Phosphorylation of caveolin-1 regulates oxidant-induced pulmonary vascular permeability via paracellular and transcellular pathways

Abstract

Rationale

Oxidants are important signaling molecules known to increase endothelial permeability, although the mechanisms underlying permeability regulation are not clear.

Objective

To define the role of caveolin-1 in the mechanism of oxidant-induced pulmonary vascular hyperpermeability and edema formation.

Methods and Results

Using genetic approaches, we show that phosphorylation of caveolin-1 Tyr¹⁴ is required for increased pulmonary microvessel permeability induced by hydrogen peroxide (H₂O₂). Caveolin-1 deficient mice (cav-1^-/-) were resistant to H₂O₂-induced pulmonary vascular albumin hyperpermeability and edema formation. Furthermore, the vascular hyperpermeability response to H₂O₂ was completely rescued by expression of caveolin-1 in cav-1^-/- mouse lung microvessels, but was not restored by the phosphorylation-defective caveolin-1 mutant. The increase in caveolin-1 phosphorylation induced by H₂O₂ was dose-dependently coupled to both increased ¹²⁵I-albumin transcytosis and decreased transendothelial electrical resistance in pulmonary endothelial cells. Phosphorylation of caveolin-1 following H₂O₂ exposure resulted in the dissociation of vascular endothelial cadherin/β-catenin complexes, and resultant endothelial barrier disruption.

Conclusions

Caveolin-1 phosphorylation-dependent signaling plays a crucial role in oxidative stress-induced pulmonary vascular hyperpermeability via transcellular and paracellular pathways. Thus, caveolin-1 phosphorylation may be an important therapeutic target for limiting oxidant-mediated vascular hyperpermeability, protein-rich edema formation, and acute lung injury.

Introduction

An increase in vascular permeability is a key hallmark of inflammation and has been implicated in the pathophysiology of many disease states including acute lung injury, ischemia-reperfusion injury, atherosclerosis, and diabetes.¹ Under physiological conditions, adherens junctions predominate in endothelial cell-cell contacts and control pulmonary endothelial barrier integrity. Stabilization of adherens junctions is dependent on the association of vascular endothelial (VE)-cadherin, β-catenin, p120-catenin, and α-catenin proteins and their linkage to the actin cytoskeleton² wherein VE-cadherin association with the actin cytoskeleton is thought to be dependent on the β-catenin linkage.³ Oxidants including superoxide and hydrogen peroxide (H₂O₂) generated by activated neutrophils and endothelial cells in response to inflammatory stimuli increase paracellular endothelial permeability by promoting the loss of cell–cell adhesion and activation of actin-myosin based cell retraction.^1,4,5

We recently demonstrated that caveolae-mediated transendothelial transport (transcytosis) of macromolecules through the microvascular endothelial barrier is also an important mechanism responsible for inflammation-evoked pulmonary vascular hyperpermeability and protein-rich edema formation.⁶ We showed that an increase in transcellular (caveolae-mediated) permeability, triggered by the binding of neutrophils to endothelial cell surface intercellular adhesion molecule-1 (ICAM-1), was mediated by Src activation and phosphorylation of caveolin-1 (Cav-1)⁶ which thereby stimulates caveolae formation and trafficking.^7-11 ICAM-1 expression and activation is mediated by oxidant signaling in lung endothelial cells.¹ In addition, oxidants have also been shown to directly induce Src-dependent phosphorylation of Cav-1 at tyrosine 14 (Tyr¹⁴) in endothelial cells.^12-14 Thus, oxidants generated by neutrophils and/or endothelial cells may serve as signal transduction mediators that regulate transcellular permeability through a Cav-1 dependent mechanism.

Cav-1 expression is required for caveolae-mediated endocytosis and transcytosis in endothelial cells.^{4,6,7-11,15-17} Recent evidence also points to the potential role of Cav-1 in cell-cell adhesion and thus paracellular permeability regulation. Cav-1 co-localizes with adherens junction proteins E-cadherin, β-catenin and γ-catenin in MDCK epithelial cells.¹⁸ Downregulation of Cav-1 leads to a loss and redistribution of tight junction proteins (occludin and ZO-1) in brain microvascular endothelial cells resulting in an increase in paracellular permeability.¹⁹ Furthermore, siRNA-mediated depletion of Cav-1 in the mouse lung induced an increase in the number of interendothelial gaps in pulmonary capillaries and veins.²⁰ In contrast, deletion of Cav-1 attenuated protein kinase C (PKC)-induced inter-endothelial gap formation in myocardial microvascular endothelial cells.²¹ Whether and how Cav-1 regulates endothelial barrier function during oxidative stress remains an important question.

In the present study, using genetic approaches, we investigated the role of Cav-1 in pulmonary microvascular permeability regulation through both transcellular and paracellular pathways. We found that Cav-1 phosphorylation is required for H₂O₂-induced stimulation of transcytosis and destabilization of cell-cell junctions, and hence propose that tyrosine phosphorylation of Cav-1 plays an important role in the pathogenesis of oxidant-induced pulmonary vascular hyperpermeability.

Materials and Methods

An expanded Materials and Methods section is available in the online supplement.

Briefly, Cav-1 null (cav-1^-/-) mice, wild-type B6/129SJ2 mice, and rat lung microvascular endothelial cells (RLMVECs) were used. Animal protocols received institutional review and committee approval. Endocytosis and transendothelial transport of ¹²⁵I-albumin, fluorescent albumin uptake, transendothelial electrical resistance (TER), siRNA transfection, Western blotting and immunoprecipitation were performed as described previously.⁶ Rescue studies were made in mouse lungs from cav-1^-/- mice by liposome-mediated plasmid cDNA transfection.^{6 125}I-albumin permeability-surface area (PS) product was measured in lungs perfused with Krebs solution.

Results

H₂O₂ Induced Cav-1 Phosphorylation via Src and c-Abl Kinases

H₂O₂ in a concentration-dependent manner (0.05∼0.8 mmol/L) increased Cav-1 Tyr ¹⁴ phosphorylation (p(Y14)-Cav-1; Figure 1A). Quantitative analysis revealed that the level of phospho-Cav-1 was increased by 2- to 9-fold (Online Figure IA) following exposure to H₂O₂. Furthermore, Cav-1 phosphorylation increased within 5 min, peaked at 30 min, and then returned to basal levels 60 min after H₂O₂ treatment (Figure 1B and Online Figure IB). Coincident with the increase in p(Y14)-Cav-1 levels, we also observed activation of c-Src in endothelial cells following treatment with H₂O₂ as measured by p(Y418)-Src phospho-immunoblot (Figure 1A and Online Figure IA). Pretreatment of endothelial cells with the Src inhibitor PP2 (4-amino-5-(4-chlorophenyl)-7-(t-butyl) pyrazolo[3,4-d] pyrimidine) completely blocked Cav-1 phosphorylation following exposure of cells to 0.2 mmol/L H₂O₂, but it only blocked ∼ 65% of the effect of higher concentrations of H₂O₂ (0.6 mmol/L) (Figure 1C and Online Figure IC) suggesting Src-independent pathways may also be involved in the mechanism of Cav-1 phosphorylation induced by high concentration of H₂O₂.

Figure 1. H₂O₂-induced activation of Src and c-Abl and resultant phosphorylation of Cav-1.

(A) H₂O₂ induced the activation of Src and phosphorylation of Cav-1 in a concentration-dependent manner. RLMVECs were exposed to different concentrations of H₂O₂ (0.01-0.8 mmol/L) for 30 min. (B) Time course of H₂O₂-induced Cav-1 phosphorylation. Cells were stimulated with 0.6 mmol/L H₂O₂ for indicated times. (C) Effect of PP2 on Cav-1 phosphorylation induced by different concentrations of H₂O₂. (D) Dose-dependent c-Abl phosphorylation induced by H₂O₂. Cells were exposed to vehicle or H₂O₂ (0.1-0.6 mmol/L) for 30 min. c-Abl phosphorylation was measured by immunoprecipitation with c-Abl antibody and immunoblotting with phosphotyrosine antibody. Immunoblot of total c-Abl is shown as a loading control. (E) Effect of c-Abl siRNA on Cav-1 phosphorylation induced by the different concentrations of H₂O₂. (F) Effect of PP2 on H₂O₂-stimulated c-Abl activation. Cells were pretreated with PP2 or vehicle for 15 min and then stimulated with H₂O₂ (0.6 mmol/L) for 30 min. All blots are representative of 3 separate experiments.

H₂O₂ has been reported to induce Cav-1 phosphorylation via a c-Abl-dependent but Src-independent mechanism in fibroblasts.²² Therefore, we examined the potential role of c-Abl kinase in the regulation of Cav-1 phosphorylation in pulmonary endothelial cells. As shown in Figure 1D and Online Figure ID, H₂O₂ induced the phosphorylation (activation) of c-Abl kinase in a concentration-dependent manner. Downregulation of c-Abl with specific siRNA reduced Cav-1 phosphorylation induced by high but not low concentration of H₂O₂ (Figure 1E and Online Figure IE). To determine whether Src and c-Abl are activated by H₂O₂ independently of one another, we measured H₂O₂-induced c-Abl activation in the presence and absence of PP2. As shown in Figure 1F, PP2 significantly blocked H₂O₂ (0.6 mmol/L)-induced activation of c-Abl kinase (Online Figure IF) suggesting that c-Abl activation by oxidants occurs at least in part via activation of Src kinase.

Tyrosine Phosphorylation of Cav-1 Signals H₂O₂-Induced Transcellular Albumin Hyperpermeability

To evaluate whether H₂O₂ stimulates transcellular albumin permeability, we first addressed the possibility that H₂O₂ facilitates caveolae-mediated endocytosis of albumin in endothelial cells, the initial step in albumin transport via transcytosis.^{1,6,7-11,15-17} As shown in Figure 2A and 2B, H₂O₂ caused a concentration-dependent increase in Alexa 488 albumin and ¹²⁵I-albumin endocytosis in endothelial cells. Confocal images (Figure 2A) indicated that albumin was internalized by endocytic vesicles and that exposure of endothelial cells to H₂O₂ induced a significant increase in albumin uptake. In parallel, tracer ¹²⁵I-albumin uptake by endothelial cells following exposure to low dose H₂O₂ (0.05-0.2 mmol/L) was significantly increased (Figure 2B). We next determined the effect of H₂O₂ on transcytosis of ¹²⁵I-albumin using a well-established technique.^6,23 The concentration range of H₂O₂ (0.05-0.2 mmol/L) was chosen because at these levels, H₂O₂ did not disrupt the endothelial barrier (Figure 3A). Consistent with the activation of endocytosis, H₂O₂ induced a concentration-dependent increase in transendothelial ¹²⁵I-albumin transport (Figure 2C). These data suggest that H₂O₂ stimulates transcellular albumin transport via a vesicular pathway.

Figure 2. Phosphorylation of caveolin-1 is required for H₂O₂-induced endocytosis and transendothelial albumin transport.

(A) Confocal images showing H₂O₂-induced concentration-dependent increase in the uptake of Alexa 488-labeled albumin (green). The nucleus (blue) was stained with DAPI. Scale bars = 10 μm. Results are typical of 3 experiments. (B) H₂O₂ increased ¹²⁵I-albumin endocytosis in a concentration-dependent manner. (C) H₂O₂ induced a concentration-dependent increase in transendothelial transport of ¹²⁵I-albumin. (D,E) Effect of H₂O₂ on ¹²⁵I-albumin endocytosis (D) and transendothelial albumin permeability (E) in cells stably expressing WT and phosphorylation-defective Y14F-Cav-1 mutant. (F,G) Effect of Src inhibition and downregulation of c-Abl kinase on H₂O₂-induced increase in ¹²⁵I-albumin endocytosis (F) and transcytosis (G). n = 4-6 for each group. The baseline permeability values for WT and control groups were 7.8±1.2 μl·min^-1·cm^-2×10^-2 (E) and 6.1±1.1 μl·min^-1·cm^-2×10^-2 (G), respectively. *P <0.05 compared with control (B, C, F, and G) and WT (D, E) groups, †P <0.05 compared with respective (D,E) and control (F,G) groups.

Figure 3. Phosphorylation of Cav-1 is required for signaling H₂O₂-induced endothelial barrier disruption.

RLMVECs were grown to confluence, treated with H₂O₂ at the concentration indicated, and TER was recorded. In the panels on the left, the original TER recordings are shown, and in the panels on the right, the mean value (±SEM) of the peak TER responses to H₂O₂ (relative to the starting value) is plotted. (A) Dose-response relationship of H₂O₂-induced decrease in TER; note there was no effect of ∼0.2 mmol/L H₂O₂. (B) Phospho-defective Y14F Cav-1 mutant blocks the effect of H₂O₂ on TER. Over-expression of WT-Cav-1 in RLMVECs reduced the threshold of H₂O₂ needed to decrease TER. (C) Effect of Src inhibitor PP2 and downregulation of c-Abl kinase on H₂O₂-induced decrease in TER. n = 4-6 for each group. *P < 0.05, compared to control (untreated) groups; †P < 0.05 compared with respective H₂O₂ groups.

To further assess the mechanism of vesicular albumin transport, cells were pretreated with the caveolae disrupting agent methyl-β-cyclodextrin and Cav-1 siRNA. As shown in Online Figure IIA and IIB, methyl-β-cyclodextrin prevented H₂O₂-induced increase in endocytosis and transcytosis of ¹²⁵I-albumin. Furthermore, H₂O₂ increased albumin endocytosis and transendothelial ¹²⁵I-albumin flux in endothelial cells transduced with scrambled siRNA, whereas Cav-1 knockdown by ∼90% abolished these effects.

To address the role of Cav-1 phosphorylation in H₂O₂-induced increase in transendothelial albumin permeability, we measured H₂O₂-induced endocytosis and transendothelial transport of ¹²⁵I-albumin in RLMVEC lines overexpressing phosphorylation-defective Y14 F Cav-1 mutant (Y14F-Cav-1) and as a control, wild-type caveolin-1 (WT-Cav-1).^6,24 H₂O₂ (0.2 mmol/L)-induced increase in endocytosis and transcytosis of ¹²⁵I-albumin was significantly attenuated by Y14F-Cav-1 expression compared to WT-Cav-1 expressing cells (Figure 2D and 2E). Moreover, inhibition of Cav-1 phosphorylation by Src inhibitor PP2 reduced H₂O₂-stimulated endocytosis and transendothelial albumin permeability (Figure 2F and 2G) whereas downregulation of c-Abl kinase did not affect the endocytosis or transcytosis of albumin (Figure 2F and 2G).

Tyrosine Phosphorylation of Cav-1 Signals H₂O₂-Induced Paracellular Hyperpermeability

H₂O₂ at higher concentrations (0.4-0.8 mmol/L) induced inter-endothelial cell gap formation, detected as a reduction in TER, whereas H₂O₂ at lower concentrations (0.05-0.2 mmol/L) did not affect TER (Figure 3A). The effect of 0.4 and 0.6 mmol/L H₂O₂ was reversible in that TER returned to basal levels (4.7±0.7 Ω·cm²) within 5 hours, whereas endothelial cells treated with 0.8 mmol/L H₂O₂ did not fully recover (Figure 3A). Accordingly, 0.6 mmol/L H₂O₂ was chosen for subsequent experiments to explore the mechanism of H₂O₂-induced loss of monolayer integrity (increase in paracellular permeability).

To determine whether Cav-1 phosphorylation signals H₂O₂-induced increase in paracellular permeability, we measured H₂O₂-induced changes in TER in WT-and Y14F-Cav-1 expressing RLMVEC lines. The effect of low and high levels of Cav-1 phosphorylation on TER, induced by 0.2 and 0.6 mmol/L H₂O₂ respectively, was evaluated in endothelial monolayers. As shown in Figure 3A, 0.2 mmol/L H₂O₂ did not alter TER in native endothelial cells (non-transfected), but remarkably decreased TER in cells overexpressing WT-Cav-1 (Figure 3B). Importantly, 0.2 mmol/L H₂O₂ had no effect on TER in cells expressing the same level of mutant Y14F-Cav-1. The decrease in TER induced by 0.6 mmol/L H₂O₂ in endothelial cells expressing WT-Cav-1 was also significantly greater than the response observed in native cells and endothelial cells expressing Y14F-Cav-1 (Figure 3A and 3B). Inhibition of Cav-1 phosphorylation by Src inhibitor PP2 or c-Abl siRNA similarly reduced the magnitude of the H₂O₂ (0.6 mmol/L)-induced decrease in TER and resembled the response observed in cells expressing the phosphorylation-defective Cav-1 mutant. One difference noted was that the response was more transient in cells treated with c-Abl siRNA compared to PP2-treated cells (Figure 3C).

Cav-1 Phosphorylation Mediates H₂O₂-Induced Dissociation of VE-cadherin and β-catenin

To gain further insight into the regulatory mechanism of oxidant-induced endothelial barrier disruption and increase in paracellular permeability, we investigated the effect of H₂O₂-induced Cav-1 phosphorylation on the stability of adherens junctions by assessing VE-cadherin/β-catenin complexes using confocal microscopy and immunoprecipitation analysis. Confocal images showed significant co-localization of Cav-1 and β-catenin at cell-cell borders (Figure 4A) and immunoprecipitation studies also demonstrated an association between Cav-1 with β-catenin at baseline (Figure 4B). However, the association between Cav-1 and β-catenin at the cell borders observed by immunostaining and as well as the co-immunoprecipitated proteins measured in the immunoblots were reduced by ∼ 25% and 75% following exposure of endothelial cells to 0.2 and 0.6 mmol/L H₂O₂, respectively (Figure 4B).

Figure 4. Caveolin-1 phosphorylation mediates H₂O₂-induced dissociation of caveolin-1 from β-catenin.

(A) Effects of H₂O₂ on the co-localization of caveolin-1 (green) and β-catenin (red) in naïve RLMVECs. Cells were exposed to different concentrations of H₂O₂ for 30 min. The nucleus (blue) was stained with DAPI. Scale bars =10 μm. (B, C) The association of endogenous caveolin-1 (B) or over-expressed Myc-tagged WT-Cav-1 or Y14F-Cav-1 mutant (C) with β-catenin in naïve endothelial cells and stable endothelial cell lines was determined by immunoprecipitation and immunoblot (IP/IB) with anti-caveolin-1, anti-Myc, or anti-β-catenin antibodies. Left, representative Western blots for β-catenin and caveolin-1 (or Myc); Right, protein quantification by densitometry. The density of proteins in each untreated control group was used as a standard (1 arbitrary unit) to compare the relative density of the other groups. *P < 0.05 compared to control (untreated) groups; † P < 0.05 compared with H₂O₂ (0.2 mmol/L) groups.

To assess the role of Cav-1 phosphorylation in the mechanism of H₂O₂-induced dissociation of Cav-1 and β-catenin, RLMVEC lines expressing Myc-tagged WT-Cav-1 or Y14F-Cav-1 were stimulated, lysed, and immunoprecipitated with anti-Myc or anti-β-catenin antibodies. As shown in Figure 4C, β-catenin was associated with WT-Cav-1 and to a greater extent with Y14F-Cav-1 in untreated cells, suggesting that β-catenin binds Cav-1 in the non-phosphorylated state. Stimulation of cells with 0.2 mmol/L H₂O₂ led to a decrease in association between β-catenin and WT-Cav-1 whereas H₂O₂ had no effect on the association between β-catenin and Y14F-Cav-1 (Figure 4C). These data strongly argue that phosphorylation of Cav-1 leads to its dissociation from β-catenin.

We next determined whether the dissociation of Cav-1 from β-catenin induced by Cav-1 phosphorylation also led to the disruption of VE-cadherin/β-catenin complexes. As shown in Figure 5A, association between VE-cadherin and β-catenin was reduced by 15% and 80% upon exposure to 0.2 and 0.6 mmol/L H₂O₂ compared to untreated endothelial cells. Furthermore, H₂O₂- induced dissociation of VE-cadherin and β-catenin observed in WT-Cav-1 expressing cells was abolished in Y14F-Cav-1 expressing cells (Figure 5B).

Figure 5. Caveolin-1 phosphorylation mediates H₂O₂-induced dissociation of VE-cadherin and β-catenin.

The association of VE-cadherin with β-catenin was determined by immunoprecipitation and immunoblot analysis using anti-β-catenin or anti-VE-cadherin antibodies. Left, representative Western blots for VE-cadherin and β-catenin; Right, protein quantification by densitometry. The density of proteins in each untreated (control) group was used as a standard (1 arbitrary unit) to compare the relative density in the other groups. (A) Effect of H₂O₂ on the association of VE-cadherin with β-catenin in naïve endothelial cells. Note dissociation upon exposure to 0.6 mmol/L H₂O₂. (B) Effect of H₂O₂ on the association of VE-cadherin with β-catenin in cells stably expressing WT and phosphorylation-defective Y14F-Cav-1 mutant. *P < 0.05, compared to control (untreated) group (A) or WT alone group (B); n = 3/each group.

To further assess whether β-catenin translocates from the membrane following H₂O₂ stimulation at the higher concentration (0.6 mmol/L), cells were treated, lysed, fractionated, and immunoblotted. Figure 6A shows β-catenin predominantly localized in the membrane basally and then in the cytosolic compartment after treatment of endothelial cells with 0.6 mmol/L H₂O₂. Moreover, we observed H₂O₂-induced β-catenin translocation in WT-Cav-1 expressing cells but not in Y14F-Cav-1 expressing cells (Figure 6B). Similarly, confocal imaging showed wide-spread gap formation, increased β-catenin translocation into the cytosol, and a loss of β-catenin staining in membranes at sites of cell-cell contact in WT-Cav-1 expressing cells whereas β-catenin localization was unaffected and gaps in the monolayer did not form in Y14F-Cav-1 mutant expressing cells following exposure to 0.2 mmol/L H₂O₂ (Online Figure III). Thus, Cav-1 phosphorylation leads to its dissociation from β-catenin, disruption of β-catenin/VE-cadherin complexes, and β-catenin translocation from the membrane to cytosolic compartments.

Figure 6. Caveolin-1 phosphorylation is required for H₂O₂-induced β-catenin redistribution.

Cells were exposed to the indicated concentrations of H₂O₂ for 30 min and lysates separated into cytosolic and membrane fractions followed by Western blotting with β–catenin antibody to determine β–catenin localization. Top panel shows a representative Western blot for β-catenin and β-actin loading control; bottom panels show protein quantification by densitometry. The density of proteins in each untreated control group was used as a standard (1 arbitrary unit) to compare relative densities in the other groups. (A) H₂O₂ at 0.6 mmol/L concentration induced β-catenin redistribution from the membrane to cytosol in RLMVECs. (B) H₂O₂ (0.2 mmol/L) induced the translocation of β-catenin from cytosol to membrane in a manner dependent on Cav-1 phosphorylation in Cav-1 over-expressing cells. The redistribution of β-catenin was observed in cells stably expressing WT but not the phosphorylation-defective Y14F-Cav-1 mutant. n = 3. *P < 0.05, compared to control (untreated) group (A) or WT alone group (B).

H₂O₂-Induced Vascular Albumin Hyperpermeability and Lung Edema Formation Requires Cav-1 Phosphorylation in Mouse Lungs

To address whether phosphorylation of Cav-1 mediates H₂O₂-induced increase in pulmonary vascular hyperpermeability and edema formation, we infused H₂O₂ into isolated, perfused mouse lungs. As shown in Figure 7A, H₂O₂ (0.5 mmol/L) induced a robust increase in Cav-1 phosphorylation in wild type (WT) mouse lungs. We chose 0.5 mmol/L H₂O₂ as this dose reliably induced an increase in ¹²⁵I-albumin permeability (PS product) (Figure 7B) and lung edema formation (wet/dry ratio) (Figure 7C) in the isolated lung. In contrast, in cav-1^-/- mouse lungs, which exhibit reduced basal albumin PS (Figure 7B), H₂O₂ did not induce an increase in albumin and fluid permeability. To assess the role of Cav-1 phosphorylation in these permeability enhancing-effects of H₂O₂ in vivo, we used liposome-mediated Cav-1 cDNA delivery to rescue vascular endothelial Cav-1 expression in cav-1^-/- mouse lungs^6,25,26 with either WT-Cav-1 or the phosphorylation-defective Y14F-Cav-1 mutant (Figure 7A). Consistent with the in vitro data, H₂O₂ induced an increase in albumin PS product and lung wet/dry weight ratio in WT-Cav-1 expressing mouse lungs, whereas these effects were not reconstituted upon expression of Y14F-Cav-1 in cav-1^-/- lungs (Figure 7B and 7C).

Figure 7. H₂O₂-induced increase in lung microvessel albumin permeability and pulmonary edema formation requires caveolin-1 phosphorylation.

Cav-1^-/- mice were injected intravenously with liposomes containing WT- or Y14F-Cav-1 cDNA. After 48 h, lungs were isolated and perfused with H₂O₂ (0.5 mmol/L) for 30 min. (A) Western blots show increase in Cav-1 phosphorylation in whole lung homogenates induced by H₂O₂, the absence of Cav-1 in null mice, and exogenous expression of Myc-tagged WT-Cav-1 and Y14F-Cav-1 in cav-1^-/- lungs. Note that H₂O₂ induced phosphorylation of reconstituted WT-Cav-1 in the isolated mouse lung. (B) Effect of H₂O₂ on pulmonary microvessel ¹²⁵I-albumin permeability (PS product) in wild-type (cav-1^+/+) and cav-1^-/- lungs with or without rescue with WT-Cav-1 or Y14F-Cav-1 mutant cDNA. (C) H₂O₂ induced and increase in lung wet/dry (W/D) weight ratio in cav-1^+/+ lungs but not in cav-1^-/- lungs. Rescue of Cav-1 expression with WT-Cav-1 but not Y14F-Cav-1 cDNA restored the lung edema response to H₂O₂. n = 6/each group. * P < 0.05 compared with control groups (cav-1^+/+ mouse without H₂O₂ treatment); †P < 0.05 compared with cav-1^+/+ mouse with H₂O₂ treatment group.

Discussion

This study provides strong evidence indicating that Cav-1 phosphorylation in endothelial cells plays a fundamental role in the mechanism of oxidant-induced pulmonary vascular hyperpermeability. Oxidant-induced increase in caveolae-mediated albumin transport and decrease in endothelial barrier integrity were both dependent on tyrosine phosphorylation of Cav-1, suggesting Cav-1 phosphorylation is a common signal regulating transcellular and paracellular permeability pathways in lung microvessels. The molecular mechanisms regulating transcellular and paracellular permeability induced by Cav-1 phosphorylation following oxidant stress appear to be different. The increase in transcellular permeability in response to low concentration of oxidant is entirely mediated by Src-dependent Cav-1 phosphorylation, whereas increased paracellular permeability induced by high concentration of oxidant depends on both Src- and c-Abl-mediated Cav-1 phosphorylation and subsequent regulation of β-catenin localization.

We established that exposure of pulmonary microvascular endothelial cells to pathophysiological concentrations of H₂O₂ (0.05-0.8 mmol/L)²⁷ stimulated Cav-1 phosphorylation in a concentration-dependent manner. Our results are distinct from previous reports which showed that higher concentrations of H₂O₂ (≥ 1 mmol/L) were required to induce an increase in Cav-1 phosphorylation in cultured bovine pulmonary artery,¹² aortic,^13,14 and human umbilical vein¹⁴ endothelial cells.

Cav-1 Tyr¹⁴ is a principal target for Src kinase phosphorylation in response to oxidative stress,^12-14,24,28 and thus we predicted that treatment of endothelial cells with Src inhibitor PP2 to inactivate Src kinase⁶ would block H₂O₂-induced Cav-1 Tyr¹⁴ phosphorylation. However, our results demonstrated that the level of phospho-Cav-1 following PP2 treatment was only decreased by approximately 65%, suggesting the existence of a Src-independent pathway, particularly when higher concentrations of H₂O₂ were used to stimulate Cav-1 phosphorylation. Depletion of c-Abl with specific siRNA demonstrated that c-Abl also contributed to H₂O₂-induced Cav-1 phosphorylation in pulmonary microvascular endothelial cells. In contrary to previous findings that c-Abl phosphorylates Cav-1 independently of Src,^22,29 our results showed that c-Abl activation and subsequent Cav-1 phosphorylation was in part dependent on Src activity. It is likely that Src family kinases directly phosphorylate tyrosine residues in the kinase domain of c-Abl, leading to enhanced activity.³⁰

Studies from our laboratory^6,7,9-11 and others^16,17 have suggested that Cav-1 phosphorylation plays an essential role in the mechanisms regulating caveolae formation and caveolae-mediated endocytosis and transcytosis of albumin in microvascular endothelial cells. Using ¹²⁵I-albumin and Alexa 488-albumin tracers, we observed that H₂O₂ (0.05-0.2 mmol/L) increased both the endocytosis and transcytosis of albumin. Cholesterol-depleting agent methyl-β-cyclodextrin and siRNA-induced depletion of Cav-1 prevented H₂O₂-induced endocytosis and transcytosis of albumin, indicative of a caveolae-dependent mechanism. Furthermore, the H₂O₂-induced increase in endocytosis and transcytosis of ¹²⁵I-albumin in WT-Cav-1 expressing cells was abolished in phospho-defective Cav-1 mutant over-expressing endothelial cell line. Noteworthy is that Src inhibitor PP2 but not c-Abl siRNA inhibited endocytosis and transcytosis of albumin induced by H₂O₂. These findings provide further support for the concept that Src-dependent phosphorylation of Cav-1 mediates oxidant-induced transcellular transport via caveolae in pulmonary microvascular endothelial cells.

The present data for the first time show that Cav-1 phosphorylation, which is an important mechanism regulating caveolae-mediated transcellular permeability,^7-11 is also a crucial regulator of oxidant-induced paracellular hyperpermeability. Expression of a Cav-1 mutant lacking the tyrosine phosphorylation site resulted in the inhibition of H₂O₂-induced decrease in TER, a measure of loss of endothelial junctional integrity. Furthermore, overexpression of wild type Cav-1 in endothelial cells reduced the threshold concentration of H₂O₂ needed to disrupt endothelial cell-cell junctions, presumably due to an increase in the total amount of phosphorylated Cav-1.^6,22 In agreement with these findings, inhibition of Cav-1 phosphorylation by either Src inhibitor PP2 or c-Abl siRNA partially attenuated H₂O₂-induced decrease in TER. These results highlight the importance of Cav-1 Tyr¹⁴ phosphorylation in the regulation of H₂O₂-induced endothelial barrier disruption. A recent study showed that oxidized phospholipid-induced Cav-1 phosphorylation is associated with sphingosine 1-phosphate receptor signaling in caveolae that leads to endothelial barrier enhancement.³¹ However, it remains unclear whether Cav-1 phosphorylation per se plays an important signaling role in strengthening the endothelial barrier in this scenario.

In cav-1^{-/- 32} and Cav-1 siRNA-treated²⁰ mice, an increase in the number of open inter-endothelial junctions in pulmonary capillaries and veins was observed, indicating there is a compensatory mechanism that increases basal paracellular permeability in the absence of the vesicular (transcellular) permeability pathway.³³ In our study, however, both transcellular and paracellular permeability pathways were stimulated upon Cav-1 phosphorylation, suggesting these two pathways are affected in a coordinate-manner in response to oxidative stress. Consistent with previous findings,^6,34 the present study demonstrated that the increase in transendothelial permeability occurred prior to an increase in paracellular permeability. These data suggest that the mechanism mediating increased transcellular permeability, namely Src phosphorylation of Cav-1, is also a trigger signal for increasing paracellular permeability. Thus, it is likely that the cooperative mechanism between the two pathways via Cav-1 phosphorylation may play an important role in the development and progression of vascular hyperpermeability and inflammatory lung injury.

Recent studies indicate the important role of Cav-1 in stabilization of adherens junctions.^19,35 Knockdown of Cav-1 was shown to be associated with a significant decrease in both VE-cadherin and β-catenin localized at inter-endothelial junctions.¹⁹ In our studies with confluent endothelial cell monolayers, we observed that Cav-1 co-localized with adherens junction-associated protein β-catenin, consistent with a previous study conducted in MDCK cells.¹⁸ Confocal imaging and immunoprecipitation studies indicated that exposure of endothelial cells to higher concentrations of H₂O₂ caused a marked dissociation of β-catenin from Cav-1 as well as VE-cadherin. H₂O₂-induced Src activation may also promote direct phosphorylation of VE-cadherin and/or β-catenin, inducing their dissociation from cytoskeletal anchors.^1,4,36 However, using a Cav-1 overexpressing endothelial cell line, we showed that dissociation of β-catenin from Cav-1 and VE-cadherin following exposure to H₂O₂ was dependent on Cav-1 Tyr¹⁴ phosphorylation as these changes did not occur in endothelial cells expressing the phosphorylation-defective Cav-1 mutant. Furthermore, Cav-1 phosphorylation promoted the translocation of β-catenin from the membrane to cytosol. These data support a model in which, when phospho-Cav-1 is present in cell-cell junctions at significant levels, it negatively regulates adherens junction stability by decreasing the association between VE-cadherin and β-catenin. At this stage, we cannot rule out the possibility that Cav-1-mediated endocytosis of VE-cadherin and/or β-catenin may lead to the disassembly of adherens junction complexes and thereby contribute to the increase in paracellular permeability.³⁷ While Cav-1 phosphorylation is known to regulate caveolae-mediated endocytosis, as also shown herein, this explanation of β-catenin internalization is less plausible knowing that the co-immunoprecipitation of β-catenin with Cav-1 decreased upon Cav-1 phosphorylation.

The significance of Cav-1 phosphorylation as a molecular signal in endothelial permeability regulation is not clear. Phosphorylation of Cav-1 may reduce interactions between Cav-1 and other signaling proteins that are negatively regulated by association with Cav-1.³⁸ Increased Cav-1 phosphorylation may also be necessary for translocation of proteins normally associated with Cav-1 at the plasma membrane.³⁹ Our data support these hypotheses, as indicated by the dissociation of Cav-1 and β-catenin and subsequent translocation of β-catenin from the membrane to the cytosol in a manner dependent on increased Cav-1 phosphorylation induced by high concentration of H₂O₂.

We also found that H₂O₂-induced increase in pulmonary vascular permeability was dependent on Cav-1 Tyr¹⁴ phosphorylation in vivo, consistent with our findings from pulmonary endothelial monolayers. Mice lacking Cav-1 expression were resistant to H₂O₂-induced increase in pulmonary vascular hyperpermeability and edema formation. The pulmonary vascular response to H₂O₂ was rescued by adding back WT-Cav-1 to the cav-1^-/- pulmonary vasculature, but not by expression of the phospho-defective Cav-1 mutant. These findings, together with those from pulmonary endothelial cells, strongly suggest that Cav-1 phosphorylation plays a pivotal role in the mechanism of increased pulmonary vascular permeability and edema formation evoked by oxidants.

In summary, our findings implicate Cav-1 phosphorylation as a critical mechanism mediating oxidant-induced pulmonary vascular hyperpermeability. H₂O₂ stimulated caveolae-mediated transcellular transport and the opening of inter-endothelial junctions in a manner dependent on Cav-1 phosphorylation. Although these studies focused on oxidative stress as a means of increasing endothelial permeability and inducing acute lung injury, it is possible that our findings have broader applicability in understanding the mechanisms and development of inflammatory pulmonary vascular hyperpermeability. Therefore, therapeutic inhibition of Cav-1 phosphorylation may be an effective means of limiting lung vascular injury by preventing increased transcellular albumin permeability and stabilizing the endothelial junctional barrier.

Supplementary Material

supp

Legends to Supplemental Figures

Online Figure I. H₂O₂ induced Src and c-Abl activation and caveolin-1 (Cav-1) phosphorylation. Densitometric analysis of phosphorylated caveolin-1 (p(Y14) Cav-1), Src (p(Y418) Src) and c-Abl (c-Abl-pY) in RLMVEs after different treatments (n = 3 samples). The density of proteins in each control (untreated) group was used as a standard (1 arbitrary unit) to compare relative densities in the other groups. *P < 0.05, compared to control groups (untreated); †P < 0.05, compared to H₂O₂ (0.6 mM) groups. (A) Relative densities of the bands of p(Y14) Cav-1and p(Y418) Src expression, as shown in Figure 1A. (B) Relative densities of the bands of p(Y14) Cav-1 expression, as shown in Figure 1B. (C) Relative densities of the bands of p(Y14) Cav-1 expression, as shown in Figure 1C. (D) Relative densities of the bands of c-Abl-pY expression, as shown in Figure 1D. (E) Relative densities of the bands of p(Y14) Cav-1 expression, as shown in Figure 1E. (F) Relative densities of the bands of c-Abl-pY expression, as shown in Figure 1F.

Online Figure II. H₂O₂ increases caveolae-mediated transendothelial ¹²⁵I-albumin permeability in Rat lung microvascular endothelial cell (RLMVEC) monolayers. (A) Effects of pretreatment of RLMVECs with methyl-β-cyclodextrin (MβCD) and downregulation of caveolin-1 (Cav-1) with siRNA on transendothelial ¹²⁵I-albumin permeability. (B) Effects of pretreatment of RLMVECs with MβCD and downregulation of Cav-1 with siRNA on endothelial endocytosis of ¹²⁵I-albumin. (C) Western blots showing Cav-1 knockdown. Endothelial cells were transfected with Cav-1 siRNA or scrambled (Sc) siRNA. Results are typical of 3 experiments. n = 6-8/group (A and B). *P < 0.05, compared to control (untreated) groups; †P < 0.05 compared with respective H₂O₂ (0.2 mM) groups.

Online Figure III. Immunofluorescent staining of β-catenin and F-actin in RLMVECs over-expressing WT vs. Y14F-Cav-1. Cells were grown to confluence and treated with H₂O₂ (0.2 mM) for 30 min. After fixation and permeabilization, cells were incubated with anti-β-catenin primary antibody followed by FITC-conjugated secondary antibody (green). The F-actin and nucleus were stained with Alexa 546 phalloidin (red) and DAPI (blue), respectively. In the third panel (WT + 0.2 mM H₂O₂), note gap formation (arrow), increased β-catenin translocation into the cytosol (arrow head), and absence of β-catenin in areas where gaps in the monolayer have formed. The fourth panel (Y14F-Cav-1 + 0.2 mM H₂O₂) shows no gap formation as well as β-catenin translocation in response to H₂O₂. Scale bars =10 μm. WT = wild type-caveolin-1 expressing cells; Y14F = Y14F-caveolin-1 mutant expressing cells

Non-standard Abbreviations and Acronyms

Cav-1

caveolin-1

H₂O₂

hydrogen peroxide

ICAM-1

intercellular adhesion molecule-1

MDCK

Madin-Darby canine kidney

PKC

protein kinase C

PP2

4-amino-5-(4-chlorophenyl)-7-(t-butyl) pyrazolo[3,4-d] pyrimidine

RLMVEC

rat lung microvascular endothelial cells

TER

transendothelial electrical resistance

Tyr

tyrosine

VE

vascular endothelial

WT

wild type