HCC cell lines HCCLM3 and Huh7 were obtained from the Cell Bank at the Chinese Academy of Sciences. Cells were cultured in a humidified incubator at 37°C with 5% CO₂. The DMEM medium (C11995500BT, Gibco, United States) was supplemented with 10% fetal bovine serum (FSP500, ExCell Biotech, China) and 1% penicillin/streptomycin (C100C5, NCM Biotech, China) for optimal growth.

Donafenib (HY-10201S), Alisertib (HY-10971), and Ferrostatin-1 (HY-100579) were acquired from MedChemExpress (MCE) Corporation located in Shanghai, China. Z-VAD-FMK (S7023), 3-MA (3-Methyladenine, S2767), Necrostatin-1 (Nec-1, S8037), Belnacasan (VX-765, S2228), and PMA (S7791) were sourced from Selleck Chemicals in Houston, TX. Dimethyl sulfoxide (DMSO) was supplied by Sigma-Aldrich, also in Shanghai, China.

Total protein extraction utilized the Protease and Phosphatase Inhibitor Cocktail (P002, NCM Biotech, China). Cytoplasmic and nuclear proteins were isolated using the Nuclear and Cytoplasmic Protein Extraction Kit (KGB5302, KGI Biotechnology) according to the manufacturer’s instructions. Protein concentrations were measured with the BCA assay (ZJ102 kit, Epizyme Biomedical Technology). Samples were mixed with a 5 × loading buffer (LT103, Epizyme Biomedical Technology) and heated at 95°C for 10 min. We loaded 30 µg of total proteins and 60 µg of phosphorylated proteins onto 12.5% SDS-PAGE gels, followed by transfer to 0.2 µm PVDF membranes (Merck, Millipore). Membranes were blocked with 5% skim milk or BSA for 1 h to reduce non-specific binding, then incubated overnight at 4°C with primary antibodies (see Supplementary Table S1). Species-specific secondary antibodies were applied for 1–2 h at room temperature. Finally, protein bands were visualized using Enhanced Chemiluminescent detection (P10060, NCM, China).

Total RNA was extracted from biological samples using the RNAex Pro Reagent (AG21101, Accurate Biology, China), followed by a purification step in accordance with the protocols outlined in the SteadyPure RNA Extraction Kit (AG21024, Accurate Biology, China). The quality of the isolated RNA was assessed by quantifying both concentration and purity using a NanoDrop One spectrophotometer (Thermo Fisher Scientific, MA, United States). For complementary DNA (cDNA) synthesis, 2 µg of total RNA was reverse transcribed using the Evo M-MLV RT Premix (AG11706, Accurate Biology, China). Subsequently, quantitative real-time PCR (qPCR) analysis was performed with 100 ng of the synthesized cDNA as the template for amplification. mRNA expression levels were evaluated using the SYBR Green Premix Pro Taq HS qPCR Kit (Rox Plus, AG11718, Accurate Biology, China) on a QuantStudio 1 Real-Time PCR System developed by Applied Biosystems, a subsidiary of Thermo Fisher Scientific, MA, United States. Detailed information regarding the primer sequences used in this study is provided in Supplementary Table S2.

Tumor-bearing mice were euthanized, and tumors and surrounding organs were collected to assess systemic toxicity. Specimens were fixed in 4% paraformaldehyde for at least 24 h, then embedded in paraffin and sectioned to 3–5 μm for H&E staining (BL700A, Biosharp, China). For IHC, sections were deparaffinized, rehydrated, and subjected to antigen retrieval. After blocking to reduce non-specific binding, sections were incubated with primary and secondary antibodies, and nuclei were counterstained with hematoxylin.

HCC cells were seeded in confocal dishes and allowed to adhere. After 48 h of drug treatment, the medium was discarded. Cells were fixed with 4% paraformaldehyde for 20 min, permeabilized with 0.3% Triton X-100 for 15 min, and blocked with goat serum for 1 h. They were incubated overnight at 4°C with an anti-p65 antibody (1:200, #8242, Cell Signaling Technology, United States), followed by a 1-hour incubation with Alexa Fluor® 488-AffiniPure goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, United States). Nuclei were stained with DAPI (C1002, Beyotime, China) for 5 min.

HCC cells treated with Donafenib and/or Alisertib were fixed overnight at room temperature using a fixative (G1102, Servicebio, China). This was followed by fixation with osmium tetroxide, dehydration, and embedding in resin. Ultrathin sections (60–80 nm) were prepared using an ultramicrotome (Leica EM UC7) and stained with heavy metals (lead and uranium).

Cell viability was assessed using the CCK8 kit (BS350, Biosharp, China), with HCC cells seeded in 96-well plates and treated with the respective treatments for 0–72 h. Every 24 h, the medium was replaced with CCK8 reagent, and after a 1.5-hour incubation, optical density was measured at 450 nm to evaluate cell viability.

After a 48-hour exposure to pharmacological agents, cells were harvested from 6-well plates and stained with Annexin V and propidium iodide (PI) (KGA1102, KeyGEN, China) for flow cytometry analysis on a Beckman Coulter instrument. Inhibitors used included Z-VAD-FMK (20 µM) for apoptosis, 3-MA (10 µM) for autophagy, Nec-1 (10 µM) for necroptosis, VX-765 (20 µM) for pyroptosis, and Ferrostatin-1 (10 µM) for ferroptosis.

Based on the established IC₅₀ values for Donafenib and Alisertib, a 7 × 7 concentration gradient matrix was formulated. The concentrations for Donafenib were 0, 0.625, 1.25, 2.5, 5, 10, and 20 μM, while the concentration gradient for Alisertib included 0, 1.25, 2.5, 5, 10, 20, and 40 μM. HCCLM3 and Huh7 cells were cultured in 96-well plates according to this matrix design. Subsequently, drug synergy scores were calculated using SynergyFinder software, and a heatmap representing the synergistic interactions was generated using the built-in zero interaction potency (ZIP) approach (Ianevski et al., 2022).

HCCLM3 and Huh7 cells were seeded in six-well plates at a density of 600 and 800 cells per well, respectively, with medium changes every 2–3 days until colonies exceeded 50. After treatment, the colonies were fixed with 4% paraformaldehyde and stained using a 0.1% crystal violet solution.

The EdU incorporation assay assessed cellular proliferation differences by treating adherent cells in six-well plates with compounds for 48 h, followed by 10 µM EdU for 2 h. Staining was performed per the manufacturer’s instructions (C0078, Beyotime, China), with nuclei counterstained using Hoechst 33,342.

HCCLM3 and Huh7 cells were cultured in 6-well plates at 4 × 10⁵ cells per well and treated with pharmacological compounds for 48 h. Lipid peroxides (LPO) and intracellular reactive oxygen species (ROS) levels were measured using the C11-BODIPY 581/591 and DCFH-DA probes, respectively, following the manufacturer’s instructions (Elabscience, China).

After 48 h of treatment with pharmacological agents, approximately one million cells from each group were collected and incubated with the iron probe following the manufacturer’s guidelines (E-BC-K880-M, Elabscience, China). Absorbance was measured at 593 nm using a microplate reader to evaluate the results.

Cells were plated in 6-well culture plates at 4 × 10⁵ cells per well and incubated overnight for adherence before undergoing specified treatments. Following treatment, total intracellular glutathione (GSH/GSSH) was quantified using an assay kit (E-BC-K097-M, Elabscience, China) following the manufacturer’s protocol.

The intracellular concentration of malondialdehyde (MDA) was measured using a Colorimetric Assay Kit from Elabscience (E-BC-K028-M, China), with optical density assessed at 532 nm using a microplate reader for precise quantification.

HCCLM3 cells were treated with Donafenib and/or Alisertib for 48 h. Total RNA was extracted using Trizol reagent from Invitrogen Life Technologies, following the manufacturer’s instructions. Sequencing libraries were constructed according to standardized protocols and subsequently subjected to high-throughput sequencing on the Illumina NovaSeq 6,000 platform, operated by Novogene Co., Ltd., China.

Differentially expressed genes (DEGs) were identified from transcriptomic data collected from drug-treated HCC tissues, as well as from adjacent non-tumorous samples, using information sourced from The Cancer Genome Atlas (TCGA). The analytical process was conducted using the R package DESeq2, which is well-regarded for its robustness in comparing gene expression levels across different conditions. Subsequently, we conducted Gene Set Enrichment Analysis (GSEA) (Subramanian et al., 2005) and single-sample Gene Set Enrichment Analysis (ssGSEA) (Hänzelmann et al., 2013) to evaluate the enrichment of specific biological pathways. This analytical approach was instrumental in illuminating the biological significance of our findings, allowing us to gain deeper insights into the underlying mechanisms associated with the differentially expressed genes identified in our study.

HEK293T cells were seeded in 24-well plates at 70%–90% confluency and transfected in Opti-MEM medium with a total of 0.5 μg of NFE2L2 firefly luciferase reporter plasmid, 0.5 μg of RELA overexpression plasmid (or an empty vector control), and 0.02 μg of pRL-TK Renilla luciferase internal control plasmid using Lipofectamine 3000. After 6 h, the medium was replaced with complete DMEM containing 10% FBS. Cells were harvested 24 h post-transfection for luciferase activity measurement using the Dual-Luciferase Reporter Assay System (Promega). Firefly luciferase activity was normalized to Renilla luciferase values to facilitate data analysis.

The ChIP assay was performed using the ChIP-IT® Express Enzymatic Kit (Active Motif, United States) with the ChIP-IT® Control Kit—Human to ensure experimental validity. Briefly, cells were crosslinked with 1% formaldehyde, lysed, and enzymatically sheared to generate DNA fragments (200–1,500 bp). Chromatin complexes were immunoprecipitated with anti-NF-κB/p65 antibody or control rabbit IgG, followed by qPCR analysis using specific primers targeting the κB1 (−820) and κB2 (−220) sites: κB1 forward 5′-TGCACTCGGTAATCGGCTACA-3′, reverse 5′-GGGAGCTAACGGAGACCT-3’; κB2 forward 5′-ACTCCCACGTGTCTCCATTC-3′, reverse 5′-CGATTACAGCATGTTGTGGTATT-3’.

All animal care and experimental procedures were approved by the Comparative Medicine Center of the Eastern Theater General Hospital, strictly adhering to the International Guiding Principles for Animal Research. The study involved female BALB/c mice, aged 5–6 weeks, sourced from GemPharmatech, China. To establish a tumor model, HCCLM3 cell lines were cultured in the logarithmic growth phase, and approximately 5 × 10⁶ cells were subcutaneously injected into the right anterior axilla of each mouse. Once tumors reached an estimated volume of 100 mm³ (calculated using V = ½ ab², where ‘a’ is the longest diameter and ‘b’ is the shortest), treatment began.

Data analysis and visualization were conducted using R software (version 4.2.1) alongside GraphPad Prism 9.0 (version 9.5.1). Numerical results are presented as mean ± standard deviation (SD). Statistical significance was evaluated using t-tests, unpaired t-tests, and one-way analysis of variance (ANOVA), with equal variances confirmed. To enhance reliability, all analyses included biological replicates performed in triplicate. Statistical significance thresholds were set at *P < 0.05; **P < 0.01; ***P < 0.001; and ****P < 0.0001.

In vitro experiments demonstrated that Alisertib synergistically enhances the cytotoxicity of Donafenib against HCC cells. The molecular structures of Donafenib and Alisertib are shown in Figures 1A,B. Initial biological assessments were performed using the CCK-8 assay on HCCLM3 cells, generating dose-response curves (Figure 1C), with a parallel analysis on the Huh7 cell line (Figure 1D). Both drugs exhibited concentration-dependent anti-proliferative effects, and IC₅₀ values were calculated as illustrated in Figures 1C,D. Based IC₅₀ values, we established a concentration gradient and cellular viability matrix to compute drug ZIP synergy scores using the online SynergyFinder software (Figures 1E,F). The average synergy scores were 14.15 for HCCLM3 cells and 12.83 for Huh7 cells, with peak synergy regions marked by white dashed boxes achieving scores of 26.09 and 30.37, respectively. ZIP scores above 10 indicate a strong synergistic effect in inhibiting tumor proliferation. In the peak synergy region, a concentration of 2.5 µM for Alisertib was identified as the optimal dose. To determine the optimal Dosage of Donafenib, we assessed cellular viability (Figure 1G) and colony-forming efficacy (Figures 1H,I) across various treatment regimens. Optimal concentrations were established at 10 µM for HCCLM3 cells and 5 µM for Huh7 cells, both showing significant anti-proliferative effects (p < 0.05). After 72 h of combined treatment, we observed a marked reduction in growth rates for both cell lines (Figure 1J, P < 0.001). These findings indicate that Alisertib, at concentrations well below its IC₅₀, significantly enhances the cytotoxic effects of Donafenib in HCC cells.

The combination of Donafenib and Alisertib significantly enhances the sensitivity of hepatocellular carcinoma (HCC) cells to Donafenib, resulting in reduced cell proliferation and viability. Flow cytometry analyses demonstrated that co-treatment with Donafenib and Alisertib substantially increased the percentage of Annexin-V + PI + cells in the HCCLM3 and Huh7 cell lines (Figures 2A,B). To elucidate the mechanisms underlying this cell death, we utilized various inhibitors to assess cell viability (Figures 2C,D; Supplementary Figure S1). Inhibitors of autophagy (3-MA), necroptosis (Nec-1), and pyroptosis (VX-765) did not reverse cell death. Conversely, Z-VAD-FMK, an apoptosis inhibitor, slightly mitigated cell death, while Ferrostatin-1 (Fer-1; 10 µM), a ferroptosis inhibitor, significantly restored cell viability compromised by the drug combination (Figures 2C,D). These findings suggest that ferroptosis is the primary mode of cell death.

Furthermore, co-treatment led to a significant accumulation of reactive oxygen species (ROS), which was effectively inhibited by Fer-1 (Figures 2E,F). Notably, levels of total iron and malondialdehyde (MDA) were upregulated, whereas glutathione was downregulated in the combination treatment group (Figures 2G,H). DCFH-DA fluorescence analysis revealed a marked increase in intracellular ROS levels following co-treatment (Figures 2I,J). Transmission electron microscopy (TEM) further illustrated mitochondrial shrinkage, loss of cristae, and increased electron density in membranes, all characteristic of ferroptosis (Figure 2K) (Dixon et al., 2012). Additionally, we observed significant downregulation of key ferroptotic defense molecules, including xCT, GPX4, and NRF2 pathways, indicating a diminished anti-ferroptotic response in the combined treatment group (Figures 2L,M). In brief, our findings establish that the synergistic application of Alisertib and Donafenib effectively induces ferroptosis in HCC cells, underscoring a promising therapeutic strategy for HCC.

To explore the effects of Donafenib and Alisertib on ferroptosis induction in HCC cells, we performed RNA sequencing to identify differentially expressed genes (DEGs) in HCCLM3 cells treated with each agent alone and in combination. The principal component analysis (PCA) revealed strong intra-group consistency and significant inter-group differentiation (Figure 3A). This strongly suggests that the different treatment regimens considerably altered the global gene expression profiles of the cells.

In the combination group, we identified 1,555 differentially expressed genes (DEGs) excluding genes from individual treatments (p < 0.05, |log₂FC|>1, Figure 3B; Supplementary Figure S2). The volcano plot highlighted significantly regulated genes in the combination group (|log₂FC| > 2, p < 0.05, Figure 3C). KEGG enrichment analysis revealed significant enhancement in pathways related to reactive oxygen species metabolism, glutathione metabolism, and the pentose phosphate pathway (Figure 3D), suggesting a close connection to the regulatory mechanisms of ferroptosis (Yao et al., 2021; Liang et al., 2022). Quantitative assessment of ferroptosis-related pathways employed ssGSEA scoring with five established programmed cell death (PCD) gene sets (Figure 3E) (Liu et al., 2022). Ferroptosis scores showed maximal activation in the combination group versus monotherapies (p < 0.05), while apoptosis remained minimally active across treatments. TCGA validation (n = 371 LIHC) confirmed ferroptosis as the dominant PCD mechanism (Figure 3F). Pan-cancer analysis (33 tumor types) revealed NF-κB pathway’s universal correlation with ferroptosis, particularly in HCC (LIHC: r = 0.489, p < 0.05; Figures 3G,H) (Guo et al., 2023). Western blot analyses revealed significantly reduced phosphorylation of both p-p65 (Figures 3I,M) and pIκBα (Figure 3N) in the combination therapy group versus single-agent treatments. Total p65 levels remained stable across groups (Figure 3K), while IκBα protein demonstrated marked elevation in the combination cohort (Figure 3L). Both p-p65/p65 and pIκBα/IκBα ratios exhibited significant reductions (Figures 3O,P). Subcellular localization analysis showed cytoplasmic accumulation alongside nuclear depletion of p65 in combination-treated cells (Figures 3J,Q,R). Immunofluorescence validation confirmed potent suppression of p65 nuclear translocation (Figures 3S,T). Collectively, the Donafenib-Alisertib combination suppresses NF-κB signaling by inhibiting p65 activation and nuclear translocation while stabilizing IκBα, mechanistically driving ferroptosis in HCC.

Subcutaneous HCCLM3 xenograft models were established to assess the antitumor effects of Donafenib (30 mg/kg) combined with Alisertib (30 mg/kg). Tumor-bearing mice (baseline volume: 100 mm³) received monotherapy or combination therapy (Figure 4A). The combination group demonstrated significantly reduced tumor volume compared to monotherapy groups (Figures 4B,D) and suppressed tumor growth relative to controls (Figure 4C). Body weights remained stable across all groups (p > 0.05, Figure 4E). Tumors from the combination group exhibited elevated 4HNE immunostaining (Figure 4F), indicating enhanced lipid peroxidation. H&E staining confirmed increased tumor necrosis, while Ki-67 staining revealed reduced proliferation. Consistent with in vitro findings, p-p65 immunostaining showed decreased NF-κB activation in the combination group.

Co-treatment with the NF-κB activator PMA (100 nM, 24 h) rescued NF-κB pathway inhibition caused by Donafenib and Alisertib in HCCLM3 and Huh7 cells. Western blot analysis showed PMA increased phosphorylation of p65 and IκBα while reducing total IκBα levels (Figures 5A–E), confirming restoration of NF-κB activity that reversed drug-induced blockade of p65 nuclear translocation and IκBα degradation.

The ferroptosis phenotype was systematically reversed by PMA: Both cytosolic ROS (DCFH-DA assay) and lipid ROS (BODIPY-C11) levels decreased significantly (Figures 5F–I), with concurrent upregulation of ferroptosis defense proteins xCT, GPX4, and NRF2 (Figures 5J–M). PMA also reduced total iron content and malondialdehyde (MDA) levels while increasing the GSH/GSSG ratio (Figures 5N,O). Functionally, PMA restored cell viability (Figures 5P,Q), increased EdU + proliferating cells (Figures 5R,S), and enhanced colony formation capacity (Figures 5T,U). Therefore, Pharmacological NF-κB reactivation via PMA counteracted the pro-ferroptotic synergy effects of Donafenib and Alisertib, demonstrating that synergistic ferroptosis induction by these agents requires NF-κB pathway inhibition.

To better understand the mechanisms behind the synergistic effects of donafenib and alisertib on HCC, we conducted a detailed analysis of the top 50 differentially expressed genes (DEGs) identified from RNA sequencing data, focusing specifically on ferroptosis-related markers. Notably, we found that NRF2 expression was significantly downregulated after combination therapy, which is closely linked to the induction of ferroptosis (Figure 6A). Additionally, GPX4, a key enzyme involved in NRF2-regulated glutathione (GSH) synthesis, showed altered expression, indicating a reduced capacity for cellular antioxidant defense (Anandhan et al., 2020).

Given that NRF2 is transcriptionally activated by RELA/NF-κB (Rushworth et al., 2012; Wang et al., 2021), we performed a correlation analysis using data from the TCGA-HCC dataset. This analysis revealed a significant positive correlation between the mRNA levels of NRF2 and RELA (NF-κB/p65) (Figure 6B), suggesting an interplay between the NRF2 and NF-κB signaling pathways in the context of ferroptosis. To further validate this interaction, we overexpressed RELA in HCC cell lines and evaluated its impact on ferroptosis-related gene expression through qPCR and Western blot analyses. Notably, the upregulation of NRF2 correlated with RELA induction, an effect that was significantly reduced in the presence of the drug combination (Figures 6C–E). Subsequent mechanistic studies using dual-luciferase reporter assays confirmed that p65 directly regulates NRF2 promoter activity (Figure 6F). Additionally, chromatin immunoprecipitation (ChIP) assays (Figure 6G) showed that combination treatment significantly decreased RELA/NF-κB occupancy at the NRF2 promoter in HCC cells, particularly at the kB2 site (Figures 6H,I), providing direct evidence of transcriptional regulation. Collectively, these findings underscore the critical role of the NF-κB/NRF2 axis in regulating ferroptosis, highlighting its potential as a promising therapeutic target in HCC.

To further investigate the role of the NF-κB signaling cascade in mediating the synergistic effects of this combination treatment, we conducted rescue experiments using PMA (300 μg/kg) in mice with subcutaneously implanted tumors. These mice were treated either with Donafenib monotherapy or with Donafenib in combination with Alisertib, as shown in Figure 7A. The results demonstrated that the antitumor efficacy observed in both the Donafenib and combination treatment groups was significantly reversed following PMA exposure (Figure 7B). Specifically, PMA administration led to a substantial increase in tumor volume in both treatment regimens, highlighting a marked reduction in the synergistic antitumor effects (Figures 7C,D).

Body weight measurements across the groups showed no statistically significant differences, indicating an acceptable tolerance profile (Figure 7E). Furthermore, histopathological analysis revealed that PMA treatment resulted in a significant decrease in the necrotic rate within the tumor tissue (Figure 7F). Collectively, these in vivo rescue experiments suggest that the efficacy of the combination therapy heavily relies on the inhibition of NF-κB signaling pathway activity to facilitate ferroptosis. The sensitization of HCC cells to Donafenib by Alisertib may pave the way for novel therapeutic strategies in the treatment of HCC.

The systemic toxicity of the different treatment groups was evaluated through monitoring body weight changes and assessing the histopathological integrity of major organs in nude mice. Over time, all treatment groups exhibited an increase in body weight, attributed to tumor enlargement. Importantly, our findings indicated that neither Donafenib administered alone nor the combination of Donafenib with Alisertib resulted in significant weight loss, and intergroup differences were not statistically significant. This suggests that the dosages of Donafenib and Alisertib utilized in our study are within a safe range. Furthermore, we conducted thorough dissection, weighing, and H&E staining of the principal organs of the mice to evaluate any potential adverse effects. As depicted in Figures 8A–E, we calculated organ indices using the formula: Organ Index (%) = [(Organ Weight)/(Body Weight of Mouse)] × 100%. Optical microscopy was utilized to examine and capture images of the organ sections, as illustrated in Figure 8F. The findings revealed normal structural morphology of the major organs, devoid of any discernible pathological changes, indicating that the administered drugs did not elicit systemic toxicity. Additionally, we performed biochemical analyses on blood samples collected from the orbital sinus of the mice, as illustrated in Figures 8G–J. The serum levels of ALT, AST, CREA, and BUN remained within normal reference ranges, indicating that hepatic and renal functions were unaffected by the treatments. Collectively, these findings substantiate the robust safety profile of our therapeutic regimen.