Fig 1.
DENV NS1 increases the permeability of endothelial cells.
(A) HMEC-1 cells were treated with different doses of 293T-NS1 for 6 h, and the endothelial permeability was determined by the transwell permeability assay. n = 4, triplicated. (B) HMEC-1 cells were treated with 20 μg/ml of 293T-NS1 or the same volume of PBS as the vehicle control for the indicated periods, and the endothelial permeability was determined by the transwell permeability assay. n = 3, triplicated. (C) HMEC-1 cells were treated with 20 μg/ml 293T-NS1 or S2-NS1 for 6 h, and the endothelial permeability was determined by the transwell permeability assay. The same volume of PBS was used as a control. n = 3, triplicated. (D) BALB/c mice were i.p. injected with 50 μg BSA or S2-NS1 for 6 h. After sacrifice, the abdominal cavity was washed with PBS, and the protein concentration in the peritoneal lavage was determined by the BCA method. n = 5, protein quantification was duplicated. *P<0.05, **P<0.01, ***P<0.001.
Fig 2.
DENV NS1-induced vascular leakage is inhibited by anti-NS1 antibodies.
(A) HMEC-1 cells were grown on a 96-well E-plate. When the cells grew to confluence, 293T-NS1 was added with or without anti-NS1 antibodies. PBS was added as a vehicle control, and CTRL mIgG was added as a negative control. The electrical resistance over a period of 24 h was measured and normalized as the cell index. n = 3, duplicated. (B) HMEC-1 cells were treated with PBS, 293T-NS1 with or without anti-NS1 antibodies. After 6 h, the endothelial permeability was determined by the transwell permeability assay. n = 4, triplicated. (C) BALB/c mice were i.p injected with PBS, S2-NS1 with or without anti-NS1 antibodies. After 6 h, the mice were sacrificed, and the abdominal cavities were washed with PBS. The abdominal lavage was collected, and the protein concentration was determined by the BCA method. n = 4, protein quantification was duplicated. *P<0.05, **P<0.01, ***P<0.001.
Fig 3.
MIF secretion is induced by DENV NS1.
(A) HMEC-1 cells were treated with PBS, 293T-NS1 with or without anti-NS1 antibodies for 6h. The supernatants of the cells was collected, and the MIF concentration in the supernatant was determined by ELISA. n = 3, MIF quantification was duplicated. (B) BALB/c mice were treated with S2-NS1 with or without anti-NS1 antibodies by i.p. injection. After 6 h, the peritoneal lavage was collected, and the murine MIF concentration in the peritoneal lavage was measured by ELISA. n = 5, protein quantification was duplicated. (C) BALB/c mice were i.v. injected with PBS or S2-NS1 for 48 h. Mice blood was collected by orbital sinus sampling at indicated time after i.v. injection. MIF concentration in the plasma was determined by ELISA. n = 3, duplicated. *P<0.05, **P<0.01, ***P<0.001.
Fig 4.
Inhibition of MIF prevents DENV-NS1-induced endothelial hyperpermeability.
(A) HMEC-1 cells were grown on a 96-well E-plate. When the cells grew to confluence, 293T-NS1 was added with or without the MIF inhibitors. The electrical resistance over a period of 24 h was measured and normalized as the cell index. n = 3, duplicated. (B) HMEC-1 cells were treated with PBS, 293T-NS1, 293T-NS1 mixed with MIF inhibitors or 293T-NS1 mixed with anti-MIF pAb. CTRL RaIgG was also added as a negative control of anti-MIF pAb. After 6 h, the endothelial permeability was determined by the transwell permeability assay. n = 4, triplicated. ***P<0.001.
Fig 5.
DENV NS1 induces autophagy formation of endothelial cells, which is subdued by MIF inhibitor.
(A) HMEC-1 cells were treated with PBS or 20 μg/ml 293T-NS1 for 12 hours. The cell lysates were collected, and the relative protein levels of VE-cadherin, p62, and LC3 were determined by Western blot with specific antibodies. (B) HMEC-1 cells were treated with PBS, 293T-NS1, 293T-NS1 mixed with ISO-1, or ISO-1 only. After 6 h, the cells were fixed, and IFA was performed using specific antibodies. The number of LC3 puncta was quantified in (C), and the colocalization of LC3 and VE-cadherin was quantified in (D) by using FV1000 software. Bar: 10 μm. ***P<0.001.
Fig 6.
Inhibition of autophagy avoids DENV NS1-induced vascular leakage.
(A) HMEC-1 cells were treated with PBS, 293T-NS1 or 293T-NS1 mixed with 3-MA or NAC for 24 h. The relative cell index was measured every hour with RTCA. n = 3, duplicated. (B) HMEC-1 cells were treated with PBS, 293T-NS1, or 293T-NS1 mixed with 3-MA or NAC for 6 h. The relative permeability was determined by the transwell permeability assay. n = 3, triplicated. (C) HMEC-1 cells were transfected with luciferase or Atg5 shRNA. After selection with puromycin, the resultant stable clones were treated with PBS or 20 μg/ml S2-NS1 for 6 h, and the endothelial permeability was determined by the transwell permeability assay with streptavidin-HRP and TMB. The knockdown efficiency is shown in the right panel with the results of Western blot analysis. n = 3, triplicated. (D) BALB/c mice were intraperitoneal injected with BSA, S2-NS1 with or without ISO-1, p425, 3-MA, or NAC. After 6 h, the mice were sacrificed, and the abdominal cavities were washed with 5 ml PBS. Protein concentration of the peritoneal lavage was quantified by BCA method. n = 3, duplicated. *P<0.05, **P<0.01, ***P<0.001.
Fig 7.
DENV NS1 induces VE-cadherin disarray, which is rescued by inhibiting MIF or autophagy.
(A) HMEC-1 cells were treated with PBS, 293T-NS1, 293T-NS1 mixed with ISO-1, 293T-NS1 mixed with 3-MA, ISO-1 only or 3-MA only. After 6 h, the cells were fixed, and IFA was applied by anti-VE-cadherin antibodies. The fluorescence images were acquired by a FV1000 confocal microscopy. The fluorescent intensity was quantified by Image J software and cytosolic/perinuclear versus barrier/marginal VE-cadherin ratio is shown in (B) by measuring 50 cells for each condition. *P<0.05, **P<0.01, ***P<0.001.
Fig 8.
Hypothetical model of DENV NS1-induced vascular hyperpermeability.
DENV infection increases the NS1 level in circulation. When NS1 binds to TLR4 on endothelial cells, the secretion and expression of MIF are induced. MIF then binds to receptor CXCR2/CXCR4 and CD74 on endothelial cells through paracrine or autocrine function. The receptor mediates the signal, which induces the formation of autophagy through PI3K. The autophagosomes mediate the degradation of junction proteins, which results in dysfunction of endothelial barrier and vascular leakage.
Table 1.
List of accession numbers for genes and proteins.