Ox (cat. no. sc-215880) was purchased from Santa Cruz Biotechnology, Inc., whilst iohexol (Omnipaque^™), ethiodized oil (Lipiodol^®) and polyvinyl alcohol (PVA) foam embolization particles (Ivalon^®; 180–300 mm) were purchased from Guerbet Laboratories Ltd., B. Braun Melsungen AG and Cook Medical, respectively. All reagents were of chemical pure or analytical grade, and were used as purchased without any further purification.

New Zealand white rabbits (n=48; weight, 2.5-3 kg; 24 male and 24 female rabbits; age, 10–12 weeks) were obtained from Liaoning Changsheng Biotechnology Co., Ltd. All animal experiments were carried out according to the Chinese National Animal Law on the use of laboratory animals. All animal welfare considerations were taken, including efforts to minimize suffering and distress, use of analgesics or anaesthetics and special housing conditions [20°C; ventilation rate, 2–3 m³/h; 14/10-h light/dark cycle; humidity, 60–65%; pellet feed, 150 g/day, twice a day; automatic water supply system (free access)]. All rabbits were euthanized by intraperitoneal injection of 100 mg/kg pentobarbital sodium. At 20 min after the injection, death was confirmed by checking respiration, heartbeat, pupil and nerve reflex. All procedures were approved by the Animal Care and Use Committee of Huazhong University of Science and Technology (Wuhan, China; 2023; approval no. 3380).

All interventional procedures were performed by a radiologist with ≥15 years of experience. The rabbit VX2 liver tumor model was established as previously described (24). A 12-week-old tumor-bearing male rabbit was purchased from the Animal Center of Huazhong University of Science and Technology; the VX2 tumor was collected at 14 days since xenograft implantation, and the tumor was collected after the rabbit was sacrificed by intraperitoneal injection of 100 mg/kg pentobarbital sodium. At 20 min after the injection, death was confirmed by checking respiration, heartbeat, pupil and nerve reflex. Then, the VX2 tumor was sliced into 1.0 mm³ pieces and, under gas anesthesia (3–4% isoflurane used for induction and 2% for maintenance), a piece of the VX2 tissue was implanted into the left liver lobe of the rabbits (24). To avoid infection, the wounds were topically treated with 2.0×10⁴ IU gentamycin, while rabbits were also administered an intramuscular injection of 1.0×10⁴ IU ampicillin (25). No adverse effects were observed after injection of ampicillin.

A total of 24 VX2-tumor-bearing rabbits were randomly divided into the following four groups (n=6 rabbits/group; treatment duration, 14 days): Control group, normal saline was injected into the tumor-feeding arteries (control group); TACE group, TACE was performed via delivering a mixture of 0.3 ml Lipiodol mixed with 6 mg doxorubicin into the tumor-feeding arteries. The arteries were then occluded with PVA foam embolization particles; Ox group Ox, the tumor-feeding arteries were injected with an intra-arterial bolus of Ox saline (5 mg/kg; 5 min); and TACE + Ox group, the tumor-feeding arteries were injected with an intra-arterial bolus of Ox saline (5 mg/kg; 5 min), followed by TACE. When the VX2 tumor volume reached 1.61 cm³, TACE was performed in the TACE and TACE + Ox groups using a 2.7F microcatheter (Terumo Corporation), guided by digital subtraction angiography (Siemens Healthineers). The tumor-feeding arteries were continuously monitored during TACE until the tumor blood supply was completely occluded (no delivery of contrast agent was visible in tumoral and peritumoral vessels), with the normal hepatic artery unobstructed; the contrast agent used for angiography was Iohexol.

A 320-row spiral computed tomography (CT; Siemens Healthineers) was used for plain scan and contrast-enhanced CT. The contrast-enhanced CT was performed to assess the therapeutic effects of different treatments at 14 days. CT was repeated at days 3, 7, 10 and 14 to record the metastasis in the four groups. Tumor volume was calculated using CT (80 kV; 100 mA; 1-mm slice thickness; 1.1 pitch; 200×200-mm² field of view) with the following formula: V=a × b × c × π/6, where a, b and c indicate the anteroposterior, transverse and axial diameters, respectively. In addition, the survival of all treated rabbits (an additional 6 rabbits/group) was recorded. Animal health and behaviour were monitored every day and the rabbits were euthanized when the tumor weight reached ≤10% of their weight or when pain could not be effectively controlled during observation. The rabbits were considered to be in pain when they were abnormally vocal, or when they licked, bit, scratched or shook the painful area. The time of euthanasia or death was recorded, which was used to indicate mortality.

Blood samples were collected from the auricular vein of the rabbits under anesthesia (3–4% isoflurane used for induction and 2% for maintenance) in each treatment group (2 ml/collection; once daily on days 0, 1, 3, 7 and 14). The 14-day blood samples were collected before euthanasia. Subsequently, the samples were centrifuged (4°C, 1,000 × g, 10 min) to obtain serum, which was then stored at −80°C until use. The tumor and peritumoral tissues from each group were fixed in 4% formalin overnight at 25°C and paraffin-embedded at 52–54°C. The paraffin-embedded tissues were then cut into 5.0-µm slices for histological analysis.

To measure the levels of biochemical markers in serum in vivo, the blood samples from all four groups were collected and centrifuged at 1,000 × g for 10 min at 4°C to collect serum. The levels of the biochemical markers alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN) and creatinine (CRE) in serum were measured using a biochemical auto-analyzer (Model DXC8000; Beckman Coulter, Inc.).

For rehydration, tissue sections were immersed in a descending ethanol series. Subsequently, antigen retrieval in 1 mM EDTA (pH 8.0) was performed at 98°C for 15 min followed by washing twice in PBS and distilled water. For permeabilization, the Permeabilization Wash Buffer (cat. no. 40403ES64; Shanghai Yeasen Biotechnology Co., Ltd.) was used, after which, tissues were blocked in 10% normal goat serum (cat. no. C0265; Beyotime Institute of Biotechnology) for 20 min at 25°C. H₂O₂ (3%, 50 µl) was used to block endogenous peroxidase. The tissue sections were then incubated with anti-matrix metalloproteinase-9 (MMP-9; 1:1,000; cat. no. GB12132-100; Wuhan Servicebio Technology Co., Ltd.), anti-CD3 (1:100; cat. no. GB12014-100; Wuhan Servicebio Technology Co., Ltd.), anti-CD8 (1:100; cat. no. GB12068-100; Wuhan Servicebio Technology Co., Ltd.) and anti-vascular endothelial growth factor (VEGF; 1:150; cat. no. AB1876-I; MilliporeSigma) primary antibodies at 4°C overnight. Subsequently, the sections were incubated with an anti-mouse HRP (IgG H&L) secondary antibody (cat. no. ab97040; Abcam; 1:1,000) at 37°C for 1 h. DAB was used for chromogen detection. Tumor tissues also underwent hematoxylin and eosing (H&E) staining, as follows: Sections were stained with hematoxylin at 25°C for 15 min and eosin at 25°C for 5 min. In addition, the tumor necrosis ratio (TNR) was calculated using the following formula: TNR=N/(N + T), where N indicates the necrotic areas of the tumor and T indicates the living tumor areas of the tumor (detected by H&E assay). A total of six sections per tumor tissue were observed and the integrated optical density (IOD) sum in five randomly selected fields (magnification, ×400; light microscope) was calculated using ImageJ version 22 (National Institutes of Health). The mean values from the five randomly selected fields were calculated and used for statistical analysis.

Cell apoptosis was assessed using a Terminal Deoxynucleotidyl Transferase mediated dUTP Nick-End Labeling (TUNEL) assay (cat. no. 12156792910; MilliporeSigma) according to the manufacturer's instructions. In addition, a Ki67 assay (dilution, 1:200; Wuhan Servicebio Technology Co., Ltd.) was used to evaluate tumor cell proliferation according to the manufacturer's instructions. Tumor cell nuclei were stained with DAPI at 25°C for 20 min. In both fluorescence assays, images were captured by excitation at 488 nm and emission at 560 nm (BX53; Olympus Corporation). A total of six sections were assessed per tumor sample. The IOD sum of the images was calculated using ImageJ. The mean results from five randomly selected fields were calculated and statistical analysis was performed.

All statistical analyses were performed using SPSS v24.0 (IBM Corp.) and GraphPad Prism (version 8.0; Dotmatics) software. All data are expressed as the mean ± standard deviation. Kaplan-Meier analysis was used to plot the overall survival curves, and the significance was calculated using the log-rank test. Comparisons among multiple groups were performed using one-way ANOVA, followed by Tukey's Honest Significant Difference test to assess the differences between groups. P<0.05 was considered to indicate a statistically significant difference.

The liver tumor masses were clearly visible on CT images (Fig. 1A). Prior to treatment, tumor volumes were 1,628.5±20.3, 1,601.8±22.3, 1,608.9±29.8 and 1,630.2±19.2 mm³ in the control, TACE, Ox and TACE + Ox groups, respectively. No statistically significant difference in tumor volume was demonstrated among the different groups. The TACE procedure is shown in Fig. 1B and C. No rabbits were euthanized or found dead before the end of the experiment and all tumor tissues and serum samples were successfully collected.

A representative CT image from each treatment group at 14 days after the operation is shown in Fig. 2A. Tumor growth rate was notably reduced in the TACE (202.7±46.23%) and Ox (203.0±54.47%) groups, and significantly reduced in the TACE + Ox (96.8±20.31%) group compared with that of the control group (248.2±45.50%; Fig. 2B). Rabbits in the TACE + Ox group demonstrated a significantly improved tumor response compared with those in the TACE (P<0.001) and Ox (P<0.001) groups. No statistically significant difference was observed between the TACE and Ox groups (P>0.999; Fig. 2B). Representative tumor tissues of each group after treatment for 14 days are presented in Fig. 2C. The majority (94.7%) of tumor tissues in the TACE + Ox group were necrotic. Furthermore, compared with in the control group (28.12±8.91%), TNR was significantly enhanced in the TACE (60.25±12.25%; P<0.001), Ox (62.23±9.32%; P<0.001) and TACE + Ox (90.62±4.96%; P<0.001) groups. Notably, the combination demonstrated a significantly better TNR than that of the Ox group; the combination group demonstrated the greatest TNR (Fig. 2D and E). Subsequently, to assess the survival benefits of TACE, Ox and TACE + Ox therapy in VX2-bearing rabbits, the ability of the different treatments to promote survival was compared. Significant survival benefits were observed in rabbits in the combination group compared with those in the control (P<0.001), TACE (P<0.001) and Ox (P<0.001) groups (Fig. 2F). The number of metastases during follow-up is demonstrated in Fig. 2G. The metastases occurred at day 7 and gradually increased in the control and TACE groups. Conversely, Ox exhibited an inhibition of metastasis. The volume of metastasized tumors was 263.7±44.4, 342.1±32.2, 52.1±23.8 and 51.6±18.9 mm³ in the control, TACE, Ox and TACE + Ox groups after 14 days of treatments (P<0.001), repectively.

Semi-quantitative analysis of IHC staining revealed that the administration of Ox significantly enhanced the infiltration of CD3⁺ and CD8⁺ cells in tumor tissues compared with that in the control group. Although TACE alone also significantly increased the infiltration of the aforementioned cells compared with that in the control group, its capacity was notably attenuated compared with the Ox and TACE + Ox groups (Fig. 3A-C). In addition, although the application of TACE significantaly upregulated MMP-9 in the tumor tissues, treatment with Ox significantly decreased the expression of MMP-9 in the combination group compared with that in the control group (Fig. 3A and D). Moreover, the results indicated that Ox significantly decreased the expression of VEGF compared with that in the control group, and VEGF expression was significantly reduced in the TACE + Ox group compared with that in the control group (Fig. 3A and E). Furthermore, the immunofluorescence staining results demonstrated that the protein expression of Ki67 was significantly decreased in tumors in the TACE + Ox group compared with that in the control, TACE and Ox groups (Fig. 4A and C). Finally, the number of TUNEL⁺ cells was markedly increased in the TACE + Ox group compared with that in the other groups (Fig. 4A and D).

The changes in hepatorenal function are demonstrated in Fig. 5. Except from in the control group, rabbits in the TACE and combination groups had transient liver function impairment; the levels of AST and ALT were significantly higher than control on days 1, 3 and 7, and they then returned to normal levels at day 14 (Fig. 5A and B). However, kidney function was only slightly affected (Fig. 5C and D). The levels of ALT, AST, BUN and CRE reached their peaks on day 1 after treatment and gradually recovered (Fig. 5). These findings indicated that combination therapy may cause a transient impairment in hepatorenal function.