Overdistention of lung tissue during mechanical ventilation may be one of the factors that initiates ventilator-induced lung injury (VILI). We hypothesized that cyclic mechanical stretch (CMS) of the lung epithelium is involved in the early events of VILI through the production of reactive oxygen species (ROS). Cultures of an immortalized human airway epithelial cell line (16HBE), a human alveolar type II cell line (A549), and primary cultures of rat alveolar type II cells were cyclically stretched, and the production of superoxide (O2−) was measured by dihydroethidium fluorescence. CMS stimulated increased production of O2− after 2 h in each type of cell. 16HBE cells exhibited no significant stimulation of ROS before 2 h of CMS (20% strain, 30 cycles/min), and ROS production returned to control levels after 24 h. Oxidation of glutathione (GSH), a cellular antioxidant, increased with CMS as measured by a decrease in the ratio of the reduced GSH level to the oxidized GSH level. Strain levels of 10% did not increase O2− production in 16HBE cells, whereas 15, 20, and 30% significantly increased generation of O2−. Rotenone, a mitochondrial complex I inhibitor, partially abrogated the stretch-induced generation of O2− after 2 h CMS in 16HBE cells. NADPH oxidase activity was increased after 2 h of CMS, contributing to the production of O2−. Increased ROS production in lung epithelial cells in response to elevated stretch may contribute to the onset of VILI.
- ventilator-induced lung injury
in patients with acute hypoxemic respiratory failure, mechanical ventilation strategies employed to improve arterial oxygenation can initiate or exacerbate lung injury. Mechanisms implicated in ventilator-induced lung injury (VILI) include high inspiratory volumes or “volutrauma,” mechanical stress-induced inflammation or “biotrauma,” and cyclical airway collapse and reopening. Low tidal volume ventilation strategies that minimize overdistention of the lung tissue decrease mortality in patients with acute respiratory distress syndrome (ARDS), presumably by limiting or preventing additional VILI (40). Recent animal studies in a mouse model of VILI demonstrated that high tidal volume ventilation without underlying lung injury led to increased neutrophil infiltration (5) and proinflammatory cytokines in the lungs (46), indicating a link between volutrauma and biotrauma. However, the factors leading from mechanical stress to the initiation and propagation of VILI remain uncertain. One potential initiating signal may be increased production of reactive oxygen species (ROS) in the lung epithelium in response to increased stretch.
Increased ROS production in response to mechanical stress has been described in a variety of cell types. Endothelial cell production of ROS has been shown to increase in response to shear stress (4, 6, 7, 10, 19, 30, 45, 50) and to cyclic mechanical stretch (CMS; see Refs. 2, 29, 47, 48). CMS also stimulates ROS production in vascular smooth muscle cells (13, 22, 31) and cardiac myocytes (1). Although superoxide (O2−) appears to be the initial species generated in these cell types, there is disagreement as to the source of ROS production. The NADPH oxidase system (7, 13, 29, 30), mitochondrial production (2, 21), and the xanthine oxidase system (30) have all been implicated as potential sources for increased O2− production in response to mechanical stress. However, whether these potential sources are activated directly or indirectly by mechanical stress is unclear. Little data are available demonstrating increased ROS production by stretched alveolar or airway epithelial cells. A 25% increase in ROS production, as measured by 2,7-dichlorodihydrofluorescein (DCFH) fluorescence caused by interaction with H2O2, was measured in response to 30% biaxial stretch of A549 cells, a human alveolar type II-like cell line (42). In another study, A549 cells subjected to cyclic stretch produced significantly increased levels of isoprostane (a marker of oxidant injury) after 0.5 h, and levels of reduced glutathione (GSH, an endogenous antioxidant) were significantly decreased after 1 h of CMS (23). Pretreatment with antioxidants prevented the change in isoprostane levels in this study. However, the source(s) of increased ROS production and species generated in pulmonary epithelial cells exposed to mechanical stretch have not been determined.
In the current study, we examined whether cyclic mechanical strain induced production of O2− in the main epithelial constituents of the lungs, airway and alveolar epithelial cells. O2− production was significantly enhanced in an SV40-transformed human airway epithelial cell line (16HBE), A549 cells, and isolated primary rat alveolar type II (rat AT II) epithelial cells undergoing CMS. Increased ROS production was dependent on the time and magnitude of stretch. The source of the O2− was a combination of the NADPH oxidase system and the mitochondria (complex I). Mitochondrial involvement may be initiated by a direct distention of the mitochondria resulting from mechanical stretch.
MATERIALS AND METHODS
Unless otherwise noted, all chemicals were purchased from Sigma-Aldrich Chemical (St. Louis, MO).
16HBE cells were generously provided by Dr. Deiter Gruenert and were cultured in DMEM (GIBCO, Carlsbad, CA) with 4 mM glutamine supplemented with 10% FBS (Cellgro, Herndon, VA or GIBCO), 20 mM HEPES (Sigma), and 2.5 μM Plasmocin (Invivogen, San Diego, CA). These cells were first treated for 2 wk with 25 μM Plasmocin. A549 cells were cultured in DMEM with 4 mM glutamine supplemented with 10% FBS (Cellgro) and 1% penicillin/streptomycin solution (GIBCO). Both cell types were cultured until 85–100% confluence before being seeded on the BioFlex plates coated with collagen type I (Flexcell International, Hillsborough, NC) for studies on confluent monolayers. The medium was changed every other day for all cultures.
Isolation and culture of rat AT II cells.
Primary rat AT II cells were isolated according to established procedures (8, 26, 32). Briefly, male Sprague-Dawley rats were anesthetized with phenobarbital, killed by exsanguination, and their lungs were excised. The trachea was catheterized, and the pulmonary vasculature was perfused via the pulmonary artery with solution II (in mM: 140 NaCl, 5 KCl, 2.5 Na2HPO4, 10 HEPES, 1.3 MgSO4, and 2.0 CaCl2; pH 7.4) to remove the blood. The air spaces were then lavaged with solution I (in mM: 140 NaCl, 5 KCl, 2.5 Na2HPO4, 6 glucose, 0.2 EGTA, and 10 HEPES; pH 7.4) to remove free, nonepithelial cells. Elastase (4.3 U/ml in solution II; Worthington Biochemical, Lakewood, NJ) was instilled in the air space and incubated at 37°C for 10 min. This was repeated, and the large airways and heart were removed. The remaining lung tissue was then minced in 5 ml FBS and 250 μl of 250 μg/ml DNase (Sigma) per 4 lungs. The minced lungs were filtered through gauze followed by a nitrocellulose membrane, and the cell suspension was collected. The suspension was centrifuged and resuspended in AT II culture medium [DMEM with 10% heat-inactivated FBS (HyClone, Logan, UT), 4 mM glutamine, 1% penicillin/streptomycin, and 0.25 μM amphotericin B (Sigma)] and plated on untreated petri dishes coated with IgG. The plates were incubated for 1 h to allow nonepithelial cells such as macrophages to bind to the IgG. The plates were “panned” to loosen nonspecifically bound cells, pooled, and counted. BioFlex plates coated with collagen type I by the manufacturer were then coated with 32.3 μg/ml human fibronectin (Roche Life Sciences, Indianapolis, IN), and cells were seeded to confluence at 3.0 × 106/well in AT II culture medium. Experiments were performed on day 2 after isolation. AT II cells were identified using Nile Red (Sigma) staining of lamellar bodies, and >90% of the cells were Nile Red positive on day 2.
Application of mechanical strain.
Strain was applied using a Flexercell strain unit 4000 (Flexcell International), a vacuum-driven device that applies biaxial strain to cells cultured on Silastic-bottomed well plates. The device allows control of the frequency of strain and magnitude up to 30% linear strain. 16HBE and A549 cells were strained at 20% elongation and 30 cycles/min (cpm) for the initial studies. This strain profile was previously found to maximally inhibit wound closure in airway epithelial cells (38). Rat AT II cells were strained at 15% elongation and 15 cpm because higher levels of strain and frequency resulted in increased cell detachment from the culture plate.
Measurement of O2− production.
Intracellular O2− production was measured by the fluorescent intensity of oxoethidium after dihydroethidium (DHE; Molecular Probes, Eugene, OR) was oxidized by O2−. Oxoethidium binds to DNA, becoming highly fluorescent. After treatment, cells were rinsed two times with warm serum- and phenol red-free (PR-Free) DMEM and loaded with 5 μM DHE in PR-Free DMEM at 37°C in the incubator for 30 min. The cells were then rinsed one time with room temperature PR-Free DMEM, fresh PR-Free DMEM at room temperature was added, and the cells were imaged immediately. Images were taken on an inverted fluorescent microscope (TE300; Nikon Instruments, Lewisville, TX) outfitted with a ×20 PlanFluor objective (Nikon Instruments), rhodamine filter set for excitation and emission (Chrorma Technology, Rockingham, VT), and digital CCD camera (Roper Scientific, Tucson, AZ). Images were acquired using MetaMorph software (Universal Imaging, Downingtown, PA) for 10 s with a ×4 neutral density filter engaged. Four separate images were taken per well, avoiding areas of low confluence and areas near the edge of the well. The average pixel intensity per image was calculated, and an average intensity per well was determined. To assess the involvement of the mitochondria, treatment with the pharmacological inhibitor of the mitochondrial complex I, rotenone (25 nM), was initiated at the beginning of the stretch protocol.
Measurement of reduced and oxidized glutathione levels.
A protocol from Kamencic et al. (25) adapted for a fluorescence plate reader was used to determine levels of reduced and oxidized GSH (GSSG). 16HBE cells were stretched or held static for various periods of time and then placed on ice. The cells were rinsed two times with ice-cold Ca2+/Mg2+-free Dulbecco's phosphate-buffered saline (DPBS). Ice-cold cell lysate buffer (0.5 ml, 0.2 mM EDTA in Ca2+/Mg2+-free DPBS) was added to each well. The cells were scraped and transferred to centrifuge tubes at 4°C. Samples were sonicated on ice for 20 s. The samples were centrifuged for 10 min at 10,000 g and 4°C. Two 100-μl aliquots of each sample were diluted with 200 μl deionized H2O/aliquot. One diluted aliquot from each sample was completely reduced by the addition of 15 μl of 4 M triethanolamine (TEA) to determine the total amount of GSH. A standard curve of GSH with and without TEA was created. Wells of a black 96-well plate were loaded with 50 μl of diluted sample and 150 μl assay solution (100 μM monochlorobimane, 1 U/ml glutathione-S-transferase, and 0.2 mM EDTA in Ca2+/Mg2+-free DPBS), and the plate was incubated in the dark for 30 min at room temperature. The plate was then read in a fluorescence plate reader (Molecular Devices, Sunnyvale, CA) with excitation at 380 nm and emission at 470 nm. To calculate GSSG, the amount of reduced GSH was subtracted from the TEA-treated sample and divided by two to account for the dimerization. The ratio of reduced GSH to GSSG was then calculated to monitor the redox state of the cells.
NADPH oxidase assay.
The assay used to determine NADPH oxidase activity was adapted from previous studies (7, 51). Cells were stretched or held static, placed on ice, and rinsed three times in ice-cold DPBS. Cells were placed in 0.5 ml homogenization buffer (1 mM EGTA, 10 μg/ml aprotinin, 0.5 μg/ml leupeptin, and 0.5 mM phenylmethanesulfonyl fluoride in Ca2+/Mg2+-free DPBS), scraped, homogenized manually, and placed in tubes. Cell homogenate (50 μl) was added to 3 ml warm chemiluminescence buffer (10 μM Lucigenin, 1 mM EGTA, and 150 mM sucrose in Ca2+/Mg2+-free DPBS) containing NADPH (50 μM) in a scintillation vial. Chemiluminescence was measured (in duplicate) for 3 min in a Packard scintillation counter. Diphenylene iodonium (DPI, 10 μM) was included in the chemiluminescence buffer for selected static and stretched samples.
Staining and imaging of mitochondria of stretched cells.
16HBE cells were cultured on collagen I-coated Silastic membranes for the StageFlexer (Flexcell International), which is a single membrane stretching device adapted for use on a microscope stage. Live 16HBE cells were stained with 500 nM MitoTracker Orange (Molecular Probes) for 30 min and rinsed two times with warm DPBS. Membranes were placed in the StageFlexer, and images were acquired with a ×60 objective (Nikon) before and after the application of ∼17% biaxial strain. Because images included contributions of structures from multiple planes, we selected structures that were likely to be in thin or confined regions of the cells such as beneath the nucleus and near the edges of the cells. The total length of each mitochondrion was measured by tracing a line down the center line of each structure to determine the overall length before and during strain.
All values are presented as means ± SE. All statistical analyses were performed with the SigmaStat statistical package (version 2.03; Jandel Scientific, San Rafael, CA). One-way ANOVA was performed for comparisons of multiple treatments. If this showed significant difference between the treatments, a Student's t-test with the Bonferroni correction was performed to determine significant differences between the individual conditions. All of the ANOVA tests were run with multiple comparisons against the static control except for the data collected using the mitochondrial inhibitor rotenone. These experiments employed multiple comparisons between all of the different treatments. Significant differences are based on a threshold of P ≤ 0.05.
CMS induces O2− production in pulmonary epithelial cells.
To determine if CMS increased O2− production in airway and alveolar epithelial cells, we subjected 16HBE cells, A549 cells, and primary rat AT II cells to 2 h of CMS and measured the O2− production via DHE fluorescence. Figure 1A shows a region of representative images of 16HBE cells showing the changes in DHE fluorescence intensity after 2 h of CMS. O2− production was significantly increased in each of the cell types that we tested compared with static controls (see Fig. 1B). The A549 cells exhibited the strongest response to CMS (142.5 ± 0.5%; P < 0.05, n = 3) after 2 h of stretch, but 16HBE and rat AT II cells displayed the same order of magnitude increase in O2− production (126.7 ± 1.9 and 114.9 ± 2.8%; P < 0.05, n = 9 and 6, respectively). However, rat AT II cells were stretched at a lower frequency and magnitude because of cell detachment at the higher levels.
Time course of O2− production in response to stretch.
We further investigated the time course of O2− production in 16HBE cells in response to CMS. Figure 2 shows that there was no significant change after 0.5 h of stretch, but the O2− production increased significantly after 2 h, achieving the maximal response after 6 h (157.2 ± 9.2%; P < 0.05, n = 9), and remained significantly elevated up to 18 h (118.4 ± 5.4%; P < 0.05, n = 9) before returning to baseline after 24 h. These results suggest an initial increase in O2− production in response to the initiation of stretch followed by a later adaptation to increased O2− production.
Oxidation of GSH increased in 16HBE cells exposed to CMS.
To determine whether the measured O2− production was increased because of a change in the redox state of the cells, we measured the relative levels of oxidized (GSSG) and reduced (GSH) glutathione. Reduced glutathione is an important cellular defense against ROS that forms a dimer with a disulfide bond (GSSG) upon oxidation. The ratio of GSH to GSSG is a marker of the redox state of the cell and is reduced when the cells experience an oxidative stress. We measured this ratio to determine if the ROS generated by CMS caused changes in the redox state of 16HBE cells. Figure 3 shows that 0.5 h of CMS caused a significant reduction in the normalized GSH-to-GSSG ratio to 50.8 ± 4.4% of the 0-h control (P < 0.05, n = 4). The GSH-to-GSSG ratio remained significantly reduced through 24 h, reaching 40.5 ± 7.5% of control at its lowest point after 24 h of CMS indicating a persistent oxidative stress. Thus, although O2− production had returned to control levels after 24 h of CMS (Fig. 2), the ratio of GSH to GSSG remained significantly reduced.
O2− production is dependent on the magnitude of CMS.
To determine whether increased O2− production was dependent on the magnitude of CMS, 16HBE cells were stretched for 2 h at 10, 15, 20, and 30% strain, and the O2− production was measured by DHE fluorescence. Figure 4 shows that 10% CMS did not significantly increase O2− production, but levels of strain 15% and higher resulted in significant increases after 2 h of strain.
Mitochondrially derived O2− is partially responsible for cyclic mechanical strain-induced production of O2−.
Mitochondria are the largest source of intracellular ROS production from normal cell respiration and have been implicated in ROS production stimulated by mechanical deformation of endothelial cells (2, 21). To investigate whether stretch-induced O2− production was the result of increased mitochondrial generation, we treated 16HBE cells with the complex I inhibitor rotenone. As shown in Fig. 5, rotenone (25 μM) partially but significantly inhibited the enhanced O2− production after 2 h of CMS from 143.7 ± 4.8 to 120.7 ± 2.5% (P < 0.05, n = 6), but the response was still significantly greater than control (P < 0.05, n = 6). Cells treated with the vehicle used to solubilize the rotenone (chloroform) demonstrated a response similar to that of untreated controls after 2 h of CMS (138.2 ± 6.6%, P < 0.05, n = 3). These results suggest that at least part of the increased O2− production was generated by the mitochondria.
Mitochondria are deformed by mechanical strain.
Although increased mitochondrial production of O2− has been previously reported in endothelial cells (2, 21), the mechanism of stimulation is not known. To investigate whether increased mitochondrial production of O2− was initiated by direct deformation of the mitochondria, we examined mitochondria in live cells labeled with a mitochondrial specific dye, MitoTracker Orange. Labeled cells were imaged before and during the application of strain. As indicted in Fig. 6, mitochondria were observed to distend as the Silastic membrane was strained biaxially by 17%. Measurements made of the total length of selected mitochondria (6 mitochondria from 4 different fields) increased from 3.1 ± 0.6 to 4.1 ± 0.7 μm (32.1% linear strain), indicating local deformation when a global strain was applied. The observed mitochondrial deformation may contribute to initiation of strain-induced O2− production.
NADPH oxidase activity is increased by cyclic mechanical strain.
We next investigated whether NADPH oxidase contributed to the stretch-induced production of ROS in pulmonary epithelial cells. NADPH oxidase has previously been implicated in the ROS produced by CMS of endothelial cells (3, 24, 29, 48) and vascular smooth muscle cells (22, 31). As shown in Fig. 7, cyclic stretch of 16HBE cells significantly increased the activity of NADPH oxidase to 204.8 ± 27.8% relative to unstretched controls (P < 0.05, n = 6). Also, the normalized activity of NADPH oxidase was reduced to near background levels when 10 μM DPI was added to the chemiluminescence reaction solution of cell homogenates exposed to either 0 or 2 h of CMS. The chemiluminescence signal was the same as background if cell homogenate was added to the reaction cocktail without NADPH (data not shown).
A recent study demonstrated that mechanical ventilation with lower tidal volumes reduced the mortality of ARDS patients (40), and these findings are supported by studies demonstrating that lower tidal volume settings decreased injury to epithelial and endothelial cells in ventilated rats (12) and decreased inflammation in a mouse model of VILI (5, 46). Thus higher tidal volumes may contribute to VILI through mechanisms related to overdistention of the lungs. We hypothesized that overdistention of the lung epithelium leads to increased production of ROS that may contribute to VILI. We exposed pulmonary epithelial cells to CMS and measured the production of O2− with DHE. We found that O2− production was significantly increased after 2 h of CMS of 16HBE, A549, and primary rat AT II cells compared with static controls (see Fig. 1B). We examined O2− production to focus our study on the potential initial events stimulated by mechanical stretch.
CMS of A549 cells has been shown previously to increase generation of ROS. Upadhyay et al. (42) demonstrated that A549 cells stretched for 1 h at 30% biaxial elongation and 30 cpm increased generation of H2O2 by 25% compared with static controls (42). In that study, the cells were loaded with DCFH, which is oxidized to the fluorescent compound dichlorofluorescein by H2O2. H2O2 is one of several downstream metabolites of O2− generation. We measured a greater increase in ROS production by A549 cells using DHE after 2 h of CMS (Fig. 1B). The lower increase in ROS production detected by Upadhyay et al. (42) may be because of the metabolism of O2− to other pathways not detected by DCFH. In addition, DCFH has been reported to leak out of cells (37) and in preliminary experiments we too observed leakage of DCFH from cells. Also, stretching cells loaded with this dye may cause further leakage through transient, nonlethal strain-induced membrane breaks (44). Oxidized DHE binds to DNA and fluoresces, and thus the fluorescent dye remains inside the cells. We loaded the dye after the stretch concluded, reducing potential leakage of the DHE dye. Differences between our study and previous studies could also be because of the duration of the stretch. We measured changes in ROS production over time in 16HBE cells (Fig. 2) and found that CMS of 0.5 h did not elicit an increase in O2− production, but 2 h caused a significant increase. Although we did not measure ROS production after 1 h of CMS in A549 cells, the response may be similar to that reported by Upadhyay et al. (42).
Other investigators have used DHE to monitor changes in O2− production in human lung endothelial cells exposed to hyperoxia (35), nutrient-deprived human aortic endothelial cells (28), and ANG II treatment of bovine aortic endothelial cells (9). Lopez et al. (28) used images of DHE staining to quantify changes in the generation of O2− with nutrient deprivation. We used a similar method to quantify changes in the current study. Parinandi et al. (35) used cell lysate fluorescence from DHE-loaded cells to measure changes in O2− production. O2− production has also been measured by an HPLC method in cells and intact tissues (9) based on the fluorescence of oxoethidium. This study also demonstrated the specificity of DHE for O2− by showing that fluorescence was not increased after treatment with H2O2, peroxynitrite, or hypochlorous acid.
ROS generation caused by CMS resulted in a significant reduction of the ratio of GSH to GSSG in 16HBE cells, indicating an oxidant stress that occurred as early as 0.5 h, and the ratio remained significantly reduced through 24 h (see Fig. 3). An increase in ROS production as a result of CMS was not observed in 16HBE cells until 2 h of strain, as measured by DHE fluorescence (see Fig. 4), but the GSH-to-GSSG ratio indicated an oxidant stress after only 0.5 h of stretch. This implies that the cellular antioxidant defenses were able to absorb the excess ROS before being overwhelmed by 2 h, resulting in an increase in DHE fluorescence. However, after 24 h, DHE measurements returned to basal levels, whereas the GSH-to-GSSG ratio remained significantly reduced. Jafari et al. (23) measured GSH levels in A549 cells and found that levels were significantly reduced after 1 h of CMS at 15% strain and 20 cpm, returned to basal levels after 2 and 3 h, but showed a significant increase after 4 h of CMS. Our results corroborate their findings of an early oxidant stress caused by CMS. However, in our study, although the ratio of GSH to GSSG decreased, reduced GSH levels were maintained at a constant level in 16HBE cells subjected to CMS with no significant changes over 24 h, whereas total GSH levels increased significantly after 24 h, resulting in the continued reduction in the ratio (data not shown). This may not fully explain the reduction in O2− after 24 h, but we did not investigate activity or protein levels of antioxidants more specific to O2− such as superoxide dismutase. Increased levels of superoxide dismutase after 24 h might account for the decrease in O2− production with the sustained decrease in the GSH-to-GSSG ratio. Differences between antioxidant systems in 16HBE and A549 cells may also contribute to the observed differences. However, measurement of both GSH and GSSG levels is important to assess the redox state of the cells.
We explored the effects of varying the magnitude of strain on ROS production of pulmonary epithelial cells and found that application of strain levels <15% elongation resulted in O2− production that was not significantly different from control. Application of strain levels ≥15% caused significant increases in ROS production (Fig. 4). To our knowledge, this is the first time ROS production has been studied as a function of the magnitude of strain in pulmonary epithelial cells. These results suggest that a threshold level of strain exists above which elevated ROS production occurs. The strain levels experienced by the lungs in vivo during either spontaneous breathing or mechanical ventilation are difficult to determine. Tschumperlin and Margulies (41) measured the changes in epithelial basement membrane surface area (EBMSA) of lungs fixed at different inflations and found the expected nonlinear response. Inflation from 24% to total lung capacity (TLC) caused a 40% increase in EBMSA, whereas inflation from 42% of TLC to TLC resulted in a 34% increase. This change in surface area corresponds to a linear strain of ∼18%. In injured lungs, some regions may be collapsed or fluid-filled, functionally reducing the volume of the lungs. Therefore, during mechanical ventilation, the remaining volume of the lungs may be significantly overinflated, causing even greater distention and injury (20). However, the actual levels of strain in the airways and alveoli in injured lungs have not been well characterized. Some authors have suggested that linear strain of ∼10% occurs during normal breathing and that higher levels occur during high tidal volume mechanical ventilation, but there are no direct measurements of this. Our results showed that ROS production was significantly increased at strains of 15% or greater. During mechanical ventilation at high tidal volumes or even during lower tidal volumes, if the lung volume is effectively reduced by injury, the average distention of the tissue may be above this threshold and cause generation of ROS from the epithelium.
In this study, we were unable to stretch the primary rat AT II cells at 20% elongation and 30 cpm because significant detachment of the cells from the membrane occurred. Therefore, the values reported in Fig. 1B for rat AT II cells are from cells stretched at 15% strain and 15 cpm on day 2 of culture, and the cells remained attached under these conditions. Oswari et al. (33) showed a substantial loss of viability (49%) in rat AT II cells cultured for 1 day on fibronectin-coated Silastic and exposed to 1 h of CMS of 25% change in surface area (∼12% linear strain) and 15 cpm. However, treatment with keratinocyte growth factor or culture of the cells on a matrix deposited by AT II cells for 5 days resulted in levels of viability similar to static controls. We are currently adapting our culture methods to better promote attachment and survival at increased strain and frequency. Despite the lower strain and frequency of stretch, we observed a significant stimulation of ROS production, albeit at a lower level than that seen in the cell lines tested.
The source of mechanical strain- or shear stress-induced ROS is currently debated, and potential sources in different cell types include NADPH oxidase (7, 13, 29, 30), xanthine oxidase (30), nitric oxide synthase, and the mitochondria (2, 21). Previous studies with endothelial cells concluded that the source of ROS was primarily NADPH oxidase, since treatment with the flavoprotein inhibitor DPI abolished the response. However, DPI is a nonspecific inhibitor of NADPH oxidase that also inhibits other potential sources of ROS generation, including the enzymes of the electron transport chain in mitochondria. Recent studies have demonstrated that mitochondria are involved in strain-induced production of ROS in endothelial cells. Ichimura et al. (21) and Ali et al. (2) have shown that the complex I inhibitor rotenone eliminated strain-induced ROS production. Ali et al. also used ethidium-treated human umbilical vein endothelial cell cultures to eliminate mitochondrial DNA and the electron transport chain proteins and found that the CMS-induced increase in ROS production was eliminated. A more specific pharmacological inhibitor of NADPH oxidase, apocynin, was employed and was found to have no effect on CMS-induced ROS generation. We found that CMS of 16HBE cells stimulated NADPH oxidase activity and that ROS production was partially sensitive to inhibition of mitochondrial complex I. The activity of NADPH oxidase, measured by Lucigenin chemiluminescence, increased significantly after 2 h of CMS (Fig. 6). Treatment of the cell lysates with DPI (10 μM) inhibited the increase in activity and reduced the activity in static cell lysates to similar levels. The specificity of the response was tested by comparing results with and without an excess of NADPH (50 μM) in the reaction mixture. Chemiluminescent readings were similar to background when NADPH was not included in the reaction cocktail (data not shown). Rotenone partially inhibited CMS-induced increases in DHE fluorescence (Fig. 5). Taken together, our results suggest that CMS-induced increases in ROS production are the result of both NADPH oxidase and mitochondrial generation in pulmonary epithelial cells.
If mitochondrial production of O2− is stimulated by mechanical stretch, then what is the initiating event? We hypothesized that global strain applied to the cells may lead to direct distention of the mitochondria, since mitochondria are attached to both the actin (39) and microtubule (16, 17) components of the cytoskeleton. We observed substantial deformation of mitochondrial structures when cells were stretched (Fig. 7). We speculate that mechanical distention of mitochondrial structures may activate pathways that lead to generation of ROS. For example, mitochondrial ATP-sensitive K+ () activation has been linked to generation of ROS (11, 15), and activation of the mitochondrial channel has been linked to mitochondrial swelling (18). However, the mechanism of ROS production through activation of the mitochondrial channel has not been determined, and no studies have connected the distention caused by mitochondrial swelling to ROS production. One of the major limitations of our measurements of mitochondrial deformation is that the images that were analyzed (Fig. 6) were from a single plane of focus. Because we did not collect images at different z-planes, we cannot exclude the possibility that the change in length that was observed was due to the appearance of mitochondrial structures from out-of-focus planes in the images of stretched cells. However, these images were not taken with a confocal microscope, and thus contributions from multiple planes are more likely to appear in the image. In addition, 16HBE cells are quite thin, and we specifically identified mitochondrial structures that were in more confined areas near the cell boarder or beneath the nucleus, making the possibility of out-of-plane regions appearing in the focal plane less likely. These considerations, however, still do not rule out the possibility that apparent deformation was the result of the appearance of out-of-focus structures.
One proposed mechanism for VILI is that increased mechanical deformation of lung tissue stimulates an acute inflammatory response characterized by increased levels of proinflammatory cytokines and inflammatory cell infiltration. In previous studies, high tidal volume mechanical ventilation without underlying lung injury led to both increased cytokine production (46) and enhanced neutrophil infiltration (5). CMS has been shown to stimulate production of proinflammatory cytokines in both airway and alveolar cells. CMS of BEAS-2B, an airway epithelial cell line, resulted in an increased production of interleukin (IL)-8 (34). A549 cells when subjected to CMS increased production of proinflammatory cytokines IL-8 (23, 27, 43, 49), IL-6 (23), and transforming growth factor-β (49). Primary cultures of rat AT II also exhibited CMS-induced production of IL-8 (36). Exposure to H2O2 or exogenous O2− production by xanthine/xanthine oxidase and exposure of rat AT II cells to hyperoxia also caused increased production of the proinflammatory cytokines IL-1β, IL-6, and tumor necrosis factor-α (14). Ali et al. (2) demonstrated that stretch-induced expression of vascular cell adhesion molecule-1 and activation of NF-κB could be abrogated by blocking mitochondrial generation of ROS in endothelial cells. Thus inflammatory responses may be initiated by stretch-induced production of ROS. We demonstrated in this study that stretch stimulated increased ROS production in both airway and alveolar epithelial cells, that the response was both time varying and magnitude dependent, and that ROS are derived from both NADPH oxidase and mitochondria.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-064981, HL-004479, HL-63886, and HL-72902.
We gratefully acknowledge Dr. Harry Ischiropoulos and Dr. Navdeep Chandel for helpful discussions regarding this study and Dr. Karen Ridge for discussions regarding type II cell isolations. We also thank Charlean Luellen for expert technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society