Nitric Oxide Interacts with Caveolin-1 to Facilitate Autophagy-Lysosome-Mediated Claudin-5 Degradation in Oxygen-Glucose Deprivation-Treated Endothelial Cells

Abstract

Using in vitro oxygen-glucose deprivation (OGD) model, we have previously demonstrated that 2-h OGD induces rapid, caveolin-1-mediated dissociation of claudin-5 from the cellular cytoskeletal framework and quick endothelial barrier disruption. In this study, we further investigated the fate of translocated claudin-5 and the mechanisms by which OGD promotes caveolin-1 translocation. Exposure of bEND3 cells to 4-h OGD, but not 2-h OGD plus 2-h reoxygenation, resulted in claudin-5 degradation. Inhibition of autophagy or the fusion of autophagosome with lysosome, but not proteasome, blocked OGD-induced claudin-5 degradation. Moreover, knockdown of caveolin-1 with siRNA blocked OGD-induced claudin-5 degradation. Western blot analysis showed a transient colocalization of caveolin-1, claudin-5, and LC3B in autolysosome or lipid raft fractions at 2-h OGD. Of note, inhibiting autophagosome and lysosome fusion sustained the colocalization of caveolin-1, claudin-5, and LC3B throughout the 4-h OGD exposure. EPR spin trapping showed increased nitric oxide (NO) generation in 2-h OGD-treated cells, and inhibiting NO with its scavenger C-PTIO or inducible nitric oxide synthase (iNOS) inhibitor 1400W prevented OGD-induced caveolin-1 translocation and claudin-5 degradation. Taken together, our data provide a novel mechanism underlying endothelial barrier disruption under prolonged ischemic conditions, in which NO promotes caveolin-1-mediated delivery of claudin-5 to the autophagosome for autophagy-lysosome-dependent degradation.

Keywords: Autophagy; Caveolin-1; Claudin-5; Lysosome; Oxygen-glucose deprivation.

Fig. 1

Exposure of bEND3 cells to 4-h OGD induces and causes the degradation of claudin-5. The bEND3 cells were subjected to the following treatments: 2-h OGD (O2h), 2-h OGD plus 2-h reoxygenation (O2R2), 4-h OGD (O4h), 2-h OGD plus 24-h reoxygenation, and 4-h OGD plus 24-h reoxygenation before collecting total cell lysates or sub-cellular fractions for Western blot analysis. a A significant reduction in claudin-5 protein levels was detected in bEND3 cells exposed to O4h, but not O2h or O2R2. Upper panel: representative immunoblots for claudin-5 and β-actin; bottom panel: quantitative data of claudin-5 protein band intensity after normalization to β-actin. Data were expressed as mean± SEM, n=4, *P<0.05 versus control (Ctrl), ANOVA. b Representative immunoblots showed that the claudin-5 contents in the cytosolic fraction (CF), membranous fraction (MF), and actin cytoskeleton fraction (ACF) were significantly reduced in O4h-treated cells, and there was a redistribution of claudin-5 in O2h- or O2R2-treated cells, with reduced protein levels in the ACF and increased protein levels in the CF and MF. Experiments were repeated 3 times with similar results. c After 24 h of reoxygenation, the total protein level of claudin-5 remained significantly reduced in O4h-treated bEND3 cells when compared with control cells or O2h-treated cells. *P<0.05 versus Ctrl, ANOVA; n=4. d Representative immunoblots showed that after 24-h reoxygenation, the reduction of claudin-5 was recovered in CF and MF, but not in ACF of O4h–treated cells; however, O2h–induced redistribution of claudin-5 was completely recovered. Experiments were repeated 3 times with similar results

Fig. 2

Lysosome pathway is responsible for claudin-5 degradation in 4-h OGD-treated endothelial cells. bEND3 cells were treated with lysosome inhibitor chloroquine (100 µM) or proteasome inhibitor MG132 (20 µM) 1 h before and during 4-h OGD treatment (O4h) before collecting indicated cellular extracts for claudin-5 detection by Western blot. The degradation of claudin-5 was blocked by chloroquine (a), but not MG132 (b), in O4h-treated cells. Upper panels: representative immunoblots for claudin-5 and β-actin; bottom panels: quantitative data of claudin-5 protein band intensity after normalization to β-actin. Data were expressed as mean±SEM, n=5 for chloroquine and n=4 for MG132, *P<0.05 versus control (Ctrl)+vehicle (0.1 % DMSO), ANOVA. c Representative confocal microphotographs showed circumcellular staining of claudin-5 in control bEND3 cells (Ctrl), which was significantly reduced after O4h treatment. The addition of chloroquine prevented claudin-5 reduction and resulted in its accumulation in the cytosol. Experiments were repeated 3 times with similar results. Scale bar, 40 µm. d Western blot analyses showed that 2-h OGD (O2h) induced the translocation of claudin-5 from the ACF to the CF and MF. When OGD was prolonged to 4 h (O4h), claudin-5 levels were decreased in all three subcellular fractions. The presence of chloroquine prevented claudin-5 degradation and rendered its redistribution comparable to that of O2h-treated cells. Experiments were repeated 3 times with similar results

Fig. 3

Autophagy is involved in 4-h OGD-induced claudin-5 gradation in bEND3 cells. a A significant increase in the level of LC3B (both LC3B-I and LC3B-II), the marker of autophagosome, was detected in bEND3 cells exposed to 2-h OGD (O2h) or 4-h OGD (O4h), with a greater increase for O4h. Experiments were repeated 3 times with similar results. b, c bEND3 cells were treated with or without autophagosome-lysosome fusion inhibitor BFA1 (100 nM) and autophagy inhibitor 3-MA (5 mM) 1 h before and during O4h treatment before collecting total cellular extracts for claudin-5 detection by Western blot. Both inhibitors blocked O4h–induced claudin-5 degradation. β-Actin served as a loading control. Data were expressed as mean±SEM, n=4, *P<0.05 versus control (Ctrl)+vehicle, ANOVA

Fig. 4

There was a transient colocalization of LC3B, claudin-5, and Cav-1 in bEND3 cells exposed to 4-h OGD treatment. Following indicated treatments, OptiPrep (60 % w/v iodixanol in H₂O) gradients and 5–40 % discontinuous sucrose gradient were used to isolate autolysosome-enriched fractions or caveolins/caveolae-enriched fractions from bEND3 cells, respectively, and Cav-1, claudin-5, and LC3B in each fraction were analyzed by Western blot. a A transient colocalization of Cav-1, claudin-5, and LC3B was detected in autophagosome-enriched fractions (fractions 2, 3, and 4) after 2-h OGD treatment (OGD2h); when OGD was prolonged to 4 h (OGD4h), claudin-5 and Cav-1 disappeared from these fractions. b A similar colocalization of Cav-1, LC3B, and claudin-5 was transiently seen in caveolins/caveolae-enriched fractions (lipid raft, fractions 3, 4, 5) after 2-h OGD treatment. When OGD was prolonged to 4 h, the protein bands of claudin-5 and LC3B vanished from these fractions, while the presence of autophagosome-lysosome fusion inhibitor BFA1 (O4h+BFA1) sustained the colocalization of these three proteins. Both experiments were repeated for 3 times with similar results

Fig. 5

Cav-1 is involved in OGD-induced claudin-5 degradation in bEND3 cells. bEND3 cells were incubated with control siRNA (Ctrl siRNA) or Cav-1 siRNA for 48 h before exposing to 2-h OGD (O2h) or 4-h OGD (O4h). a Western blot analysis showed Cav-1 siRNA effectively knocked down Cav-1 protein expression (~90 % reduction) in bEND3 cells compared to Ctrl siRNA. b Knockdown of Cav-1 with siRNA prevented O2h-induced claudin-5 redistribution (i.e., increased contents in the cytosolic and membranous fractions — CF and MF, and reduced contents in the actin cytoskeleton fraction — ACF) as well as O4h-induced claudin-5 reduction in all three subcellular fractions. Experiments were repeated 3 times with similar results. c Knockdown of Cav-1 with siRNA abolished O4h–induced claudin-5 degradation. Upper panels: representative immunoblots for claudin-5 and β-actin; bottom panels: quantitative data of claudin-5 protein band intensity after normalization to β-actin. Data were expressed as mean±SEM, n=5, *P<0.05 versus Ctrl+Ctrl siRNA, ANOVA

Fig. 6

Exposure to 2-h OGD increased iNOS-derived NO production in bEND3 cells. a bEND3 cells were incubated with spin trapping agent Fe(MGD)₂ (1 mM) during 2-h OGD (O2h) treatment with or without the presence of iNOS specific inhibitor 1400W. At the end of exposure, NO in 0.5 ml media was immediately measured by EPR. A typical NO signal was detected in the media of O2h–treated cells, but not in normoxic cells, and the presence of iNOS inhibitor 1400W completely abolished NO’s signal. Experiments were repeated for 3 times with the same results. b, c After treatment, total RNA and protein were extracted from bEND3 cells for real-time PCR and Western blot analysis, respectively. Comparing to normoxia (control), OGD significantly increased iNOS mRNA (b) and protein expression (c). Upper panels: representative immunoblots for claudin-5 and β-actin; bottom panels: quantitative data of claudin-5 protein band intensity after normalization to β-actin. Data were expressed as mean±SEM, n=4, *P<0.05 versus control, Student’s t test

Fig. 7

iNOS-derived NO mediated OGD-induced Cav-1 translocation in bEND3 cells. bEND3 cells were exposed for 2-h OGD (O2h) with or without the presence of NO scavenger C-PTIO and iNOS inhibitor 1400W before analyzing subcellular redistribution of Cav-1 by Western blot or immunostaining. Following OGD treatment, Cav-1 protein levels were significantly increased in the cytosolic fraction (CF), which was accompanied by a significant reduction in the actin cytoskeleton fraction (ACF) (a, b). Scavenging NO with C-PTIO (a) or inhibiting iNOS with 1400W (b) completely abolished these changes. Upper panels: representative immunoblots of Cav-1 and β-actin; bottom panel: quantitative data of Cav-1 protein band intensity after normalization to β-actin. Data were expressed as mean±SEM, n=4; *P<0.05 versus vehicle+control (Ctrl), ANOVA. c Confocal micrograph showed that 2-h OGD significantly increased Cav-1 immuno-staining in the cytosol, which was blocked by C-PTIO. Experiments were repeated 3 times with similar results. Scale bar, 40 µm

Fig. 8

NO scavenger C-PTIO or iNOS inhibitor 1400W inhibited 4-h OGD-induced claudin-5 degradation in bEND3 cells. bEND3 cells were exposed to OGD treatment for 4 h (O4h) with or without the presence of C-PTIO or 1400W before collecting total cellular extracts for claudin-5 detection by Western blot. Both C-PTIO (a) and 1400W (b) inhibited claudin-5 degradation induced by 4-h OGD. Upper panel: representative immunoblots of claudin-5 and β-actin; bottom panel: quantitative data of claudin-5 protein band intensity after normalization to β-actin. Data were expressed as mean±SEM, n=4, *P<0.05 versus vehicle+control, ANOVA