The subjects involved in the present study were included healthy volunteers and patients with a diagnosis of sepsis, who were admitted to the intensive care unit (ICU) of Chongming Branch, Xin Hua Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China) within 12 h of onset. Sepsis was defined as SIRS caused by infection, which was confirmed by the presence of bacteria or clinically suspicious foci. Disease severity was classified as sepsis, severe sepsis or septic shock, according to the Surviving Sepsis Campaign (SSC) guidelines (20). The investigation period ran between August 2012 and July 2013. The inclusion criteria were also defined according to SSC guidelines (20): Patients with the following characteristics were excluded from the present study: (1) A history of coronary heart disease, chronic obstructive pulmonary disease, diabetes, hypertension, blood diseases, hyperlipidemia, autoimmune diseases, cancer, rheumatism/connective tissue disease, chronic renal insufficiency, liver cirrhosis; (2) unable to participate to the end of the clinical trial when serum samples were collected; (3) patients <18 years or >75 years of age; (4) smoking or (and) alcohol consumption; (5) patients with a medication history of >1 month. According to the above standards, 38 patients with sepsis (21 men and 17 women; age, 45±3.4 years) and 32 healthy volunteers (17 men and 15 women; age, 48±4.2 years) were enrolled in the present study, which was approved by the ethics committee of the ChongMing branch of the Xin Hua hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China). All the subjects provided written informed consent prior to commencement of the investigation.

Peripheral venous blood (5 ml) was collected from the subjects and healthy volunteers at 6, 12, 18, 24, 36, 48 and 96 h following hospital admission. All blood samples were subjected to natural clotting at room temperature, followed by centrifugal serum separation (2,000 × g for 8 min) and complement inactivation at 56°C. The aliquots were stored at −70°C until further use.

A human umbilical vein EC (HUVEC) line was purchased from Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). The cell line was originally isolated from human embryonic umbilical vein ECs. Ethyl pyruvate (EP), fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse and anti-rabbit secondary antibodies, and tetramethylrhodamine (TRITC)-phalloidin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Anti-vascular endothelial (VE)-cadherin antibodies (polyclonal goat anti-human; cat. no. sc-6458), anti-B-cell lymphoma 2 (BCL-2; monoclonal mouse anti-human; cat. no. sc-7382) and anti-BCL-2-associated X protein (BAX; monoclonal mouse anti-human; cat. no. sc-70407) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Polyclonal rabbit anti-human HMGB1 antibody (cat. no. LS-C31535-100) was purchased from LifeSpan Biosciences, Inc. (Seattle, WA, USA). Invitrogen Annexin V-FITC/propidium iodide (PI) was purchased from Thermo Fisher Scientific, Inc. (Waltham, MA, USA).

The HUVECs (1×10⁶ cells) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Biochrom AG, Berlin, Germany), 150 U/ml penicillin (Beyotime Institute of Biotechnology, Hangzhou, China) and 150 U/ml streptomycin at 37°C in an atmosphere containing 5% CO₂. Cell morphology was observed under a microscope (IX70; Olympus Corporation, Tokyo, Japan) and identified using factor VIII immunofluorescence staining (2%) with anti-factor VIII rabbit polyclonal primary antibody (Hua Yi Biotechology, Shanghai, China; cat. no. hy032974). Briefly, ECs were seeded onto slides and cultured for 24 h, and the slides were subsequently fixed with 4% paraformaldehyde at room temperature for 15 min. The slides were then washed with phosphate-buffered saline (PBS) and treated with 0.2% Triton X-100 (Liansuo Biological Technology Co., Ltd., Shanghai, China), prior to being blocked with 1% bovine serum albumin (BSA; LiRui Biological Technology Co., Ltd., Shanghai, China) at room temperature for 30 min, and incubated with rabbit anti-human factor VIII polyclonal primary antibody (100 µl; 1:50) at 4°C for 12–14 h. Following washing with PBS, the slides were incubated with PE-labeled mouse anti-rabbit IgG 100 µl (1:50) in the dark at room temperature for 1 h. Images of the slides were captured using a IX71 fluorescence microscope (Olympus, Tokyo, Japan). Once grown to 80% confluence, the cells were pretreated with EP (5 µM) in DMEM with 10% FBS for 24 h at 37°C. The culture medium was subsequently removed by washing twice with PBS (pH 7.4), and the cells were exposed to 20% septic serum obtained from patient blood samples, diluted in culture medium for 6 h at 37°C. The present study included the following experimental groups: A 20% septic serum-treated group and 20% septic serum + EP (5 µM) pre-treated group; and the following control groups: A 20% control serum-treated (healthy volunteers) group and a 20% control serum + EP (5 µM)-treated group. Cell viability was determined using trypan blue staining (40%; Beyotime Institute of Biotechnology).

The ECs (1×10⁵ cells) were grown on 3 µm pore Transwell filters (Costar^®, Corning Incorporated, Corning, NY, USA) until confluent, and were then transferred to starvation medium (Biochrom AG) containing 1% FBS for 2 h. FITC-dextran (Mr, 42,000; Sigma-Aldrich) was applied apically at 1 µg/ml, and allowed to equilibrate for 30 min at 25°C prior to 200 µl sample of the medium being removed from the lower chamber, to measure basal permeability. The monolayers were pretreated with EP (5 µM) for 24 h at 37°C, and were then either untreated (control) or were stimulated in triplicate with 20% septic serum for 6 h at 37°C. Samples (200 µl) were removed from the lower chamber for fluorescence measurements, which were compared with those from the control monolayers. The fluorescence intensity of FITC was detected using a fluorescence spectrometer (LS-50B; PerkinElmer, Waltham, MA, USA), operating at an excitation wavelength of 492 nm and a detection wavelength of 520 nm.

The ECs (1×10⁴ cells) were serum-starved for 2 h, pretreated with EP for 24 h, and then either remained untreated or were stimulated with 20% septic serum for 6 h. For F-actin staining, the ECs were fixed in 3.7% formaldehyde (Liansuo Biological Technology Co., Ltd.) at 4°C for 10 min. The cells were permeabilized in 0.2% Triton-X-100 for 5 min and then blocked in 1% BSA in PBS, prior to being incubated with TRITC-phalloidin (2 kµ/l) at room temperature in the dark for 1 h. For VE-cadherin staining, following dewaxing, samples were heated to 92–98°C for 15–20 min in 0.01 M sodium citrate buffer solution (pH 6.0), and then cooled to room temperature for 20–30 minutes. Finally, the slice was rinsed with distilled water and PBS, respectively. The ECs were fixed and blocked, followed by incubation with mouse anti-VE-cadherin antibody (1:100 dilution) overnight at room temperature. The culture medium was removed by washing twice with PBS (pH 7.4) and the ECs were subsequently exposed to FITC-labeled goat anti-mouse IgG secondary antibody (1:200 dilution; cat. no. sc-2010; Santa Cruz Biotechnology, Inc.) at room temperature in the dark for 1 h. Images were captured using a confocal laser scanning microscope (LSM410; Carl Zeiss AG, Oberkochen, Germany).

The ECs (3×10⁶ cells) were serum-starved for 2 h, pretreated with EP for 24 h, and either remained untreated or were stimulated with 20% septic serum for 6 h. The ECs were subsequently lysed with SDS sample buffer (Beijing Taize Technology Development Co., Ltd., Beijing, China). Protein concentration was determined using a Bicinchoninic acid assay method, and the supernatants (100 µl protein) were separated by 10% SDS-PAGE. The proteins were then transferred to nitrocellulose membranes (Mai Bio Co., Ltd., Shanghai, China), which were blocked with 10% non-fat dry milk in Tris-buffered saline and Tween-20 (Double Helix Biotechnology Co., Ltd., Shanghai, China), containing 20 mmol/l Tris (pH 8.0); 137 mmol/l NaCl and 10% Tween-20, and incubated for 12 h at 4°C with monoclonal mouse anti-human BAX, monoclonal mouse anti-human BCL-2 antibody, and polyclonal rabbit anti-human HMGB1 primary antibodies, and then incubated for 2 h at 25°C with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (cat. no. B1706516; ShangHaiyelibio) and HRP-conjugated goat anti-rabbit IgG (cat. no. 96692; Baijibio Beijing). Bound proteins were detected using enhanced chemiluminescence (Beyotime Institute of Biotechnology), according to the manufacturer's instructions. The chemiluminescence signal was detected using a Quantitative Gel and Western Blot Imaging system, and the results were analyzed using FluorChem Q software (ProteinSimple, San Jose, CA, USA).

The ECs were serum-starved for 2 h, pretreated with EP (5 µM) for 24 h, and then either remained untreated or were stimulated with 20% septic serum for 6 h. The ECs were then diluted to a suspension of 1×10⁶ cells/ml in 1X binding buffer (Ebioeasy Co., Ltd., Shanghai, China) and were lysed twice with cold PBS buffer. Purified recombinant Annexin V (5–15 µg) was mixed gently with the cell suspension (100 µl) at room temperature for 15 min. Subsequently, Annexin V-FITC (5 µl) and PI (10 µl) were added to the cell suspension and mixed gently for 15 min at room temperature in the dark, prior to adding 1X binding buffer (400 µl). The rate of EC apoptosis was measured using flow cytometry (Epics XL; Beckman Coulter, Inc., Brea, CA, USA) within 1 h.

Statistical analysis was performed using SPSS 13.0 (SPSS, Inc., Chicago, IL, USA). Data are expressed as the mean ± standard deviation. Statistically significance differences between groups were identified using one-way analysis of variance, and a least significant difference test was used for inter-group comparisons. P<0.05 was considered to indicate a statistically significant difference.

The serum expression levels of HMGB1 in patients with sepsis were detected using western blot analysis. As shown in Fig. 1, HMGB1 was detected in the sera of the patients with sepsis; however, it was not expressed until 24 h post-onset (Fig. 1), following which its expression gradually increased, peaking at 48 h (Fig. 1). The expression levels of HMGB1 remained elevated for a long period of time, prior to gradual weakening. The elevated expression levels were ultimately sustained for ~96 h. These results suggested that HMGB1 was not released until sepsis is fairly well established (24 h), and continued to be expressed from that point, maintaining and extending the pathological process of sepsis.

Treatment of ECs with septic patient sera increased cell permeability, as determined using immunofluorescence. As shown in Fig. 2, septic serum induced EC permeability, compared with control serum, after 18 and 24 h; however, the difference was only significant after 36 h (Fig. 2; P<0.05), and peaked at 48 h (Fig. 2; P<0.01). These results suggested that HMGB1 was released in late sepsis and was involved in the activation and injury of ECs, predominantly by inducing endothelial permeability. In addition, as shown in Fig. 3, treatment with septic serum significantly increased endothelial monolayer permeability after 6 h, compared with cells treated with control serum (P<0.01). However, pretreatment with 5 µM EP (an inhibitor of HMGB1) for 24 h prior to treatment with serum significantly reduced serum-induced permeability. HMGB1 was detected readily in the sera from the patients with sepsis, and its inhibition decreased permeability. These results indicated that the functions of HMGB1 were likely responsible for the increased EC permeability

In the control cells, the majority of F-actin was localized to the cell periphery, parallel to cell-cell junctions, with no clear gaps between the cells (Fig. 4A). When the ECs were exposed to septic serum for 6 h, almost all of cells became elongated, with thick actin stress fibers that traversed the cells in the direction of cell elongation (F-actin reorganization), increasing the intercellular gap distances (Fig. 4B). However, pretreatment of the cells with EP (5 µM) for 24 h inhibited F-actin reorganization, as shown in Fig. 4C. Treatment with EP alone did not cause F-actin reorganization, and the morphology of endothelial F-actin in these cells was similar to the normal morphology (Fig. 4D). Similarly, in the control cells VE-cadherin was located at the cell periphery and predominantly formed a continuous line at the cell-cell junctions with occasional gaps (Fig. 5A). However, upon exposure to septic serum for 6 h, the junctions became segmented and discontinuous, despite the fact that VE-cadherin remained localized to areas of cell-cell contact, indicating that the adhesive connection integrity was disrupted (Fig. 5B). However, pretreatment with EP partially inhibited the diffuse redistribution of VE-cadherin (Fig. 5C), whereas EP treatment alone had no effect on the morphology of VE-cadherin (Fig. 5D).

As shown in Fig. 6A, BAX, a pro-apoptotic protein, was not expressed in the ECs treated with control serum; however, its expression was markedly upregulated following treatment with septic serum (Fig. 6B). In addition, the upregulation in the expression of BAX was inhibited by pretreatment with 5 µM EP (Fig. 6C), whereas treatment with EP alone had no effect on the expression levels of BAX (Fig. 6D). As shown in Fig. 7, the protein expression levels of BCL-2, an inhibitor of apoptosis, were markedly reduced when the ECs were stimulated with septic serum, compared with those treated with control serum (Fig. 7A and B). However, the protein expression levels of BCL-2 were restored following EP pretreatment (Fig. 7C). These results suggested that HMGB1-rich serum induced EC apoptosis by downregulating the expression of BCL-2 and upregulating the expression of BAX.

The effects of septic serum on endothelial apoptosis were evaluated using flow cytometry. In Fig. 8, the lower left quadrant represents normal cells, whereas the upper left quadrant represents dead cells, and the lower right quadrant represents early apoptotic cells, whereas the upper right quadrant represents late apoptotic cells. Treatment with septic serum markedly increased the rate of endothelial apoptosis; with early apoptotic cells accounting for 45.9% of the total cells and late apoptotic cells accounted for 11.9%. In addition, compared with the control group (Fig. 8A), necrotic cells accounted for 12.9% of the cells (Fig. 8B). However, endothelial apoptosis was markedly reduced following EP pretreatment, particularly the early apoptotic cells, which accounted for only 13.9% of the cells (Fig. 8C). However, treatment with EP alone had no effect on endothelial apoptosis (Fig. 8D). These results indicated that high levels of HMGB1 in the septic serum may be involved in inducing endothelial apoptosis and EC permeability, particularly early apoptosis. However, with prolonged exposure, EC death occurs.