INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

REACTIVE OXYGEN SPECIES (ROS) appear to mediate specific responses to changes in PO₂ in oxygen-sensitive tissues such as carotid body cells, pulmonary vascular smooth muscle, and erythropoietin (EPO)-producing hepatocytes. On the other hand, ROS are involved in the degradation of biological material, and elevated levels of ROS can cause significant cell damage, apoptosis, and cell death (14). In these oxygen-sensing cells, H₂O₂ seems the most suitable of all ROS to serve as an intracellular messenger. In comparison to other ROS such as superoxide anion or hydroxyl radical, H₂O₂ is characterized by high stability and free membrane permeability. Production of H₂O₂ proportional to changes in PO₂ could be demonstrated in HepG2 hepatocytes (11). Acker et al. (1) showed a decreasing nerve discharge of rat carotid bodies by increasing concentrations of H₂O₂, thus mimicking hyperoxia, whereas hypoxic stimulation was suppressed during the period of H₂O₂ addition. In pulmonary vascular smooth muscle, Archer et al. (2) showed a PO₂-dependent ROS production and subsequent changes in vascular tone that appeared to depend on ROS concentration in isolated rat lungs. Fandrey et al. (11) demonstrated inhibition of hypoxia-induced EPO production by H₂O₂ in HepG2 hepatocytes. Moreover, H₂O₂-induced inhibition of EPO production could be antagonized by incubation with catalase (31), and a relationship between the oxygen dependence of EPO expression induced by hypoxia-inducible factor-1 and ROS has been discussed (31, 35). Molecular analysis of EPO expression showed great similarities with other known hypoxia-responsive genes (31), thus supporting the hypothesis of a highly conserved oxygen sensor. However, a link between ROS and the regulation of ion transport of alveolar epithelial cells in hypoxia has not yet been demonstrated.

Both hypoxia and hyperoxia were shown to affect ion transport of alveolar epithelial cells (for a review, see Ref. 6). Alveolar epithelial cells respond to hypoxia with decreased activity of the Na/K pump, Na/K/2Cl cotransport (NKCC), and other Na-transporting systems (22, 27, 28), which was discussed as one mechanism crucial for developing and aggravating hypoxia-induced pulmonary edema (30). Whereas acute hyperoxia appears to inhibit the Na/K pump, prolonged exposure activates and thus helps clear hyperoxia-induced edema (6, 25, 26). Alterations of transport might also occur after a respiratory burst of macrophages and other cells of the immune system that liberate ROS for host defense.

Although a variety of systems exist that affect formation of ROS inside the cell, it appears that ROS concentration also depends on available oxygen. This is the basis for the assumption of a ubiquitous, ROS-dependent cellular oxygen-sensing mechanism. The present study was performed to investigate the role of ROS in regulating alveolar epithelial cell ion transport and to test whether changes in ROS are involved in mediating the hypoxic inhibition of ion transport (21, 22, 28). Effects of modulating cellular ROS by pro- and antioxidative substances on ion transport were measured on A549 cells, a human lung-derived carcinoma cell line with conservation of many functions of alveolar type II cells, in particular the hypoxic inhibition of ion transport (22). The results indicate that cellular ROS depletion modulates cation transport of A549 cells similar to hypoxia.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents, oxidants and antioxidants. All reagents were of analytic grade, and stock solutions were prepared from deionized water. Phosphate-buffered saline (PBS), Ham's F-12 medium, penicillin-streptomycin, FCS, and HEPES were from GIBCO BRL (Life Technologies). H₂O₂ (30% stock solution), menadione sodium bisulfite (Men), and 3-amino-1,2,4-triazole (ATZ; aminotriazole), Cu/Zn superoxide dismutase (SOD; from bovine liver), catalase (from bovine liver), diphenyleneiodonium chloride (DPI), N-acetyl-L-cysteine (NAC), and deferoxamine (DFO; desferrioxamine mesylate), tetramethylpiperidine-N-oxyl (TEMPO), and diethyldithiocarbamate (DETC) were from Sigma.

A549 cells. A549 cells, a human pulmonary carcinoma cell line with many characteristics of alveolar type II cells (20), were from American Type Culture Collection. The cells were cultured on untreated 24-well plates (Costar) in Ham's F-12 medium supplemented with 7% FCS, 100 U/ml penicillin-0.1 mg/ml streptomycin, 10 mM HEPES, and 10 mM sodium bicarbonate. Confluence was reached 3-4 days after the cells were seeded, and the experiments were carried out on 2- to 8-day-old confluent monolayers passaged 3-12 times after purchase from American Type Culture Collection.

Hypoxia was applied by an initial change to a gas-equilibrated culture medium. In most experiments, oxidants and antioxidants were added at this point. Culture plates then were transferred to a CO₂-O₂-controlled incubator (NUNC) adjusted to 3% O₂, 5% CO₂ and 92% N₂ at 37°C as described previously (22). For exposure to hyperoxia, gas mixtures containing 30 or 40% O₂, 5% CO₂ and balance N₂ were used. Control cells were treated similarly under the usual tissue conditions (5% CO₂ and 37°C) in a tissue culture incubator (Heraeus).

Flux measurements. The activity of ion transport pathways was determined by unidirectional tracer uptake measurements (22). After exposure to normoxia or hypoxia, the cells were washed two times with washing medium [150 mM NaCl and 2 mM HEPES, pH 7.4, at room temperature (RT)] and incubated for 15 min at RT with the flux medium equilibrated with the respective CO₂-free gas. The flux medium contained (in mM) 140 NaCl, 20 HEPES, 10 glucose, 5 KCl, 1 NaH₂PO₄, 1 MgCl₂, and 0.2 CaCl₂, pH 7.4, at RT. Fluxes were started by adding flux medium containing ⁸⁶Rb (as a tracer for K) or ²²Na at final activities of 0.5-2 µCi/ml and the respective inhibitors. Ouabain (100 µM) and bumetanide (50 µM) were added to quantify the activity of the Na/K pump and NKCC, respectively. Equivalent volumes of the solvent were added to the control cells. Tracer uptake was stopped, and contaminant tracer was removed with five washes with ice-cold washing medium. The cells were lysed with 0.1 M NaOH. Radioactivity of the lysate was determined in a -counter (model TR 2100, Canberra Packard). The protein concentration in the lysate was measured photometrically with a test kit from Bio-Rad, with human serum albumin and globulin standards in saline diluted with 0.1 M NaOH.

Photometric measurement of H₂O₂ concentration. To measure the capacity of A549 cells to metabolize H₂O₂, its concentration was measured photometrically according to the method previously described (11). Briefly, one volume of supernatant was mixed with four volumes of 2,2'-azino-di(3-ethylbenzthiazoline-sulfonate) reagent (100 mg/l) in 100 mM NaCl and 50 mM NaH₂PO₄, pH 4.4. Horseradish peroxidase (Boehringer Mannheim) was added to start the reaction. Absorption was measured after 30 min of incubation at RT, and H₂O₂ concentration was calculated from known standards. The lower detection limit of this assay was ~4 µM H₂O₂.

Measurements of ATP and ADP. ATP and ADP concentrations were measured as described by Weicker et al. (33). Briefly, the cells were washed with ice-cold PBS, and protein was denatured by adding ice-cold 0.6 M perchloric acid onto the cell layer. The cells were scrubbed from the tissue culture plate and vortexed, and the precipitate was sedimented by a 3-min centrifugation at 10,000 g in a microfuge. The pH of the clear supernatant was neutralized before measurement with HPLC with an ultraviolet detection system.

Statistics. All measurements were repeated on several batches of cells from different passages, and fluxes were determined in triplicate. Results are means ± SD. The level of significance was P  0.05. Graphs were designed with SigmaPlot version 2.01 (Jandel Scientific).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

H₂O₂ elimination by A549 cells. Taking into consideration the potential cytotoxicity of oxidants and especially of H₂O₂, it was important to evaluate the capability of A549 cells to handle a H₂O₂ charge. Due to the limited sensitivity of the assay used, neither basal levels of H₂O₂ in A549 cells nor PO₂-dependent changes in H₂O₂ could be detected. Elimination of H₂O₂ was tested after known amounts of H₂O₂ were added to the cells kept in culture medium. Figure 1A shows the high capacity of A549 cells to eliminate within 2 min H₂O₂ that amounted to ~40% of the H₂O₂ added initially. The H₂O₂ concentration fell almost instantly when NAC (10 mM) was present. The rate of H₂O₂ degradation was not affected by hypoxia (Fig. 1B). The catalase inhibitor ATZ did not significantly influence H₂O₂ elimination (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Degradation of H2O2 by A549 cells in normoxia and hypoxia. A549 cells were kept in culture medium (7% FCS) equilibrated to normoxic (A) and normoxic and hypoxic (3% O2; B) gases containing 5% CO2 for 4 h. A: 10 (), 50 (), or 100 () µM H2O2 with () and without N-acetyl-L-cysteine (NAC; 10 mM) was added to culture medium at 0 min, and samples for measurement of H2O2 were taken at indicated time points. Data points are means from triplicate measurements from 1-3 (A) and 2 (B) experiments with similar results.

Effects of oxidants and hyperoxia on ion uptake of A549 cells. Cellular levels of ROS were increased by the addition of H₂O₂ to test whether elevated ROS concentrations would abolish the decrease in ion uptake by A549 cells in hypoxia (22). It has to be noted that incubation with H₂O₂ at a concentration of 1 mM caused cell damage, indicated by floating cells (only ~50-75% of the cells remained adherent to tissue culture plates as quantified by measuring the protein content per culture well). This cytotoxic effect was quite variable because it was not observed in ~40% of all experiments. It was less pronounced at a H₂O₂ concentration of 100 µM (85% adherent cells) but was not observed at 10 µM. Occurrence of lysis was not related to passage number, the age of the cells, or the feeding cycle. However, changes in transport activity were found regardless of whether cell damage was observed or not. Figure 2 shows that total ²²Na uptake of adherent cells was affected by H₂O₂ in a dose- and time-dependent manner: a 1-h incubation with 1 mM H₂O₂ caused a sixfold stimulation of ²²Na uptake; at a concentration of 100 µM, activation was about twofold, whereas no significant effect was seen at lower concentrations. After a 4-h incubation, H₂O₂ affected ²²Na uptake only when applied at 1 mM. Total ⁸⁶Rb uptake was inhibited by H₂O₂ when applied at concentrations > 100 µM (Fig. 3A). This was due to an inhibition of the Na/K pump (Fig. 3B). NKCC was not affected, whereas ouabain- and bumetanide-insensitive (OBI) ⁸⁶Rb uptake was increased significantly (Fig. 3B). Concentrations of 100 and 10 µM H₂O₂ did not affect ⁸⁶Rb uptake components.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Dose and time dependency of H2O2 effects on 22Na uptake by A549 cells. Unidirectional 22Na uptake was measured after exposure of A549 cells for 1-4 h to indicated concentrations of H2O2 in 0.5 ml culture medium (7% FCS)/well of a 24-well culture plate. Fluxes were measured in presence of H2O2 in a flux medium composed of (in mM) 135 NaCl, 5 KCl, 5 glucose, 1 MgCl2, 1 Na2HPO4, and 20 HEPES-Tris, pH 7.4, at room temperature. Prot, protein. Values are means ± SD of triplicate flux measurements from 1 of 6 experiments. * Significant difference between control and experimental samples, P  0.05.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   H2O2 effects on 86Rb uptake by A549 cells. After 4-h exposure to indicated concentrations of H2O2, unidirectional 86Rb uptake was measured as described in Fig. 2. A: dose dependency of H2O2 on total 86Rb uptake. B: activity of Na/K pump and Na/K/2Cl cotransport (NKCC) measured as ouabain (0.1 mM)- and bumetanide (50 µM)-sensitive 86Rb uptake components, respectively, after treatment with 1 mM H2O2. Residual, ouabain+bumetanide-insensitive flux. Dotted line, activity in control cells. Values are means ± SD from 20 experiments. * P  0.05 compared with control value.

To test whether H₂O₂ actually simulates the effects of an increased level of oxygen, the cells were also exposed to hyperoxia. Four hours of exposure to 30% O₂ had no effect on either component of ⁸⁶Rb uptake, whereas exposure to 40% O₂ inhibited total ⁸⁶Rb uptake by ~28%, which was mainly due to inhibition of the Na/K pump. The slight decrease in NKCC was not significant. In contrast to H₂O₂, hyperoxia did not affect OBI ⁸⁶Rb uptake (Fig. 4A) nor did it affect total ²²Na uptake (Fig. 4B).

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effects of hyperoxia and NAC on 86Rb (A) and 22Na (B) uptake by A549 cells. Fluxes were measured after 4-h exposure of A549 cells to hyperoxia (40% O2-5% CO2) in absence and presence of NAC (2 mM). Fluxes were measured as described in Fig. 2. Values are means ± SD from triplicate flux measurements of 3 experiments. * P  0.05 compared with control value.

ATZ and Men were applied to increase ROS generation by inhibiting catalase and affecting electron shuttling between different electron acceptors, respectively. The cells were incubated with 50 mM ATZ and 0.6 mM Men over 2 h before exposure to hypoxia to establish their effects (11). DETC was used to increase cellular superoxide anion levels by inhibiting SOD. Figure 5A shows that ATZ and Men together inhibited ⁸⁶Rb uptake in normoxic cells. This effect can mainly be attributed to the inhibition of NKCC. Also, DETC inhibited total ⁸⁶Rb uptake by ~40%, which is brought about by a 20% inhibition of the Na/K pump and a 60% inhibition of NKCC (Fig. 5B). Pretreatment with DETC for at least 30 min was required to detect significant inhibition (time course not shown).

View larger version (38K): [in this window] [in a new window]   Fig. 5.   Effects of oxidants on 86Rb uptake by A549 cells. To increase intracellular formation of reactive oxygen species (ROS), cells were kept in normoxia in presence of aminotriazole (AZT; 50 mM) plus menadione (Men; 0.6 mM) for 2 h before flux measurements (A) and exposure to diethyldithiocarbamate (DETC) for 4 h (B). Dotted lines, activity ± SD in control cells. Values are means ± SD from results of triplicate flux measurements on 5 different batches of cells. * P  0.05 compared with control value.

Effects of antioxidants on ion uptake by A549 cells. Application of antioxidants to normoxic cells should lower the cellular ROS levels thought to mimic a decrease in ROS as in hypoxia. SOD and catalase added to A549 cells had no effect on ⁸⁶Rb uptake regardless of the concentration applied and duration of treatment (data not shown).

NAC and DFO were added at different concentrations to the culture medium, and the cells were exposed for 1 and 4 h before the flux measurement. NAC caused a dose-dependent inhibition of total ⁸⁶Rb uptake (Fig. 6A) that was complete at ~10 mM NAC and was independent of the duration of the treatment. However, at this high concentration of NAC, some cell damage was observed as indicated by floating cells. NAC (2 mM) also caused a slight inhibition of total ²²Na uptake (Fig. 4B). DFO had no significant effect on ⁸⁶Rb uptake (Fig. 6B).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effects of antioxidants NAC (A) and deferoxamine (DFO; B) on total 86Rb uptake of A549 cells. Cells kept in normoxia were incubated with NAC and DFO for 1 and 4 h before flux measurements (see Fig. 2). Values are means ± SD from 6 flux measurements. * P  0.05 compared with control value.

DPI reduces superoxide anion production and ROS generation by inhibition of NAD(P)H oxidase and other flavine-containing enzymes. TEMPO has been used as a ROS scavenger that appears to be more selective for the superoxide anion (29). Figure 7A shows that 10 µM DPI inhibited total ⁸⁶Rb uptake, which was independent of the duration of treatment (10 min to 4 h; data not shown). One micromolar DPI was ineffective; a greater variability of results was observed at 100 µM (data not shown). The inhibition of total ⁸⁶Rb uptake by 10 µM DPI was caused by inhibition of the Na/K pump. NKCC and OBI ⁸⁶Rb uptake did not change significantly. Because DPI might also affect mitochondrial NADPH oxidases, it might exert its action by energy depletion. Figure 8 shows that DPI did not affect cellular ATP and ADP concentrations. Figure 7B shows that TEMPO (3 mM) inhibited total ⁸⁶Rb uptake by ~20% due to inhibition of the Na/K pump and NKCC.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Effects of antioxidants diphenyleneiodonium chloride (DPI; A) and tetramethylpiperidine-N-oxyl (TEMPO; B) on 86Rb uptake by A549 cells. Cells kept in normoxia were exposed to DPI (10 µM) and TEMPO (3 mM) for 4 h before flux measurements. Fluxes were started by adding tracer together with ouabain and bumetanide. Values are means ± SD from 6 flux measurements. * P  0.05 compared with control value.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Effects of DPI on ATP (A) and ADP (B) content of A549 cells. Normoxic cells were exposed to DPI for indicated times and washed with ice-cold PBS. Protein was precipitated with ice-cold 0.6 M perchloric acid, and ATP and ADP were measured after HPLC separation with ultraviolet detection and appropriate standards. Values are means ± SD from duplicate measurements on 3 sets of cells. * P  0.05 compared with 0 min.

Combined effects of oxidants, NAC, and hypoxia. To establish a role for ROS in sensing cellular oxygen levels and their involvement in the modulation of ion transport by hypoxia, oxidants and antioxidants were applied to cells exposed to normoxia and hypoxia because clamping intracellular ROS levels should render transport insensitive to hypoxia if ROS are involved in oxygen-dependent transport modulation. Figure 9A indicates that H₂O₂ stimulated ²²Na uptake in normoxic and hypoxic A549 cells, but the effect was smaller in hypoxia. In both oxygenation states, 2 mM NAC had no effect on ²²Na uptake. Figure 9B shows that 1 mM H₂O₂ inhibited ⁸⁶Rb uptake in normoxia and hypoxia but did not prevent the hypoxia-induced transport inhibition. Also, in the presence of NAC, hypoxic inhibition of ⁸⁶Rb uptake was detectable. Figure 9C shows that the SOD inhibitor DETC reduced transport activity in normoxia. However, in the presence of DETC, no significant inhibition of transport by hypoxia could be seen.

View larger version (32K): [in this window] [in a new window]   Fig. 9.   Combined effects of H2O2, NAC, DETC, and hypoxia on transport activity of A549 cells. Cells were exposed to H2O2 (1 mM), NAC (2 mM), or DETC (10 mM) for 4 h while being kept in normoxia or hypoxia (3% O2) before flux measurements (see Fig. 2). A and B: effect of H2O2 and NAC on total 22Na and 86Rb uptake, respectively. C: effect of DETC on 86Rb uptake components. Dotted line, activity in control cells. Values are means ± SD from 6 measurements. * P  0.05 compared with control value. # P  0.05 compared with corresponding value in normoxia.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Our results indicate that ROS and their scavengers as well as hyperoxia and hypoxia have profound effects on ion transport of A549 cells but that different transport systems are affected in different ways: hypoxia and scavengers of ROS inhibit Na and Rb uptake, whereas hyperoxia and increased levels of ROS activate Na uptake but inhibit Rb uptake, particularly the Na/K pump. Increasing cellular ROS levels, however, did not prevent transport inhibition by hypoxia.

These results contrast to some extent with the hypothesis of a sensor detecting hypoxia by a decrease in cellular ROS levels, which might also be involved in the modulation of cation transport of alveolar epithelial cells by hypoxia, analogous to results obtained on peripheral chemoreceptor glomus cells, pulmonary artery smooth muscle, and EPO-producing hepatocytes (24, 31, 34). This hypothesis predicts that varying the cellular levels of ROS by the addition of ROS scavengers and oxidants would mimic the respective effects of hypoxia and hyperoxia on ion transport (6). Addition of oxidants should then prevent hypoxic transport inhibition, and transport should be independent of PO₂ at any preset concentration of ROS adjusted by adding and scavenging ROS.

If the above-discussed arguments held true, a direct relationship might be expected between ROS concentration and cellular ion transport. ROS depletion should inhibit transport, whereas increased ROS levels should activate ion transport of A549 cells. Lung epithelial cells are constantly exposed to high levels of ROS from inhaled oxidants or formation by phagocytic and inflam-matory cells inside the lung as well as by alveolar epithelial cells (32). These cells require an effective protection system consisting of enzymes eliminating ROS (10) and ion transport pathways that adjust cellular solute and water contents on oxidative cell damage. Increased levels of ROS have been shown to modulate ion transport of lung alveolar cells (3, 7, 13, 15, 16, 25). Acute exposure to increased ROS concentrations induced by severe hyperoxia and addition of H₂O₂ was shown to inhibit Na-K-ATPase (7, 13), whereas long-term hyperoxia activates ion transport (4, 13). A brief pulse of externally added ROS was sufficient for long-term transport modulation despite rapid H₂O₂ degradation (13). Whereas the initial inhibition of the Na/K pump by increased ROS might be a consequence of oxidative cell damage as well as of direct damage of the transport protein and/or its regulating systems, stimulation of the Na/K pump on prolonged ROS exposure by increasing its expression might represent a mechanism to reduce edema formation caused by oxidative lung damage (6, 23). In contrast to the above-mentioned prediction, this time-dependent diversification of the response of the Na/K pump to increased levels of ROS indicates that no direct relationship exists between the actual ROS concentration and transport activity. Our results indicate that short-term exposure of cells to increased ROS affects different transporters in different ways. Although ²²Na uptake was activated by H₂O₂ but not by short-term hyperoxia, the Na/K pump (and, to some extent, also NKCC) were inhibited by 40% O₂, whereas ATZ plus Men and DETC affected mainly NKCC (Figs. 4 and 5). Clerici et al. (7) also found inhibition with 2.5 mM H₂O₂ of the Na/K pump and Na-coupled transport of phosphate and amino acids of alveolar type II cells in primary culture. The authors did not comment on total Na uptake. The mechanisms involved in ROS activation of total ²²Na uptake and inhibition of the Na/K pump and NKCC are not understood. It appears that increased ROS might modulate ion transport by alveolar epithelial cells not by action on a common signaling pathway but by transporter-specific effects because some transporters are more sensitive to oxidant stress than others. In alveolar epithelial cells, Ca-dependent mechanisms seem not to be involved in mediating oxidant-induced transport inhibition. In contrast to other cell types, which respond to micromolar concentrations of H₂O₂ with a pronounced increase in intercellular Ca (Ca_i) (9, 19), primary cultured rat AII cells as well as A549 cells do not increase Ca_i unless concentrations of H₂O₂ > 2 mM are applied (results not shown). In the latter case, the increase in Ca_i is certainly a consequence of H₂O₂-induced cell damage.

Scavengers of ROS are well known to prevent oxidative tissue injury. However, the effects of decreased ROS levels on ion transport have not been studied in great detail. Our results indicate a pronounced decrease in ion transport by alveolar epithelial cells when cellular ROS levels were decreased by the addition of ROS scavengers (TEMPO and NAC) and an inhibitor of ROS formation (DPI; Figs. 6, 7, and 9). This result is consistent with the notion that hypoxia causes transport inhibition by decreasing the concentration of ROS. It is, however, not in line with results on B cells where oxygen sensitivity was found to be independent of NAD(P)H oxidase, the enzyme thought to be mainly responsible for formation of the superoxide radical from oxygen (36). Inhibitors of NAD(P)H oxidase, such as DPI, were also shown to affect mitochondrial function by inhibition of flavine-containing enzymes, which might impair ATP synthesis (8, 12). Our results on cellular ATP and ADP concentrations in cells treated with low doses of DPI show no indications for an impairment of mitochondrial function (Fig. 8). Energy depletion can therefore be ruled out as a reason for transport inhibition by DPI.

Recently, Rafii et al. (29) reported activation of Na transport and epithelial Na-channel expression when fetal lung distal epithelial cells cultured in 3% O₂ were exposed to increased PO₂ (normoxia). This effect was prevented only by ROS scavengers that appear to be more specific in binding and neutralizing the superoxide radical rather than peroxides (29). The similarity between hypoxia- and scavenger-related changes in transcription factor nuclear factor-B led to the conclusion that in this case, an oxygen-induced increase in cellular ROS levels might mediate upregulation of the epithelial Na channel. Our results contrast with this finding because we could not demonstrate an upregulation of Na transport with hyperoxia but only with H₂O₂.

Do changes in ROS mediate transport inhibition by hypoxia? The inhibition of transport by free radical scavengers in normoxic cells is consistent with the notion of an ROS-dependent oxygen-sensing mechanism that measures the degree of hypoxia as the decrease in cellular levels of ROS. This is in line with findings on EPO-producing hepatocytes (11). It is contrasted by recent results (5) on a hypoxia-induced increase in dichlorofluorescin fluorescence, which was interpreted to show an increase in mitochondrial ROS formation in hypoxia. With the methods applied, we were unable to detect changes in ROS of A549 cells during hypoxia. Our results taken together indicate no direct relationship between ROS concentration and transport activity. Although transport inhibition by depletion of cellular ROS supports the model, the lack of H₂O₂ and DETC to prevent hypoxia-induced transport inhibition contrasts with the notion that a ROS-dependent oxygen sensor mediates inhibition of ion transport in hypoxia. It might be, however, that the high capacity of A549 cells to remove added H₂O₂ effectively still allows hypoxia to exert its effects. This argument is weakened because the H₂O₂-induced inhibition of ⁸⁶Rb transport was stable over several hours despite the rapid removal of H₂O₂ by A549 cells. Also, in hypoxic hepatocytes, an initial pulse of H₂O₂ was sufficient to prevent the hypoxia-induced EPO formation (11). In those experiments, a constantly elevated level of ROS could only be maintained in the presence of inhibitors of catalase (11). It appears, therefore, that hypoxic effects can occur even though cellular ROS levels are increased, which makes their involvement in an oxygen-sensing process unlikely unless the signal consists of a (transient) change in ROS rather than absolute values. The fact that hypoxic inhibition of transport is also seen in the presence of NAC (Fig. 9), when cellular ROS concentrations are already low, further supports this notion.

Cellular concentrations of ROS vary due to alterations in ambient oxygen and changes in metabolic activity. They indicate, therefore, the metabolic state of a cell. On the basis of this assumption, ROS were discussed as "sensors" for altered oxygen availability in various cell types that show oxygen-dependent changes in cell functions. The alveolar epithelium, however, is constantly exposed to high levels of oxidants from, e.g., the high oxygen content of inspired air, pollutants like NO₂, and a respiratory burst of activated neutrophils and macrophages. Alveolar epithelial cells were even shown to be capable of initiating respiratory bursts themselves due to a highly active NADPH oxidase (18, 32). As a protective mechanism, they also have high-capacity ROS eliminating pathways such as extracellular catalase (10, 17). It is, therefore, questionable whether in cells that have a high capacity to both produce and eliminate ROS, an oxygen-sensing system based on changes in ROS would be very sensitive unless the oxygen sensor consists of a distinct ROS species that is not directly affected by cellular ROS detoxification systems and that we were unable to modulate with the above-described maneuvers.