The orthotopic liver tumor model was established in nude mice (male, 4–6 weeks old, 18–24 g) obtained from the Lab Animal Unit of the University of Hong Kong (22). MHCC-97L cells (6×10⁵; Liver Cancer Institute, Fudan University, China) in 0.1 ml culture medium were injected subcutaneously into the right flank of the nude mice. Once the subcutaneous tumors reached 0.8 to 1 cm in diameter, they were removed and cut into cubes 1 to 2 mm³ in size, which were subsequently implanted into the left liver lobes of another group of nude mice. Mice were housed at room temperature (20–25°C) in a standard animal laboratory with free activity and access to water and food. The mice were maintained under constant environmental conditions with a 12-h light-dark cycle. All operations were performed in clean conditions. The present study was licensed according to Animal Ordinance Chapter 340 by the Department of Health, Hong Kong Special Administrative Region (ref. 13–501 in DH/HA&P/8/2/3 Pt. 52).

Two weeks after tumor implantation, a single dose of SL7207 (5×10⁷ CFU; lab stock; Stanford University School of Medicine, Stanford, CA, USA), YB1 (5×10⁷ CFU; modified from SL7207) (20) or saline (control group; Thai Otsuka. Pharmaceutical Co., Ltd., Bangkok, Thailand) was injected via the tail vein. Five mice each from the SL7207 and YB1 groups were used to observe the effect of bacteria on mouse survival. The control group mice (n=9) and the remaining nude mice from the YB1 group (n=14) were used to monitor liver tumor growth and metastasis using a Xenogen in vivo imaging system (PerkinElmer, Inc., Waltham, MA, USA) through the detection of luminance signals from tumor cells. The mice were sacrificed 3 weeks after the YB1 or saline injection. The tumors, livers and lungs were sampled for further research.

The histological changes were assessed by H&E staining (Vector Laboratories, Inc., Burlingame, CA, USA), and the infiltration of neutrophils was assessed by IHC. The tissue samples were fixed in 10% formalin and embedded in paraffin (Sigma-Aldrich, St. Louis, MO, USA). Paraffin sections were dewaxed with xylene (Merck Millipore, Darmstadt, Germany), rinsed in graded alcohol (Merck Millipore) and rehydrated in water, and subsequently were stained with H&E for histological examination. The distribution of YB1 in liver tissue was detected by IHC staining. The initial step was the same as H&E staining, but the paraffin sections then underwent antigen retrieval with citrate buffer (pH 6.0; Abcam, Cambridge, UK). Subsequently, the sections were blocked using 10% fetal bovine serum (Life Technologies Ltd.; Thermo Fisher Scientific, Inc., Waltham, MA, USA) for 30 min, and rat anti-mouse monoclonal primary antibody (dilution, 1:200; catalog no., 560454; BD Biosciences, Franklin Lakes, NJ, USA) was applied and incubated for 1 h. Following incubation, the sections underwent blocking with 3% peroxidase (Dako, Glostrup, Denmark) for 30 min, and then goat anti-rat polyclonal secondary antibody (dilution, 1:100; catalog no., 559286; BD Biosciences) was applied and incubated for 30 min. The details of the H&E and IHC staining processes are described in a previous study by the same group (22).

Human HCC cell lines MHCC-97L and PLC (Japanese Cancer Research Bank; National Institutes of Biomedical Innovation, Health and Nutrition, Osaka, Japan) were maintained in high-glucose Dulbecco's modified Eagle's medium (Life Technologies Ltd.; Thermo Fisher Scientific, Inc.) supplemented with 10% heat-inactivated fetal bovine serum (Life Technologies Ltd.; Thermo Fisher Scientific, Inc.), 100 mg/ml penicillin G and 50 µg/ml streptomycin (Life Technologies Ltd., Thermo Fisher Scientific, Inc.) at 37°C in a humidified atmosphere containing 5% CO₂. MHCC-97L cells were labeled via stable transfer of the luciferase gene in chromosomes. Briefly, cells were transfected with pGL3 vector (Promega Corporation, Madison, WI, USA), and positive clones were selected according to luciferase activity in the Xenogen IVIS 100 imaging system (PerkinElmer, Inc., Waltham, MA, USA).

YB1-incubated MHCC-97L and PLC cells were prepared and co-cultured at a ratio of 1,000:1 for 2 h under anaerobic (O₂<0.5%) or aerobic conditions. Extracellular bacteria were then removed by washing with phosphate-buffered saline (USB Corporation; Affymetrix, Inc., Santa Clara, CA, USA) and the cells were further cultured with gentamycin medium (Life Technologies Ltd., Thermo Fisher Scientific, Inc.) for 24 h. Cells were fixed in paraformaldehyde (4%) and stained with a rabbit polyclonal anti-Salmonella antibody (dilution, 1:500; catalog no., ab35156; Abcam) for 12 h at 4°C. Goat anti-rabbit polyclonal immunoglobulin G Alexa Fluor^® 488-conjugated (dilution, 1:100; catalog no., ab150077) and polyclonal horseradish peroxidase-conjugated (dilution, 1:100; catalog no., ab6721) secondary antibodies (Abcam) were added and the cells were incubated for 1 h at room temperature. Then CytoPainter Phalloidin-iFluor 488 Reagent (dilution, 1:1,000; catalog no., ab176753; Abcam) was applied to indicate cell boundaries. CytoPainter Phalloidin-iFluor 555 Reagent (dilution, 1:1,000; catalog no., ab176756; Abcam) was also applied to demonstrate cell boundaries. The cells were observed under a confocal microscope (LSM 710; Carl Zeiss AG, Oberkochen, Germany).

Cancer cell apoptosis and death induced by YB1 under anaerobic conditions was detected using an Annexin V-propidium iodide (PI) kit (BioVision, Inc., Milpitas, CA, USA) according to the manufacturer's protocol. As demonstrated using flow cytometry (FACSCalibur; (BD Biosciences) Annexin V1/PI- cells were apoptotic and Annexin V1/PI1 cells were dead.

Continuous variables were expressed as the median with a range. The t-test was used for statistical comparison. The χ² test was used to compare the incidence of lung metastasis. P<0.05 was considered to indicate a statistically significant difference. Calculations were performed using the SPSS computer software version 16 (SPSS, Inc., Chicago, IL, USA).

Orthotopic liver tumor-bearing mice were treated with parental strain SL7207 or engineered strain YB1 for 3 weeks. All mice in the SL7207 group died within 11 days of treatment. All of the mice were alive and healthy 3 weeks after YB1 injection (Fig. 1). To monitor the healthy condition of YB1 treated mice, the body weights were recorded after injection. The results revealed that the body weight decreased by ~9% on day 1 after YB1 injection and recovered afterwards (Fig. 1B). Notably, we observed that YB1 had low-level distribution in the tumors. This result is different from that observed in our previous study in a breast tumor model (20). On day 21 after treatment, only low amounts of YB1 were identified in the shrunken tumor by paraffin section staining (Fig. 1C). In addition, YB1 was entirely eliminated from the kidney, lymph nodes and spleen 3 weeks subsequent to injection. YB1 was undetected in the blood throughout the experiment.

The effect of YB1 on liver tumor growth and metastasis was studied using an orthotopic liver tumor model in nude mice. Tumor-bearing mice were grouped randomly, and there were no significant differences in the average size of liver tumors between the treatment and control group when the experiment started. Longitudinal monitoring of tumor growth and metastasis was achieved by Xenogen IVIS and the tumor final volume was measured at 3 weeks after treatment. The results indicated that YB1 significantly reduced the size of the primary liver tumor from 1 week to 3 weeks after treatment (Fig. 2A-B). After 3 weeks, the tumors in the control and YB1-treated groups demonstrated significant differences (580.1 mm³ vs. 61.2 mm³; P=0.000; Table I). Lung metastasis was totally repressed by YB1 treatment (0 of 14), however, the control group still demonstrated a 55.6% (5 of 9; P=0.004) lung metastasis rate (Table I). These results were additionally confirmed by histological examination (Fig. 2C).

During inflammation of bacterial treatment, circulating neutrophils are recruited to the site of inflammation, which releases a number of various cytokines and enhances the innate immune response (19). In order to investigate the mechanism of YB1 suppression of tumor growth and metastasis, the present study compared the infiltration of neutrophils in the liver between the YB1 treatment and control group. In the control group, only a small number of Gr1-positive cells were detected during the entire experiment (Fig. 3). Large amounts of Gr1-positive cells were detected on day 1 to day 21 following YB1 treatment (Fig. 3).

To test the interaction between YB1 and liver cancer, MHCC-97L and PLC cell lines were incubated with YB1 or SL7207 in vitro under anaerobic conditions (oxygen level below 0.5%). Following the removal of extracellular bacteria and further culturing, it was demonstrated that SL7207 and YB1 effectively invaded the liver cancer cells (Fig. 4A). The effect of SL7207 and YB1 on tumor cells was further investigated by measuring tumor cell apoptosis and death. The results of the present study revealed that an increasing number of dying and apoptotic MHCC-97L cells were detected following co-culture with SL7207 or YB1, relative to the control. The YB1 and SL7207 groups exhibited significant damage to cancer cells under hypoxic conditions (Fig. 4B). Similar results were observed in the PLC cell line (Fig. 4B).