The effect of GM-CSF on endothelial permeability. (a) Dose-dependent changes in HRP flux. One-way ANOVA was used for repeated measurements. (b) Dose-dependent changes in TEER. HBMECs were plated in 24-well fibronectin-coated Transwell insert plates (pore size: 0.4 µm) and grown for four days. Next, GM-CSF (0, 10, 20 ng/ml) was added to the HBMEC monolayer for 24 h. Dose-dependent changes in the TEER and HRP fluxes of the HBMEC were subsequently measured. One-way ANOVA was used for repeated measurements. * p < 0.05 vs. normal HBMEC, the experiment was independently performed three times (n = 3). (c) ZO-1 expression in GM-CSF-stimulated HBMECs. HBMECs were stimulated with GM-CSF (0, 20 ng/ml) for the indicated time, after which tight junction protein expression was examined. The results represent three independent experiments. Data are shown as means ± standard deviations. One-way ANOVA was used to compare ZO-1 expression. *p < 0.05 vs. the 48 h-GM-CSF-treated HBMECs. (d) Changes in the location of ZO-1 (green) in HBMECs and GM-CSF-stimulated HBMECs were visualized by immunofluorescence. DAPI was used to visualize cell nuclei (blue). The disassembly of ZO-1 (not continuous) was indicated by white arrow. Scale bar: 50 µm. (e) Claudin-5 expression in GM-CSF-stimulated HBMECs. (f) Occludin distribution in GM-CSF stimulated HBMECs. Western blotting detected a shift in endothelial occludin protein from the insoluble to soluble phase. One-way ANOVA was used for repeated measurements. * p < 0.05 vs. normal HBMEC, the experiment was independently performed three times.