Superoxide Flux in Endothelial Cells via the Chloride Channel-3 Mediates Intracellular Signaling

Abstract

Reactive oxygen species (ROS) have been implicated in both cell signaling and pathology. A major source of ROS in endothelial cells is NADPH oxidase, which generates superoxide (O₂^.−) on the extracellular side of the plasma membrane but can result in intracellular signaling. To study possible transmembrane flux of O₂^.−, pulmonary microvascular endothelial cells were preloaded with the O₂^.−-sensitive fluorophore hydroethidine (HE). Application of an extracellular bolus of O₂^.− resulted in rapid and concentration-dependent transient HE oxidation that was followed by a progressive and nonreversible increase in nuclear HE fluorescence. These fluorescence changes were inhibited by superoxide dismutase (SOD), the anion channel blocker DIDS, and selective silencing of the chloride channel-3 (ClC-3) by treatment with siRNA. Extracellular O₂^.− triggered Ca²⁺ release in turn triggered mitochondrial membrane potential alterations that were followed by mitochondrial O₂^.− production and cellular apoptosis. These “signaling” effects of O₂^.− were prevented by DIDS treatment, by depletion of intracellular Ca²⁺ stores with thapsigargin and by chelation of intracellular Ca²⁺. This study demonstrates that O₂^.− flux across the endothelial cell plasma membrane occurs through ClC-3 channels and induces intracellular Ca²⁺ release, which activates mitochondrial O₂^.− generation.

INTRODUCTION

Reactive oxygen species (ROS) have been implicated in cellular signaling processes as well as a cause of oxidative stress (Taniyama and Griendling, 2003). It is now appreciated that a major source of ROS in the vasculature is through one or more isoforms of the phagocytic enzyme NADPH oxidase, a membrane-localized protein which generates the superoxide (O₂^.−) anion on the extracellular surface of the plasma membrane (Lambeth, 2004). As a charged and short-lived anion, it is believed that O₂^.− flux is insufficient to initiate intracellular signaling due to the combination of poor permeability through the phospholipid bilayer (Tanabe et al., 2005) and a rapid dismutation to its uncharged and more stable derivative, hydrogen peroxide (Finkel, 2003). However, recent evidence has indicated discrete signaling roles for both O₂^.− and H₂O₂ (Madesh and Hajnoczky, 2001; Devadas et al., 2002). In our studies, we have found that extracellular O₂^.−, but not H₂O₂, leads to Ca²⁺ signaling and apoptosis in pulmonary endothelial cells (Madesh et al., 2005). This indicates that extracellular O₂^.− produced by NADPH oxidase or other sources either crosses the plasma membrane or modifies cell surface proteins to mediate cell signaling.

Previous studies of erythrocytes and amniotic cells have provided evidence for O₂^.− transport through anion channels, which could be effectively blocked by 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid (DIDS; Lynch and Fridovich, 1978; Ikebuchi et al., 1991). DIDS also effectively blocked release of O₂^.− from mitochondria into the cytosol (Han et al., 2003) without affecting ROS production (Korchak et al., 1980). Despite these reports, whether O₂^.− crosses the cell membrane to elicit a discrete intracellular signal remains controversial (Babior, 1999; Mikkelsen and Wardman, 2003).

The present study evaluated the response of pulmonary microvascular endothelial cells (PMVECs) to extracellular O₂^.−. Our findings using a fluorophore trap demonstrate that O₂^.− enters the cell through a chloride channel-3 (ClC-3)-dependent mechanism. Further, extracellular O₂^.−, through a Ca²⁺-mediated signaling event, stimulates the production of O₂^.− by the mitochondria. This observation provides a model by which extracellular O₂^.− can propagate intracellular ROS signaling.

MATERIALS AND METHODS

Materials

Hydroethidine (HE), MitoSOX Red, propidium iodide (PI), rhodamine 123, Fluo-4/AM, and BAPTA-AM were purchased from Invitrogen (Carlsbad, CA). The Basic Nucleofector Kit for primary mammalian endothelial cells was purchased from Amaxa Biosystems (Gaithersburg, MD). Silencer Predesigned ClC-3 (ID 60947 and 60858), negative control 1, and Cy3-labeled GAPDH small interfering RNA (siRNA) were purchased from Ambion (Austin, TX). Primers for β-actin, ClC-3, and ClC-4 were obtained from Operon Biotechnologies (Huntsville, AL). Rabbit polyclonal anti-ClC-3 antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). pEYFP-Mito was purchased from Clontech (BD Biosciences, Mountain View, CA). KO₂, apocynin, angiotensin II, thapsigargin, and all other chemicals were purchased from Sigma (St. Louis, MO). Mice were obtained from The Jackson Laboratory (Bar Harbor, ME).

Cell Culture

Immortalized human pulmonary microvascular endothelial cells (HPMVEC clone ST1.6R) were generated as described previously (Krump-Konvalinkova et al., 2001) and cultured in Medium-199 supplemented with 15% FBS, glutamax, antibiotics, and endothelial cell growth supplement. Isolation, characterization, and propagation of mouse pulmonary microvascular endothelial cells (MPMVEC) from wild-type (C57BL/6) and gp91^phox gene-targeted mice have been previously described (Milovanova et al., 2006). Primary cells were cultured in DMEM supplemented with 10% FBS, nonessential amino acids, endothelial cell growth supplement, and antibiotics and used between passages 6 and 20. For some experiments cells were transfected with various DNA constructs by electroporation (Amaxa Biosystems, Gaithersburg, MD) using programs T-23 or S-05 according to the manufacturer's instructions.

Imaging of O₂^.− Flux

PMVECs cultured on 0.2% gelatin-coated 25-mm diameter glass coverslips were loaded with the O₂^.−-sensitive dye HE (10 μM) in DMEM for 10 min at 37°C. Cells were then placed on a temperature-controlled stage, and images were recorded every 5 s for 5 min using LaserSharp software (Bio-Rad Laboratories, Hercules, CA) on a Bio-Rad Radiance 2000 imaging system (Bio-Rad Laboratories) equipped with a Kr/Ar-ion laser source at 568- and 605-nm excitation and emission, respectively, using a 60× oil objective. A bolus of KO₂ was added after 1 min of baseline recording. KO₂ was prepared in a 1.8 mM concentration as described previously (Reiter et al., 2000). KO₂ was not directly applied to the image field to avoid alterations in microscope focus. For inhibitor studies, DIDS (200 μM) was present during HE loading and KO₂ addition. Antioxidant enzymes were added immediately before imaging and mitochondrial inhibitors were added similarly as KO₂. HE fluorescence was quantified by nuclear masking of all cells in the field. For angiotensin II (Ang II) and thrombin experiments, HPMVECs were cultured on coverslips, and the medium was replaced with M-199 containing 2% FBS 18 h before study. HPMVECs were pretreated with DIDS (300 μM) or apocynin (Apo; 2 μM) for 10 min before addition of 2 μM Ang II or 0.5 U/ml thrombin. For imaging, cells were loaded with HE (10 μM) and five independent fields were recorded by confocal microscopy.

Cell-free HE Oxidation Measurement

HE fluorescence (40 μM) in a 2 ml solution of PBS was monitored in a multiwavelength-excitation dual wavelength-emission fluorimeter (Delta RAM, PTI, Birmingham, NJ) using 510- and 568-nm excitation and emission, respectively. Briefly, KO₂ or the xanthine/xanthine oxidase (X/XO; X-100 μM; XO-50 mU/ml) O₂^.−-generating system was added to the solution after 60 s of baseline recording. Total recording time was 3 min. DMSO, H₂O₂, and KOH were added in a similar manner. For dismutation studies, KO₂ was added to a solution of PBS containing 1000 U superoxide dismutase (SOD), mixed briefly, and then added to the HE solution. Results were normalized to the baseline fluorescence before addition of O₂^.−. The stable oxidation product was assessed in intact MPMVECs loaded with HE. Briefly, cells were treated for 20 min with either antimycin A or a 10 μM bolus of KO₂ and a spectral scan of emission wavelengths was performed using an excitation wavelength of 494 nm.

O₂^.− -induced HE Fluorescence and Mitochondrial ROS Production

PMVECs cultured on coverslips were loaded with HE and mounted on a confocal microscope stage as described earlier. After measurement of HE baseline fluorescence, KO₂ (10 μM) or X/XO (X-100 μM; XO-20 mU/ml) was added to the medium evenly across the coverslip and gently agitated to mix the solution. After 20 min, five fields were chosen for imaging and quantitation. To measure O₂^.− in mitochondria, MPMVECs were transfected with 2 μg/ml pEYFP-Mito (Clontech, BD Biosciences) and cultured in complete medium. Colonies were selected and passaged to increase the number of green fluorescent protein (GFP)-positive cells and plated on gelatin-coated coverslips. Cells after loading with the mitochondrial-O₂^.−–sensitive fluorophore MitoSOX Red (Molecular Probes; 1.25 μM) were exposed to KO₂, Tg, and DIDS as described above. In some experiments, MPMVECs were pretreated with BAPTA-AM (50 μM) for 30 min before KO₂ application.

ClC-3 Knockdown

Confluent MPMVECs (5 × 10⁶) were washed and placed in serum-free DMEM before transfection by electroporation with 250 pmol of either ClC-3 or negative control siRNA. To establish transfection efficiency, PMVECs were also transfected with Cy3-labeled GAPDH siRNA. Cells were then transferred to the appropriate culture vehicle and cultured in RPMI medium supplemented with 10% FBS, essential amino acids, endothelial cell growth supplement, and antibiotics. To confirm transfection, cells at 24 h after transfection were counterstained with the nuclear marker DAPI and images were acquired using MetaMorph software (Molecular Devices, Downingtown, PA) via epifluorescence microscopy (TE2000U, 10× objective; Nikon, Melville, NY). After 24 h, medium was replaced with standard growth medium and changed daily for an additional 48 h. Cells at 60 h after transfection were lysed and evaluated for ClC-3 mRNA by RT-PCR or imaged via confocal microscopy, respectively. Cells at 72 h after transfection were lysed and ClC-3 protein level was assessed by Western blotting using a rabbit polyclonal anti-ClC-3.

Total RNA Extraction and RT-PCR

Total RNA was prepared from wild-type and siRNA transfected MPMVECs using an RNeasy Mini Kit (Qiagen, Valencia, CA). The Transcriptor first-strand cDNA synthesis kit (Roche Applied Science, Indianapolis, IN) was used to reverse transcribe cDNA from 2 μg of RNA using both random hexamer and anchored-oligo(dT)₁₈ primers. For ClC-3, the forward and reverse primers were GCGTGAGAACCGCGTTACT and GCTTTCAGGAGAGGTTACGT, respectively. For ClC-4, the forward and reverse primers were GATGGGCATTATTTTGAGAAG and CAGTAGCATGCGAATACCCC, respectively. For β-actin, the forward and reverse primers were ATGGATGACGATATCGCTGC and CTTCTGACCCATACCCACCA, respectively. The PCR amplification profile consisted of an initial denaturation at 95°C for 2 min followed by 35 cycles of a 30-s denaturation at 95°C 30 s, annealing at 55°C for 30 s, and 1-min extension at 72°C, followed by a final 10-min extension step at 72°C using GoTaq DNA Polymerase (Promega, Madison, WI). PCR products were separated by electrophoresis on a 2% agarose/TBE gel and visualized by ethidium bromide staining.

Mitochondrial Membrane Potential

Cells cultured on coverslips were incubated with the cationic potentiometric fluorescent dye rhodamine 123 (25 μM) for 20 min at 37°C. After dye loading, the cells were washed and resuspended in DMEM. Images were recorded every 5 s for 5 min using the Bio-Rad Radiance 2000 imaging system with excitation at 488 nm. A decrease in mitochondrial membrane potential (ΔΨ_m) results in loss of rhodamine 123 from the mitochondria into the cytoplasm and the nucleus. Quantitation of the ΔΨ_m change was determined by nuclear masking for fluorescence of all cells in the field. Treatment with DIDS and other agents was performed as described above.

Measurement of [Ca²⁺]_i Mobilization

Endothelial cells adherent to 25-mm-diameter glass coverslips were loaded with the cytosolic Ca²⁺ indicator Fluo-4/AM (5 μM; Invitrogen, Carlsbad, CA) at room temperature for 30 min in extracellular medium (ECM) containing 121 mM NaCl, 5 mM NaHCO₃, 10 mM Na-HEPES, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 2 mM CaCl₂, 10 mM glucose, and 2.0% bovine serum albumin (BSA), pH 7.4, in the presence of 100 μM sulfinpyrazone and 0.003% pluronic acid. After dye loading, the cells were washed and resuspended in the experimental imaging solution (ECM containing 0.25% BSA) and images recorded every 3 s at 488-nm excitation using the Bio-Rad Radiance 2000 imaging system.

Annexin V Imaging

To determine phosphatidylserine externalization as an indication of early apoptosis, cells were exposed to KO₂ for 3 h and incubated with the conjugate annexin V Alexa-Fluor-488 (Molecular Probes) for 15 min in annexin V binding buffer. PI (0.5 μg/ml) was added 5 min before imaging. After treatment, annexin V– and PI-positive cells were excited at 488 and 568 nm, respectively, and were counted in 10 independent fields. The normally impermeable PI is internalized as the plasma membrane loses integrity. Positive PI staining indicates either late stage apoptosis or necrosis.

Data Analysis

Either nuclear (HE, rhodamine 123) or perinuclear (MitoSOX Red) masking of all cells in a given field was used to quantitate the cellular response using Spectralyzer (custom software provided by Paul Anderson, Thomas Jefferson University) image analysis software. Tracings indicate the mean fluorescence value of all cells in one field and are indicative of n independent experiments. Multiple experiments were normalized to baseline average and expressed as fold change. Data are expressed as mean ± SEM for n independent experiments.

RESULTS

Extracellular O₂^.− Causes Rapid and Transient Intracellular HE Oxidation

To evaluate transmembrane flux of O₂^.−, we measured the effects of the addition of a single extracellular bolus of O₂^.− (10 μM) to the cell culture medium, a concentration within the range produced by activated macrophages or by the granulocyte respiratory burst (Nathan and Root, 1977; Johnston et al., 1978). The bolus of O₂^.− caused rapid and transient HE oxidation in MPMVECs (Figure 1A), which was eliminated by pretreatment with DIDS (200 μM; Figure 1B). HPMVECs responded similarly to bolus O₂^.− addition (Figure 1C). Subsequent experiments with MPMVECs exposed to varied O₂^.− concentrations revealed a concentration-dependent response that was blocked either by SOD (2500 U/ml) or by DIDS (200 μM); catalase (1,000 U/ml) had relatively little effect (Figure 1D). Application of O₂^.− also resulted in a dip in fluorescence below baseline values after the transient peak (Figure 1B). Because this dip corresponded to brief gap formation in the endothelial monolayer as observed by simultaneous recording of differential interference contrast (DIC; data not shown), it most likely represents a transient alteration of focus. We postulate that the transient gaps in the endothelial monolayer are associated with intracellular Ca²⁺ release as described below. H₂O₂ (500 μM) added as a bolus had no effect on HE fluorescence (Figure 1D).

Figure 1.

Hydroethidine (HE) oxidation with addition of extracellular O₂^.−. (A) A single bolus of O₂^.− was delivered to HE-loaded MPMVECs and recorded every 5 s. (B) Tracing indicating mean nuclear fluorescence of all cells in the field after addition of DMSO vehicle (control) or O₂^.− with or without preincubation with DIDS (200 μM). (C) Same experiment as B using HPMVECs. (D) Peak fluorescence change in MPMVECs (fold change normalized to baseline) in response to DMSO (vehicle control, n = 5), O₂^.− at 0.5 μM (n = 5), 2 μM (n = 5), 5 μM (n = 4), and 10 μM (n = 5), and H₂O₂ (500 μM; n = 3). The effects of SOD (2500 U/ml; n = 5), catalase (1000 U/ml; n = 3), and DIDS (200 μM; n = 5) were evaluated in the presence of 10 μM O₂^.−. (E) Mean cellular nuclear HE fluorescence after treatment with mitochondrial complex III inhibitor antimycin A (20 μM; n = 3) or the uncoupler FCCP (2 μM; n = 3).

To evaluate the effects of O₂^.− from an intracellular source, the mitochondrial complex III inhibitor antimycin A (AA; 20 μM) and the mitochondrial uncoupler carbonyl cyanide p-[trifluoromethoxy]-phenyl-hydrazone (FCCP; 2 μM) were applied to HE-loaded MPMVECs because these compounds are known to lead to rapid generation of intracellular O₂^.− (Koopman et al., 2005). Mitochondrial-derived O₂^.− resulting from either FCCP or AA treatment resulted in a rapid and progressive increase in HE fluorescence (Figure 1E). The fluorescence spectrum at 494-nm excitation was similar after treatment with either extracellular O₂^.− or AA, suggesting the same HE oxidation product (Supplementary Figure 1). The present studies do not allow us to differentiate between oxyethidium and ethidium in these experiments although previous studies suggest oxyethidium as the major stable metabolite (Zhao et al., 2005).

O₂^.− Causes Rapid and Transient HE Oxidation in a Cell-free System

The reaction of HE with O₂^.− creates a stable product in a multistep process (Fink et al., 2004; Zhao et al., 2005). We therefore hypothesized that the HE fluorescence transient (Figure 1) may be an HE oxidation intermediate. A cell-free system was used to investigate the chemical nature of the transient response of HE to O₂^.− observed in PMVECs. The HE fluorescence changes were monitored after delivery of either a bolus of KO₂ (200 μM) or a bolus of X/XO (50 mU/ml) delivered to HE (40 μM) dissolved in PBS (Figure 2A). Similar to findings with PMVECs, a rapid HE fluorescence transient was observed after KO₂ application, whereas the X/XO O₂^.− generating system resulted in a progressive increase. HE fluorescence was unaltered after addition of DMSO vehicle (200 μl), KOH (200 μM), H₂O₂ (5 mM), or KO₂ (100 μM) that had been predismutated into H₂O₂ by SOD (1000 U/ml). Increasing concentrations of O₂^.− correlated with the magnitude of both the initial peak and the stable postpeak HE fluorescence (Figure 2B).

Figure 2.

HE oxidation transient in a cell-free system. (A) Fluorescence change was measured in a fluorimeter after addition of agent to PBS containing 40 μM HE. Additions were DMSO vehicle, KO₂ (200 μM), xanthine/xanthine oxidase (X/XO; 50 mU), H₂O₂ (5 mM), KOH (200 μM), and KO₂ predismutated into H₂O₂ with SOD. (B) Quantitation of the normalized peak and postpeak HE fluorescence increase over baseline after addition of varying concentrations of KO₂. Linear regression lines were calculated from the mean values of three independent experiments.

Extracellular O₂^.− Leads to Progressive HE Fluorescence Increase

After the rapid HE transient, we consistently observed a gradual increase in HE fluorescence over the subsequent 300 s of imaging (Figure 1, B and C). We therefore monitored the effect of extracellular O₂^.− on stable (i.e., not transient) nuclear HE fluorescence in MPMVECs. Examination by microscopy showed a pronounced increase in nuclear HE fluorescence at 20 min after O₂^.− exposure that was prevented by pretreatment with DIDS (200 μM; Figure 3A). Quantitation of the images showed a 5.7 ± 1.5-fold increase over baseline in nuclear HE fluorescence with KO₂ addition versus a 1.0 ± 0.1-fold increase with DIDS pretreatment. Similarly, extracellular O₂^.− derived from X/XO resulted in a profound increase in HE fluorescence (6.2 ± 0.7-fold) versus DIDS-pretreated cells (1.7 ± 0.3-fold). The fluorescence in cells without addition of O₂^.− was essentially unchanged (Control, Figure 3B). Because O₂^.− added as a bolus to the aqueous medium of a cell monolayer would be rapidly dissipated, we postulated a secondary source of oxidants for the delayed increase in endothelial cell nuclear HE fluorescence. MPMVECs lacking the gp91^phox catalytic subunit of endothelial NADPH oxidase showed no difference from wild-type cells in the response of nuclear HE fluorescence to a bolus addition of O₂^.− (Figure 3B). This result indicates that the progressive HE oxidation triggered by extracellular O₂^.− is not the direct result of plasma membrane NADPH oxidase O₂^.− production, but rather suggests an intracellular source.

Figure 3.

The stable increase of nuclear HE fluorescence by extracellular O₂^.− is blocked by DIDS but is gp91^phox independent. (A) HE fluorescence in MPMVECs before and 20 min after exposure to extracellular O₂^.− (10 μM KO₂) and in the absence (top panels) and presence of DIDS (200 μM). (B) HE fluorescence expressed as fold change versus baseline (zero time) was measured in untreated MPMVECs at 20 min after addition of O₂^.− (10 μM KO₂ and 20 mU X/XO) with or without preincubation with DIDS (200 μM) and in gp91^phox null cells. Control represents no additions. Results are mean ± SE; n = 3.

Receptor-mediated Increase in HE Fluorescence

To determine whether “physiological” stimulation of O₂^.− generation could be detected by measurement of HE fluorescence, we used angiotensin II (Ang II; 2 μM) and thrombin (0.5 U/ml), two agonists known to increase endothelial ROS production through receptor-mediated signaling cascades (Takano et al., 2002; Li and Shah, 2004). O₂^.− production was monitored by the increase in HE fluorescence after 1 h of agonist stimulation. Both Ang II and thrombin increased O₂^.− production in HPMVECs (Supplementary Figure 2 and Figure 4). Pretreatment with the NADPH oxidase inhibitor Apo (2 μM) or with DIDS (300 μM) decreased HE oxidation in Ang II–stimulated cells (Supplementary Figure 2 and Figure 4), compatible with Ang II–mediated extracellular O₂^.− generation via NADPH oxidase. Thrombin-stimulated O₂^.− release, on the other hand, was insensitive to Apo and DIDS, suggesting an intracellular source of O₂^.− for this agonist.

Extracellular O₂^.− Triggers Mitochondrial O₂^.− Production

To assess mitochondria as a secondary source of O₂^.− in this model system, MPMVECs transfected with mitochondrial-targeted GFP were loaded with the O₂^.−-sensitive fluorophore MitoSOX Red (1.25 μM). O₂^.− added as a bolus to the cell culture medium evoked a large increase in MitoSOX Red fluorescence that was prevented by pretreatment of the cells with DIDS (Figure 5, A and B). The increase in MitoSOX Red fluorescence was ∼5.5-fold higher for O₂^.− versus untreated (vehicle only) cells (Figure 5B). Treatment with the sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase inhibitor thapsigargin (Tg) in the absence of extracellular O₂^.− resulted in transient elevation of intracellular Ca²⁺ and also triggered mitochondrial ROS production. Chelation of intracellular Ca²⁺ by BAPTA abolished the increase in MitoSOX Red fluorescence in response to bolus O₂^.− (Figure 5B). This result provides evidence that mitochondrial ROS generation occurs as a result of intracellular Ca²⁺ mobilization (Madesh et al., 2005). These data indicate that inhibition of O₂^.− flux across the plasma membrane by DIDS or ablation of O₂^.− -triggered Ca²⁺ release by chelation prevents mitochondrial ROS production associated with extracellular O₂^.−.

Figure 5.

Mitochondrial O₂^.− production in response to extracellular O₂^.−. (A) Images of MitoGFP-transfected MPMVECs incubated with the mitochondrial O₂^.−-sensitive fluorescent dye MitoSox Red (1.25 μM) after extracellular bolus of O₂^.− with or without DIDS (200 μM). Images were taken 20 min after KO₂ addition. All cells in the field show increased MitoSox Red fluorescence with addition of O₂^.−. Cells where MitoGFP is expressed show colocalization with MitoSox Red (merge panels). (B) Quantitation of MitoSox Red fluorescence at 20 min expressed as fold change versus baseline (zero time) for untreated cells (control) or cells treated with Tg (2 μM) or O₂^.− (10 μM) with or without DIDS (200 μM) or BAPTA (50 μM). Results are mean ± SE; n = 3.

O₂^.− Effect Can Be Blocked by Selective ClC-3 Knockdown

The effect of DIDS on HE and MitoSOX Red fluorescence in PMVECs suggests the requirement of an anion channel in O₂^.− flux across the plasma membrane. Because ClC-3 is the most abundant anion channel in endothelial cells (Lamb et al., 1999) and is inhibitable by DIDS (Hirotsu et al., 1999), we chose to assess whether knockdown of ClC-3 elicits a reduction in HE fluorescence similar to that demonstrated by DIDS. Electroporation efficiently delivered siRNA to MPMVECs as evidenced by microscopic evaluation of Cy3-labeled GAPDH siRNA (Figure 6A). siRNA sequences targeting exon 10 (sequence 1) and exons 2 and 3 (sequence 2) resulted in a marked decrease in ClC-3 mRNA expression by RT-PCR at an equivalent cycle amplification as compared with wild-type and negative control siRNA-transfected MPMVECs (Figure 6B). No differences in β-actin or ClC-4 expression were observed, indicating specificity of the siRNA effect. A reduction in ClC-3 protein expression was also noted at 72 h after delivery of ClC-3 siRNA (Figure 6B, bottom panel). Both the rapid peak (Figure 6C) and the subsequent increased nuclear HE fluorescence at 20 min after O₂^.− exposure (Figure 6, D and E) were markedly inhibited in ClC-3 siRNA treated cells (sequence 1) compared with negative control–transfected cells. These siRNA knockdown experiments provide additional evidence that O₂^.− membrane flux is mediated by ClC-3 and are consistent with the effect of DIDS on O₂^.−-mediated HE oxidation.

Figure 6.

ClC-3 knockdown attenuates intracellular HE oxidation. (A) Electroporation of siRNA effectively delivers siRNA to MPMVECs. Images represent cy3-labeled GAPDH siRNA counterstained with the nuclear marker DAPI. (B) CIC-3, CIC-4, and β-actin mRNA expression in MPMVECs 60 h after transfection with one of two different CIC-3 sequences or negative control siRNA (250 pmol). Bottom, Western blot of ClC-3 protein expression in wild-type (WT) MPMVECs and at 72 h after transfection with ClC-3 #1 or negative control siRNA (control). (C) HE fluorescence transient in WT (n = 5), ClC-3 #1 (n = 5) and #2 (n = 5) and negative control (n = 5) siRNA transfected MPMVECs after addition of 10 μM KO₂ normalized to fold change versus baseline. (D) HE fluorescence at baseline and at 20 min after exposure to extracellular O₂^.− (10 μM) in ClC-3 siRNA #1 and negative control siRNA-transfected MPMVECs. (E) Quantitation of HE fluorescence normalized to fold change versus baseline for control and ClC-3 #1-treated MPMVECs (n = 3).

Mitochondrial O₂^.− Production Is Associated with Ca²⁺-dependent Changes in ΔΨ_m

The effect of extracellular O₂^.− on mitochondrial membrane potential was evaluated as a potential mechanism for inducing mitochondrial O₂^.− production. In rhodamine 123–loaded MPMVECs, addition of O₂^.− (10 μM) to the medium caused a rapid leakage of dye from the mitochondria compatible with mitochondrial membrane depolarization (Figure 7A). Depolarization was associated with a dramatic alteration in mitochondrial morphology (Figure 7A, 105 s compared with zero time). At later times, mitochondrial depolarization was propagated to adjacent cells (Figure 7A, 200 and 300 s). Mitochondrial depolarization was blocked by pretreatment of cells with DIDS (Figure 7B). Reversible mitochondrial depolarization without major alterations in mitochondrial morphology was observed after low levels of O₂^.− (2 μM; data not shown). HPMVECs and MPMVECs demonstrated similar biphasic ΔΨ_m changes after addition of 5 μM O₂^.− (Figure 7, D and E). However, HPMVECs appeared to be more sensitive to O₂^.−, as addition of a 10 μM bolus produced irreversible ΔΨ_m loss (data not shown). ΔΨ_m changes were consistently delayed in comparison to the HE fluorescence transient. However, the biphasic phenomenon of ΔΨ_m alteration is similar to that observed with HE after O₂^.− exposure (see Figures 2 and 7).

Figure 7.

Effect of extracellular O₂^.− on ΔΨ_m. (A) Time-lapse images of MPMVECs loaded with the mitochondrial potentiometric dye rhodamine 123 (25 μM) before and after addition of KO₂ (10 μM). (B) Representative tracing of nuclear rhodamine 123 fluorescence after addition of DMSO and KO₂ (10 μM) with or without DIDS (200 μM). (C) Representative tracing of the cytosolic Ca²⁺ indicator dye Fluo4 after application of KO₂ (10 μM) in the absence and presence of the Ca²⁺ chelator BAPTA (50 μM) normalized to baseline fluorescence. (D) Representative tracing of nuclear rhodamine 123 fluorescence after addition of DMSO and KO₂ (10 μM) with or without pretreatment with thapsigargin (Tg; 2 μM). Dissipation of ΔΨ_m is demonstrated by addition of the mitochondrial uncoupler FCCP (2 μM). (D) Representative tracings of rhodamine 123 fluorescence in HPMVECs after addition of KO₂ (5 μM) or KCl (20 mM). Representative tracings are indicative of three independent experiments.

Bolus addition of O₂^.− triggered rapid mobilization of intracellular Ca²⁺ that could effectively be abolished by BAPTA (50 μM; Figure 7C). Because Ca²⁺ mobilization demonstrated a similar transient to that of the ΔΨ_m, we investigated the causal role of intracellular Ca²⁺ in ΔΨ_m alterations in MPMVECs pretreated with Tg (2 μM). This pretreatment with Tg prevented mitochondrial depolarization after bolus addition of O₂^.− (Figure 7D) without having a direct effect on ΔΨ_m (Supplementary Figure 3), indicating that the effect of increased extracellular O₂^.− on ΔΨ_m requires Ca²⁺ derived from intracellular stores. The specificity of the Ca²⁺ effect was evaluated by measuring the responses to an uncoupler of mitochondrial respiration and to depolarization of the plasma membrane. Addition of the mitochondrial uncoupler FCCP (2 μM) facilitated irreversible ΔΨ_m loss in contrast to the biphasic response to extracellular O₂^.−. To investigate a possible interaction between plasma membrane and mitochondrial membrane potentials, 20 mM KCl was added to rhodamine 123–loaded HPMVECs in order to partially depolarize the plasma membrane (Zhang et al., 2005). KCl addition had no effect on ΔΨ_m (Figure 7E). This excludes a possible effect of this cation when added with O₂^.−.

Anion Channel Blockade Prevents O₂^.−-induced Apoptosis

Addition of O₂^.− as a single bolus led to a subsequent significant increase in the number of MPMVECs that stained positively for annexin V (Figure 8, A and B). Many of the annexin V–positive cells also stained positively for PI. There was no significant population of PI-positive but annexin V–negative cells. These results are compatible with early to later events of apoptosis of MPMVECs associated with the O₂^.− bolus.

Figure 8.

Extracellular O₂^.− leads to apoptosis in MPMVECs. MPMVECs were grown on coverslips and subjected to an extracellular bolus of KO₂ (10 μM) with or without DIDS pretreatment (200 μM). (A) Cells were stained 3 h post-O₂^.− exposure for annexin V and propidium iodide. (B) The percentage of annexin V–positive cells was determined for 10 fields in each of 3 independent experiments.

DISCUSSION

Pathophysiological models indicate that endothelial oxidative stress can lead to vascular dysfunction and damage (Taniyama and Griendling, 2003). However, specific oxidants may result in discrete effects on endothelial cells (Finkel, 2001; Devadas et al., 2002). Endogenous O₂^.− produced by NADPH oxidase has been implicated in cell proliferation (Milovanova et al., 2006), whereas phagocyte-derived O₂^.− is associated with endothelial apoptosis (Madesh et al., 2005). Although it is possible that the disparate functions of extracellular O₂^.− may be attributed to varying mechanisms including rapid dismutation to H₂O₂, modification of cell-surface proteins, or release of cytokines secondary to endothelial cell activation, we hypothesize a unique signaling cascade specific to the extracellular presence of this anion. The biological effects of extracellular ROS have been studied primarily by adding exogenous H₂O₂ to target cells (Wright et al., 1994) or by examining autocrine effects on the ROS-producing cells themselves (Thannickal and Fanburg, 2000). No studies, to our knowledge, have demonstrated a role for paracrine-derived O₂^.− in triggering intracellular ROS generation in a physiological/pathophysiological relevant context. In the present study, we demonstrate that extracellular O₂^.− can cross the cell membrane via ClC-3 and initiate an intracellular signaling cascade resulting in Ca²⁺-mediated O₂^.− production by the mitochondria. These results suggest an important role for extracellular O₂^.− in endothelial cell biology.

In this study, we investigated O₂^.− membrane flux by utilizing a previously unpublished property of the O₂^.−-sensitive fluorophore HE. The transient intracellular fluorescence peak associated with HE oxidation was O₂^.− concentration-dependent in both live cell and cell-free models. Specificity of this transient to O₂^.− is indicated by its inhibition with SOD, whereas catalase had no effect. Addition of KO₂-treated HE to the extracellular milieu did not alter HE fluorescence (data not shown), excluding the possibility that HE oxidized outside the cell rapidly traverses the plasma membrane. Abrogation of the effect by DIDS suggests that this intracellular HE fluorescence transient results from membrane flux of O₂^.− through an anion channel. ClC-3 is the most abundant chloride channel in endothelial cells (Lamb et al., 1999) and knockout of this channel results in compensatory changes in cell membrane protein expression and function (Yamamoto-Mizuma et al., 2004). Selective knockdown of ClC-3 using siRNA resulted in a significant reduction in the HE transient similar to that observed by anion channel inhibition with DIDS. Therefore, we conclude that ClC-3 is the primary channel that supports transmembrane O₂^.− flux in endothelial cells.

After the HE transient with addition of O₂^.−, we observed a progressive increase in nuclear HE fluorescence that was blocked by DIDS and ClC-3 knockdown. It seems unlikely that this delayed response is due to the extracellular O₂^.− because of the expected short lifetime of O₂^.− in solution. A possibility for this finding is that extracellular O₂^.− triggered a secondary response in the cells leading to O₂^.− generation from a cellular source. Intracellular O₂^.− production by the mitochondrial inhibitor AA and the uncoupler FCCP resulted in progressive increase of nuclear HE fluorescence. This led us to hypothesize that the mitochondria may be a secondary source of O₂^.− after addition of O₂^.− to the extracellular medium. Measurement of nuclear HE fluorescence has been suggested as an indicator for O₂^.− derived from NADPH oxidase (Sun et al., 2005). However, O₂^.− generated by the mitochondria elicits a similar response (Becker et al., 1999). The failure of NADPH oxidase deficient cells to change the response and the use of the mitochondrial O₂^.− specific dye MitoSOX Red in the present experiments indicate that mitochondrial production of O₂^.− is primarily responsible for the progressive increase in nuclear HE fluorescence associated with extracellular O₂^.−. These results suggest that nuclear HE fluorescence associated with activation of NADPH oxidase and consequent extracellular O₂^.− generation may actually reflect mitochondrial-derived ROS resulting from intracellular Ca²⁺-mediated signaling.

Addition of Ang II or thrombin was used to initiate endogenous NADPH oxidase activity in endothelial cells in order to test whether mitochondrial O₂^.− production was activated by physiological levels of extracellular O₂^.−. We observed that Ang II triggered a significant increase in HE fluorescence that was blocked by both Apo and DIDS. Ang II–induced endothelial cell O₂^.− production has been linked to endothelial dysfunction and associated hypertension (Lassegue et al., 2001), and a link has been demonstrated between Ang II stimulated NADPH oxidase-derived O₂^.− and mitochondrial ROS production (Kimura et al., 2005). It has been proposed that ROS produced by mitochondria in endothelial cells serve an intracellular signaling function (Quintero et al., 2006). Oscillations in mitochondrial ROS production due to a localized production of ROS by a small number of mitochondria have provided evidence for mitochondrial-mediated signaling via ROS (Zorov et al., 2000; Aon et al., 2003). However, the possibility that extracellular ROS also could stimulate intracellular ROS production by the mitochondria has not been previously reported. The present study demonstrates that extracellular O₂^.− produced by NADPH oxidase can permeate the cell membrane to trigger intracellular (mitochondrial) ROS production.

Transmembrane O₂^.− flux has previously been shown in membranes highly enriched with anion channels, such as the erythrocyte (Lynch and Fridovich, 1978). However, under normal conditions, the diffusion distance of O₂^.− before spontaneous dismutation to H₂O₂ is estimated at 0.5 μm (Mikkelsen and Wardman, 2003). The rate of dismutation would be increased within the cell by cytosolic SOD (Fridovich, 1995). This precludes extracellular O₂^.− from traveling much beyond the plasma membrane to react with potential intracellular signaling proteins (Finkel, 2001). Nonetheless, we demonstrate that O₂^.−-mediated signaling can be attenuated by both molecular inhibition of ClC-3 and anion channel blockade by DIDS, indicating a discrete role for O₂^.− membrane flux in endothelial function. The question therefore arises as to the mechanism through which the short-lived O₂^.− anion leads to cell signaling. The experimental findings are that extracellular O₂^.− triggered rapid Ca²⁺ mobilization and that mitochondrial ROS production was preceded by Ca²⁺-dependent changes in ΔΨ_m and was prevented by passive depletion of ER Ca²⁺ stores. These results are in agreement with studies using activated macrophages or the X/XO O₂^.−-generating system (Madesh et al., 2005). Loss of ΔΨ_m in isolated mitochondria as a result of Ca²⁺ overload has been demonstrated previously (Galindo et al., 2003). Thus, we propose that the mechanism by which extracellular O₂^.− triggers mitochondrial O₂^.− production is through cell signaling secondary to Ca²⁺ release from intracellular stores. The present evidence supports this hypothesis, because mitochondrial O₂^.− generation occurred in association with increased intracellular Ca²⁺ after Tg treatment and was abolished by chelation of the increased Ca²⁺ mediated by extracellular O₂^.−. Based on our previous studies, O₂^.−-mediated Ca²⁺ release occurs via an inositol trisphosphate receptor-dependent mechanism (Madesh et al., 2005).

In summary, transmembrane O₂^.− flux occurs in PMVECs through ClC-3 channels and results in ΔΨ_m alterations and mitochondrial O₂^.− production. This novel finding elucidates a potential mechanism by which extracellular O₂^.− is propagated to the intracellular milieu to trigger endothelial cell signaling or dysfunction associated with oxidative stress. We postulate that endothelial cell injury via paracrine O₂^.− signaling may represent a basis for pulmonary vascular remodeling.

Figure 4.

Receptor-mediated endothelial cell O₂^.− generation results in a stable increase of nuclear HE fluorescence. HPMVECs were loaded with HE (10 μM) and stimulated with Ang II (2 μM) or thrombin (0.5 U/ml) for 1 h with or without pretreatment with the NADPH oxidase inhibitor apocynin (Apo; 2 μM) or the anion channel blocker DIDS (300 μM). Nuclear HE fluorescence was quantitated from the confocal microscopic images. Data represent mean ± SE of five independent fields (n = 3).

Abbreviations used:

ROS

reactive oxygen species

SOD

superoxide dismutase

ClC-3

chloride channel-3

ΔΨ_m

mitochondrial membrane potential

DIDS

4,4′-diisothiocyanostilbene-2,2′-disulfonic acid

HE

hydroethidium

HPMVEC

human pulmonary microvascular endothelial cells

Ang II

angiotensin II

MPMVECs

murine pulmonary microvascular endothelial cells

Apo

apocynin

FCCP

carbonyl cyanide p[trifluoromethoxy]-phenyl-hydrazone

Tg

thapsigargin.