From January 2008 to December 2013, a total of 77 endoscopic biopsy specimens from patients who were diagnosed with esophageal squamous carcinoma and received chemoradiotherapy (CRT) were obtained at The Fourth Hospital of Hebei Medical University (Hebei, China). We acquired the approval from patients and the Ethics Committee of the Fourth Hospital of Hebei Medical University for the usage of the specimens for research. We sectioned 4-µm-thick slides from paraffin blocks, deparaffinized the slides in alcohol solutions with gradient concentrations, rehydrated the slides with distilled water, and then soaked the slides with citrate buffer and boiled them for 3 min for heat-induced antigen retrieval. After blocking the endogenous peroxidases with hydrogen peroxide solution for 15 min, normal goat serum (5%) was used to block non-specific antibody binding for 30 min, and then the sections were incubated with anti-HMGB1 monoclonal antibody (dilution 1:400; cat. no. ab79823; Abcam, Cambridge, MA, USA) at 4°C overnight. Then, goat anti-rabbit polyclonal antibody (dilution 1:100; cat. no. SP-9000; Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd., Beijing, China) and horseradish peroxidase-conjugated streptavidin working solution were each used to cover the histological sections at 37°C for 30 min successively. Subsequently, the DAB substrate-chromogen solution was applied to the sample to produce a color reaction. According to the scoring standard of previous studies (12,13), 2 pathologists who were blinded to the clinical parameters of the patients interpreted the immunohistochemical results for the final analysis.

TE13, KYSE180, KYSE30, YES2 and ECA109 human esophageal cancer cell lines were cultured in RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) with the addition of 10% fetal bovine serum (FBS; Invitrogen; Thermo Fisher Scientific, Inc.) and 1% penicillin/streptomycin (Gibco; Thermo Fisher Scientific, Inc.). The TE13 cell line is same with TE2, TE3, TE7 and TE12 (14), and it has been authenticated by STR profiling.

All cells were maintained in Thermo Fisher Scientific, Inc., CO₂-incubators with the temperature set at 37°C and the concentration of carbon dioxide set to 5%.

Esophageal tumor cells were irradiated by a 6-MV Siemens linear accelerator (Siemens, Buffalo Grove, IL, USA) with a source-skin distance (SSD) of 100 cm and dose rate maintained at 5 Gy/min. Then, the cells were collected at certain time-points for further study.

Cell transfection was performed with human HMGB1 small interfering RNA (shRNA) and negative control shRNA (PPL, Genebio Technology, Inc., Nanjing, China) using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) under the guidance of the manufacturer's instructions. The concentrated virus solution and esophageal cancer cells were co-cultured, and the fluorescence of the cells was observed by optical microscopy to confirm successful transfection. The sequences were as follows: HMGB1-shRNA sense, 5′-GGGAGGAGCAUAAGAAGAATT-3′ and antisense, 5′-UUCUUCUUAUGCUCCUCCCTT-3′; NC shRNA sense, 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense, 5′-ACGUGACACGUUCGGAGAATT-3′.

RNA was extracted from esophageal tumor cells with TRIzol reagent (Takara Bio, Inc., Shiga, Japan), and then reversed-transcribed to cDNA with a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Inc.). Then, RT-PCR was applied to analyze the expression of the HMGB1 gene using Platinum SYBR-Green qPCR SuperMix-UDG (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. First, the synthetic cDNA was denatured at 94°C for 30 sec, then used for the following cycle: Denaturing cDNA at 94°C for 5 sec, annealing cDNA at 56°C for 15 sec and extending cDNA at 72°C for 10 sec, 40 cycles in total. The GAPDH gene was used for normalization of RT-qPCR data. The 2^−ΔΔCT method was applied to analyze HMGB1 gene expression (15).

We conducted western blot analysis following a method previously described (16). The relevant antibodies were anti-HMGB1 (dilution 1:10,000; cat. no. ab79823), anti-γH2AX (dilution 1:1,000; cat. no. ab26350; Abcam), anti-MMP-2 (dilution 1:1,000; cat. no. 10373-2-AP), anti-MMP-9 (dilution 1:1,000; cat. no. 10375-2-AP), anti-p16 (cat. no. 10883-1-AP), anti-caspase-9 (dilution 1:1,000; cat. no. 10380-1-AP), anti-Bcl-2 (dilution 1:2,000; cat. no. 12789-1-AP), anti-CDK4 (dilution 1:2,000; cat. no. 11026-1-AP), anti-cyclin D1 (dilution 1:5,000; cat. no. 60186-1-lg), anti-Bax (dilution 1:5,000; cat. no. 50599-2-lg; ProteinTech Group, Inc., Chicago, IL, USA) and anti-β-actin (dilution 1:10,000; cat. no. AP0060; Bioworld Technology, Inc., St. Louis Park, MN, USA). The blotted protein bands were revealed by an Odyssey system (LI-COR Biosciences, Lincoln, NE, USA). Finally, the intensity of the protein bands was assessed with ImageJ (National Institutes of Health, Bethesda, MD, USA) and the ratio of the protein to corresponding β-actin was calculated to reflect the changes in expression levels. All western blot analyses were performed independently at least 3 times.

After preparing a cell suspension with a concentration of up to 5×10⁴ cells/ml, 100 µl/well was seeded into 96-well plates with 5 duplicates for each sample. At certain time-points after irradiation, the viability of the cells under the different treatments was determined by incubation with 10 µl Cell Counting Kit-8 (CCK-8) [MedChem Express (MCE) Princeton, NJ, USA] for 2 h. Then, the cell content was calculated by detecting the absorbance of each well at 450 nm via a Multiskan microplate. The experiment was repeated at least 3 times.

Esophageal cancer cells irradiated with a set of graded doses were cultured in 6-well culture plates for 15 days in triplicate. After the formation of cell clones, crystal violet (0.6%) was applied to stain the cells for 15 min. Colonies containing 50 cells or more were taken into account. Basic data was input into GraphPad Prism version 5.0 (GraphPad Software, Inc., La Jolla, CA, USA), and the biological parameters of radiation and cell survival curves were obtained.

Esophageal tumor cells were plated in 6-well plates. When the cells grew to confluence >80%, a straight cell-free zone was created by a 200-µl pipette tip/well. Then the scratched areas were marked as 0 h and photographed using computer-assisted fluorescent microscopy. Cells were cultured in 2 ml of media without FBS, then images were captured to observe the scratched areas at 16 and 32 h separately. The migration of cells with various treatments was presented as the percentage of the cell-free zone at the aforementioned time-points compared to 0 h and related to the control group.

The invasion ability of tumor cells was detected using Transwells coated with Matrigel (both from Corning Inc., Corning, NY, USA) before use. Cells were diluted to 5×10⁵/ml with serum-free medium and 200 µl was transferred to the top chambers. Medium containing 10% FBS was placed in the lower chamber. The cells on the upper side of the chamber were slightly wiped with a cotton swab after penetrating for 24 h. Concurrently, the cells passing through the filter were fixed with formaldehyde and stained with crystal violet for 10 min. Then, the number of stained cells was photographed and counted with a fluorescent microscope from 5 randomly selected fields (magnification, ×200).

Cell cycle analysis used esophageal tumor cells that were harvested and fixed with 70% precooled ethanol and maintained at 4°C. The following day, the cells were stained with propidium iodide probe solution (BD Biosciences, San Jose, CA, USA) in the dark for 30 min. Cell apoptosis was detected with an Annexin V-FITC apoptosis detection kit (BD Biosciences). In accordance with the kit's instructions, cells were collected and stained with propidium iodide (PI) and Annexin V-FITC. Flow cytometry was applied to detect the cell cycle distribution and apoptosis rates, encompassing both the early apoptosis (Annexin V⁺/PI⁻) and late apoptosis/necrosis (Annexin V⁺/PI⁺) phases.

Twelve 6-week-old male BALB/c nude mice (weight, 19±1 g) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China), and maintained in specific pathogen-free conditions with controlled temperature (23±2°C), humidity (55±5%) and light (12 h light/dark cycle). The mice were provided with sterile food and water ad libitum. All experiments with animals were carried out with the approval of the Animal Care and Use Committee of the Fourth Hospital of Hebei Medical University. A total of 4×10⁶ HMGB1-shRNA or NC tumor cells were injected into the left hind paw of the mice. Three weeks after injection, irradiation (5 Gy) was performed each day for 3 days with a collimator container to protect the normal tissue. Calipers was used to assess the tumor diameter (mm) twice a week, and the tumor volume (TV) was calculated by the following formula: TV = AB²/2, where A represents the long diameter and B represents the short diameter.

The data collected in the present study were analyzed using SPSS software package version 21 (SPSS, Inc., Chicago, IL, USA) and recorded as the mean ± standard deviation (SD). The endpoint of progression-free survival (PFS) was assessed from the beginning of CRT to progression, death or the date of the last follow-up. The survival analysis was carried out by Kaplan-Meier method with log-rank test and the χ² or Fisher's exact tests was used to analyze the association between clinical parameters and HMGB1 expression. The comparisons of data between different groups was performed using ANOVA with least significant difference (LSD) test. If the P-value of a two-sided statistic test was <0.05 or <0.01, the result was considered to be statistically significant.

The relationship between the expression level of HMGB1 and the clinical parameters of esophageal cancer patients was analyzed. Table I revealed that the associations between the expression level of HMGB1 and sex, age, lesion location, or N classification (P>0.05) had no statistical significance. Conversely, the accumulation of HMGB1 in esophageal cancer was significantly associated with gross tumor volume (GTV) (P=0.014), TNM stage (7th AJCC) (P<0.001), T classification (P=0.003), relapse (P=0.003) and distant metastasis (P<0.001). Immunohistochemical staining revealed that the expression of HMGB1 in ESCC samples (Fig. 1Ab, c and d) was higher in comparison with adjacent tissues with no tumor complication (Fig. 1Aa). After the generation of survival curves, the HMGB1 expression level in ESCC was significantly associated with overall survival (P=0.007; Fig. 1B) and progression-free survival (P=0.008; Fig. 1B), which were both significantly shorter in patients with positive expression of HMGB1 compared with patients with negative expression of HMGB1 according to the results of the log-rank test. In summary, statistical analysis revealed the critical role that HMGB1 played in the progression of esophageal cancer, and high levels of HMGB1 indicated poor clinical prognosis.

The expression of HMGB1 in the TE13, KYSE180, KYSE30, YES2 and ECA109 esophageal squamous carcinoma cell lines was determined in vitro by western blot analysis, and the results revealed higher expression levels of HMGB1 in the ECA109 and TE13 cell lines compared with the other cell lines (Fig. 1C). Thus, ECA109 and TE13 cells were chosen for further study, and the expression of HMGB1 after transfection with an shRNA targeting HMGB1 (HMGB1-shRNA) was determined through western blot and RT-qPCR analyses. Finally, no difference was revealed in HMGB1 expression between ECA109-NC or TE13-NC cells and their corresponding parental cells. However, HMGB1 expression in HMGB1-shRNA cells relative to their corresponding parental cells was significantly reduced (Fig. 1D and E). These data indicated that HMGB1-shRNA significantly downregulated HMGB1 in esophageal cancer ECA109 and TE13 cells.

To investigate the biological function of HMGB1 on the cell viability of esophageal cancer cells, cells under different conditions were collected at 24, 48, 72 and 96 h after transfection and irradiation. The results of the CCK-8 assay indicated that the proliferation rates of the HMGB1-shRNA groups were significantly lower at each time-point than those of the corresponding control and NC groups with or without irradiation (Fig. 2A). In addition, a colony formation assay was applied to detect the effect of the alteration of HMGB1 expression on the radiosensitivity of ECA109 and TE13 cells. After the cell survival curves were plotted based on the formed clones, and the radiobiological parameters of each group were input into statistical analysis, it was revealed that the radiosensitivity of the HMGB1-shRNA group was higher compared with those of the control and NC groups (Fig. 2B). To explore the function of HMGB1 in tumor formation in vivo, the HMGB1-shRNA and NC cells of the ECA109 cell line were implanted into nude mice. The results revealed that the tumor volume (Fig. 2C and D) and weight (Fig. 2E) in the HMGB1-shRNA group were notably smaller than those of the NC group before and after irradiation (P<0.05). These findings indicated that transfection with HMGB1-shRNA induced proliferation inhibition with or without irradiation and increased the radiosensitivity of esophageal cancer cells both in vitro and in vivo.

In the present study, whether HMGB1 was involved in radiation-induced DNA damage repair was assessed by analyzing the variations in HMGB1 and γH2AX expression after irradiation. The results of the western blot analysis revealed that ionizing radiation induced the protein expression of γH2AX and HMGB1 in a dose-dependent manner (Fig. 2F). In addition, the changes in expression of these 2 proteins were consistent over the course of time after irradiation (Fig. 2G), and both of them were highest at 1 and 2 h after exposure to 6 Gy irradiation, and then their expression gradually reduced. These findings prompted us to suppose that HMGB1 is associated with the generation of γH2AX. Then, the level of phosphorylated H2AX was detected to evaluate the influence of HMGB1 deficiency with and without irradiation. The results revealed that knockdown of HMGB1 protein in ECA109 and TE13 cells inhibited accumulation of γH2AX following irradiation treatment, while there were no notable changes in γH2AX concentrations in the non-irradiated groups (Fig. 2H). The western blot analysis demonstrated that HMGB1 deficiency inhibited the phosphorylation of H2AX induced by irradiation.

To observe the migration and invasion abilities of esophageal carcinoma cells, a wound-healing experiment and Transwell assay, respectively, were applied to ECA109 and TE13 cells. The results revealed that blocking the expression of HMGB1 significantly decreased the migration activity of ECA109 and TE13 cell lines both with (Fig. 3Ac, d, f and h) and without irradiation (Fig. 3Aa, b, e and g). In addition, the invasive ability of esophageal carcinoma cells transfected with HMGB1-shRNA was significantly inhibited compared with those of the control and NC groups (P<0.05) both with and without irradiation (Fig. 3B). These results indicated that HMGB1 deficiency decreased the migration and invasion abilities of esophageal tumor cells. Furthermore, to analyze the potential mechanism underlying the HMGB1-shRNA-mediated blockade of invasion and migration, MMP-2 and MMP-9 protein expression was examined, and the expression levels of both were reduced in the HMGB1-shRNA group before and after irradiation in comparison to those of the control and NC groups (Fig. 3C).

The apoptosis rate of the HMGB1-shRNA group was higher than that of the NC group before irradiation as evidenced by Annexin V and PI staining (Fig. 4A). In addition, the HMGB1-shRNA group exhibited more apoptosis after irradiation than the NC group in ECA109 and TE13 cells (Fig. 4A). To thoroughly assess the mechanism, the expression of apoptosis-related proteins was detected, including Bcl-2, Bax and caspase-9, before and after irradiation by western blot analysis. The results revealed decreased expression of Bcl-2 and increased Bax and caspase-9 levels in the HMGB1-shRNA group with and without irradiation compared with the NC group (Fig. 4B). These results revealed that downregulation of HMGB1 increased the apoptosis of esophageal carcinoma cells by affecting the expression levels of Bcl-2, Bax and caspase-9.

According to flow cytometric results, silencing of HMGB1 caused G₀/G₁ arrest of esophageal carcinoma cells with and without irradiation (Fig. 4C and D). A western blot assay that explored the expression levels of proteins related to the cell cycle revealed downregulation of cyclin D1 and CDK4 and upregulation of p16 in the HMGB1-shRNA groups with and without irradiation compared with the levels in the NC group (Fig. 4E). The cyclin D1 expression was lower in the irradiated group than the non-irradiated group. Furthermore, cyclin D1 expression was significantly decreased in the HMGB1-shRNA group with irradiation treatment. These data demonstrated that suppressing the expression of HMGB1 arrested esophageal tumor cells in the G₀/G₁ phase by upregulating the expression of p16 and decreasing the expression cyclin D1 and CDK4.