdehyde (MDA) content, an indicator of lipid peroxidation, was measured using the Beyotime MDA Assay Kit (S0131) following the manufacturer’s protocol with modifications. Briefly, 1 × 10⁶ cells were seeded in 10-cm culture dishes 24 h prior to treatment. After treatment, cells were lysed with 200 µL of ice-cold lysis buffer and scraped using a cell scraper. The lysates were centrifuged at 12,000×g for 10 min at 4 °C. The supernatant was divided into two 100 µL aliquots for parallel measurements, while the remainder was used for protein concentration determination by BCA assay. Each aliquot was mixed with 200 µL of 0.37% thiobarbituric acid (TBA) working solution and heated at 100 °C for 15 min in a metal bath to develop the color reaction. After cooling to room temperature on ice, samples were centrifuged at 1,000×g for 10 min to remove precipitates. The absorbance of the supernatant was measured at 532 nm using a microplate reader. MDA concentrations were calculated based on a standard curve and normalized to protein content (nmol/mg protein). All experiments were performed in triplicate and repeated in three independent biological replicates to ensure data reliability.

Glutathione (GSH) assay

Total glutathione (GSH) levels were determined using a commercial GSH assay kit (#BC1175, Solarbio) following the manufacturer’s instructions. Briefly, cells or tissues were homogenized in cold assay buffer and centrifuged at 10,000 × g for 10 min at 4 °C to obtain clear supernatants. The reaction system containing 5,5’-dithiobis-(2-nitrobenzoic acid) (DTNB) and glutathione reductase was prepared, and samples were incubated with the working reagent at 25 °C for 5 min. The enzymatic recycling reaction was initiated by adding NADPH, and the absorbance at 412 nm was monitored every 30 s for 5 min using a microplate reader (Molecular Devices). A standard curve was generated with known GSH concentrations (0–50 µM) for quantification. All samples were analyzed in triplicate, and GSH levels were normalized to total protein content determined by BCA protein assay.

ROS levels

Intracellular ROS levels were measured using the Reactive Oxygen Species Assay Kit (S0033M, Beyotime Biotechnology) according to the manufacturer’s protocol. Briefly, cells were washed twice with PBS and incubated with 10 µM 2’,7’-dichlorodihydrofluorescein diacetate (DCFH-DA) in serum-free medium at 37 °C for 30 min in the dark. After incubation, cells were washed three times with PBS to remove excess probe. Fluorescence intensity was immediately measured using a microplate reader (excitation 488 nm, emission 525 nm) or observed under a fluorescence microscope (Nikon Eclipse Ti). For quantitative analysis, fluorescence values were normalized to cell number determined by parallel CCK-8 assay. All experiments were performed in triplicate under light-protected conditions.

flow cytometry

Cells were inoculated into a 6-well plate. After 72-h transfection, the cells were gathered and washed with cold PBS. Next, they were resuspended in staining buffer, followed by detection utilizing Annexin VFITC/PI Apoptosis Detection Kit (Yeasen, #40302ES60) away from light for 30 min. The results were analyzed by FlowJo (V10.6).

Chromatin Immunoprecipitation (ChIP) and ChIP‒qPCR assays

The ChIP assay was performed using the EZ-Magna ChIP Kit (Millipore) according to the manufacturer’s protocol with modifications. Cells were cross-linked with 1% formaldehyde for 10 min at room temperature, followed by quenching with 125 mM glycine. Chromatin was sheared to 200–500 bp fragments using a Bioruptor sonicator (Diagenode) with 15 cycles of 30 s ON/30 sec OFF at high intensity. After pre-clearing with Protein A/G beads, 100 µg of chromatin was immunoprecipitated overnight at 4 °C with 5 µg of specific antibodies: anti-EZH2 (ab191250, Abcam), anti-H3K27me3 (ab6002, Abcam), or normal IgG (negative control). Immune complexes were captured with Protein A/G magnetic beads for 2 h at 4 °C, followed by extensive washing. After reverse cross-linking and proteinase K treatment, precipitated DNA was purified using spin columns. Enrichment of target genomic regions was analyzed by quantitative PCR (qPCR) using specific primers, with results normalized to input DNA and expressed as percentage of input. The sequences of the primers used in this study are listed in Supplementary Table S3.

Animal experiments

All the animal experiments were performed following the policies of the Animal Ethics Committee of AffiliatedCancer Hospital and Institute of Guangzhou Medical University and conducted in strict compliance with ethical guidelines. For tumor xenograft establishment, Huh7LR-shCtrl or Huh7LR-shEZH2 cells (5 × 10^6 cells in 100 µL PBS per mouse) were injected subcutaneously into the right flank of 5-week-old BALB/c nude mice. The tumor volume was measured every three days and calculated as volume (mm3)=(length × width²)/2. After the tumor volume reached ~ 100 mm3, the mice were randomly lenvatinib (30 mg/kg, once daily, oral gavage). After the experiment, the tumor-bearing mice were euthanized by intraperitoneal injection of sodium pentobarbital (50 mg/kg), the tumor tissues were collected and weight was measured.

Bioinformatics analyses of TCGA dataset

Gene expression profiles and clinical information of patients were obtained from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). RNA-seq data (transcripts Per Million, TPM) were downloaded and log2-transformed for downstream analyses. For differential gene expression analysis, samples were stratified based on clinical or molecular characteristics of interest. Expression values were compared using the limma package in R, and results were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR). For correlation analyses, pairwise associations between genes were calculated using Spearman’s correlation coefficient with the cor.test function in R. Correlation strength and significance were presented as correlation coefficients (ρ) and adjusted P-values. Data preprocessing, normalization, and visualization were conducted in R software (version 4.x) using the packages limma, edgeR, ggplot2, and corrplot. A two-sided P-value < 0.05 was considered statistically significant unless otherwise specified.

Lipid ROS measurement by C11-BODIPY 581/591 fluorescence

Lipid peroxidation was assessed using the fluorescent probe C11-BODIPY 581/591 (Thermo Fisher Scientific, Cat. #D3861). Briefly, cells were incubated with 2 µM C11-BODIPY for 30 min at 37 °C in the dark, washed with PBS, and analyzed immediately by flow cytometry (BD FACSCanto II). The green (oxidized, 510 nm) to red (reduced, 590 nm) fluorescence intensity ratio was calculated to quantify lipid ROS levels.

FerroOrange staining

A FerroOrange probe (F374, Dojindo, Shanghai, China) was employed to detect intracellular Fe²⁺. Following the indicated treatments, HCC cells were washed with PBS and incubated with FerroOrange working solution (1 µmol/L) for 30 min. Subsequently, the stained cells were visualized using confocal laser scanning microscopy.

Statistical analysis

All statistical analyses were conducted utilizing SPSS version 17.0 (SPSS, Inc., Chicago, IL, USA). For pairwise comparisons, the Student’s t-test (or paired t-test) was employed, while for comparisons involving more than two groups, a one-way ANOVA was utilized, supplemented by posthoc analyses to discern intergroup differences. A P-value below 0.05 was deemed indicative of statistical significance.

Results

Elevated EZH2 expression correlates with poor prognosis in HCC

While EZH2 overexpression has been widely observed in HCC, its exact contribution to disease pathogenesis requires further elucidation. Our comprehensive investigation revealed that EZH2 expression was significantly upregulated in HCC tissues compared with adjacent normal liver tissues in the TCGA-LIHC dataset (Fig. 1A). This finding was consistently validated at both transcriptional and translational levels in our validation cohort of 12 matched HCC samples, demonstrating significantly higher EZH2 mRNA levels and protein expression in 10 out of 12 HCC tissues compared to their normal counterparts (Fig. 1B-D). Immunohistochemical analysis further confirmed the pronounced overexpression of EZH2 protein in HCC specimens, showing strong nuclear staining in tumor cells but minimal expression in adjacent normal hepatocytes (Fig. 1E). Most clinically relevant, elevated EZH2 expression was strongly associated with poor patient outcomes in the TCGA cohort (n = 360), demonstrating significantly reduced overall survival (HR = 2.1, 95% CI: 1.5–2.9, p = 7.8e-05) and disease-free survival (HR = 1.8, 95% CI: 1.3–2.5, p = 0.00012) in high expressers (Fig. 1F-G). These results provide compelling evidence that EZH2 overexpression plays a critical role in HCC progression and serves as a robust predictor of unfavorable clinical outcomes.

Fig. 1.

Elevated EZH2 expression correlates with poor prognosis in hepatocellular carcinoma (HCC). A Expression of EZH2 in the TCGA-HCC dataset, comparing HCC tissues with peri-tumoral liver tissues. B RT-qPCR assays demonstrating mRNA levels of EZH2 in 12 HCC patient samples and their corresponding adjacent normal tissues. C, D Western blot demonstrating protein levels of EZH2 in 12 HCC patient samples and their corresponding adjacent normal tissues. E Representative immunohistochemistry (IHC) images showing EZH2 expression in HCC tissues versus paired adjacent normal tissues. Statistical analysis comparing IHC scores of 12 HCC tissues with their paired normal tissues. F, G Kaplan-Meier overall survival curve and disease-free survival curve of HCC patients correlated with the expression of EZH2. All data are presented as the means ± SDs, n = 5 per group, * (p < 0.05), ** (p < 0.01), ns = no significance

EZH2 expression is increased in lenvatinib-resistant HCC cells and contributes to resistance development

To investigate the role of EZH2 in lenvatinib resistance, we established two resistant HCC cell line models (Huh7-LR and Hep3B-LR) through chronic lenvatinib exposure. These resistant lines demonstrated significantly elevated IC50 values (14.82 µM and 11.13 µM, respectively) compared to their parental counterparts (Huh7: 4.512 µM; Hep3B: 1.999 µM; Fig. 2A-B). Notably, western blot analysis revealed a consistent upregulation of EZH2 protein expression in both resistant cell lines (Fig. 2C-D). To functionally characterize its contribution to resistance, we generated stable EZH2 knockdown cell lines in Huh7-LR and Hep3B-LR using shRNA (Fig. 2E-F). Importantly, EZH2 depletion significantly restored lenvatinib sensitivity in both resistant lines compared to scramble controls (Fig. 2G-H). These results provide compelling evidence that EZH2 overexpression drives lenvatinib resistance in HCC, while its targeted suppression may represent a viable strategy to overcome therapeutic resistance.

Fig. 2.

EZH2 expression is increased in lenvatinib-resistant HCC cells and contributes to resistance development. A Schematic representation of the establishment of lenvatinib-resistant (LR) cells. B Cell viability assessed by CCK-8 assay following exposure to varying concentrations of lenvatinib for 48 h. Proliferation of both Huh7-LR and Hep3B-LR cells was greater than that of their parental cells at different lenvatinib concentrations. C, D mRNA and protein levels of EZH2 in Huh7-LR and Hep3B-LR cells, as well as their parental cells, detected by RT-PCR and western blot. E, F mRNA and protein levels of EZH2 in Huh7-LR and Hep3B-LR cells transfected with shEZH2#1, shEZH2#2, or shCtrl,, detected by RT-PCR and western blot. G, H Proliferation of both Huh7-LR and Hep3B-LR cells transfected with shEZH2#1 compared to those transfected with shCtrl at varying concentrations of lenvatinib. All data are presented as the means ± SDs, n = 5 per group, * (p < 0.05), ** (p < 0.01), ns = no significance

EZH2 inhibits ferroptosis in HCC cells

To investigate the mechanism of lenvatinib-induced cell death, we first analyzed cell death patterns using flow cytometry. Treatment with lenvatinib resulted in a modest increase in cell death in Huh7-LR and Hep3B-LR cells, with no significant difference in early apoptotic populations, suggesting the occurrence of non-apoptotic cell death (Fig. 3A). To identify the specific death modality, Huh7-LR and Hep3B-LR cells were pretreated with inhibitors targeting major cell death pathways: Nec-1 (necroptosis inhibitor, 10 µM), Z-VAD-FMK (pan-caspase inhibitor, 10 µM), CQ (autophagy inhibitor, 20 µM), and Fer-1 (ferroptosis inhibitor, 10 µM). Notably, neither Nec-1, Z-VAD-FMK, nor CQ attenuated lenvatinib-induced cytotoxicity, whereas Fer-1 treatment significantly rescued cell viability in both Huh7-LR and Hep3B-LR cells (Fig. 3B). These findings indicate that lenvatinib primarily triggers ferroptosis. Furthermore, we evaluated the sensitivity of resistant cell lines (Huh7-LR and Hep3B-LR) to the ferroptosis inducer RSL3. Consistent with their lenvatinib-resistant phenotype (Fig. 2G), both cell lines exhibited signifcant resistance to RSL3 (Fig. 3C), reinforcing the link between ferroptosis and drug resistance.

Fig. 3.

EZH2 inhibits ferroptosis in HCC cells. A Analysis of cell death and apoptosis in Huh7-LR and Hep3B-LR cells after Lenvatinib (Lenv) treatment via Annexin V-FITC/PI staining. B Inhibitory ratio that was measured in Huh7-LR and Hep3B-LR using CCK-8 assay 24 h after treatment with Lenvatinib with or without Z-VAD-FMK, CQ, Nec-1 or Fer-1. C Proliferation of both Huh7-LR and Hep3B-LR cells transfected with shCtrl compared to those transfected with shEZH2#1 at varying concentrations of RSL3. D-F Relative levels of reactive oxygen species (ROS) (D), glutathione (GSH) (E), and malondialdehyde (MDA) (F) in Huh7-LR and Hep3B-LR cells transfected with shEZH2 #1 compared to shCtrl treated with Lenvatinib, RSL3 and Fer-1. (G-H) Intracellular lipid peroxidation and iron accumulation in Huh7-LR and Hep3B-LR cells transfected with shEZH2 #1 compared to shCtrl treated with Lenvatinib, RSL3 and Fer-1 were determined by C11-BODIPY and FerroOrange staining. All data are presented as the means ± SDs, n = 5 per group, * (p < 0.05), ** (p < 0.01), ns = no significance

To further explore how EZH2 regulates ferroptosis in lenvatinib-resistant HCC, we first assessed key indicators of oxidative stress, including ROS, GSH and MDA levels, following EZH2 knockdown. Mechanistically, EZH2 silencing exacerbated lenvatinib- and RSL3-induced ferroptotic events, as evidenced by elevated ROS and MDA levels, along with diminished GSH content—all of which were reversed by Fer-1 (Fig. 3D-F). In addition, EZH2 knockdown promoted lenvatinib- and RSL3-induced increases in intracellular lipid peroxidation and iron accumulation, as evidenced by C11-BODIPY and FerroOrange staining (Fig. 3G-H). Collectively, our data demonstrate that EZH2 ablation sensitizes HCC cells to lenvatinib by potentiating ferroptosis, highlighting its role as a critical regulator of cell death in drug-resistant contexts.

EZH2 suppression induced ferroptosis via enhancing ACSL1 expression through H3K27me3

To elucidate the molecular mechanisms underlying EZH2 downregulation-induced ferroptosis, we systematically analyzed ferroptosis-related proteins in lenvatinib-resistant Huh7-LR and Hep3B-LR cells following stable EZH2 knockdown. Quantitative RT-PCR analysis demonstrated a significant increase in ACSL1 mRNA levels, while other key ferroptosis regulators (GPX4, SLC3A2, ACSL4, AIFM2, and DHODH) [6] remained unaffected (Fig. 4A). Intriguingly, western blot analysis revealed a concomitant increase in ACSL1 protein expression upon EZH2 depletion (Fig. 4B, C). Bioinformatics analysis of TCGA dataset further substantiated a strong negative correlation between EZH2 and ACSL1 expression in HCC (Fig. 4D, E). Consistent with these observations, integrated analysis of TCGA and GEO datasets revealed consistent downregulation of ACSL1 in both HCC tissues and lenvatinib-resistant HCC cells (Figure S2A, B). These findings collectively suggest that EZH2 downregulation preferentially modulates ACSL1 expression to potentiate ferroptosis in HCC cells. ACSL1 silencing completely abrogated the chemosensitizing effects of EZH2 knockdown, restoring lenvatinib resistance in HCC cells (Fig. 4F). Functional validation experiments demonstrated that ACSL1 knockdown effectively attenuated the shEZH2-induced elevation of reactive oxygen species (ROS) (Fig. 4G) and normalized the altered glutathione (GSH) and malondialdehyde (MDA) levels (Fig. 4H, I). In addition, the shEZH2-induced increases in intracellular lipid peroxidation and iron accumulation were strongly blocked by ACSL1 knockdown, as evidenced by C11-BODIPY and FerroOrange staining (Fig. 4J-K). These data establish ACSL1 as the critical downstream effector through which EZH2 deficiency promotes ferroptosis.

Fig. 4.

EZH2 suppression induced ferroptosis via enhancing ACSL1 expression through H3K27me3. A, B RT-qPCR assays demonstrated mRNA levels of indicated genes in in Huh7-LR (left panel) and Hep3B-LR (right panel) cells transfected with shEZH2 compared to shCtrl. C, D Expression levels of EZH2 in Huh7-LR (C) and Hep3B-LR (D) cells transfected with shEZH2 compared to shCtrl, detected by western blot. (E)Pearson correlation analysis of TCGA dataset revealed a significant negative correlation between EZH2 and ACSL1 expression. (F)Proliferation of Huh7-LR and Hep3B-LR cells transfected with shEZH2 or shACSL1. G-I Relative levels of ROS (G), GSH (H), and MDA (I) in Huh7-LR and Hep3B-LR cells transfected with shEZH2 or shACSL1. J-K Intracellular lipid peroxidation and iron accumulation in Huh7-LR and Hep3B-LR cells transfected with shEZH2 or shACSL1 were determined by C11-BODIPY and FerroOrange staining. L The protein levels of EZH2 and H3K27me3 in Huh7-LR cells transfected with shEZH2 or shCtrl were determined by western blot. M The enrichments of EZH2 and H3K27me3 on promoters of ACSL1 were determined by ChIP assays. N The altered enrichments of EZH2 and H3K27me3 on promoters of ACSL1 after EZH2 downregulatiton were determined by ChIP-qPCR assay. Normal IgG served as a negative control. All data are presented as the means ± SDs, n = 5 per group, * (p < 0.05), ** (p < 0.01), ns = no significance

Given the well-established role of EZH2 as a histone methyltransferase mediating gene silencing via trimethylation of histone H3 at lysine 27 (H3K27me3) [8], we investigated whether this epigenetic mechanism governs ACSL1 regulation. Western blot analysis confirmed marked reduction of H3K27me3 levels in EZH2-deficient Huh7^LR (Fig. 4L). To directly assess whether EZH2 regulates ACSL1 expression through H3K27me3 modification, we performed chromatin immunoprecipitation (ChIP) assays to confirm H3K27me3 enrichment at the ACSL1 promoter region. The ChIP results demonstrated that EZH2 and H3K27me3 were enriched in the promoter region of ACSL1 in Huh-7-LR cells (Fig. 4M). Consistently, ChIP-qPCR further demonstrated significant downregulation of both EZH2 and H3K27me3 at the ACSL1 promoter region following EZH2 knockdown (Fig. 4N).

To explore whether ACSL1 repression is directly dependent on the methyltransferase activity of EZH2, we performed ChIP-qPCR and western blotting using the selective EZH2 inhibitor Tazemetostat, which specifically blocks the methyltransferase activity of EZH2. Treatment with Tazemetostat significantly reduced H3K27me3 enrichment at the ACSL1 promoter and concomitantly restored ACSL1 expression (Figure S3A-B), indicating that the repression of ACSL1 is indeed dependent on EZH2’s methyltransferase activity. Furthermore, knockdown of core PRC2 component SUZ12 produced similar effects (Figure S3C-D), suggesting that the regulation occurs through the canonical EZH2-PRC2 chromatin-modifying complex.

Taken together, these results demonstrate that EZH2 epigenetically represses ACSL1 expression through H3K27 trimethylation in HCC cells, thereby establishing a novel mechanistic link between epigenetic regulation and ferroptosis susceptibility in HCC.

EZH2 reduction enhances lenvatinib efficacy in treating resistant HCC in vivo

To evaluate the in vivo therapeutic implications of EZH2 downregulation in promoting ferroptosis via ACSL1 induction, we established a xenograft model by subcutaneously injecting Huh7-LR cells transfected with different constructs into BALB/c nude mice. The mice were subsequently treated with lenvatinib (30 mg/kg every day for three weeks) (Fig. 5A). Tumor growth analysis revealed that EZH2 knockdown significantly suppressed tumor progression; however, this anti-tumor effect was effectively abrogated upon concurrent ACSL1 silencing (Fig. 5B-D). Notably, immunohistochemical analyses of xenograft tissues demonstrated that the shEZH2 group exhibited markedly reduced levels of EZH2 and its epigenetic marker H3K27me3, concomitant with a significant upregulation of ACSL1 compared to the control group. In contrast, the shEZH2 + shACSL1 group showed restored ACSL1 expression to baseline levels without significant alterations in EZH2 or H3K27me3 (Fig. 5E). Collectively, these in vivo findings provide compelling evidence that targeting the EZH2-H3K27me3-ACSL1 axis represents a promising therapeutic strategy to overcome lenvatinib resistance in HCC by modulating ferroptosis susceptibility.

Fig. 5.

EZH2 reduction enhances lenvatinib efficacy in treating resistant HCC in vivo. A Treatment schedule illustrating the timing of tumor inoculation and administration of treatments in BALB/c nude mice (n = 6 per group). B, C Representative images of tumors (B) harvested on day 27 posttreatment, and corresponding average tumor weights (C) in each group. D Tumor growth curves representing the progression of Huh7-LR cell-induced tumors in mice treated with lenvatinib. E Expression levels of EZH2, H3K27me3 and ACSL1 in tumors. All data are presented as the means ± SDs, n = 6 per group, * (p < 0.05), ** (p < 0.01), ns = no significance

Discussion

In this study, we identified EZH2 as a critical epigenetic regulator that promotes lenvatinib resistance in HCC by suppressing ferroptosis. Clinically, elevated EZH2 expression was significantly associated with poor overall survival and disease-free survival in HCC patients, suggesting its role as a potential prognostic marker. Mechanistically, we demonstrated that EZH2 was markedly upregulated in lenvatinib-resistant HCC cells and that its inhibition restored sensitivity to lenvatinib treatment. Further investigations revealed that EZH2 repressed ferroptosis by transcriptionally silencing the ferroptosis-associated enzyme ACSL1 through H3K27 trimethylation. Importantly, both in vitro and in vivo assays confirmed that suppression of EZH2 not only enhanced ACSL1 expression and ferroptotic cell death but also significantly improved the therapeutic efficacy of lenvatinib in resistant tumor models. These findings collectively highlight a novel EZH2–ACSL1–ferroptosis regulatory axis and provide mechanistic insight into the epigenetic basis of lenvatinib resistance in HCC.

EZH2, the catalytic subunit of the PRC2, is a well-characterized histone methyltransferase responsible for trimethylation of histone H3 at lysine 27 (H3K27me3), a repressive chromatin mark that silences gene transcription [33]. Increasing evidence has implicated EZH2 in oncogenesis and drug resistance across various cancer types, including breast, prostate, and gastric cancers [34–36]. In the context of HCC, EZH2 has been reported to drive tumor progression, promoting stemness, epithelial-mesenchymal transition, and immune evasion [37]. However, its role in therapeutic resistance, particularly to tyrosine kinase inhibitors such as lenvatinib, has remained underexplored. Our study demonstrates for the first time that EZH2 is markedly upregulated in lenvatinib-resistant HCC cells compared with their parental counterparts. Functional inhibition or knockdown of EZH2 not only restored drug sensitivity but also suppressed resistant cell proliferation, strongly implicating EZH2 as a key driver of lenvatinib resistance. These findings extend the oncogenic repertoire of EZH2 by highlighting its previously unrecognized role in mediating acquired resistance to targeted therapy in HCC, and position EZH2 as a potential therapeutic target to overcome treatment failure in this setting.

Ferroptosis is a distinct form of regulated cell death characterized by iron-dependent lipid peroxidation, and has recently emerged as a critical vulnerability in cancer cells, particularly under therapeutic stress. Unlike apoptosis or necroptosis, ferroptosis is tightly linked to metabolic rewiring, redox imbalance, and lipid homeostasis, making it a promising target in tumors with high oxidative pressure [38]. Accumulating evidence suggests that evasion of ferroptosis may contribute to therapeutic resistance [11, 12]; however, the underlying mechanisms remain incompletely understood. In this study, we reveal that EZH2 plays a pivotal role in suppressing ferroptosis in hepatocellular carcinoma, thereby conferring resistance to lenvatinib. Notably, genetic inhibition of EZH2 reactivated ferroptosis in resistant cells, as evidenced by increased lipid ROS accumulation and enhanced cell death upon ferroptotic induction. This effect was abrogated by ferroptosis inhibitors, confirming the specificity of the pathway involved. Our findings are consistent with recent reports in other malignancies where epigenetic regulators modulate ferroptosis-related genes, suggesting that ferroptosis suppression may be a generalizable mechanism of drug resistance. More importantly, our data position ferroptosis not merely as a passive bystander but as an actively suppressed death program during lenvatinib resistance, providing a novel therapeutic entry point. These results underscore the therapeutic potential of targeting epigenetic blocks to ferroptosis in overcoming drug resistance in HCC.

Our mechanistic investigations identified ACSL1 as a critical downstream effector of EZH2 in regulating ferroptosis and lenvatinib resistance. ACSL1 is a key enzyme involved in the activation of long-chain fatty acids by catalyzing their conversion to acyl-CoA esters, which are essential substrates for lipid metabolism and peroxidation processes [39]. Recent studies have suggested that ACSL1 contributes to ferroptosis by enriching polyunsaturated fatty acids (PUFAs) in membrane phospholipids, thereby sensitizing cells to lipid peroxidation [40]. In our study, EZH2 suppression markedly increased ACSL1 expression at both the mRNA and protein levels, and this upregulation was shown to be directly mediated through loss of H3K27me3 deposition at the ACSL1 promoter region. These results implicate ACSL1 as an epigenetically silenced target of EZH2. Functionally, rescue experiments demonstrated that ACSL1 knockdown significantly reversed the ferroptotic phenotype induced by EZH2 inhibition, confirming its essential role in mediating this process. Importantly, these findings establish a mechanistic link between epigenetic repression and metabolic vulnerability in HCC, whereby EZH2 suppresses ACSL1 to inhibit ferroptosis and promote drug resistance. While prior work has primarily focused on other ferroptosis regulators such as GPX4 and SLC7A11, our study identifies ACSL1 as a novel and functional mediator in this context. This expands the repertoire of ferroptosis-related targets and highlights ACSL1 as a potential biomarker or therapeutic sensitizer in lenvatinib-resistant HCC. Targeting the EZH2–ACSL1 axis may thus represent a promising strategy to restore ferroptotic sensitivity and enhance anti-tumor efficacy.

Lenvatinib is a frontline treatment for advanced HCC, but resistance frequently develops, limiting long-term benefits [41]. Our in vivo data demonstrate that EZH2 inhibition enhances the antitumor effect of lenvatinib in resistant HCC models. Combination treatment significantly suppressed tumor growth and increased ACSL1 expression and ferroptotic activity, supporting the role of ferroptosis reactivation in overcoming drug resistance. Notably, we further confirmed this mechanism through dual knockdown experiments, where ACSL1 silencing reversed the tumor-suppressive effects induced by EZH2 inhibition. This provides direct evidence that EZH2 mediates lenvatinib resistance by repressing ACSL1 and blocking ferroptosis. These findings suggest that targeting the EZH2–H3K27me3-ACSL1–ferroptosis axis may offer a promising strategy to restore lenvatinib sensitivity. EZH2 inhibitors, such as tazemetostat, have already been approved for other cancer types and are under investigation in clinical trials for solid tumors, raising the possibility of rapid clinical translation of this combination strategy [42]. Patients with high EZH2 expression or suppressed ferroptotic activity may derive particular benefit. A schematic representation of the effect of EZH2 inhibition on lenvatinib efficacy by promoting ACSL1-Mediated Ferroptosis via H3K27me3 is shown in Fig. 6. Further validation in patient-derived models and the development of ferroptosis-related biomarkers will be essential for clinical implementation.

Fig. 6.

Inhibition of EZH2 enhances lenvatinib efficacy in treating hepatocellular carcinoma by promoting ACSL1-Mediated Ferroptosis via H3K27me3

While our study reveals a critical role for the EZH2–H3K27me3–ACSL1–ferroptosis axis in lenvatinib resistance, several limitations remain. First, the broader epigenetic network regulating ferroptosis was not fully explored, and EZH2 may interact with additional chromatin modifiers. Second, although ACSL1 is validated as a key effector, ferroptosis involves multiple regulators, and the potential influence of EZH2 on other pathways warrants further investigation. Third, our findings were based on xenograft models, and additional validation in patient-derived or genetically engineered mouse models is needed to confirm clinical relevance. Future work should focus on elucidating the full ferroptotic regulatory landscape, evaluating predictive biomarkers such as EZH2 and ACSL1, and assessing the translational potential of EZH2-targeted combination therapies in clinical settings.

Conclusions

This study reveals that EZH2 drives lenvatinib resistance in liver cancer by epigenetically suppressing ACSL1 via H3K27me3, inhibiting ferroptosis. EZH2 inhibition restores ACSL1, promotes ferroptosis, and enhances lenvatinib efficacy. Targeting the EZH2–H3K27me3–ACSL1–ferroptosis axis may overcome resistance, and combining EZH2 inhibitors with current therapies could offer a new strategy for drug-resistant liver cancer.

Abbreviations

HCC

hepatocellular carcinoma

ROS

reactive oxygen species

GSH

glutathione

MDA

malondialdehyde

TKIs

tyrosine kinase inhibitors

FGFRs

fibroblast growth factor receptors

PDGFRα

platelet-derived growth factor receptorα

PRC2

Polycomb repressive complex 2

ABCG2

ATP-binding cassette subfamily G member 2

MRPs

multidrug resistance proteins

TCGA

The Cancer Genome Atlas

ACSL1

acyl-CoA synthetase long-chain family member 1

FPS

fetal bovine serum

IHC

immunohistochemical

PUFAs

polyunsaturated fatty acids

Authors’ contributions

Conceptualization, Tongchong Zhou; Formal analysis, Huitang Cai and Hongwei Yu; Funding acquisition, Yucong Lin; Investigation, Yibin Zhang and Yucong Lin; Methodology, Huitang Cai and Hongwei Yu; Project administration, Yibin Zhang; Resources, Huitang Cai and Hongwei Yu; Validation, Yibin Zhang and Yucong Lin; Writing – original draft, Yibin Zhang and Tongchong Zhou; Writing – review & editing, Yibin Zhang and Tongchong Zhou.

Funding

This work was supported by the Postdoctoral Foundation from the Ministry of Human Resources and Social Security of China (Grant No. 110215901). The authors sincerely appreciate the financial support for this research.

Data availability

The data used to support the findings of this study are available upon reasonable request from the corresponding author.

Declarations

Ethics approval and consent to participate

All the animal experiments were performed following the policies of the Animal Ethics Committee of AffiliatedCancer Hospital and Institute of Guangzhou Medical University and conducted in strict compliance with ethical guidelines. Following the Declaration of Helsinki, this study acquired the approval by the Medical Ethics Committee of AffiliatedCancer Hospital and Institute of Guangzhou Medical University. Informed consent was acquired from all the patients.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.