Autophagy inhibitor 3-methyladenine protects against endothelial cell barrier dysfunction in acute lung injury

Reagents.

Lipopolysaccharide (LPS) of Escherichia coli origin, the autophagy inhibitor 3-MA, and the phosphoinositide 3-kinase (PI3K) inhibitor LY294002 were obtained from Sigma-Aldrich (St. Louis, MO); human thrombin from Enzyme Research Laboratories (South Bend, IN); antibodies to VE-cadherin from Abcam (Cambridge, MA) and BD Biosciences (San Jose, CA); and antibodies to ATG5 and β-actin from Cell Signaling Technology (Beverly, MA) and Santa Cruz Biotechnology (Santa Cruz, CA), respectively. All other materials were obtained from Thermo Fisher Scientific (Waltham, MA).

Murine model of ALI.

Male 8- to 10-wk-old wild-type C57BL/6 mice (Jackson, Bar Harbor, ME) were exposed to an aerosol of saline alone or saline containing Escherichia coli LPS (0.5–0.75 mg/ml, 12 ml) for 30 min, as described elsewhere (2, 39). All animal care and treatment procedures were approved by the University of Rochester Committee on Animal Resources and performed in accordance with National Institutes of Health guidelines.

Evaluation of lung inflammation and injury.

After the experiment, bronchoalveolar lavage (BAL) fluids and lungs were collected as described elsewhere (2, 40). BAL fluids were analyzed for total protein levels using the bicinchoninic acid kit (Pierce, Rockford, IL) and for albumin levels using ELISA kits (R & D Systems, Minneapolis, MN). Lung tissues were homogenized in radioimmunoprecipitation (RIPA) buffer [50 mM Tris·HCl, pH 7.4, 150 mM NaCl, 0.25 mM EDTA, pH 8.0, 1% deoxycholic acid, 1% Triton X, 5 mM NaF, and 1 mM sodium orthovanadate supplemented with protease inhibitor cocktail (Sigma-Aldrich)] or in hexadecyltrimethylammonium bromide buffer (pH 6.0), as described elsewhere (2, 10). The levels of proinflammatory mediators (vascular cell adhesion molecule-1, monocyte chemoattractant protein-1, and IL-1β) in mouse lung homogenates were determined using ELISA kits (R & D Systems), as described elsewhere (2). Myeloperoxidase (MPO) activity in the lung tissues was monitored to determine sequestration of polymorphonuclear leukocytes in the lung, as described elsewhere (2, 3, 10).

Wet-to-dry lung weight ratio and Evans blue-conjugated albumin (EBA) were measured in mice that were not subjected to BAL. At 1 h before they were euthanized, these mice were injected retroorbitally with EBA (30 mg/kg) to monitor vascular leak, as described elsewhere (31, 45). Blood was removed from the lungs by gentle infusion of 10 ml of phosphate-buffered saline containing 5 mM EDTA through the right ventricle. The lungs were excised en bloc and blotted dry, and the right lungs were snap-frozen in liquid nitrogen. The left lungs were weighed before and after being dried at 60°C for 24 h to calculate lung wet-to-dry weight ratio. Dried lungs were incubated with formamide at 60°C for 24 h and centrifuged, and supernatant absorbance at 620 and 740 nm was recorded. Tissue heme pigment contamination was corrected using readings of absorbance at 740 nm. The extravasated EBA concentration in lung homogenate was calculated against a standard curve and expressed as micrograms of EBA per milliliter.

Lung histopathology and immunohistochemistry.

After the experiment, lungs were harvested from mice that did not undergo BAL. The lungs were fixed in 10% formalin, embedded in paraffin, cut into 5-μm sections, mounted onto slides, and stained with hematoxylin-eosin for histological analysis (10). Sections were imaged with a Nikon Eclipse E400 microscope, and the representative images were taken at ×400.

Endothelial cells.

Human pulmonary artery EC (HPAEC) were obtained from Lonza (Walkersville, MD) and cultured as described elsewhere (30) in gelatin-coated flasks with use of endothelial basal medium 2 (EBM2) with BulletKit additives (BioWhittaker, Walkersville, MD). Experiments were performed in HPAEC below passage 6.

RNAi knockdown.

Predesigned short-interfering RNA (siRNA) specific for human ATG5 (si-ATG5) and a nontargeting siRNA control (si-Con) were purchased from Dharmacon (Lafayette, CO). HPAEC were transfected with si-ATG5 or si-Con using DharmaFect1 siRNA transfection reagent (Dharmacon) essentially as described elsewhere (2). Briefly, 50–100 nM siRNA and DharmaFect1 were mixed together and then added to cells that were 50–60% confluent. At 24–36 h after transfection, cells were used for experiments in which transendothelial electrical resistance (TER) was measured or cells were lysed to monitor ATG5 levels.

Immunoblot analysis.

Immunoblotting was performed as described previously (9). Briefly, equal amounts of protein from cell lysates or lung homogenates prepared in RIPA buffer were subjected to SDS-PAGE and then transferred onto nitrocellulose membranes for Western blotting, as described elsewhere (9). Representative blots were from the same membrane, which may have more samples in various groups.

Immunofluorescence.

Confluent HPAEC monolayers grown on gelatin-coated coverslips were subjected to immunofluorescence staining, as described elsewhere (2, 11). Alexa Fluor 488-phalloidin (Fisher Scientific, Pittsburg, PA) and VE-cadherin antibody (BD Biosciences, San Jose, CA) were used to visualize F-actin filaments and AJs, respectively. DNA staining with Hoechst dye was used to visualize nuclei. Images were obtained using a fluorescence microscope (Nikon Instech, Tokyo, Japan).

Measurement of endothelial permeability by TER.

The integrity of the endothelial barrier was determined using an electrical cell-substrate impedance-sensing system (Applied Biophysics, Troy, NY), as described elsewhere (2, 36). Briefly, confluent HPAE monolayers grown on gelatin-coated gold microelectrodes were treated with thrombin or LPS in the presence or absence of 3-MA (2 mM) and/or LY294002 (30 μM) as indicated. TER was measured over time and normalized to baseline resistance.

Statistical analysis.

Values are means ± SE. Data were analyzed by standard one-way ANOVA. Significance between the groups was determined using Tukey’s test (Prism 5.0, GraphPad Software, San Diego, CA). In some cases (Figs. 1, 5, 6, and 7), Student’s t-test was performed for comparisons between experimental groups. P < 0.05 was considered statistically significant.

Protective and therapeutic utility of the autophagy inhibitor 3-MA prevents and treats LPS-induced lung vascular inflammation and injury.

We tested the possibility that autophagy plays an important role in lung vascular inflammation and injury. We first determined if autophagy is induced in an aerosolized bacterial LPS inhalation mouse model of ALI (2, 39). A marked increase in autophagy, as evidenced by an increase in the levels of ATG5 [conjugated with ATG12 (ATG5-ATG12), ~55 kDa], an essential regulator of autophagy (33, 35, 53), was noted in the lungs of mice exposed to LPS (Fig. 1). Intriguingly, lungs from mice exposed to the autophagy inhibitor 3-MA alone also showed an increase in ATG5 levels; however, lungs from mice exposed to 3-MA before LPS (3-MA + LPS) showed a significant decrease in ATG5 levels compared with those exposed to LPS alone. These data suggest that while 3-MA increases the constitutive (basal) autophagy, it protects against LPS (stress)-induced autophagy. We next determined if the induction of autophagy contributes to lung inflammatory injury evoked by LPS. We performed a “protection study (3-MA+LPS),” in which mice were exposed to the autophagy inhibitor 3-MA both before and after LPS challenge (Fig. 2A; hereafter referred to as “before”). We also evaluated the therapeutic utility of autophagy inhibition against ALI by performing a “treatment study (LPS + 3-MA),” in which mice were first challenged with LPS to induce lung injury and then treated with 3-MA (Fig. 2B; hereafter referred to as “after”). Histological examination of the lung showed increased interstitial thickening and neutrophil infiltration in mice exposed to LPS (Fig. 2, Ca vs. Cc). By contrast, in mice treated with 3-MA before or after LPS challenge, these markers of lung inflammatory injury were substantially reduced (Fig. 2, Cc vs. Cd and Ce). Consistent with this finding, LPS induced an increase in the levels of proinflammatory mediators (vascular cell adhesion molecule-1, monocyte chemoattractant protein-1, and IL-1β) and neutrophil infiltration in the lung, but these responses were significantly inhibited in mice treated with 3-MA before or after LPS challenge (Fig. 3). Analysis of BAL fluid showed that LPS induced a nearly fourfold increase in BAL protein and albumin levels; however, this response was inhibited in mice treated with 3-MA before or after LPS challenge (Fig. 4, A and B). We also determined the effect of 3-MA on LPS-induced EBA extravasation and increase in lung wet-to-dry weight ratio, important indexes of lung vascular leak and tissue edema. We found a significant decrease in LPS-induced lung vascular leak and tissue edema in mice treated with 3-MA before or after LPS challenge (Fig. 4, C and D). Because the levels of VE-cadherin, a major regulator of endothelial barrier integrity (14, 50, 51), are reduced in inflamed lungs (13, 16), we tested the possibility that autophagy regulates lung vascular leak by decreasing VE-cadherin levels. Indeed, we observed a significant decrease in VE-cadherin levels in the lungs of mice challenged with LPS (Fig. 5); however, this response was inhibited in mice treated with 3-MA (Fig. 5). Collectively, these data identify a critical role of autophagy in the mechanism of lung vascular inflammation and injury. These data also underscore the importance of targeting autophagy via 3-MA as a viable therapeutic approach to control lung injury.

Inhibition of autophagy protects against EC barrier dysfunction by suppressing VE-cadherin disassembly and actin stress fiber formation.

We investigated the possibility that 3-MA exerts its protective effect on lung vascular injury by its ability to prevent EC barrier disruption caused by thrombin, an edemagenic and proinflammatory agonist, the concentration of which is elevated in plasma and lavage fluids of patients with ALI/acute respiratory distress syndrome (12, 20). We determined the effect of 3-MA on thrombin-induced change in TER as measured by electric cell-substrate impedance sensing. A characteristic time-dependent decrease in TER with maximal change at ~0.5 h was observed upon thrombin challenge of confluent EC monolayers (Fig. 6A) (2, 28). Treatment of cells with 3-MA significantly increased the baseline TER (Fig. 6B) and was also effective in reducing and reversing the thrombin-induced decrease in TER (Fig. 6A). We also evaluated if 3-MA can restore EC barrier integrity after it is disrupted. Indeed, 3-MA was effective in restoring EC barrier function when added after thrombin or LPS challenge (Fig. 6, C–E). To exclude the possibility that the protective effect of 3-MA is mediated by a PI3K-inihibitable pathway other than autophagy, we assessed the effect of the PI3K inhibitor LY294002 on thrombin-induced EC barrier disruption. Interestingly, we found that, unlike 3-MA, LY294002 not only failed to protect the barrier from disruption but also inhibited recovery of barrier function after thrombin challenge (Fig. 7A). These data are consistent with an earlier report showing a role of PI3K in barrier enhancement (21). To further reinforce the conclusion that the effect of 3-MA on EC barrier disruption is uncoupled from its ability to inhibit PI3K, we determined if 3-MA can exert its protective effect in the presence of LY294002. Indeed, 3-MA was effective in protecting against thrombin-induced barrier disruption, even in the presence of LY294002 (Fig. 7B). To definitively establish the role of autophagy in EC barrier disruption, we determined the effect of siRNA-mediated knockdown of ATG5, an essential regulator of autophagy (17, 34, 35), on EC barrier disruption. Depletion of ATG5 by this approach (Fig. 8, A and B) also attenuated thrombin-induced EC barrier disruption (Fig. 8C). These data indicate an important role of autophagy in EC barrier disruption and the utility of 3-MA in restoring the barrier integrity of EC.

Because formation of actin stress fibers and loss of VE-cadherin at AJs are the key determinants of EC barrier dysfunction (4, 7, 8, 27, 29), we examined if 3-MA protects EC barrier integrity by preventing these events (actin stress fiber formation and VE-cadherin loss at AJs). Visualization of actin stress fibers showed that thrombin challenge increased stress fiber formation and that pretreatment of cells with 3-MA attenuated this this response (Fig. 9A). Similarly, we observed that thrombin caused a marked decrease in immunostaining of VE-cadherin at AJs and that this response was inhibited in 3-MA-treated cells (Fig. 9B). Notably, 3-MA also increased the basal immunostaining of VE-cadherin at AJs (Fig. 9B), consistent with the 3-MA-mediated increase in baseline TER (Fig. 6B). Studies have shown that thrombin and other stimuli induce cleavage of VE-cadherin, which also contributes, at least in part, to disassembly of AJs (7, 15, 42). Therefore, we determined the possibility that inhibition of autophagy promotes AJ integrity by preventing cleavage of VE-cadherin. Thrombin induced cleavage of VE-cadherin in a time-dependent manner (Fig. 9C). Importantly, the time course of VE-cadherin cleavage was consistent with changes in TER caused by thrombin (Fig. 9C, Fig. 6, A, C, and D, and Fig. 7, A and B). We found that thrombin-induced VE-cadherin cleavage was significantly inhibited in 3-MA-treated cells (Fig. 9D). Together, these data show a role of autophagy in regulating EC barrier dysfunction by virtue of promoting actin cytoskeleton rearrangement and AJ disruption via VE-cadherin cleavage.