TrxR1 inhibition sensitizes hepatocellular carcinoma to Motesanib via an autophagy-ROS-JNK/ER stress axis Authors: Shengwei Du, Peisen Zheng, Wen Li, Guorong Chen, Hanbin Chen Published: 2026 ABSTRACT Hepatocellular carcinoma (HCC) remains a highly aggressive malignancy with a dismal prognosis, largely due to the limited effectiveness of current therapeutic interventions. Although Motesanib (MOT) is a clinically studied VEGFR-centered inhibitor, VEGFR-targeted strategies remain limited by innate or acquired resistance. In this study, we explored a novel pharmacological strategy combining MOT with Auranofin (AF) to enhance therapeutic outcomes in HCC. Our data demonstrate that this dual-targeting approach yields robust synergistic anticancer activity across diverse experimental models. Mechanistically, we discovered that TrxR1 inhibition is critical for this sensitization, which is closely associated with autophagy-related reactive oxygen species (ROS) accumulation. Specifically, genetic depletion of Atg5 effectively suppressed ROS accumulation and attenuated cytotoxicity, suggesting that autophagy contributes to the synergistic oxidative damage. This orchestrated stress response subsequently led to sustained endoplasmic reticulum (ER) stress and JNK pathway activation, culminating in DNA damage, impaired proliferation, and reduced cell viability. Taken together, our results identify the combination of MOT and AF as a promising and mechanistically grounded treatment regimen for advanced HCC (Graphical Abstract).Graphical Abstract Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s13062-026-00874-5. Introduction As one of the leading causes of cancer-related death globally, hepatocellular carcinoma (HCC) is a highly common and lethal malignancy [1]. As the primary form of liver cancer, HCC typically arises in patients with chronic liver injury or cirrhosis. Known risk factors include hepatitis B or C virus infection, long-term alcohol intake, and non-alcoholic steatohepatitis [2]. Patients with early HCC generally receive the current mainstream treatment, such as surgical removal of tumors [3], liver transplantation [4], or local ablation therapies [5]. However, as early-stage HCC often has no clear symptoms, many patients are not diagnosed until the disease has progressed, making systemic therapy the principal treatment option [6]. In recent years, targeted drugs and immune checkpoint inhibitors have demonstrated efficacy in advanced HCC, though the overall survival benefit remains modest [7]. Single-agent therapy is often limited by drug resistance, tumor heterogeneity, and toxic side effects [8], therefore, combination therapy may help improve therapeutic outcomes in patients with HCC.VEGFR signaling pathway plays an important role in tumor angiogenesis, which is necessary for tumor growth and development [9, 10]. VEGFR binds to its ligands to activate a series of downstream signals in endothelial cells, promoting cell proliferation, migration and survival [11]. VEGFR has been considered as an important target for cancer therapy, and several inhibitors targeting this pathway have been developed, including Motesanib (MOT) [12]. Monotherapy with VEGFR inhibitors is frequently limited by tumor resistance via alternative angiogenesis or hypoxia adaptation, mechanisms that may paradoxically drive progression [13]. Oxidative stress is triggered by the disproportion of reactive oxygen species (ROS) and is closely related to the development of tumors [14]. Compared to normal cells, cancer cells typically maintain higher basal ROS levels balanced by a more robust antioxidant system. This equilibrium sustains ROS at pro-survival rather than cytotoxic levels [15]. However, this balance is fragile and can be exploited by further elevating ROS or inhibiting antioxidant defenses. The resulting oxidative stress can trigger cancer cell death via apoptosis, autophagy, or ferroptosis [16, 17]. Therefore, regulating intracellular ROS levels to disrupt redox balance has been considered a potential strategy for cancer therapy. Emerging evidence suggests that instead of merely being a secondary response, hyperactivated autophagy can serve as a pro-oxidant engine that actively amplifies ROS production, a mechanistic hierarchy that remains poorly defined in HCC treatment. Thioredoxin reductase 1 (TrxR1) is an important component of the cellular redox system [18]. TrxR1 is involved in several basic cellular activities, such as DNA synthesis, cell death and the regulation of transcription factors (including NF-κB), mainly through the reduction of thioredoxin (Trx) [19]. In many cancers, including HCC, elevated TrxR1 levels are associated with tumor growth, chemotherapy resistance, and poor prognosis, supporting its role as a key therapeutic target [19]. Auranofin (AF), an approved rheumatoid arthritis drug, inhibits thioredoxin reductase 1 (TrxR1). This attenuation elevates intracellular ROS, disrupts redox balance, and can ultimately trigger endoplasmic reticulum stress and cancer cell death [20]. Several studies have reported that AF shows antitumor effects when used alone in different cancer models [21]. Furthermore, by disrupting the cancer cell antioxidant system, AF sensitizes them to other anticancer therapies, including chemotherapy and targeted agents [22]. These findings suggest that AF could be effective in combination therapy, particularly with VEGFR inhibitors like MOT, as it targets a distinct pathway to help overcome resistance. In this study, we investigated whether the TrxR1 inhibitor AF could enhance the cytotoxicity of the VEGFR inhibitor MOT against HCC cells. Mechanistically, the synergistic effect of the MOT/AF combination is driven by autophagy-mediated ROS accumulation, which subsequently triggers endoplasmic reticulum (ER) stress and activates the JNK signaling pathway in HCC cells. Our in vitro findings showed that co-administration of MOT and AF synergistically suppressed HCC cell growth. Collectively, our experiments elucidate a novel mechanistic basis for the synergistic antitumor efficacy of the MOT-AF combination and emphasize its promise as a viable treatment approach for human hepatocellular carcinoma. Materials and methods Cell lines and chemicals Two human hepatocellular carcinoma cell lines, PLC (RRID: CVCL_0485) and Huh7 (RRID: CVCL_0336), were used in this study. Both cell lines were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, CN). Cells were maintained in DMEM (11965084, Gibco). MOT (GC13012) and AF (GC15905) were purchased from GLP Bioscience. The inhibitors and pretreatment conditions were as follows: N-acetyl-L-cysteine (NAC, 5 mM, 1 h, A408248, Aladdin, CN), 3-methyladenine (3-MA, 2 mM, 24 h, M486160, Aladdin, CN), ciclopirox (CPX, 5 µM, 1 h, 54029ES50, Yeasen, CN), deferoxamine (DFO, 100 µM, 1 h, 70-51-9, MCE, USA), ferrostatin-1 (Fer-1, 5 µM, 1 h, Y240805, Beyotime, CN), hydroxychloroquine (HCQ, 5 µM, 24 h, 118-42-3, MCE, USA), Z-VAD (20 µM, 1 h, 161401-82-7, MCE, USA), necrostatin-1 (Nec-1, 20 µM, 1 h, S8037, Selleck, CN), MG-132 (5 µM, 4 h, 52801ES08, Yeasen, CN), ML385 (10 µM, 24 h, HY-100523, MCE, USA). After pretreatment, inhibitors were maintained in the culture medium during subsequent MOT, AF, or MOT + AF treatment unless otherwise indicated. Drug configuration MOT was reconstituted in DMSO to a final concentration of 120 mmol/L stock solution. AF powder was solubilized in DMSO to obtain a 10 mmol/L stock solution. Stock solutions were stored at − 80 °C and formulated in DMSO to the working concentration before use. Cell viability assay Cells were collected and dissociated with 0.25% trypsin, then plated into 6-well plates (2 × 10⁵ cells/well). Following attachment, MOT was added at the indicated concentrations either alone or together with AF. Cell viability was determined the next day by trypan blue staining. Each experiment was performed in three independent experiments. Synergism determination To quantitatively assess the nature of drug–drug interactions, the combination index (CI) was calculated based on the results of the cell viability assays. Following drug treatment, cell viability was measured and normalized to the untreated control group to obtain the fraction of affected cells for each single agent and their combinations. CI values were generated by CompuSyn software (USA) [23]. CI < 1.0 indicates synergism, the lower the CI value, the stronger the synergistic effect. Colony formation assay PLC and Huh7 cells were seeded separately into 6‑well plates at densities of 1 × 103 cells/well. The plates were gently swirled to ensure uniform cell distribution and then incubated at 37 °C in 5% CO₂. After 24 h, the cells were subjected to the indicated treatments. After 10 days, the cells were washed with PBS. The colonies were fixed with methanol, then staining with crystal violet solution. 5-ethynyl-2’-deoxyuridine labelling assay We seeded 6-well plates with cells at 1 × 10⁵ cells per well. After cell attachment, the indicated treatments were applied. EdU working solution was then added directly to the culture medium, followed by incubation as required. After labeling, cells were processed with the detection reagent provided in the EdU kit (Beyotime, C0071S). Imaging was performed with a fluorescence microscope to assess fluorescence signals. Each experiment was performed in three independent experiments. Immunofluorescence staining After drug treatment, cells were processed for immunofluorescence analysis. Cells were washed, fixed, and permeabilized by sequential treatment with 4% paraformaldehyde and 0.5% Triton X-100. Cells were then exposed to an anti-53BP1 primary antibody (#DF7472, Affinity; 1:2000) at 4 °C. After the incubation and wash steps, the matching secondary antibody was applied. Fluorescence signals were observed and recorded using a fluorescence microscope. Each experiment was performed in three independent experiments. Assessment of ROS production Cells in the logarithmic growth stage were plated into 6-well plates. After allowing them to adhere, the cells were treated with the specified drugs. Intracellular total ROS levels were detected using the DCFH-DA fluorescent probe (Beyotime, S1105S, CN). The probe was then diluted in culture medium and added to the wells according to the manufacturer’s protocol. Finally, fluorescence signals were visualized and documented using a fluorescence microscope. Each experiment was performed in three independent experiments. Western blotting analysis For protein extraction, the cells were carefully harvested on ice. Proteins were denatured at 100 °C for 10 min and then separated by 10% SDS-PAGE. The proteins were then electrotransferred to PVDF membranes, which were blocked with 5% non-fat milk. This membrane underwent overnight incubation with primary antibodies (p-JNK: 1:1000, #4668, CST, USA; JNK:1:1000, #9252, CST, USA; LC3:1:6000, #14600, Proteintech, CN; ATF4:1:600, #10835, Proteintech, CN; CHOP:1:1000, #15204, Proteintech, CN; GAPDH:1:10000, #60004, Proteintech, CN; TrxR1:1:1000, #11117, Proteintech, CN; p62: 1:10000, EPR4844, Abcam, USA; calnexin: 1:1000, #2433, CST, USA; NRF2: 1:4000, 16396-1-AP, Proteintech, CN; HO-1: 1:5000, 10701-1-AP, Proteintech, CN; cleaved PARP: 1:5000, 60555-1-Ig, Proteintech, CN; cleaved caspase-3: 1:500, 25128-1-AP, Proteintech, CN). Subsequently, it was rinsed thrice with TBST. A secondary antibody (#7054, CST, MA, USA) at a 1:20000 dilution was applied, complemented with the addition of the ECL substrate. The intensity of the target proteins and reference bands was quantified using the ImageJ software (V1,8.0.172). Each experiment was performed in three independent experiments. siRNA transfection Cells seeded in six-well plates were cultured at 37 °C with 5% CO₂ for 24 h before siRNA transfection with lipo8000 (C0533, Beyotime, CN). After 24 h, further experiments were conducted. These siRNA sequences were used: siAtg5-1: 5’- GCA ACA UAG UGA GAU AUG UTT − 3’, siAtg5-2: 5’- GAC GGU GAA CUC UUU UCC UTT − 3’, siTrxR1-1: 5’- GCA AGA CUC UCG AAA UUA UTT − 3’, siTrxR1-2: 5’- CUU UGC AGC UGC GCU CAA ATT − 3’, siJNK1: 5’- GGA GGA GAG AAC CAA GAA UTT − 3’, siJNK2: 5’- CGU GCA CUA ACU UCA UGA UTT − 3’. Lentivirus transfection CHOP-targeting lentiviral vectors (shCHOP) and a negative control (shNC) were obtained from GeneChem (Shanghai, CN). For transduction, PLC cells were first plated in 12-well plates and grown to about 70–80% confluence. The medium was then replaced with fresh medium containing the lentiviral particles at a predetermined MOI. After 12 h, the virus-containing medium was removed and replaced with complete fresh medium to limit toxicity. Following a 48-hour recovery period to allow transgene expression, puromycin selection was initiated. Cells were passaged and maintained in complete medium containing an appropriate concentration of puromycin to eliminate uninfected cells. Selection continued for at least one week until stable polyclonal populations emerged. The resulting stable shCHOP and shNC cell lines were then used for downstream experiments. Xenografts The animal experiment was conducted from June 1, 2023, to June 20, 2023. The animal study protocol was approved by the Ethics Committee of Wenzhou Medical University (Approval code: KY2022-R130; Approval date: 14 August 2022). All procedures were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. PLC cells were used to establish the xenograft model. A total of 1 × 106 cells suspended in 100 µL PBS were subcutaneously administered to nude mice. Male athymic BALB/c nude mice (5 weeks old, WFY-BCNU-4w-m, Vital River Laboratories, CN) were used for xenograft experiments. When tumor volumes reached approximately 100 mm3, mice were randomly assigned to four treatment groups (n = 5 per group, the sample size was chosen to provide adequate statistical power, in line with preliminary data and established practices in comparable xenograft studies [24]): (1) vehicle control (10% PEG300 + 10% Tween-80 in saline), (2) MOT (20 mg/kg), (3) AF (5 mg/kg), and (4) MOT (20 mg/kg) combined with AF (5 mg/kg). All drugs were given by intraperitoneal (i.p.) injection every two days [24]. Tumor growth was monitored using a vernier caliper throughout the 10-day treatment period. Tumor volume was calculated using the formula: V = (length × width2)/2. At the experimental endpoint, animals were euthanized by overdose of pentobarbital sodium, following the American Veterinary Medical Association (AVMA) Guidelines for the Euthanasia of Animals (2020 edition). The study was conducted in compliance with the ARRIVE guidelines. Tumor tissues were harvested for further analysis. The investigators were blinded to group allocation during drug administration and outcome assessment (tumor measurement, endpoint analysis). Serum biochemical analysis Blood samples were collected from five mice per group via cardiac puncture at the end of the 10-day treatment period. Serum was isolated by centrifugation at 3000 rpm for 15 min at 4 °C. The serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and serum creatinine (Cr) were quantified using an automated biochemical analyzer according to the manufacturer’s instructions. Immunohistochemistry Tissues from three randomly chosen mice per group (out of the four groups above) were sectioned. Tumor tissues were first fixed, embedded, and sectioned for histological staining. To prepare for antibody-based detection, tissue sections underwent antigen retrieval and were treated to quench endogenous enzyme activity. After blocking, the sections were incubated with primary antibodies against Ki-67, p-JNK, or γ-H2AX, followed by corresponding secondary reagents. Signals were developed using a chromogenic substrate, and nuclei were lightly counterstained. Finally, the stained slides were examined and imaged under a light microscope. Hematoxylin and Eosin (H&E) staining Liver and kidney samples were prepared for routine histological examination. After paraffin embedding, tissue sections were processed for staining. Sections were treated to remove paraffin and rehydrated, followed by nuclear staining and cytoplasmic counterstaining. After completion of staining, sections were dehydrated, cleared, and mounted. Tissue morphology of the liver and kidney was examined under a light microscope by an experienced pathologist who was blinded to the treatment conditions, with the aim of assessing potential tissue injury. Statistical analysis Unless stated otherwise, all quantitative analyses are based on independent experimental observations. Statistical analyses were performed using GraphPad Prism software (11.0 version). Comparisons between two groups were assessed using unpaired two-tailed Student’s t-tests. Data are presented as mean ± SD. P value < 0.01 was considered statistically significant. Exact sample sizes are given in the corresponding figure legends, and any repeat is indicated as an independent experiment or observation where relevant. All experiments were independently repeated thrice. Unless indicated by significance bars, significance annotations refer to comparisons with the corresponding control group, and only statistically significant differences (P < 0.01) are shown. Results MOT suppresses HCC cell proliferation and induces DNA damage via ROS accumulation MOT treatment concentration-dependently reduced the viability of PLC and Huh7 cells, as measured by trypan blue exclusion assay (Fig. 1A-C). Further investigation into the cytotoxic mechanism using EdU incorporation assays demonstrated a dose-dependent inhibition of DNA synthesis and proliferative activity (Fig. 1D-E). 53BP1 immunofluorescence showed a marked increase in nuclear foci following MOT treatment — indicating the induction of significant DNA damage (Fig. 1F-G). Fig. 1MOT suppresses HCC cell proliferation and induces DNA damage via ROS accumulation. (A) Chemical structure of MOT. (B, C) Cell viability of PLC and Huh7 treated with different concentrations of MOT. (D, E) Cell proliferation of PLC and Huh7 treated with different concentrations of MOT. Scale bar = 50 μm. (F, G) Nuclear damage of PLC and Huh7 with a range of MOT concentrations. Scale bar = 5 μm. (H) MOT altered ROS levels in HCC cells. Scale bar = 75 μm. (I) Viability of PLC and Huh7 cells. (J-K) EdU incorporation in PLC (80µM) and Huh7 (160µM) cells treated with MOT in the presence or absence of NAC. Scale bar = 50 μm. (L-M) DNA damage of PLC (80µM) and Huh7 (160µM) cells treated with MOT in the presence or absence of NAC. Scale bar = 5 μm. ** p < 0.01 MOT suppresses HCC cell proliferation and induces DNA damage via ROS accumulation. (A) Chemical structure of MOT. (B, C) Cell viability of PLC and Huh7 treated with different concentrations of MOT. (D, E) Cell proliferation of PLC and Huh7 treated with different concentrations of MOT. Scale bar = 50 μm. (F, G) Nuclear damage of PLC and Huh7 with a range of MOT concentrations. Scale bar = 5 μm. (H) MOT altered ROS levels in HCC cells. Scale bar = 75 μm. (I) Viability of PLC and Huh7 cells. (J-K) EdU incorporation in PLC (80µM) and Huh7 (160µM) cells treated with MOT in the presence or absence of NAC. Scale bar = 50 μm. (L-M) DNA damage of PLC (80µM) and Huh7 (160µM) cells treated with MOT in the presence or absence of NAC. Scale bar = 5 μm. ** p < 0.01 Intracellular ROS levels increased in both HCC cell lines following MOT treatment, with higher concentrations inducing stronger signals (Fig. 1H). The MOT-induced ROS increase was significantly attenuated when cells were co-treated with the antioxidant NAC (Fig. 1H). In addition, NAC partially rescued the decrease in cell viability caused by MOT (Fig. 1I), increased EdU-positive cells, and decreased 53BP1 foci formation (Fig. 1J-M). Collectively, these findings demonstrate that ROS accumulation underlies the anti-proliferative and genotoxic effects of MOT in HCC cells. TrxR1 inhibitors have synergistic effects with MOT We therefore evaluated whether TrxR1 inhibition could synergize with MOT by first testing three inhibitors —AF, Piperlongumine (PL), and WZ26—in PLC cells, where all exhibited concentration-dependent cytotoxicity (Fig. 2A-C). Fig. 2TrxR1 inhibitors have synergistic effects with MOT. (A-C) Cell viability of PLC cells treated with different TrxR1 inhibitors. (D) Cell viability and the corresponding CI values of PLC cells treated with MOT in combination with AF (1.2µM), PL (2.5µM), or WZ26 (5µM). (E-F) Cytotoxicity of MOT combined with AF in PLC and Huh7 cells and the corresponding CI values. (G-H) Viability of PLC in various of groups. (I-J) Representative images and colony formation assay in PLC (MOT:20µM, AF:0.4µM) and Huh7 (MOT:40µM, AF:0.2µM) cells treated with MOT, AF, or their combination. (K-L) EdU level in PLC and Huh7 cells. Scale bar = 50 μm. (M-N) Nuclear foci of 53bp1 in PLC and Huh7 cells in each group. Scale bar = 5 μm. ** p < 0.01 TrxR1 inhibitors have synergistic effects with MOT. (A-C) Cell viability of PLC cells treated with different TrxR1 inhibitors. (D) Cell viability and the corresponding CI values of PLC cells treated with MOT in combination with AF (1.2µM), PL (2.5µM), or WZ26 (5µM). (E-F) Cytotoxicity of MOT combined with AF in PLC and Huh7 cells and the corresponding CI values. (G-H) Viability of PLC in various of groups. (I-J) Representative images and colony formation assay in PLC (MOT:20µM, AF:0.4µM) and Huh7 (MOT:40µM, AF:0.2µM) cells treated with MOT, AF, or their combination. (K-L) EdU level in PLC and Huh7 cells. Scale bar = 50 μm. (M-N) Nuclear foci of 53bp1 in PLC and Huh7 cells in each group. Scale bar = 5 μm. ** p < 0.01 The results demonstrated that AF, PL, and WZ26 all exhibited synergistic effects with MOT. Among these combinations, AF + MOT displayed the most pronounced synergism. The CI values for all three combinations were below 1 which demonstrated a favorable synergistic effect, and the AF combination yielded the lowest CI value (Fig. 2D), indicating the strongest synergistic interaction. This synergistic effect between MOT and AF was further confirmed over a range of concentrations in both PLC and Huh7 cell lines. In PLC cells, the combination of MOT at 40 µM with AF produced the lowest CI value, suggesting that this combination exhibited the strongest synergistic effect. Whereas in Huh7 cells, the lowest CI value was obtained with MOT at 80 µM in combination with AF (Fig. 2E-F). In addition, combinations of MOT with PL or WZ26 also produced significant synergistic effects in PLC cells (Fig. 2G-H). The colony formation assay also confirmed the synergistic effect of MOT and AF (Fig. 2I-J). To further assess the biological effects of the MOT and AF combination, EdU and 53BP1 assays showed that the MOT + AF combination synergistically enhanced DNA damage while more potently suppressing DNA synthesis compared to either agent alone (Fig. 2K-N). We further examined apoptosis-related markers. Cleaved caspase-3 was detectable and increased after drug treatment, particularly in PLC cells, indicating activation of apoptosis-related signaling. However, the MOT + AF combination did not further significantly increase cleaved caspase-3 compared with the single-agent groups. In addition, cleaved PARP was not robustly induced (Supplementary Fig. 1A). These results suggest that caspase-related apoptotic signaling may be activated after treatment, but classical caspase-dependent apoptosis is unlikely to be the predominant mechanism responsible for the reduced viability caused by the MOT + AF combination. Autophagy-dependent ROS accumulation contributes to MOT + AF-induced cytotoxicity To delineate the cytotoxic mechanism induced by MOT + AF, PLC and Huh7 cells were pretreated with pathway-related inhibitors before combination treatment. Direct comparisons were performed between the MOT + AF group and each MOT + AF+inhibitor group. Fer-1 modestly rescued cell viability in PLC cells, whereas DFO, Nec-1, and MG-132 slightly enhanced MOT + AF-induced cytotoxicity. In Huh7 cells, CPX, DFO, Fer-1, and MG-132 slightly increased cytotoxicity. Notably, only NAC and 3-MA substantially rescued MOT + AF-induced viability loss in both cell lines (Fig. 3A, B), supporting ROS accumulation and autophagy as major contributors under the tested condition, while other regulated cell-death pathways cannot be completely excluded. Fig. 3Autophagy-dependent ROS accumulation contributes to MOT + AF-induced cytotoxicity. (A, B) Cell survival rate in PLC and Huh7 cells caused by MOT and inhibitors. (C) ROS level in PLC and Huh7. Scale bar = 75 μm. (D, E) ROS and survival rate in PLC (MOT:40µM, AF:1.2µM) with or without silence of Atg5. (F, G) ROS and survival rate in Huh7 (MOT:80µM, AF:0.6µM) with or without silence of Atg5. (H, I) Colony formation assay and quantification in PLC (MOT:20µM, AF:0.4µM) and Huh7 (MOT:40µM, AF:0.2µM) cells treated with MOT plus AF with or without Atg5 knockdown. (J, K) EdU incorporation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF with or without ATG5 knockdown. Scale bar = 50 μm. (L, M) 53BP1 nuclear foci formation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF with or without ATG5 knockdown. Scale bar = 5 μm. **P < 0.01 Autophagy-dependent ROS accumulation contributes to MOT + AF-induced cytotoxicity. (A, B) Cell survival rate in PLC and Huh7 cells caused by MOT and inhibitors. (C) ROS level in PLC and Huh7. Scale bar = 75 μm. (D, E) ROS and survival rate in PLC (MOT:40µM, AF:1.2µM) with or without silence of Atg5. (F, G) ROS and survival rate in Huh7 (MOT:80µM, AF:0.6µM) with or without silence of Atg5. (H, I) Colony formation assay and quantification in PLC (MOT:20µM, AF:0.4µM) and Huh7 (MOT:40µM, AF:0.2µM) cells treated with MOT plus AF with or without Atg5 knockdown. (J, K) EdU incorporation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF with or without ATG5 knockdown. Scale bar = 50 μm. (L, M) 53BP1 nuclear foci formation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF with or without ATG5 knockdown. Scale bar = 5 μm. **P < 0.01 Based on this result, we further examined the relationship between autophagy and ROS. Significant intracellular ROS accumulation potently induced by the MOT + AF combination was abolished by either NAC or 3-MA (Fig. 3C). We further applied siRNA to silence the expression of Atg5 to assess the contribution of autophagy. The data suggest that autophagy is involved in the treatment-induced ROS accumulation and knockdown of Atg5 significantly reversed the ROS level and cytotoxicity of MOT and AF in both PLC (Fig. 3D-E) and Huh7 (Fig. 3F-G) cells. The colony formation experiment confirmed that siAtg5 can reverse the cytotoxicity induced by MOT and AF (Fig. 3H-I). Consistently, the resulting inhibition of DNA synthesis and elevated 53BP1 foci were also reversed by knockdown of Atg5 (Fig. 3J-M). We next examined NRF2 expression to determine whether the oxidative stress induced by MOT + AF activated an antioxidant adaptive response. It is showed that MOT or AF alone mildly increased NRF2 protein expression, whereas the MOT + AF combination markedly upregulated NRF2 (Supplementary Fig. 1B). This result was consistent with the stronger ROS accumulation observed after combined treatment. To further test the functional role of NRF2, cells were treated with the NRF2 inhibitor ML385. ML385 did not rescue MOT + AF-induced cytotoxicity but further reduced cell viability (Supplementary Fig. 1C), supporting the interpretation that NRF2 activation mainly represents a compensatory cytoprotective response rather than the primary driver of combination-induced cytotoxicity. The MOT + AF combination triggers ER stress-related cytotoxicity via the autophagy-ROS axis To explore the downstream consequences of ROS, we assessed ER stress induction by the MOT/AF combination. LC3-II levels increased in PLC and Huh7 cells after combination treatment while p62 expression showed an opposite trend to LC3-II, as shown by Western blot, demonstrating elevated autophagic flux and autophagic degradation (Fig. 4A-B). Knockdown of Atg5 by siRNA pretreatment significantly suppressed the increase in autophagic flux and autophagic degradation, confirming that it was specifically induced by the combination therapy (Fig. 4C-D). Fig. 4The MOT + AF combination triggers ER stress-related cytotoxicity via the autophagy-ROS axis. (A) LC3 and p62 expression in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells were assessed by Western blot. (B) LC3 and p62 quantification in PLC and Huh7 cells under the indicated treatments. (C) Modulation of LC3-II, p62 levels in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells following genetic silencing of Atg5. (D) LC3 and p62 quantification in PLC and Huh7 cells under the indicated conditions. (E) Protein expression of ER stress-related markers, including ATF4, CHOP, HO-1 and calnexin in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (F) Expression of ATF4, CHOP, HO-1 and calnexin was assessed after siAtg5 treatment under MOT + AF exposure in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (G) Effect of CHOP silencing on CHOP expression in drug combination. (H) Relative band intensity of CHOP normalized to GAPDH. (I) Survival rate of PLC cells following CHOP knockdown. ** p < 0.01 The MOT + AF combination triggers ER stress-related cytotoxicity via the autophagy-ROS axis. (A) LC3 and p62 expression in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells were assessed by Western blot. (B) LC3 and p62 quantification in PLC and Huh7 cells under the indicated treatments. (C) Modulation of LC3-II, p62 levels in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells following genetic silencing of Atg5. (D) LC3 and p62 quantification in PLC and Huh7 cells under the indicated conditions. (E) Protein expression of ER stress-related markers, including ATF4, CHOP, HO-1 and calnexin in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (F) Expression of ATF4, CHOP, HO-1 and calnexin was assessed after siAtg5 treatment under MOT + AF exposure in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (G) Effect of CHOP silencing on CHOP expression in drug combination. (H) Relative band intensity of CHOP normalized to GAPDH. (I) Survival rate of PLC cells following CHOP knockdown. ** p < 0.01 Analysis of ER stress markers showed that the MOT/AF combination significantly upregulated both ATF4 and its downstream target CHOP. We further found that the MOT + AF combination markedly induced HO-1. In contrast, calnexin was only mildly increased after combined treatment. When cells were pretreated with siAtg5, these increases were reduced. These results further indicate that MOT + AF activates an autophagy-dependent ER stress-related response (Fig. 4E-F). To interrogate the role of ER stress, we generated a stable CHOP-knockdown PLC line. Depletion of CHOP significantly rescued cells from MOT/AF-induced cytotoxicity (Fig. 4G-I). In summary, the MOT + AF combination induces ROS accumulation via autophagy, which in turn triggers ER stress, culminating in CHOP-dependent cytotoxicity. In addition, siRNA treatment alone did not obviously alter ATF4, CHOP, HO-1 or calnexin expression compared with the control group, indicating that the observed ER stress-related changes were mainly caused by MOT + AF treatment rather than siRNA transfection. JNK signaling pathway activation is essential for the anti-tumor efficacy of the MOT + AF combination The observed oxidative stress prompted us to test the involvement of the JNK signaling pathway, which is known to be activated by ROS [25]. Western blot analysis showed that the MOT/AF combination significantly increased JNK phosphorylation, and this activation was largely suppressed when cells were transfected with siAtg5, confirming that JNK acts downstream of autophagy in this pathway (Fig. 5A-B). Fig. 5JNK signaling pathway activation is essential for the anti-tumor efficacy of the MOT + AF combination. (A) Phosphorylated and total JNK levels were examined following treatment with MOT, AF, or their combination in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (B) Activation of the JNK signaling pathway was assessed after administration in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (C) Western blot analysis of phosphorylated JNK and total JNK levels after combined JNK1/2 knockdown in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF. (D) Cell viability of PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF with or without JNK1/2 knockdown. (E, F) EdU incorporation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells after JNK knockdown. Scale bar = 50 μm. (G, H) 53bp1 nuclear foci formation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with siJNK in the presence or absence of MOT and AF. Scale bar = 5 μm. ** p < 0.01 JNK signaling pathway activation is essential for the anti-tumor efficacy of the MOT + AF combination. (A) Phosphorylated and total JNK levels were examined following treatment with MOT, AF, or their combination in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (B) Activation of the JNK signaling pathway was assessed after administration in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells. (C) Western blot analysis of phosphorylated JNK and total JNK levels after combined JNK1/2 knockdown in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF. (D) Cell viability of PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with MOT plus AF with or without JNK1/2 knockdown. (E, F) EdU incorporation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells after JNK knockdown. Scale bar = 50 μm. (G, H) 53bp1 nuclear foci formation in PLC (MOT:40µM, AF:1.2µM) and Huh7 (MOT:80µM, AF:0.6µM) cells treated with siJNK in the presence or absence of MOT and AF. Scale bar = 5 μm. ** p < 0.01 To assess the functional role of JNK activation, JNK1 and JNK2 were simultaneously knocked down using siRNA. Combined knockdown of JNK1 and JNK2 effectively suppressed JNK phosphorylation and significantly rescued the reduction in cell viability induced by the MOT and AF combination (Fig. 5C-D). Consistent with these results, JNK knockdown markedly restored EdU incorporation and reduced the formation of 53BP1 foci (Fig. 5E-H). Together, these findings indicate that activation of the JNK pathway is required for the cytotoxic effects of the MOT and AF combination and links upstream stress signals to cell proliferation inhibition and DNA damage in HCC cells. Genetic silencing of TrxR1 recapitulates and confirms the synergistic mechanism with MOT To confirm that the observed synergy was due to TrxR1 inhibition rather than off-target effects of AF, TrxR1 expression was silenced using siRNA (Fig. 6A). Combined treatment with siTrxR1 and MOT in both PLC and Huh7 cells reproduced the synergistic effects observed with AF. This combination led to increased JNK phosphorylation and elevated expression of the ER stress markers ATF4 and CHOP (Fig. 6B-E), accompanied by a marked reduction in cell viability (Fig. 6F-G). Fig. 6Genetic silencing of TrxR1 recapitulates and confirms the synergistic mechanism with MOT. (A) TrxR1 expression following siRNA-mediated knockdown, with relative intensity quantified by western blots analysis. (B, C) JNK pathway activity and ER stress response in PLC (MOT:40µM) and Huh7 (MOT:80µM) cells treated with MOT in the presence or absence of TrxR1 knockdown. (D, E) Quantification and statistical analysis of western blots. (F, G) Cell viability was assessed in PLC (MOT:40µM) and Huh7 (MOT:80µM) cell lines after MOT treatment with or without TrxR1 silencing. (H, I) JNK pathway activity and ER stress in PLC (MOT:40µM) and Huh7 (MOT:80µM) cells treated with MOT and siTrxR1, with or without siAtg5. (J, K) Quantification and statistical analysis of western blots. (L, M) Viability of PLC (MOT:40µM) and Huh7 (MOT:80µM) cells following combined treatment with MOT and siTrxR1, with or without siAtg5. ** p < 0.01 Genetic silencing of TrxR1 recapitulates and confirms the synergistic mechanism with MOT. (A) TrxR1 expression following siRNA-mediated knockdown, with relative intensity quantified by western blots analysis. (B, C) JNK pathway activity and ER stress response in PLC (MOT:40µM) and Huh7 (MOT:80µM) cells treated with MOT in the presence or absence of TrxR1 knockdown. (D, E) Quantification and statistical analysis of western blots. (F, G) Cell viability was assessed in PLC (MOT:40µM) and Huh7 (MOT:80µM) cell lines after MOT treatment with or without TrxR1 silencing. (H, I) JNK pathway activity and ER stress in PLC (MOT:40µM) and Huh7 (MOT:80µM) cells treated with MOT and siTrxR1, with or without siAtg5. (J, K) Quantification and statistical analysis of western blots. (L, M) Viability of PLC (MOT:40µM) and Huh7 (MOT:80µM) cells following combined treatment with MOT and siTrxR1, with or without siAtg5. ** p < 0.01 To further determine whether this effect depended on autophagy, cells were pretreated with siAtg5. Inhibition of autophagy significantly reduced the activation of JNK signaling and ER stress induced by the siTrxR1 + MOT combination (Fig. 6H-K). Consistent with this result, the decrease in cell viability caused by siTrxR1 + MOT was largely reversed by inhibition of autophagy (Fig. 6L-M). Together, these results indicate that genetic inhibition of TrxR1 is sufficient to enhance the cytotoxic effect of MOT and that this synergy, similar to pharmacological inhibition, depends on autophagy-mediated activation of the JNK and ER stress pathways in HCC cells. The MOT + AF combination potently inhibits tumor growth in vivo To evaluate whether the combination treatment was effective in vivo, we tested the antitumor effect of MOT and AF in a PLC cell–derived xenograft mouse model. Consistent with the in vitro results, treatment with the MOT and AF combination significantly suppressed tumor growth compared with the vehicle group or either drug alone, as reflected by reduced tumor volume and lower final tumor weight (Fig. 7A-C). Fig. 7The MOT + AF combination potently inhibits tumor growth in vivo. (A) The volume of tumors was measured every other day starting from the onset of treatment. (B) Photos of tumors. (C) Tumor weights across experimental groups. (D) IHC staining of Ki67, P-JNK and γ-H2AX in tumors. Scale bar = 75 μm. (E) HE staining of livers and kidneys. Scale bar = 75 μm. (F) Levels of ALT, AST, BUN and Cr in serum. (G) Body weight curves. ** p < 0.01 The MOT + AF combination potently inhibits tumor growth in vivo. (A) The volume of tumors was measured every other day starting from the onset of treatment. (B) Photos of tumors. (C) Tumor weights across experimental groups. (D) IHC staining of Ki67, P-JNK and γ-H2AX in tumors. Scale bar = 75 μm. (E) HE staining of livers and kidneys. Scale bar = 75 μm. (F) Levels of ALT, AST, BUN and Cr in serum. (G) Body weight curves. ** p < 0.01 Immunohistochemical staining was then performed on tumor tissues. Tumors from the combination-treated group showed a clear decrease in the proliferation marker Ki67. In addition, the levels of p-JNK and γ-H2AX were increased, indicating activation of JNK signaling and enhanced DNA damage in tumor tissues (Fig. 7D), which was consistent with the in vitro results. To evaluate potential toxicity, histological and serological analyses of major organs were carried out. H&E staining of liver and kidney tissues did not show obvious pathological changes in any treatment group (Fig. 7E). No significant differences in serum ALT, AST, BUN and Cr levels were observed among the four groups (Fig. 7F). In addition, body weight in mice showed a similar trend throughout the experiment, with no significant differences among groups (Fig. 7G). These results indicate that the MOT and AF combination can effectively inhibit tumor growth in vivo without causing apparent liver or kidney toxicity. Discussion In this work, we show that combining MOT with AF produces stronger antitumor activity than either agent alone in HCC cell models and a PLC-derived xenograft model. Mechanistically, the combination promotes autophagy-associated accumulation of intracellular ROS, followed by JNK activation, ER stress, DNA damage and loss of cell viability. Thus, the main contribution of this study is to connect VEGFR-centered kinase inhibition with TrxR1-targeted redox stress in a single therapeutic framework for HCC. The translational framing of MOT requires caution. Several clinical trials registered on ClinicalTrials.gov, including NCT00121628, NCT00254267, NCT00356681, NCT00460317, and NCT02629848, have evaluated MOT in solid tumors, supporting its available clinical safety and pharmacological background. However, its later-stage clinical development was limited by insufficient efficacy in some tumor settings, and no established HCC-specific clinical indication has been reported. Therefore, we viewed this study as a preclinical repurposing investigation rather than a comparison with or replacement for approved HCC therapies. MOT targets VEGFR1/2/3, PDGFR and Kit, pathways linked to angiogenesis and tumor survival [26], and adaptive resistance to VEGF/VEGFR-targeted treatment remains a major therapeutic limitation [27]. In our experiments, MOT alone increased ROS in HCC cells, and NAC reduced ROS accumulation, partially restored cell viability and EdU incorporation, and decreased 53BP1 foci. AF, by inhibiting TrxR1, weakens thioredoxin-dependent antioxidant defense and can promote ROS accumulation and ER stress [28]. The synergy between MOT and AF is therefore best interpreted as complementary redox pressure: MOT creates an oxidative vulnerability, while AF limits antioxidant compensation. This interpretation is consistent with evidence that TrxR1 inhibition can cooperate with multi-targeted kinase inhibition to induce ROS-dependent cytotoxicity [29]. Autophagy appears to be an important upstream contributor to this redox response. Pharmacological inhibition with 3-MA and, more decisively, Atg5 knockdown reduced MOT + AF-induced ROS accumulation and downstream cytotoxic signaling, whereas inhibitors of apoptosis, necroptosis and ferroptosis did not show comparable rescue. Because DCFH-DA mainly reports general intracellular ROS rather than lipid ROS, we avoid interpreting the ROS signal as direct evidence of ferroptosis. Although autophagy can promote ferroptosis in some contexts and ferroptosis is defined by iron-dependent lipid peroxidation [17, 29], ferroptosis-related inhibitors showed inconsistent effects in our system. Fer-1 modestly rescued cell viability in PLC cells, but it slightly enhanced MOT + AF-induced cytotoxicity in Huh7 cells, while CPX and DFO also failed to substantially reverse the MOT + AF phenotype. Therefore, the current data do not support classical lipid peroxidation-dependent ferroptosis as the predominant mechanism under this experimental condition. Nevertheless, a partial or context-dependent contribution of ferroptosis cannot be completely excluded. In addition, the partial rescue by Atg5 knockdown indicates that autophagy is also not the only determinant of the response. Downstream of ROS, JNK signaling and ER stress contributed functionally to the cytotoxic phenotype. MOT + AF increased JNK phosphorylation and upregulated ATF4 and CHOP, and these changes were attenuated by Atg5 knockdown, placing JNK and ER stress downstream of autophagy-associated ROS accumulation. JNK1/2 knockdown or CHOP suppression partially rescued cell viability, supporting their functional involvement. In vivo, combination-treated tumors showed reduced Ki67 and increased p-JNK and γ-H2AX without obvious liver or kidney histological injury, consistent with antitumor activity accompanied by stress-pathway activation. We also found that MOT + AF induced NRF2 and HO-1, suggesting activation of an antioxidant feedback response. NRF2 is a central antioxidant-response transcription factor [30], TrxR1 expression is associated with NRF2 pathway activation [31]. However, NRF2 inhibition with ML385 further decreased cell viability rather than rescuing cells. Because NRF2 nuclear translocation and broader ARE target profiling were not fully assessed, the precise contribution of the NRF2–TrxR1 axis to MOT + AF sensitivity should be further investigated in future studies. These findings support a broader therapeutic concept in which HCC cells may be sensitized by combining anti-angiogenic pressure with disruption of antioxidant capacity. Potentially relevant subgroups for future study include tumors with high TrxR1 expression, strong antioxidant dependence, elevated oxidative-stress vulnerability or resistance to anti-angiogenic therapy. TrxR1 has been reported as an unfavorable prognostic factor in HCC and as a targetable redox vulnerability in HCC models [32], and NRF2/Keap1-dependent regulation has been linked to hepatoma growth [33]. Anti-angiogenic resistance in HCC is also associated with hypoxia and stress-response pathways [34]. These biomarker concepts remain hypotheses rather than validated patient-selection criteria. The present study is also limited by its reliance on cell lines and immunodeficient xenografts, which do not fully capture patient heterogeneity or the immune microenvironment. Future work using larger HCC cell panels, patient-derived organoids, immunocompetent models and clinical specimens will be needed to validate predictive biomarkers and define how MOT + AF perturbs autophagy to sustain cytotoxic ROS signaling. Conclusion Our study shows that MOT and AF form a strongly synergistic drug combination against HCC. This effect is closely associated with enhanced general intracellular ROS accumulation, which is functionally linked to autophagy-related signaling and accompanied by activation of the JNK pathway and ER stress response. Although our pharmacological and genetic data support the importance of the autophagy-ROS-JNK/ER stress axis, the contribution of additional regulated cell-death pathways cannot be fully excluded, particularly under the strongly cytotoxic MOT + AF condition used in the current inhibitor-rescue experiments. Our results provide preclinical evidence supporting the repurposing of AF in combination with VEGFR inhibitors for HCC treatment. Given that anti-angiogenic therapy is an established therapeutic strategy in HCC and that MOT is a clinically studied VEGFR-centered multi-kinase inhibitor, this work provides a mechanistic basis for future translational studies evaluating MOT repurposing or related VEGFR inhibitor-based redox-targeting combinations in HCC. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1