The present study was ethically approved by The First Affiliated Hospital of Zhengzhou University (Zhengzhou, China) and written informed consent was obtained from every participant. Brain tissue was collected post-surgery from both patients with intractable epilepsy (EP group) after resection, and the control brain tissue of patients with craniocerebral trauma induced intracranial hypertension. All brain tissue was frozen in liquid nitrogen.

IHC was performed as previous described (5). Brain tissue sections (5 µm) for immunohistochemical analysis were fixed in 4% paraformaldehyde (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) and paraffin-embedded. Sections of the brain regions were selected for the IHC study following fixing of the sections with acetone. Briefly, slides were incubated with 0.03% H₂O₂ in PBS for 10 min at room temperature, then incubated in 5% normal goat serum in PBS for 15 min. Sections were then incubated with primary antibodies: Anti-HMGB1 (cat. no. ab18256, 1:100; Abcam, Cambridge, UK), anti-NF-κB p65 (cat. no. ab16502, 1:5,000; Abcam) and anti-OX42 (cat. no. ab1211, 1:500; Abcam) in 3% bovine albumin serum in PBS at 4°C overnight. Immunoreactivity was detected using avidin-biotin-peroxidase technique; sections were incubated with 3, 3′-diaminobenzidine as the chromogen for 20 min at 37°C. Images were acquired using an upright optical microscope (DM4000; Leica Microsystems, Mannheim, Germany). Prior to mounting, the cell nuclei were counterstained by Hoechst 33342 fluorescent dye (1:1,000; Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA USA) for 10 min at 37°C. Samples without the addition of a primary antibody were used as the negative control. For each field, the number of positively stained capillaries were counted under a light microscope and the percentage of positive cells was calculated from the average value of five fields.

Total protein was extracted from brain tissue or cultured cells in different groups as described previously (24). Protein concentration was determined using a bicinchoninic acid protein assay kit. Proteins (30 µg/lane) were separated using 10% SDS-PAGE and transferred onto polyvinylidene difluoride membranes. The membranes were incubated overnight with the following primary antibodies: Anti-HMGB1 (cat. no. ab18256; 1:100; Abcam), anti-TLR4 (cat. no. 293072; 1:1,000; Santa Cruz, Biotechnology, Inc., Dallas, TX, USA), anti-RAGE (cat. no. AB15323; 1:1,000; Santa Cruz Biotechnology, Inc.), anti-NF-κB p65 (cat. no. ab16502; 1:5,000; Abcam), anti-inducible nitric oxide synthase (cat. no. 49055; iNOS; 1:200; Abcam) and anti-β-actin (cat. no. A1978; 1:50,000; Sigma-Aldrich; Merck KGaA) at 4°C. Subsequently, the membranes were incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit IgG secondary antibody (cat. no. 6721; 1:5,000; Abcam) at 37°C for 30 min and exposed to X-ray film using an enhanced chemiluminescence system (Thermo Fisher Scientific, Inc.). The intensity of the bands was measured using Lab-works (version 4.0; Ultra-Violet Products Ltd., Cambridge, UK).

Human microglia (HM) cells were obtained from Dr. Dai Antibody's laboratory (Case Western Reserve University, Cleveland, OH, USA) and were cultured as described previously (25). HM cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.), 100 U/ml penicillin-G, 100 µg/ml streptomycin and 0.01 M Hepes buffer (Thermo Fisher Scientific, Inc.) in a humidified 5% CO₂ atmosphere at 37°C.

HM microglia were stimulated with epileptic agent coriaria lactone (CL), and the expression levels of HMGB1, TLR4, RAGE and p-NF-κB p65 were detected by western blotting after collecting cells.

Full-length human HMGB1 cDNA was amplified by polymerase chain reaction from pCMV4-RelA plasmid (Shanghai GenePharma Co., Ltd., Shanghai, China) using the forward primer 5′-GGTCGGTACCATGGACGAACTGTTCCCCCT-3′ and the reverse primer 5′-CCATCTCGAGTTAGGAGCTGATCTGACTCA-3′, inserted into a pcDNA3.1 vector (Shanghai GenePharma Co., Ltd.) tagged with FLAG. HMGB1 small interfering (si)RNA (HMGB1-KD) and its control siRNA were purchased from Santa Cruz Biotechnology, Inc. Transient transfection of HM cells with pcDNA3.1/HMGB1 cDNA or control pcDNA3.1, and its control siRNA was carried out using Lipofectamine 2000 reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol.

The expression levels of cytokines IL-1 (cat. no. 68616), IL-6 (cat. no. 68834), TNF-α (cat. no. 55845), TGF-β (cat. no. 31366) and IL-10 (cat. no. 68540) in the cell culture supernatant were detected by ELISA (MLBIO; Shanghai Enzyme-linked Biotechnology Co., Ltd., Shanghai, China). A 96-well plate was coated with the primary monoclonal antibody of IL-1, IL-6, TNF-α, TGF-β or IL-10. Standards and supernatant samples were added to the wells. Following washing with 5% non-fat dry milk in PBS with 0.05% Tween-20 (Sigma-Aldrich; Merck KGaA), the wells were incubated with HRP-conjugated anti-rabbit IgG antibody (cat. no. 6721; 1:1,000; Abcam) for 10 min at 37°C. The plate was washed again with the aforementioned solution, followed by incubation with tetramethylbenzidene solution for 15 min at 37°C. This reaction was stopped by adding hydrochloride solution. Optical density was measured by using a microplate reader (Bio-Rad 550; Bio-Rad Laboratories, Inc., Hercules, CA, USA) set to 450 nm.

The potential toxic effect of HMGB1 on HM cells was analyzed using the 3-(4, 5-dimethylthylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method. Cells were seeded into 96-well plates at a concentration of 8×10³ cells per well. After being grown for 24 h, cells were incubated at 37°C with either HMGB1 (10, 100 and 500 ng/ml) or untreated control medium (PBS), respectively, for different periods of time (4, 8, 16 and 24 h). Subsequently, 50 µl MTT tetrazolium salt (Sigma-Aldrich; Merck KGaA) was diluted with PBS to a concentration of 5 mg/ml, then the mixture was added to each well and incubated for another 2 h at 37°C. The medium was then aspirated from each well and 200 µl dimethyl sulfoxide was added to dissolve the formazan crystals. The resulting solution (150 µl) was transferred to a 96-well plate and the absorbance of each well was determined using a Tecan infinite M200 Pro plate reader at a wavelength of 570 nm. Each experiment was replicated three times. The potential toxic effect of CL on HM cells was also analyzed with an MTT assay.

The EdU assay is a method of detecting proliferating cells with the advantage of being simple and convenient, avoiding DNA denaturation and maintaining the shape of the sample. It has replaced the traditional method of bromodeoxyuridine staining. It is a method of detection of DNA synthesis in proliferating cells and relies on the incorporation of labeled DNA precursors into cellular DNA during the S phase of the cell cycle (1). EdU staining was conducted using Cell-Light^™ EdU kit (Guangzhou RiboBio Co., Ltd., Guangzhou, China), according to the manufacturer's protocol. Paraffin sections were de-paraffinized in xylene for 10 min twice, rinsed in an alcohol gradient (100, 95, 85%) for 10 min, and rinsed in deionized water for 5 min. After washing with 2 mg/ml glycine solution diluted in double distilled water for 10 min, the sections were permeabilized with 0.5% Triton X-100 in PBS for 20 min, and then washed twice with PBS for 10 min. The Apollo reaction buffer liquid, catalyst, fluorescent dyes and buffer additives (Guangzhou RiboBio Co., Ltd.) were dissolved in deionized water, and shaken to make the Apollo 567 staining reaction solution (Guangzhou RiboBio Co., Ltd.). The sections were then incubated for 30 min without light. The sections were washed twice with PBS for 10 min at 37°C. For subsequent DNA staining, sections were counterstained with Hoechst 33342 for 30 min away from light at 37°C. The slides were then washed twice with PBS for 3 min, and observed immediately under a fluorescent microscope (magnification, ×400). All the procedures were conducted at room temperature. EdU-positive cells were detected by an Olympus BX51 microscope (Olympus Corporation, Tokyo, Japan). Images of the ApolloW 567 were captured with a ‘red’ filter set: Excitation, 550 nm; emission, 565 nm; filter, 555±15 nm. Images of the Hoechst 33342 were captured with a ‘blue’ filter set: Excitation, 350 nm; emission, 461 nm; filter, 405±15 nm.

All data are expressed as the mean ± standard deviation. The statistical analyses were performed using GraphPad Prism 6.0 software (GraphPad Prism Inc., USA). Comparison between two groups was performed by unpaired Student's t-test. Comparison among multiple groups was performed by one-way analysis of variance with a least significant difference post-hoc test. P<0.05 was considered to indicate a statistically significant difference.

To explore the expression level and changes in the distribution pattern of HMGB1 in epileptic brain tissues, resection of brain tissue after surgery was collected. The distribution of HMGB1, NF-κB p65 and OX42 in the control and EP group was observed using IHC. In the control group, HMGB1 immunoreactivity was detected mostly in the nuclei of the glial cells and pyramidal neurons. In the EP group, cytoplasmic staining of HMGB1 was significantly increased in the pyramidal neurons (Fig. 1). Next to glial cells with nuclear staining, glial cells with both nuclear and cytoplasmic staining were also present. These results indicated the nuclear to cytoplasmic translocation of HMGB1 in neurons and glial cells, and predicted its subsequent release into the extracellular space. Higher expression levels of NF-κB p65 were detected in the EP group in comparison with the control group. Positive NF-κB p65 staining was detected mainly in the cell nucleus in the EP group, suggesting the cytoplasmic to nuclear translocation of NF-κB p65 in neurons and glial cells. Next, IHC staining of the brain tissues of patients was conducted, using a marker for reactive microglia (OX42). The results reflected that the cell density of the two was similar.

To investigate the expression level of HMGB1, TLR4, RAGE, NF-κB p65 and iNOS in the control and the EP brain tissues, western blotting was performed. From the result, HMGB1 was notably increased, and the TLR4 and RAGE signaling pathway was markedly activated; their expression levels were both significantly increased. Both genetic and biochemical data support that TLR4 and RAGE signaling pathway could eventually lead to the activation of NF-κB (23). Thus, in the present study, upregulation of NF-κB was observed (Fig. 2).

Comparing the level of cytokines of IL-1, IL-6, TNF-α, TGF-β and IL-10 in the cell supernatant of the control and the CL-induced epileptic cell model. To determine the levels of cytokines of IL-1, IL-6, TNF-α, TGF-β and IL-10 in the cell supernatant of the control and the CL-induced epileptic cell model, ELISA was used to determine the concentration of each cytokine. From the result (Fig. 3), CL effectively enhanced the level of the all the above cytokines, and induced overexpression of HMGB1.

Western blotting was used to detect expression of HMGB1, TLR4, RAGE and NF-κB p65 in HM cells. From the result (Fig. 4A), CL appeared to increase the levels of HMGB1, TLR4, RAGE, NF-κB p65 and iNOS in HM cells; overexpression of HMGB1 enhanced this result (Fig. 4B), and inhibition of HMGB1 with HMGB1-KD blocked the function of CL to HM cells.

As presented in Fig. 5, the overexpression of HMGB1 did not significantly decrease the viability of HM cells at any tested concentrations or time points, compared with the control cells. Additionally, incubation with CL followed by overexpression also had no effect on the viability of HM cells.

To determine the proliferation of HM cells, an EdU assay was used. Similar to the MTT results, there were no significant differences among each group (Fig. 6), which reflects that HMGB1 and CL had no significant effect on the proliferation of HM cells.