The present study was approved by the Ethics Committee of the Affiliated Hospital of Chengde Medical University, and all the patients provided written informed consent. A total of 73 patients (31 female and 42 male; mean age, 55.24 years; age range, 26-75 years) who were pathologically diagnosed with NSCLC were enrolled into the present study between January 2017 and December 2018, among. The patients were recruited based on the following inclusion criteria: i) Aged <70 years old, and with NSCLC, confirmed by histopathology; ii) had an expected survival time >3 months; iii) had received chest radiotherapy for the first time, without concurrent chemotherapy or immunomodulator treatment; iv) had a Karnofsky performance score >70; and v) exhibited normal heart, liver and kidney function. All patients received radiotherapy with intensity-modulated radiation therapy on an Elekta Synergy linear accelerator, and the total radiation dose was 60 Gy (2 fractions at 30 Gy).

Fasting venous blood was collected from the patients 1 week prior to, and 2, 4 and 6 weeks after radiotherapy. The blood samples were centrifuged at 700 x g for 10 min at -4˚C to isolate the serum, which was then stored at -80˚C for subsequent analysis.

The HMGB1 concentrations in the blood samples were determined using an HMGB1 ELISA kit (cat. no. ARG81351; arigo Biolaboratories Corp.) according to the manufacturer's instructions.

The NSCLC cell lines, A549 and H1299, were obtained from the Shanghai Institute of Cell Biochemistry and Cell Biology (Shanghai, China). The cells were cultured in DMEM supplemented with 10% FBS (both Invitrogen; Thermo Fisher Scientific, Inc.) at 37˚C and a CO₂ concentration of 5%.

NSCLC cell irradiation was conducted at room temperature using a 6-MV linear accelerator (Siemens AG), with a source skin distance of 100 cm and an irradiation dose rate of 200 cGy/min.

The miR-107 mimics (5'-AGCAGCAUUGUACA GGGCUAUCA-3'), mimic negative control (NC; 5'-UUCUCC GAACGUGUCACGUTT-3'), miR-107 inhibitor (5'-UGAUAG CCCUGUACAAUGCUGCU-3') and inhibitor NC (5'-CAGUA CUUUUGUGUAGUACA-3') were synthesized and purchased from Shanghai GenePharma Co., Ltd. Preliminary experiments were carried out using A549 and H1299 cells transfected with 50 nM miR-107 mimics, inhibitor or the relevant NCs.

To knockdown HMGB1, shRNA HMGB1 or a scrambled shRNA sequence was sub-cloned into pGPU6/Neo plasmids (Shanghai GenePharma Co., Ltd.), referred to as sh-HMGB1 or scrambled control. A549 and H1299 cells were transfected with sh-HMGB1 or the associated scrambled control.

All transfections were performed using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. After transfection for 24 h at 37˚C, the transfection efficiency was verified using reverse transcription-quantitative PCR (RT-qPCR).

Total RNA was extracted from blood using the miRNeasy Mini Kit (Qiagen GmbH), and from NSCLC cells using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) per the manufacturers' instructions. Then, the total RNA was reverse transcribed into cDNA using the miScript II RT Kit (Qiagen GmbH) or PrimeScript™ RT Master Mix (Takara Bio, Inc.) according to the manufacturers' instructions. qPCR was performed using the SYBR Premix Ex Taq II PCR kit (Takara Bio, Inc.) according to the manufacturer's protocol. The thermocycling conditions were as follows: 95˚C for 5 min, followed by 35 cycles of 95˚C for 20 sec, 60˚C for 30 sec, and 72˚C for 20 sec, with a final extension at 72˚C for 5 min. The relative levels of miR-107 and HMGB1 expression were calculated using the 2^-ΔΔCq method (21), and U6 and GAPDH served as the internal controls for miRNA and mRNA expression, respectively. The qPCR primer sequences were as follows: miR-107 forward, 5'-GCCGAGAGCAGCAUUGUACA-3', and reverse, 5'-CTCAACTGGTGTCGTGGA-3'; HMGB1 forward, 5'-CCA GGGGCTTTTCTACCAGG-3' and reverse, 5'-AACCCCAAC AAACTGCTGGA-3'; U6 forward, 5'-GTGCTCGCTTCGGCA GCACATATAC-3' and reverse, 5'-AAAAATATGGAACGC TCACGAATTTG-3'; and GAPDH forward, 5'-ACCCAGAAG ACTGTGGATGG-3' and reverse, 5'-CACATTGGGGGTAG GAACAC-3'.

A549 and H1299 cells were seeded into 96-well plates at a density of 5x10³ cells/well and incubated at 37˚C. A total of 10 µl CCK-8 solution (Dojindo) was added to each well at 0, 24, 48 and 72 h. After incubation for 2 h, the optical density value was measured with a microplate reader at a wavelength of 450 nm.

NSCLC cells were irradiated with a set of graded doses (0, 2, 4, 6 and 8 Gy) and then cultured in 6-well plates for 14 days. The medium was replaced with fresh medium every 3 days. After the formation of cell clones (colonies containing ≥50 cells), the cells were stained with 0.5% crystal violet solution (Sigma-Aldrich; Merck KGaA) at room temperature for 20 min, and counted under a light microscope (Olympus Corporation).

Transwell invasion assays were conducted to determine the invasive ability of radioresistant cells, using precoated Matrigel Transwell chambers (BD Bioscience) with 8-µm porous membranes (Corning, Inc.). Briefly, A549 cells were seeded into 24-well plates and then exposed to a 2-Gy dose of radiation. After irradiation, 5x10⁴ A549 cells resuspended in serum-free medium were seeded into the top chamber, and complete medium was added to the lower chamber as a chemoattractant. After 48 h, invasive cells were fixed with 4% paraformaldehyde for 10 min at room temperature, and then stained with 0.1% crystal violet at room temperature for 20 min. The number of invaded cells was counted in five random independent fields per well using a light microscope (Olympus Corporation).

A computer-based algorithm, the TargetScan database (www.targetscan.org), was used to predict potential miRNAs that could target HMGB1. In the present study, miR-107 was searched for in humans following the entry of the HMGB1 gene, after which a luciferase reporter assay was conducted. Briefly, partial sequences of the HMGB1 3'-UTR with wild-type (WT) or mutant (MT) miR-107-binding sites were cloned into the pmirGLO dual-luciferase vector (Promega Corporation) to generate a WT (HMGB1 3'-UTR-WT) or an MT (HMGB1 3'-UTR-MT) reporter plasmid. Then, A549 cells were transiently transfected with the reporter plasmids along with the miR-107 mimics, mimic NC, miR-107 inhibitor or the inhibitor NC using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.); firefly luciferase activity was evaluated ~48 h later using a Dual-Luciferase Reporter Assay System (Promega Corporation) and normalized to that of Renilla luciferase.

Proteins were extracted from A549 cells using radioimmunoprecipitation assay buffer with protease inhibitor cocktail (both Cell Signaling Technology, Inc.). Protein concentration was determined using a BCA kit (CoWin Biosciences). The proteins (15 µg/lane) were separated using 10% SDS-PAGE and transferred to nitrocellulose membranes (Amersham; Cytiva). The membranes were subsequently blocked with 5% skim milk at room temperature for 2 h. Proteins were detected by western blotting using rabbit anti-HMGB1 (1:400; cat. no. ab18256; Abcam), rabbit anti-Toll-like receptor 4 (TLR4; 1:400; cat. no. ab13556; Abcam), rabbit anti-NF-κB p65(1:400; cat. no. ab207297; Abcam), and rabbit anti-GAPDH antibodies (1:1,000; cat. no. ab8245; Abcam) and were incubated at 4˚C overnight. The membranes were then incubated with the relevant goat anti-rabbit HRP-conjugated secondary antibodies (1:5,000; cat. no. ab7090; Abcam) at room temperature for 1 h, and the protein bands were visualized using ECL reagents (Amersham; Cytiva) and quantified by densitometry (ImageJ software; National Institutes of Health; http://rsbweb.nih.gov).

Statistical analyses were performed using GraphPad Prism 5.0 (GraphPad Software, Ltd.) and SPSS 23.0 (IBM Corp.). The data are presented as the mean ± SD. Patient sample data were non-normally distributed and analyzed by Friedman analysis followed by Nemenyi's post hoc test. Differences between in vitro experimental groups were analyzed by one-way ANOVA followed by Tukey's post hoc test. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the diagnostic value of HMGB1 in identifying radiotherapy-sensitive patients with NSCLC. P<0.05 was considered to indicate a statistically significant difference.

The levels of HMGB1 were quantified in the serum of patients with NSCLC who had received radiotherapy. As shown in Fig. 1A, the expression levels of HMGB1 in the serum gradually declined at 2-, 4- and 6-weeks post-radiotherapy compared with those 1 week before radiotherapy (P<0.001).

Then, the expression levels of HMGB1 before radiotherapy were evaluated for the capacity to distinguish between radiotherapy-sensitive and -resistant patients with NSCLC. As shown in Fig. 1B, ROC curve analysis revealed an area under the curve of 0.861 (sensitivity, 74%; specificity, 87%; cutoff value, 91.68 ng/ml), suggesting that the expression of HMGB1 had the diagnostic accuracy for distinguishing between radiotherapy-sensitive and -resistant patients with NSCLC.

To investigate the biological function of HMGB1 in NSCLC cell viability, A549 and H1299 cells were transfected with HMGB1 shRNA (sh-HMGB1) to knock down HMGB1expression, and cells treated under different conditions were harvested following transfection and irradiation. RT-qPCR analysis was used to evaluate transfection efficiency. The results indicated that HMGB1 mRNA expression was significantly downregulated in HMGB1-knockdown A549 and H1299 cells compared with that of the control group (P<0.001, Fig. 2A). The results of the CCK-8 assay revealed that cellular proliferation following sh-HMGB1 transfection was significantly lower than that of the corresponding control cells, both without (Fig. 2B) and with (Fig. 2C) irradiation (P<0.05). These results suggested that radiation significantly suppressed tumor cell proliferation, and that this inhibitory effect was enhanced when combined with HMGB1-knockdown.

Furthermore, a colony formation assay was conducted to determine the effects of HMGB1 on the radiosensitivity of A549 and H1299 cells. Survival curves were plotted based on the formation of clones and the radiobiological parameters of each group, and the results indicated that compared with the control, HMGB1-knockdown enhanced radiosensitivity compared with the control (Fig. 2D and E). These results indicated that transfection with sh-HMGB1 inhibited cellular proliferation with or without irradiation, and increased the radiosensitivity of A549 and H1299 cells in vitro.

Previous studies have shown that miRNA dysregulation plays a crucial role in tumor progression by negatively regulating target genes via binding to their 3'-UTRs. Thus in the present study, bioinformatics analysis was performed using the TargetScan database to predict potential upstream regulators of HMGB1. Among the conserved candidates (miR-107, miR-103 and miR-142-3p), miR-107 was selected as a putative regulator of HMGB1, which was based on the results of a study indicating that miR-107 plays a tumor suppressor role by inhibiting tumor growth and metastasis in NSCLC (19). As shown in Fig. 3A, the HMGB1 3'-UTR contains a putative binding sequence for miR-107. The serum expression levels of miR-107 were also measured by RT-qPCR in patients with NSCLC before radiotherapy, and the correlation between miR-107 expression and HMGB1 levels was assessed. The results showed that HMGB1 level was negatively associated with miR-107 expression in the serum of patients with NSCLC (r=-0.5266 and P<0.001, Fig. 3B). The results also indicated that miR-107 mimics promoted the expression of miR-107, while inhibitors suppressed expression (P<0.001, Fig. 3C). Subsequently, dual-luciferase reporter assays revealed that when A549 cells were transfected with HMGB1 3'-UTR-WT, enhanced expression of miR-107 inhibited, while miR-107-knockdown increased luciferase activity (P<0.05, Fig. 3D). However, in the cells transfected with HMGB1 3'-UTR-MT, overexpression or downregulation of miR-107 had no significant effect on luciferase activity. Furthermore, the mRNA and protein levels of HMGB1 were determined in A549 cells transfected with miR-107 mimics or inhibitor. The results indicated that miR-107 overexpression decreased the mRNA and protein levels of HMGB1, and that miR-107-knockdown promoted HMGB1 protein (P<0.05, Fig. 3E) and mRNA (P<0.001, Fig. 3F) expression.

To further investigate whether the reduction in HMGB1-enhanced radiosensitivity was regulated by miR-107, A549 cells transfected with sh-HMGB1 or a combination of sh-HMGB1 and the miR-107 inhibitor were treated with or without radiation. The RT-qPCR results showed that sh-HMGB1 downregulated the expression levels of HMGB1, and that knocking down miR-107 reversed the sh-HMGB1-mediated decrease in HMGB1 expression (P<0.001, Fig. 4A). Knocking down miR-107 with an miR-107 inhibitor reversed the inhibition in cellular proliferation without (Fig. 4B) or with (Fig. 4C) irradiation, and reduced the survival fractions (Fig. 4D and E) induced by sh-HMGB1 (P<0.05). Furthermore, Transwell invasion assay results showed that HMGB1-knockdown suppressed A549 cell invasiveness, and that downregulating miR-107 reversed the sh-HMGB1-mediated decrease in the invasion abilities of A549 cells after irradiation (P<0.001, Fig. 4F and G). In summary, these results revealed that the knockdown of HMGB1 enhanced the radiosensitivity of A549 cells regulated by miR-107 expression.

Compared with the mock group, HMGB1-knockdown and irradiation downregulated the expression of TLR4/NF-κB signaling pathway related-proteins (P<0.05, Fig. 5). Compared with the mock-radiation group, the downregulation of HMGB1 enhanced radiation-mediated decreased expression of TLR4/NF-κB signaling pathway related-proteins in A549 cells (P<0.05).