Traditional Chinese Medicine Library (cat. no. L8300), PPVI (cat. no. S9302), β-lapachone (cat. no. S7261), 3-methyladenine (3-MA, cat. no. S2767) inhibitors and Ac-DEVD-CHO (cat. no. S7901) were purchased from Selleck Chemicals; Stattic (cat. no. HY-13818) and ferrostatin-1 (Fer-1, cat. no. HY-100579) were purchased from MedChem Express. GPX4 (cat. no. ab125066), signal transducer and activator of transcription 3 (STAT3; cat. no. ab68153), phosphorylated (p-)STAT3 (pY705; cat. no. ab76315), N-cadherin (cat. no. ab18203), Vimentin (cat. no. ab92547), E-cadherin (cat. no. ab40772) and goat anti-rabbit IgG H&L (HRP; cat. no. ab6721) antibodies were purchased from Abcam. The GAPDH (2118S) antibody was purchased from Cell Signaling Technology, Inc. The goat anti-rabbit IgG (H+L) cross-adsorbed secondary antibody and cyanine3 (A-10520) were purchased from Invitrogen; Thermo Fisher Scientific, Inc.

The HCCLM3 cells were purchased from the China Center for Type Culture Collection (CCTCC), and the Huh7 cells were purchased from Suyan Biotechnology Co., Ltd. The THLE-2 cells were purchased from Meisen Cell Technology Co., Ltd. The HCCLM3 and Huh7 Cells were cultured in DMEM/high-glucose medium (Gibco; Thermo Fisher Scientific, Inc.) with 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin-streptomycin (Gibco; Thermo Fisher Scientific, Inc.) at 37°C and 5% CO₂. The THLE-2 cells require a special coating medium (CTCC-002-30, Meisen Cell Technology Co., Ltd.) and were incubated overnight in an incubator at 37°C and 5% CO₂ and cultured in BEGM™ BulletKit™ medium (cat. no. CC-3170, Lonza Group, Ltd.) with 10% FBS. The HCC cells were observed and representative images were captured under an inverted microscope (Ts2R-FL, Nikon Corporation; ×100 magnification) at 0 and 24 h following drug intervention. The HCC cells were treated with PPVI (4 µM) at 37°C for 24 h, following which different assays were performed. Inhibitors, such as 3-MA (1 mM), Ac-DEVD-CHO (50 µM) and Fer-1 (1 µM) for 2 h prior to treatment with PPVI (4 µM) for 24 h.

The plasmids that included PIRES2-EGFP (hSTAT3-IRES2-EGFP, oeSTAT3 and BW2486) and PIRES2-EGFP (No-Targeting Plasmid, the control plasmid and BVA15) were purchased from Hangzhou Youming Biotechnology Co., Ltd. For transfection, Lipofectamine 3000 transfection reagent (GlpBio) diluted in Opti-MEM (Gibco; Thermo Fisher Scientific, Inc.) was used (1:1 ratio; two tubes). After seeding, the cells were grown to 70-90% confluency and then transfected, with 5 µg/µl of the transfected DNA concentration. The DNA master mix was prepared by diluting DNA in Opti-MEM, then adding Lipofectamine 3000 reagent and mixing well. The diluted DNA was added to each tube of diluted Lipofectamine 3000 reagent and incubated for 10-15 min at room temperature. The DNA-lipid complex was added to the cells and incubated for 3 days at 37°C.

All animal experiments were approved (No. 20220070) by the Animal Ethics Committee (AEC) of the China Science and Technology Industry Holdings (Shenzhen) Co., Ltd. Male BALB/c nude mice (n=9), 4 weeks old, were purchased from the Guangdong Medical Laboratory Animal Center. All mice were housed under the following conditions: A 12-h light/dark cycle (lights on, 07:00; lights off, 19:00), a temperature of 22±2°C and a humidity of 50±10%, and were provided with free access to standard chow and water. The Huh7 cells (1×10⁷ in 0.1 ml DMEM) were injected subcutaneously into the right flanks of the nude mice. After 7 days, the mice were randomly divided into three groups (three mice per group) as follows: The control group (0.9% NaCl/day), PPVI-treated group (PPVI at 10 mg/kg/day) and the group treated with PPVI (10 mg/kg/day) + Fer-1 (10 mg/kg/day). The weight of the tumors and tumor size were measured daily. Following the intraperitoneal administration of the treatments for 10 consecutive days, the mice were sacrificed, and the serum was used to measure alanine transaminase (ALT, cat. no. C009-2-1) and aspartate aminotransferase (AST, cat. no. C010-2-1) (both from Jiancheng Bioengineering Institute) levels. The mice were euthanized by an intraperitoneal injection of pentobarbital sodium (200 mg/kg). Blood was then obtained from the orbital sinus, and the livers and subcutaneous tumors were removed. During the animal experiments, the maximum diameter of the tumors was measured daily. When the longest diameter of a single tumor was 20 mm, the experiment was terminated and the animals were euthanized. Following euthanasia, the death of the animal was confirmed by documenting the disappearance of the pain response (i.e., unresponsiveness to toe compressions with hands or forceps) and the observation of cardiac arrest and apnea. Liver and tumor tissues were collected for detection using hematoxylin and eosin (H&E) staining and immunohistochemistry, as described below.

The HCC cells were treated with the indicated concentrations of PPVI (from 1 to 10 µM) or Stattic (6 µM) for 24 h. The HCC cells were then treated with or without 3-MA (1 mM), Ac-DEVD-CHO (50 µM) and Fer-1 (1 µM) inhibitors for 2 h prior to treatment with PPVI for 24 h. Cell viability was assessed using a CCK-8 assay (cat. no. C0043, Beyotime Institute of Biotechnology). The optical density (OD) value of the sample was obtained at 450 nm using a microplate reader (BioTek Instruments, Inc.).

After the HCCLM3 or Huh7 cells were seeded in six-well plates at a density of 800 cells/well, they were subjected to PPVI (2, 4, 6, 8 and 10 µM) or PPVI (4 µM) with Fer-1 (1 µM). The medium was replaced for 3 days and the cells were cultured until the number of monoclonal cells was >50. The cells were fixed with 4% paraformaldehyde at 25°C for 30 min, stained with crystal violet (cat. no. BL802A, Biosharp Life Sciences) for 30 min at 25°C, and photographed using a digital camera (Ts2R-FL, Nikon Corporation).

The HCCLM3 or Huh7 cells were treated with or without 3-MA (1 mM), Ac-DEVD-CHO (50 µM) and Fer-1 (1 µM) inhibitors for 2 h prior to treatment with PPVI (4 µM) for 24 h. Following the instructions provided with the LDH Cytotoxicity Detection Kit (Beyotime Institute of Biotechnology), the plates were incubated at 37°C for 30 min in the dark. The OD values at 490 nm were obtained using a microplate reader (BioTek Instruments, Inc.).

The HCCLM3 and Huh7 cells were seeded in a six-well plate at a density of 4×10⁵ cells/well. When the cell density reached 90% confluency, the confluent cell monolayer was scraped using a pipette tip. The cells were incubated in serum-free medium as designed for the experiments. Images of the same scratched area at 0 and 24 h were captured (Ts2R-FL, Nikon Corporation).

The cells were starved and incubated for 12 h in advance, and 2.5×10⁵ HCC cells were seeded in the upper chamber (cat. no. 3422, Corning, Inc.) of the Transwell, followed by PPVI (4 µM) or PPVI (4 µM) + Fer-1 (1 µM) for 24 h. The lower chamber contained DMEM with 20% FBS. Following incubation of the Transwell plate at 37°C for 24 h, the invaded cells were fixed with formaldehyde for 30 min at 25°C and stained with crystal violet (cat. no. BL802A, Biosharp Life Sciences) for 30 min at 25°C. Finally, the invaded cells were observed and photographed under an inverted microscope (Nikon, Ts2R-FL).

After the HCCLM3 and Huh7 cells were treated with PPVI (4 µM) and incubated for 24 h at 37°C, the cells were collected by centrifugation (3,000 × g, 2 min, 37°C), supplemented with 2.5% Gluta fixative (Beijing Solarbio Science & Technology Co., Ltd.) at 4°C for 48 h. The cells were then dehydrated in a graded ethanol series, infiltrated with propylene oxide, embedded in an epoxy resin (cat. no. 10000418, SPI Supplies) at 37°C for 12 h and sectioned to a thickness of 70 nm. Subsequently, the ultrathin sections were observed and photographed using an HT7800 transmission electron microscope (Hitachi, Ltd.).

The HCCLM3 or Huh7 cells were seeded in six-well plates, and treated with PPVI (2, 4 and 6 µM) or PPVI (4 µM) with Fer-1 (1 µM) for 24 h. The JC-1 fluorescent probe (Beyotime Institute of Biotechnology) was then added according to the instructions of the manufacturer to observe the mitochondrial membrane potential (MMP) changes under a fluorescence microscope (DMIL4000, Leica Microsystems, Inc.).

Following 24 h of treatment with PPVI (2, 4 and 6 µM) or PPVI (4 µM) with or without Fer-1 (1 µM), the working solution was added according to the instructions provided with the GSH assay kit (cat. no. S0053, Beyotime Institute of Biotechnology) using a single-point assay, and absorbance was measured using a microplate reader (BioTek Instruments, Inc.) after 25 min.

The HCCLM3 or Huh7 cells were homogenized in MDA lysis buffer, and centrifuged at 13,000 × g for 10 min at 4°C to obtain 200 µl supernatant. The MDA assay kit (cat. no. MAK085, MilliporeSigma) was used to detect the MDA content. Finally, the absorbance value was measured at 532 nm using a microplate reader (BioTek Instruments, Inc.).

The HCCLM3 or Huh7 cells were treated with PPVI (2, 4 and 6 µM) or PPVI (4 µM) with Fer-1 (1 µM) for 24 h, and the Fe²⁺ levels were determined according to the instructions provided with the iron assay kit (cat. no. MAK025, MilliporeSigma).

FerroOrange fluorescent probes (Dojindo Laboratories, Inc.) were loaded in the HCC cells following 24 h of PPVI (2, 4 and 6 µM) or PPVI (4 µM) with or without Fer-1 (1 µM). According to the instructions provided by the manufacturer, following 30 min at 37°C of incubation, the changes in the Fe²⁺ content were observed and photographed using a fluorescence microscope (Ts2R-FL, Nikon Corporation).

Intracellular superoxide anion levels were determined using dihydroethidium (DHE; Beyotime Institute of Biotechnology). Following the manufacturer's instructions, a DHE solution was incubated with the cells at 37°C for 30 min and the cells were then observed and photographed under a fluorescence microscope (Ts2R-FL, Nikon Corporation).

The cells were lysed and homogenized using RIPA lysis buffer (cat. no. 89900, Thermo Fisher Scientific, Inc.), and the protein concentration was detected using a BCA assay kit (cat. no. 23227, Thermo Fisher Scientific, Inc.). The proteins were separated by SDS-PAGE (10% gels) and transferred onto PVDF membranes. After the membranes were incubated with 5% skim milk at 25°C for 1 h, they were incubated with the corresponding primary antibodies (1:1,000; all antibodies used are listed above in the Reagents and antibodies section) at 4°C overnight, followed by incubation with the secondary antibody (1:2,000) for 1 h at 25°C. The bands were detected using a chemiluminescent HRP substrate (WBKLS0500, MilliporeSigma) and visualized with the ChemiDoc XRS system (Bio-Rad Laboratories, Inc.). Densitometric analysis of the western blots was performed using ImageJ V1.8.0 (National Institutes of Health, Inc.).

The HCC cells were treated with PPVI (2, 4 and 6 µM) or PPVI (4 µM) with or without Fer-1 (1 µM) and incubated for 24 h at 37°C. The cells were fixed with 4% polyformaldehyde, permeabilized with 0.5% Triton X-100 (T9284, MilliporeSigma), blocked with 5% goat serum (BL210A, Biosharp Life Sciences) and incubated overnight with the required amount of primary antibody such as GPX4, N-cadherin, Vimentin and E-cadherin (1:200) in a humidified box at 4°C. The cells were incubated with the goat anti-rabbit IgG (H+L) antibody, Cy3 (1:200) at 25°C for 1 h in the dark. Nuclear DAPI (BL105A, Biosharp Life Sciences) was used for counterstaining for 15 min at 25°C in the dark with antifade mounting medium (P0126, Beyotime Institute of Biotechnology) to mount the slides and the slides were then photographed (Ts2R-FL, Nikon Corporation). The procedure was repeated with the Huh7 cells.

Total RNA was extracted from the HCC cells using the AG RNAex isolation system (Accurate Biotechnology Co., Ltd.) according to the manufacturer's protocol. Complementary DNA was synthesized using reverse transcription kits (cat. no. AG11728, Jiangsu Accuracy Biotechnology Co., Ltd.). qPCR was performed using a QuantStudio5 system (Thermo Fisher Scientific, Inc.) with SYBR-Green (cat. no. AG11718, Jiangsu Accuracy Biotechnology Co., Ltd.) and the thermo-cycling conditions were as follows: 95°C for 30 sec, 95°C for 5 sec and 60°C for 30 sec, for 40 cycles. The 2^−ΔΔCq method was used to analyze the RT-qPCR data (15). The sequences of the primers used are as follows: GPX4 forward, 5′-CCC GAT ACG CTG AGT GTG GTT TG-3′ and reverse, 5′-CCT TGC CCT TGG GTT GGA TCT TC-3′; STAT3 forward, 5′-TGT GAT GCT TCCCTG ATT GTG ACT G-3′ and reverse, 5′-TCG CTT GGT GGT GGA GGA GAA C-3′; and GAPDH forward, 5′-CAA GGC TGT GGG CAA GGT CAT C-3′ and reverse, 5′-GTG TCG CTG TTG AAG TCA GAG GAG-3′.

The tissues were paraffin-embedded, sectioned, dewaxed in xylene, soaked in 100 to 70% gradient ethanol and placed in PBS. Citric acid retrieval solution (cat. no. P0086, Beyotime Institute of Biotechnology) was then added, and the slices were placed in the microwave for 20 min at 100°C and then rinsed with PBS three times, for 5 min each. After washing, 5% blocking serum was added in a dropwise manner for blocking for 30 min at 37°C, and primary antibody such as p-STAT3, STAT3, GPX4 (1:200) with 5% BSA was then added in a dropwise manner for overnight incubation at 4°C. The following day, MaxVision™ HRP-Polymer anti-mouse/rabbit IHC kit (cat. no. KIT-5020, MXB Biotechnologies) was added 50 µl followed by incubation at 25°C for 15 min. Subsequently, 50 µl developer diaminobenzidine working solution (cat. no. DAB-0031, MXB Biotechnology Co., Ltd.) was added followed by incubation at room temperature for 3 min. The sections were counterstained with hematoxylin for 40 sec at 25°C, dehydrated, mounted, and observed under a microscope (Axio Imager M2, Zeiss AG).

The human targets of PPVI were obtained from the SwissTarget prediction database (http://swisstargetpre-diction.ch/), and the DrugBank (https://www.drugbank.ca/), GeneCards (https://www.genecards.org/), DisGeNET (http://www.disgenet.org/home), Online Mendelian Inheritance in Man (OMIM, https://www.omim.org/), TTD (http://db.idrblab.net/ttd) and PharmGKB databases (https://www.pharmgkb.org/) were then searched using the key words 'liver cancer' and 'liver carcinoma'. All obtained targets were merged and duplicate targets were removed. The targets of PPVI were used to intersect with the targets of liver cancer to obtain the anti-liver cancer targets of PPVI. The data were uploaded to the STRING 11.5 database to obtain protein-protein interaction (PPI) information and analyzed using Cytoscape 3.9.1 (https://cytoscape.org/). Cytoscape was used to calculate network topology parameters, including degree centrality, closeness centrality, betweenness centrality, and clustering coefficient. Finally, AutoDock 4.2.6 (https://autodock.scripps.edu/) was used to perform molecular docking between PPVI and key targets, and PyMOL (https://pymol.org/2/) was used to obtain a visual 2D map.

All experiments were repeated independently at least three times. GraphPad Prism 9.4 software (GraphPad Software, Inc.) was used for data analysis using an unpaired t-test and one-way ANOVA (with Tukey's multiple comparisons test). A value of P<0.05 was considered to indicate a statistically significant difference.

To identify compounds with therapeutic efficacy in HCC, a pilot library was screened which comprised ~1,444 TCM monomeric representative compounds with an ability to inhibit cell activity in HCCLM3 cells at a concentration of 10 µM. The screening process followed in the present study is illustrated in Fig. 1A. The HCCLM3 cells were treated with 1,444 monomers for 24 h and cell viability was then examined. As shown in Fig. 1B, 20 compounds (Fig. S1A and B) inhibited the growth of the HCCLM3 cells by <30%. Subsequently, the Huh7 cells were treated with these monomers 10 µM for 24 h, and a second screening was performed using CCK-8 assay (Fig. 1C). The Venn diagram shown in Fig. 1D illustrates the screening process, demonstrating monomeric molecules with toxic effects in both HCCLM3 and Huh7 cells. The intersection of the Venn diagram indicates two compounds, PPVI and β-lapachone. The Huh7 cells were then treated with PPVI and β-lapachone (10 µM) for 24 h, which was found to exert potent cytotoxic effects on both cell lines (Fig. 1E). The further comparison of the inhibitory effects of PPVI and β-lapachone on HCC cell (HCCLM3 and Huh7 cells) viability revealed that PPVI exerted the most pronounced inhibitory effect on the viability of HCC cells (Fig. 1E). The chemical structural formula of PPVI is presented in Fig. 1F.

To further analyze the cytotoxic effects of PPVI on HCC cells, the cells were treated with various concentrations of PPVI. The results revealed that PPVI inhibited the viability of HCCLM3 and Huh7 cells in a concentration-dependent manner (Fig. 1G and H). Compared with the normal liver cells (THLE-2), PPVI (4 µM) also exerted a significant inhibitory effect on HCC cell viability (Fig. 1I). The light microscopy images demonstrated that cell contraction and cell death increased with the increasing concentrations of PPVI (Fig. 1J and K). Cell cloning experiments also revealed that the proliferative capacities of the HCCLM3 and Huh7 cells were negatively associated with the concentration of PPVI (Fig. 1L and M).

The present study then examined the morphology of the treated cells in order to explore the mechanisms underlying PPVI-induced cell death. TEM analysis revealed that the HCCLM3 and Huh7 cells treated with PPVI (4 µM) for 24 h exhibited a decrease in mitochondrial volume, an increase in bilayer membrane density, and a decrease or disappearance of mitochondrial cristae (Fig. 2A and B), which are typical morphological features of ferroptosis. To verify whether this form of cell death involved ferroptosis, the HCCLM3 and Huh7 cells were treated with PPVI in the absence or presence of several cell death inhibitors. Combination treatment with the autophagy inhibitor, 3-MA, and the pyroptosis inhibitor, Ac-DEVD-CHO, did not prevent the death of the cells treated with PPVI (Fig. 2C-F), and the ferroptosis inhibitor, Fer-1, exerted the most pronounced reversal effect (Fig. 2C and D). Apoptosis or necrosis can cause damage to the cell membrane structure, and the quantitative analysis of cytotoxicity was achieved by detecting the activity of LDH released from the cells with plasma membrane rupture (Fig. 2E and F). Furthermore, in the fluorescence images of JC-1-stained mitochondria (Fig. 2G and H), red fluorescence representing an intact MMP was observed in the control group and a low MMP was observed in the PPVI-treated groups (2, 4 and 6 µM) with increasing concentrations.

To explore whether ferroptosis is a key determinant of HCC cell death induced by PPVI, several ferroptosis events were examined in HCC cells, including ROS accumulation, GSH, MDA, the Fe²⁺ content and GPX4 expression. GSH levels in the HCCLM3 and Huh7 cells were found to gradually decreased with the increasing concentrations of PPVI (2, 4 and 6 µM) (Fig. 3A and B). However, PPVI upregulated the MDA levels in a concentration-dependent manner (Fig. 3C and D). In addition, the intracellular Fe²⁺ levels significantly increased upon stimulation with 4 and 6 µM PPVI (Fig. 3E and F). Consistently, when the Fe²⁺ content was detected using a FerroOrange fluorescent probe, the accumulation of Fe²⁺ in the cells was promoted by the increase in the PPVI concentration (Fig. 3G). The level of cellular ROS was then detected using DHE. As the concentration of PPVI increased, the red fluorescence in the HCC cells was enhanced, indicating that the intracellular ROS level was positively associated with the concentration of PPVI, which suggested that PPVI induced oxidative stress in the HCC cells (Fig. 3H). The results also revealed that the GPX4 protein expression levels in the HCCLM3 and Huh7 cells were negatively associated with the concentration of PPVI (Fig. 3I and J).

To further investigate the association between mitochondrial damage and ferroptosis induced by PPVI, the ferroptosis inhibitor, Fer-1, was used. Mitochondrial fluorescence staining was performed using HCC cells treated with PPVI + Fer-1, and an increase in aggregate JC-1 and a decrease in monomeric JC-1 were observed compared to the PPVI group (Fig. 4A and B), suggesting that Fer-1 can alleviated mitochondrial damage induced by PPVI. Changes in the intracellular levels of GSH, MDA and Fe²⁺ were also examined following co-treatment with PPVI and Fer-1. Consistently, GSH expression increased in the PPVI + Fer-1 group compared with that in the PPVI group (Fig. 4C and D), whereas the contents of MDA and Fe²⁺ decreased in the PPVI + Fer-1 group (Fig. 4E-H). Furthermore, the levels of ROS and ferrous iron were decreased in the PPVI + Fer-1 group compared with the PPVI group (Fig. 4I and J). Moreover, the protein expression levels of GPX4 were measured following co-treatment with PPVI and Fer-1 using western blot analysis. Compared with the PPVI group, the protein expression level of GPX4 in the PPVI + Fer-1 group was higher (Fig. 4K and L). An analysis of the GPX4 mRNA levels also revealed that its expression was reduced by PPVI, whereas it was increased in the PPVI + Fer-1 group (Fig. 4M). These results suggest that Fer-1 reverses PPVI-induced mitochondrial damage and the ferroptosis of HCC cells.

The present study then examined the effects of PPVI on the invasion and migration of HCC cells. Following treatment of the HCCLM3 and Huh7 cells with PPVI or PPVI combined with Fer-1, the PPVI + Fer-1 combination exhibited a lower ability to inhibit colony formation than PPVI alone, indicating that Fer-1 reversed cell death induced by PPVI (Fig. 5A and B). The wound healing assay revealed that PPVI markedly inhibited the wound-healing ability of the HCC cells compared with the control group, whereas the PPVI-induced inhibition of the wound healing ability was markedly reversed by Fer-1 (Fig. 5C and D). Transwell assays revealed that PPVI markedly inhibited the number of invasive and migratory cells, and Fer-1 reversed this motility of the cells (Fig. 5E and F). As epithelial-mesenchymal transition (EMT) is one of the most critical mechanisms involved in cell migration and invasion, the effects of PPVI on the progression of EMT were investigated. The protein expression levels of E-cadherin, N-cadherin and Vimentin were examined using western blot analysis. The results revealed that PPVI increased the protein expression levels of E-cadherin in the HCCLM3 and Huh7 cells, whereas the levels of N-cadherin and Vimentin were decreased (Fig. 5G and H). The protein expression levels of E-cadherin were decreased in the PPVI + Fer-1 group, and the expression of N-cadherin and Vimentin was increased compared with that in the PPVI group (Fig. 5G and H). The results of immunofluorescence staining also revealed a similar result as those obtained for protein expression (Fig. 5I and J). These results thus suggest that PPVI inhibits the invasion and migration of HCC cells induced by the EMT mechanism.

To examine the mechanisms by which PPVI induces the ferroptosis of HCC cells, a network pharmacology-based strategy was used to obtain PPVI targets in HCC by taking the intersection of the utility targets of PPVI and the targets in liver cancer. These target proteins were analyzed using a PPI network to obtain 42 nodes and 400 edges (Fig. 6A). The target proteins in the PPI network were then selected according to the criteria of degree ≥24, betweenness centrality ≥0.011, and closeness centrality ≥0.707, and nine core target proteins were finally identified, namely PIK3CA, JUN, FGF2, VEGFA, MTOR, CASP3, SRC, TNF and STAT3 (Fig. 6A). PPVI was molecularly docked with the aforementioned target proteins, and the binding energy data indicated that PPVI had a strong binding activity with STAT3 protein (Fig. 6B).

The present study then examined whether STAT3 inhibition promoted the ferroptosis of HCC cells. The HCCLM3 and Huh7 cells were treated with 4 µM PPVI and STAT3 inhibitor (6 µM Stattic), respectively. The results of western blot analysis revealed that PPVI significantly inhibited STAT3 and GPX4 protein expression, and the phosphorylation also decreased accordingly, which followed the downward trend induced by Stattic (Fig. 6C-E). Notably, the levels of ROS, ferrous iron, MDA and GSH, and GPX4 protein expression (based on western blot analysis) also exhibited similar changes in the cells treated with PPVI or Stattic compared with the control group (Fig. S2). Moreover, the wound healing assay revealed that PPVI or Stattic inhibited the wound healing ability of the HCC cells (Fig. S1C). Furthermore, the results of immunofluorescence staining indicated that the nuclear translocation of p-STAT3 in the HCC cells was suppressed with the increasing PPVI concentrations (Fig. 6F and G). These results suggest that PPVI may regulate the protein expression of the GPX4 through STAT3. Subsequently, to investigate the mechanisms through which STAT3 regulates GPX4, STAT3 was overexpressed in the HCCLM3 and Huh7 cells, and it was confirmed that STAT3 increased the GPX4 protein and mRNA levels (Fig. 6H-J).

The present study then examined whether PPVI induces ferroptosis by regulating the STAT3/GPX4 axis in HCC cells. For this purpose, five groups were created, including one control group with nothing added. To reduce the influence of plasmid vectors on the experiment, an unloaded plasmid group and an unloaded plasmid + PPVI group were utilized. The remaining two groups were the overexpression STAT3 group and the overexpression STAT3 + PPVI group. After the HCC cells that overexpressed STAT3 were treated with 4 µM PPVI for 24 h, ROS and ferrous iron in the NT-Plasmid + PPVI group increased significantly; however, this effect was inhibited by STAT3 overexpression (Fig. 7A and B). Furthermore, the increase in MDA and the decrease in GSH levels induced by PPVI were also reversed by STAT3 overexpression (Fig. 7C and D). In addition, the PPVI-induced decrease in GPX4 expression was partially reversed by STAT3 overexpression (Fig. 7E and F). Morevoer, the phosphorylation ratio of STAT3 is presented in Fig. 7G. Similarly, the PPVI-induced decrease in the GPX4 mRNA levels was also reversed by STAT3 overexpression (Fig. 7H). Thus, PPVI induces ferroptosis by inhibiting the STAT3/GPX4 axis in HCC cells.

Subsequently, the present study investigated the mechanisms through which PPVI induces ferroptosis via regulation of the STAT3/GPX4 signaling pathway in vivo. Huh7 cells were injected subcutaneously into nude mice. Once the mice developed palpable tumors, the mice were treated for 10 consecutive days with intraperitoneal injections of PPVI, PPVI + Fer-1, or the same volume of saline (Fig. 8A and B). Fig. 8C shows that the weight of the mice in the saline group gradually decreased due to tumor consumption, while the body weight of the nude mice in the PPVI group decreased less than that in the PPVI + Fer-1 group. In the later period, the weight of the mice in the saline group exceeded that of the PPVI + Fer-1 group, which was deemed to be due to an increase in tumor net weight. PPVI suppressed tumor growth and the ferroptosis inhibitor, Fer-1, reversed this effect (Fig. 8D and E). Subsequently, liver function was examined by measuring the ALT and AST levels in mouse serum; the levels of PPVI and PPVI + Fer-1 were not significantly altered compared to the control group (Fig. 8F and G). Furthermore, H&E staining of liver and tumor tissues revealed that PPVI exerted no evident hepatic damage and was thus a safe therapeutic agent (Fig. 8H). The results of immunohistochemistry revealed that Fer-1 reversed the decrease in the expression of p-STAT3, STAT3 and GPX4 in the tumor mouse tissues (Fig. 8I). These data demonstrate the efficacy of PPVI in the treatment of liver cancer in vivo and indicate that PPVI exerts its effects via the STAT3/GPX4 axis.