AGEs conjugated to bovine serum albumin (BSA) and Calbiochem^® MMP-2/MMP-9 Inhibitor II (BiPS) were purchased from Merck KGaA (Darmstadt, Germany). Antibodies to occludin, claudin-5 and Alexa fluor^® 546 were purchased from Invitrogen Life Technologies (Carlsbad, CA, USA) and GAPDH was purchased from Proteintech Group, Inc. (Chicago, IL, USA). Dylight^® 488 was purchased from Multiscience Biotech Co., Ltd (Hangzhou, China). Millicell^®-ERS was purchased from Millipore (Bedford, MA, USA). Fluorescein isothiocyanate (FITC)-dextran was purchased from Sigma-Aldrich (St. Louis, MO, USA).

GEnC cells were purchased from ATCC (Boulevard, MA, USA). Rat GEnC cultures were established and characterized as previously described (20). Briefly, GEnCs were grown in RPMI-1640 medium (Gibco-BRL, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (Gibco-BRL) and 10% NuSerum (Sigma-Aldrich) in a cell incubator at 37°C under 5% CO₂. When the cells reached 75–80% confluence, they were incubated in low-serum medium and treated with AGEs at different doses (20, 40 and 80 μl) and for various durations (6, 12 and 24 h). In certain experiments, the MMP2/9 inhibitor BiPS (50 μM) was also added. The cells and medium were collected 24 h after the initial treatment.

The endothelial cells were treated with lysis buffer and centrifuged at 15,000 × g for 10 min at 4°C. The supernatant was collected for western blotting or stored at −80°C for future analysis. Equal quantities of protein were loaded onto a gel for 15 % SDS-PAGE (Sigma-Aldrich) in order to separate the proteins. Separated proteins were transferred onto polyvinylidene difluoride membranes (Millipore, Lyon, France) and incubated with mouse anti-claudin-5 monoclonal antibody (1:3,000; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and mouse anti-Occuludin monoclonal antibody (1:2,000; Santa Cruz Biotechnology, Inc.) overnight at 4°C. Following washing with phosphate-buffered saline (PBS), the appropriate horseradish peroxidase-conjugated rabbit anti-mouse polyclonal secondary antibody (1:1,000; Santa Cruz Biotechnology, Inc.,) was added for 1 h at room temperature. Protein immunoreactivity was detected using the enhanced chemiluminescence reaction kit (ECL; Pierce, WI, USA), with the Tanon-4200 chemiluminescence reaction detection system (Tiangen Biot Co. Ltd, Beijing China). Band intensities were quantified by Quantity One 2.0 software (Bio-Rad, Hercules, CA, USA).

The electrical resistance across the confluent cell monolayer, TEER, was measured using the Millicell^®-ERS system (Millipore) according to the manufacturer’s instructions. Briefly, the cells were grown to post-confluence on transwell filters (Corning Costar^®, Inc., Union City, CA, USA) and treated with AGEs or AGEs + BiPS. The shorter electrode was placed within the Millicell culture plate insert and the longer electrode was placed in the outer well. The resistance of a culture or cell free plate (blank) was measured at three time-points from different points across the inner and outer wells until a stable value was measured each time. The resistance of the blank was subtracted from that measured with endothelial cells (net resistance). The unit area TEER (KΩ·cm²) was calculated by multiplying the net resistance by the area of the culture plate insert.

For the FITC-dextran flux assays, the cells were grown to post-confluence on transwell filters and 20 μl FITC-dextran solution was added to the inner well insert at a final concentration of 1 mg/ml. The medium was collected from the outer well, which was continuous with the basolateral compartment after 24 h. The quantity of diffused FITC-dextran was read using a fluorometer (SpectraMax M5, Molecular Devices, Tokyo, Japan) at wavelengths of 520 nm emission and 485 nm excitation.

The endothelial cells were grown on glass cover slips to 70–80% confluence and then cultured in AGEs with or without BiPS for 24 h. The treated cells were fixed in 4% formaldehyde solution (Sigma-Aldrich) for 10 min, washed in PBS, permeabilized in 0.2% Triton X-100 (Sigma-A ldrich) plus PBS, blocked with 5% goat serum for 30 min and incubated with rabbit anti-claudin-5 polyclonal antibody (1:100; Santa Cruz Biotechnology, Inc.) and mouse anti-Occuludin monoclonal antibody (1:500; Santa Cruz Biotechnology, Inc.) overnight at 4°C in humidified chambers. The fixed cells were then probed with a fluorophore-tagged secondary antibody (1:500) for 1 h at room temperature. The slides were washed in PBS and visualized using a laser scanning confocal microscope (Zeiss LSM510 Meta; Carl Zeiss, Stuttgart, Germany).

To assess gelatinase activity, the cell medium was collected and centrifuged at 12,000 × g for 10 min at 4°C. The supernatant was collected and concentrated 20-fold using Amicon Centricon centrifugal filters (Millipore). The concentrated media was separated electrophoretically on 10% polyacrylamide gels containing 0.1% gelatin. Following electrophoresis, gels were washed for 90 min in 2.5% (v/v) Triton X-100 and incubated in enzyme buffer containing 50 mM Tris-HCl, (pH 7.5; Sigma-Aldrich), 150 mM NaCl, 5 mM CaCl₂ and 1 M ZnCl₂ for 40 h with gentle agitation at 37°C. The gel was then stained with 0.2% Coomassie brilliant blue R-250 (Sigma-Aldrich) in a mixture of methanol, acetic acid and water (2:1:7) for 1 h and then destained in destaining solution. The band density was quantified by Quantity One 2.0 software (Bio-Rad) following negative image reversal.

All the experiments were performed at least three times. The results are presented as the mean ± standard deviation of independent experiments. Statistically significant differences between treatment groups were determined by one-way analysis of variance using SPSS 19.0 software (International Business Machines, Armonk, NY, USA). P<0.05 was considered to indicate a statistically significant difference.

Elevated serum AGE levels have been linked to hyperpermeability of the microvasculature in diabetes mellitus. The altered expression of TJ proteins by endothelial cells may induce these permeability changes; therefore, the expression levels of the TJ proteins occludin and claudin-5 following AGEs exposure at different doses and treatment time periods were examined (Fig. 1). In a dose-response study, GEnCs were exposed to various concentrations of AGEs for 24 h and the total content of occludin and claudin-5 in the cell lysate was measured by western blot analysis as described previously. The results revealed that exposure of GEnCs to 80 μg/ml AGEs caused a significant decrease of occludin and claudin-5, while exposure to 40 μg/ml or less of AGEs did not cause significant changes (Fig. 1A). To examine the effect of treatment duration, endothelial cells were exposed to 80 μg/ml AGEs for 6, 12, or 24 h. The expression of occludin began to decrease after 12 h and reached the lowest level after 24 h, while claudin-5 did not decrease significantly unless treated for the full 24 h. (Fig. 1B).

The effects of AGEs on MMP-2 and MMP-9 activity were investigated using gelatin zymography analysis at different doses and durations (Fig. 2). In the time course studies, the results demonstrated that when endothelial cells were exposed to 80 μg/ml AGEs for 6, 12 and 24 h, there were significant increases in MMP-9 activity after 12 h exposure and the activities peaked when treatment was extended to 24 h, while the maximal induction of MMP-2 activity occurred at 24 h at the same concentration of AGEs (Fig. 2A). The dose-response analysis results demonstrated that the proteolytic activities of MMP-2 and MMP-9 were significantly higher at 40 μg/ml and 80 μg/ml AGEs for 24 h (Fig. 2B).

To assess the effect of AGE treatment on GEnC permeability, the TEER of the confluent GEnC monolayers was measured. After 24 h exposure to 80 μg/ml AGEs, the GEnCs exhibited a significant decrease in TEER (Fig. 3A). The transendothelial diffusion of FITC-dextran (4 kDa) after 24 h stimulation with AGEs was also examined, as this molecule may diffuse across the paracellular spaces if endothelial TJs are disrupted. The results of FITC-dextran filtration measurements mirrored the TEER results of the present study; FITC-dextran translocation across the GEnC monolayer was higher in the AGE-treated cultures than that in untreated cultures and BSA did not alter basal permeability of GEnCs (Fig. 3B). To determine the effect of MMP-2 and MMP-9 on AGE-induced GEnC hyperpermeability, BiPS, an inhibitor of MMP-2/MMP-9, was co-administered with the AGEs. The results revealed that TEER levels and FITC-dextran filtration did not change significantly when the AGE-stimulated cells were co-treated with BiPS compared with those of the control group. These findings supported the role of MMP-2 and MMP-9 as key mediators of AGE-induced hyperpermeability in GEnCs.

To examine molecular changes in the TJs following AGE stimulation, the surface expression levels of occludin and claudin-5 were measured using immunofluorescence (Fig. 4). In the untreated control cultures, simultaneous fluorescent labeling of occludin and claudin-5 revealed continuous lines surrounding the individual cell peripheries and between cell-cell borders. Exposure to AGEs caused this labeling to become weaker and discontinuous or fractured compared with untreated control cultures, particularly in the case of occludin. When AGEs were co-administered with BiPS, however, the labeling again appeared as unbroken lines surrounding the cell peripheries and few fractures (TJ gaps) were observed. Furthermore, immunostaining was more intense compared with the cells treated with AGEs alone, indicating that the surface expression of occludin and claudin-5 was rescued by the MMP-2/9 inhibitor.

Subsequently, the association between TJ protein expression and MMP-2/9 activity in the AGE-treated cultures was examined in the presence and absence of BiPS (Fig. 5). The expression levels of occludin and claudin-5 were significantly reduced following AGE treatment, while neither occludin nor claudin-5 expression were significantly different from those in controls in cultures co-treated with AGEs and BiPS (Fig. 5A). Concomitant with the AGE-induced reduction in occludin and claudin-5, MMP-2/9 activity was elevated, while co-administration with BiPS again reversed this AGE-induced increase in MMP-2/9 activity (Fig. 5B).