Fig 1.
EBO-VLPs disrupt the iBRB in the in vitro model.
(A) Schematic overview of the iBRB tri-culture model in a Transwell system of HREC, HRP and HRA. EBO-VLPs were added to the lower chamber of the tri-culture iBRB model, which represents the retinal tissue. (B) Assessment of the integrity of the in vitro barrier models by TEER every day for one week. The results are presented as the means ± standard deviation of six independent experiments. (C) Na-F permeability of iBRB tri-culture models at 6 hours and 3 days post HREC seeding. The results are presented as the means ± standard deviation of six independent experiments. (D) Images of HREC showing the expression of claudin-1, occludin, and ZO-1. Claudin-1, occludin, and ZO-1 are shown in green, and cell nuclei were stained with DAPI (blue). Representative images of three independent experiments are shown. (E) TEM images of EBO-VLPs. (F) Images of HRECs, HRPs and HRAs treated with EBO-VLPs and immunostained with anti-GP antibodies (green). Cell nuclei were stained with DAPI (blue). Representative images of three independent experiments are shown. The fluorescent images were taken with a 60× magnification objective lens under a confocal microscope. (G-H) Integrity of the tri-culture iBRB model after EBO-VLP administration. TEER values (G) and Na-F permeability (H) of the iBRB model were examined at 48 h after EBO-VLP administration. TEER values were normalized to those of iBRB models themselves before EBO-VLP administration. The box and the whisker present the median ± percentiles (25–75%) and range, respectively. The fold change of permeability compared with iBRB model itself before EBO-VLP administration is presented as the mean ± standard deviation. All values were determined in six independent experiments. Statistical analysis was performed using Student’s t test. *** p< 0.001.
Fig 2.
iBRB breakdown by EBO-VLPs is not attributable to direct cytotoxicity on HRECs.
(A) Phase contrast images of retinal endothelial cell mono-layers with or without 48 h of EBO-VLP stimulation. Representative images of three independent experiments are shown. (B) Viability of HREC treated with different concentrations of EBO-VLPs. All data were normalized to the mock group. The results are presented as the means ±standard deviation of three independent experiments. (C) Experimental schematic of EBO-VLPs addition to the mono-culture iBRB model. (D-E) Changes in the integrity of the mono-culture iBRB model after EBO-VLP administration. Na-F permeability (D) and the TEER values (E) of the iBRB model were examined 48 h after EBO-VLP administration. TEER values were normalized to those of iBRB models themselves before EBO-VLP administration. The box and the whisker present the median ± percentiles (25–75%) and range, respectively. The fold change of permeability compared with iBRB model itself before EBO-VLP administration is presented as the mean ± standard deviation. All values were determined in six independent experiments. (F) Images of HRECs in the mono-culture iBRB model showing the expression of claudin-1, occludin, ZO-1 (green) and cell nuclei stained with DAPI (blue). Representative images of three independent experiments are shown. The fluorescent images were taken with a 60× magnification objective lens under a confocal microscope. (G) Quantification of the fluorescence intensity of claudin-1, occludin, and ZO-1 after EBO-VLP administration. The regions for fluorescence intensity were determined in four independent experiments. (H) Western blot analysis of claudin-1, occludin and ZO-1 expression in HREC 48 h after EBO-VLP administration in mono-culture iBRB models. Representative images of three independent experiments are shown. Statistical analysis was performed using Student’s t test. *p < 0.05, **p < 0.01.
Fig 3.
EBO-VLPs can significantly stimulate pericytes to secrete VEGF.
(A) Representative cytokines secreted by HREC, HRP and HRA after 48 h of EBO-VLP stimulation. Representative images of three independent experiments are shown. (B) Relative cytokine levels were normalized to the mock group without EBO-VLP administration, which was set as 1. The floating bar plot shows the mean, minimum and maximum levels of each cytokine in three independent experiments. Downregulation of cytokines expression were displayed with negative numbers. (C) Changes in VEGF expression after EBO-VLP administration, as quantified by ELISA. The horizontal dashed line marks the limit of detection of the assay. The results are presented as the means ± standard deviation of three independent experiments. Statistical analysis was performed using Student’s t test. *p < 0.05, **p < 0.01.
Fig 4.
EBO-VLP causes iBRB breakdown through pericytes secretion of VEGF.
(A-B) Integrity of the tri-culture iBRB model after EBO-VLP, EBO-VLP + Avastin and VEGF treatment. Na-F permeability (A) and the TEER values (B) of the iBRB model were examined 48 h after EBO-VLP administration. TEER values were normalized to those of iBRB models themselves before EBO-VLP administration. The box and the whisker present the median ± percentiles (25–75%) and range, respectively. The fold change of permeability compared with iBRB model itself before EBO-VLP administration is presented as the mean ± standard deviation. All values were determined in ten independent experiments. (C) Experimental schematic showing the addition of EBO-VLPs to the iBRB co-culture model of HREC and HRP. (D-E) Integrity of the iBRB co-culture (HRECs and HRPs) model after EBO-VLP, EBO-VLP + Avastin and VEGF treatment. Na-F permeability (D) and the TEER values (E) of the iBRB model were examined after 48 h of treatment. TEER values were normalized to those of iBRB models themselves before EBO-VLP administration. The box and the whisker present the median ± percentiles (25–75%) and range, respectively. The fold change of permeability compared with iBRB model itself before EBO-VLP administration is presented as the mean ± standard deviation. All values were determined in ten independent experiments. (F) Experimental schematic showing the addition of EBO-VLPs to the iBRB co-culture model of HREC and HRA. (G-H) Integrity of the iBRB co-culture (HRECs and HRAs) model after EBO-VLP, EBO-VLP + Avastin, and VEGF treatment. Na-F permeability (G) and the TEER values (H) of the iBRB model were examined 48 h after EBO-VLP administration. All values were determined in ten independent experiments. (I) Changes in VEGF expression after EBO-VLP administration, as quantified by ELISA. Pericytes were transfected with si-VEGF or nontargeting control siRNA, followed by treatment with EBO-VLP for 48 h. The horizontal dashed line marks the limit of detection of the assay. The results are presented as the means ± standard deviation of three independent experiments. (J-K) Integrity of the tri-culture iBRB model after EBO-VLP, si-VEGF + EBO-VLP, nontargeting control siRNA + EBO-VLP, and nontargeting control siRNA treatment. Na-F permeability (J) and the TEER values (K) of the iBRB model were examined 48 h after administration. TEER values were normalized to those of iBRB models themselves before EBO-VLP administration. All values were determined in six independent experiments. Statistical analysis was performed using Student’s t test. *p < 0.05, **p < 0.01, ***p < 0.001.
Fig 5.
EBO-VLPs downregulate the tight junction protein Claudin-1 in the iBRB.
(A) Western blot analysis of claudin-1, occludin and ZO-1 expression in HREC at 48 h in the tri-culture iBRB model. Representative images of three independent experiments are shown. (B) Relative protein levels of claudin-1, occludin and ZO-1 were normalized to β-actin. The results are presented as the means ± standard deviation of three independent experiments. (C) Immunofluorescence images of claudin-1, occludin and ZO-1 expression in HRECs at 48 h post EBO-VLP treatment in the tri-culture iBRB model. Representative images of three independent experiments are shown. (D) Quantification of the fluorescence intensity of claudin-1, occludin and ZO-1 after EBO-VLP administration. The regions for fluorescence intensity analysis were determined in four independent experiments. **p < 0.01, ***p < 0.001.
Fig 6.
GP plays the major role in the destruction of the iBRB.
(A) Western blot analysis of VLP-ΔGP using an anti-GP antibody. Representative images of three independent experiments are shown. (B) TEM image of VLP-ΔGP. (C) Permeability and (D) TEER of the tri-culture iBRB model were measured 48 h after VLP-ΔGP, VLP-ΔGP + Avastin and VEGF treatment. TEER values were normalized to those of iBRB models themselves before EBO-VLP administration. The box and the whisker present the median ± percentiles (25–75%) and range, respectively. The fold change of permeability compared with iBRB model itself before EBO-VLP administration is presented as the mean ± standard deviation. All values were determined in six independent experiments. (E) The expression of VEGF in HRP after VLP-ΔGP treatment, as quantified by ELISA. The horizontal dashed line marks the limit of detection of the assay. The results are presented as the means ± standard deviation of three independent experiments. (F) The expression of VEGF in HRP after EBOV VP40 and GP was expressed in HRPs. The horizontal dashed line marks the limit of detection of the assay. The results are presented as the means ± standard deviation of three independent experiments. Statistical analysis was performed using Student’s t test. *p < 0.05, **p < 0.01.
Fig 7.
EBO-VLPs damage the iBRB in vivo.
(A) Immunohistofluorescence analysis of EBO-VLPs in the retinal tissue using anti-GP antibodies (red). The white arrow indicates EBO-VLPs. Representative images of three independent experiments are shown. The fluorescent images were taken with a 60× magnification objective lens under a confocal microscope. (B) H&E staining of retinas from rats treated with Avastin and/or EBO-VLPs and VEGF. Black arrows indicate the pathological destruction of the retina. Magnification: ×20. Representative images of three independent experiments are shown. (C) Evans blue assay showing rat retinal permeability at 2 days after EBO-VLP injection. The results are presented as the means ± standard deviation of four independent experiments. (D) Immunohistofluorescent staining of retinal sections was performed to analyze VEGF (red). Nuclear staining with DAPI shows the retinal layers. Representative images of three independent experiments are shown. (E) Western blot analysis of claudin-1 expression in rat retinas, which was normalized against β-actin. Representative images of three independent experiments are shown. (F) Relative protein levels of claudin-1 were normalized to β-actin. The results are presented as the means ± standard deviation of three independent experiments. (G) Immunohistofluorescent staining of claudin-1 in retinal sections. Representative images of three independent experiments are shown. (H) H&E staining of retinas from rats treated with or without VLP-ΔGP. Magnification: ×20. Representative images of three independent experiments are shown. (I) Evans blue assay showing rat retinal permeability at 2 days after VLP-ΔGP injection. The results are presented as the means ± standard deviation of four independent experiments. (J) Images of immunofluorescence staining of claudin-1 in retinal sections treated with VLP-ΔGP. Representative images of three independent experiments are shown. Statistical analysis was performed using Student’s t test. *p < 0.05, **p < 0.01.
Fig 8.
Model for the iBRB breakdown affected by EBOV.
EBOV stimulates pericytes to secrete VEGF to cause the iBRB breakdown. VEGF disrupts iBRB through downregulation of the tight junction protein claudin-1. The VEGF antibody Avastin can reduce the effect on VLP-induced iBRB breakdown.