Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Exposure to pure oxygen leads to diffuse alveolar damage, characterized by exudation of plasma into the alveolar space. Alveolar leak occurs when epithelial cells are damaged, suggesting that these cells are crucial to maintain the integrity of the alveolar-capillary barrier. Lung morphometric studies in rats have shown that endothelial cells are destroyed by oxygen exposure first, followed by epithelial cells (1). Thus, death of endothelial and alveolar epithelial cells seems to be an essential feature of oxygen-induced alveolar damage.

Two types of cell death have been recognized to date: necrosis and apoptosis, of which the latter is also known as programmed cell death (PCD). Necrosis is associated with a disruption of the cell membrane, leading concomitantly to nuclear degradation. Apoptosis typically includes the activation of specific nucleases, and occurs in a cell with an intact plasma membrane (2). A difficulty in studying the role of PCD in pathophysiologic processes is that PCD might vary according to cell type, the state of cell differentiation, and the type of injury. Different stimuli can induce apoptosis along different pathways in different cell types. Many cell types express tumor necrosis factor receptor I (TNFRI) and Fas, which contain an intracytoplasmic death domain that can trigger a cascade leading to PCD (3). The tumor repressor gene p53 is required for epithelial cell death induced by gamma irradiation, which like hyperoxia is an oxidative stress that causes DNA damage (4, 5). p53 can induce PCD via the bcl-2 and bax gene families. Proteases of the caspase superfamily (including CPP-32 and ICE), which constitute the "executioner" machinery involved in apoptosis, can be activated by the TNFRI and Fas cascades, and possibly by p53 (6, 7).

In this study, we exposed mice to 100% O₂ and studied several aspects of cell death by various methods. Both necrosis and apoptosis were evident during hyperoxia. However, interfering with genes involved in apoptotic pathways, such as with Fas-null, p53-knockout mice, or strategies that blocked specific apoptotic pathways (ICE-like proteases), did not provide a significant attenuation of the alveolar damage caused by exposure to pure O₂.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Mice

C57BL/6 and lpr mice (C57BL/6 background) were purchased from Iffa Credo (Labresle, France) and bred in our animal facility for two generations. p53-deficient mice (/) and their littermates (+/+) (C57BL/6 background) were purchased from Taconic Farms Inc. (NY). Experiments were performed with 2- to 3-mo-old mice weighing 20 to 25 g. Animals were killed with an intraperitoneal injection of pentobarbital and were bled through the abdominal aorta. The thorax was opened and the left lung was fixed by instilling 10% buffered formalin into the trachea for terminal deoxynucleotidyl transferase-uridine nucleotide end-labeling (TUNEL) assay, or 2% glutaraldehyde for electron microscopy (EM), at a hydrostatic pressure of 20 cm H₂O. The right lung was removed, weighed, and prepared for DNA extraction (upper lobe) or for measurement of enzymatic activity (lower lobe). The same procedure was performed for the preparation of protein extracts (upper lobe) or messenger RNA (mRNA) (lower lobe). Pulmonary edema was determined by measuring the wet:dry (w/d) weight ratio as described previously (8).

Hyperoxia Exposure and In Vivo Treatment

Mice were placed in a sealed Plexiglas chamber and exposed to 100% O₂ as described (8). Mice exposed to room air under the same conditions served as controls. Food and water were available ad libitum. This study protocol was approved by the local ethical committee on animal experiments (Office Vétérinaire Cantonal of Geneva). In some experiments, the caspase inhibitor Z-VAD (N-benzylcarbonyl-Val-Ala-Asp-fluoromethylketone; Bachem, Bubendorf, Switzerland), dissolved in 50% dimethylsulfoxide (DMSO), was used. A 1-mg bolus (0.1 ml) of Z-VAD was injected on Day 2, followed by a continuous infusion (30 µg/h) through Alzet osmotic minipumps (Charles River, Inc., Boston, MA). Controls received solvent only.

Assays for Apoptosis

The TUNEL assay was perfomed on formalin-fixed tissue sections (5 µ) as described previously (9, 10). TUNEL reagents were purchased from Boehringer Mannheim (Mannheim, Germany). Briefly, tissue sections were mounted on slides pretreated with 3-aminopropylethoxysilane (Merck, Darmstadt, Germany), baked overnight at 55°C, dewaxed, and rehydrated. To facilitate access of the reagent to DNA fragments, slides were treated with 30 µg/ml proteinase K for 15 min, and also incubated in a microwave oven for 1 min in citrate buffer 0.01 M, pH 6.0 (11). Subsequent end-labeling with terminal deoxynucleotyl transferase (TdT) (0.3 U/µl) in TdT buffer containing 2 µM biotin 16-uridine triphosphate (UTP) was done for 1 h at 37°C. Sections were incubated with streptavidin-biotin-horseradish peroxidase complex (Dako, Copenhagen, Denmark) and stained with diaminobenzidine (DAB). The immunostained sections were scanned with a high-sensitivity Phototonic Science Coolview color camera (Carl Zeiss, Oberkochen, Germany) as described previously (8). The whole section was analyzed (40 fields) for TUNEL staining, and the same number of fields were analyzed in the adjacent section for control staining. Three different mice were analyzed for each condition, and the percentage of positive cells was calculated for each condition. The results were expressed after subtraction of the values for the control staining (performed simultaneously but without TdT). For EM, transhilar sections were embedded in Epon and then examined with a Philips CM10 400 electron microscope (Philips SA, Zurich, Switzerland) at 70 kV.

Genomic DNA Extraction and Labeling with [-³²P]Adenosine Triphosphate

The right lung was minced with scissors and digested overnight at 50°C in 1 ml of TNE buffer (10 mM Tris-HCl, pH 7.4; 1 mM ethylenediamine tetraacetic acid [EDTA]; 0.1 M NaCl) containing 0.7% sodium dodecyl sulfate (SDS) and 0.3 mg/ml proteinase K. The samples were cooled on ice for 10 min, and one-third volume of 5M NaCl was added. After gentle mixing, samples were centrifuged at 13,000 rpm for 5 min at 4°C. The supernatant was collected and the nucleic acids precipitated with an equal volume of isopropanol. The precipitate was then rinsed in 70% ethanol and dissolved at approximately 1 mg/ml (~ 250 ml) in 10 mM Tris-HCl, pH 8.0; 0.1 mM EDTA. The RNA was digested by adding 10 µg/ml ribonuclease (RNaseA) to the sample and incubating for 1 h at 37°C. The DNA concentration was estimated by measuring the absorbance at 260 nm. Five micrograms of purified DNA were dephosphorylated with 1 U of alkaline phosphatase in 25 µl of buffer (10 mM Tris-HCl, pH 8 to 8.5; 0.1 mM EDTA) by incubation at 37°C for 2 h. The enzyme was then heat-inactivated (90°C for 5 min). The dephosphorylated DNA samples were then cooled on ice, adjusted to 50 mM Tris-HCl (pH 7.6), 10 mM MgCl₂ and 10 mM -mercaptoethanol, and incubated with 5 U of T4 polynucleotide kinase and 10 mCi [-³²P]adenosine triphosphate ([-³²P]ATP) at 37°C for 1 h. The reaction was stopped by adding 20 mM EDTA and by heating the samples at 90°C for 5 min.

Lung mRNA and Northern Blot Analysis

After removal, lungs were immediately frozen in liquid nitrogen and stored at 80°C. After homogenization of lungs in guanidine/thiocyanate, total RNA was isolated by cesium chloride centrifugation, and Northern blot analysis was performed as described previously (12). Nitrocellulose blots of total RNA were hybridized with the following complementary RNA (cRNA) probes corresponding to the complete coding sequences of the respective RNAs: mouse fas and fas ligand (13), and human bax and bcl-xs (kind gifts of J. C. Martinou, Glaxo Wellcome, Geneva, Switzerland). Quantification was achieved by PhosphorImager analysis (Molecular Dynamics, Inc., Sunnyvale, CA), using Molecular Dynamics Image Quant Software. To evaluate gel loading and membrane transfer, blots were stained with methylene blue. These blots were analyzed according to densitometry, and small differences in loading were normalized according to the density of the 18S ribosomal RNA (rRNA) bands. Results for the normalized mRNA abundance are expressed as the mean increase ± SD, relative to control mice.

Protein Extractions and Western Blot Analysis

Lungs were homogenized at 4°C for 30 min in 1 ml lysis buffer (50 mM Tris, pH 8; 250 mM NaCl; 1% Triton X 100; 0.5% sodium deoxycholate; 0.1% SDS) containing 10 µg/ml aprotinin and 1 mM phenylmethylsulfonyl fluoride (PMSF). After centrifugation at 15,000 × g, 40 µg of supernatant protein was loaded per lane onto 10% polyacrylamide gels and electrophoresed. Proteins were transferred to nitrocellulose membranes, and nonspecific binding was blocked with 5% nonfat milk in TBS (20 mM Tris-HCl; 500 mM NaCl, pH 7.5) by overnight exposure at 4°C. Membranes were washed with TBS-Tween (0.025%) and incubated for 1 h at 20°C with a rabbit antimouse p53 antibody (Ab-1; Calbiochem, La Jolla, CA) in gelatin buffer (0.24% gelatin; 50 mM Tris-HCl, pH 7.4; 5 mM EDTA; 0.15 M NaCl; 0.05% Tween) at a concentration of 2 µg/ml. Membranes were washed three times for 10 min each with TBS-Tween 0.025%. Goat antirabbit HRP (Santa Cruz Biotech, Inc.) was used as a second antibody at a concentration of 400 ng/ml. Membranes were washed again with TBS-Tween and developed with an enhanced chemiluminescence detection reagent (ECL) kit (Amersham International, Amersham, UK) at room temperature before being exposed to Hyperfilm-MP films (Amersham International). When necessary, results were quantified as described previously, using subsaturated emulsions on X-ray film.

Caspase Activity Assays

Lung-protein extracts were prepared by homogenizing 50 mg of tissue in 500 µl of hypotonic buffer (25 mM 4-[2-hydroxyethyl-1-piperazine-N'-2-ethanesulfonic acid [Hepes], pH 7.5; 5 mM MgCl₂; 1 mM ethyleneglycol-bis-[-aminoethyl ether]-N,N'-tetraacetic acid [EGTA]; 240 µg/ml 4-[2-aminoethyl]-benzenesulfonyl fluoride, hydrochloride [Pefabloc, Boehringer Mannheim, Mannheim, Germany]; 5 µg/ml leupeptin; 0.75 TIU/ml aprotinin), and were immediately frozen in liquid nitrogen. Samples were then thawed and centrifuged at 15,000 rpm for 10 min, and the supernatant was used for the assay. Twenty micrograms of extracted proteins were incubated with the different fluorescent substrates Z-YVAD-7-amino-4-trifluoromethyl coumarin (Z-YVAD-AFC), Z-DEVD-AFC, or Z-RFFL-AFC (AFC 120, 138 and 97; Enzyme System Products, Dublin, CA) at a concentration of 25 µM in 1 ml buffer, as described previously (14).

Statistical Analysis

For each parameter measured, the values for all animals in the experimental groups were averaged and the SD of the mean was calculated. The significance of differences between the values for an experimental group and those of the control group was determined with the unpaired Student's t test or the nonparametric Mann-Whitney U test when indicated. For multiple-group comparisons one-way analysis of variance (ANOVA) with a post hoc Bonferroni's correction was used. The significance level was set at P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Death during Hyperoxia

In accord with what has been reported previously in rats with EM (1), we observed extensive damage in both endothelial and epithelial cells in the lungs of animals exposed to O₂ for 4 d. The nuclei of these cell types showed condensed heterochromatin in the periphery. The plasma membrane was altered, showing blebbing of varying severity. In some cases, complete membrane disruption was also evident. The cytoplasm was not vacuolated, but in some cases was shrunken or condensed (Figure 1). Alterations were also evident in interstitial cells, but to a lesser extent than in endothelial or epithelial cells.

View larger version (77K): [in this window] [in a new window]   Figure 1.   Electron-microscopic illustration of cell damage during hyperoxia (90 h). (A) An endothelial cell (e), adjacent to red blood cells (r), shows a highly condensed and fragmented heterochromatin. Note the damage to the endothelium demonstrated by the presence of blebbing (arrow). (B) This epithelial type I cell (E ) shows complete membrane disruption (inset) and some features of nuclear condensation. Magnification: ×2,200.

To estimate the extent of apoptosis in injured lung, we utilized the TUNEL assay, which detects DNA fragmentation in tissue sections in situ. Figure 2 shows that there was an increase in the number of TUNEL-positive cells after 48 h of hyperoxic exposure. A quantitative analysis of these data showed significant increases in the number of TUNEL-positive cells, rising from 1.3 ± 1.2% in control lungs to 6.3 ± 2.9% at 48 h, 7.9 ± 3.0% at 72 h, and 15.5 ± 4.2% at 90 h. Changes in these values were significantly different among all of the different study groups (P < 0.001, ANOVA), with the exception of the comparison between 48 h and 72 h. TUNEL-positive cells were evident in the alveolar septa and in the epithelium of the bronchioles.

View larger version (152K): [in this window] [in a new window]   View larger version (150K): [in this window] [in a new window]   Figure 2.   Detection of cell death in hyperoxia-exposed lungs. Lung sections from control (A) and hyperoxia-exposed (B) lungs at 90 h were stained with TUNEL. TUNEL- positive cells and nuclei were significantly increased at 90 h of exposure compared with those of air-breathing mice (as seen by the quantification described in RESULTS). Magnification: ×60.

Under some conditions, the TUNEL assay might not distinguish between apoptotic and necrotic cells (15). However, the cleavage of genomic DNA into nucleosome-size fragments is a distinguishing feature of apoptosis. We therefore examined genomic DNA with agarose electrophoresis. Figure 3 shows that hyperoxic injury was associated with the appearance of nucleosomal ladders, as is consistent with the occurrence of apoptosis. The ladders were sometimes superimposed over a smear, suggesting concurrent necrotic cell death. In control mice, DNA fragmentation was not evident.

View larger version (119K): [in this window] [in a new window]   Figure 3.   DNA ladder detection. Lung DNA from air-breathing mice (n = 2, lanes 2 and 3) or from mice exposed to hyperoxia for 90 h (n = 4, lanes 4-7) was extracted, labeled with [-32P]ATP, and applied to a 2% agarose gel. Note the concomitant presence of DNA laddering with (lanes 6 and 7) or without DNA smear (lanes 4 and 5), showing apoptosis and/or necrosis. Thymus extract of a mouse treated with dexamethasone was used as a positive control (lane 1).

Apoptosis-related Gene Expression

On the basis of the observation that apoptosis occurs in hyperoxic lung, we explored some of the biochemical pathways to cell death. One such pathway involves the tumor-repressor gene p53. By Western blot analysis, we found that p53 increased markedly during oxygen exposure (Figure 4). The quantitative analysis showed a 5-fold increase at 72 h and a 4-fold increase at 90 h as compared with values for unexposed control lung (P < 0.05) (Figure 4). Because p53 influences expression of the apoptosis gene bax (16), we analyzed the expression of bax mRNA with the Northern blot technique. Quantitative analysis showed that bax mRNA levels did not increase until 72 h of exposure, at which time a 9-fold increase was evident (P < 0.001 versus control). This was reduced slightly to 7-fold at 90 h (P < 0.001 versus control) (Figure 5). Because bax can act either in concert with or in opposition to bcl-x, we also examined bcl-x mRNA levels during hyperoxia (17). By 72 h there was a 12-fold increase in the bcl-x mRNA level (P < 0.001). As with bax, this increase was slightly diminished (10-fold, P < 0.001 versus control) (Figure 5).

View larger version (22K): [in this window] [in a new window]   Figure 4.   Western blot analysis for the detection of p53 in lung extracts. p53 was present in control lung (C) and increased at 72 h and 90 h of hyperoxia. The film was exposed for 1 min. Quantification of five different samples for each condition was performed by scanning densitometry. Values represent mean increase above control values (C) ± SD (P < 0.05).

View larger version (39K): [in this window] [in a new window]   Figure 5.   Quantification of bax and bcl-x in murine lung during hyperoxia. Mice were exposed to air (C) or to hyperoxia for the indicated times. Total lung RNA was isolated, and 10 µg was electrophoresed, transferred onto a nylon membrane, and subsequently hybridized with 32P- labeled RNA probes. Quantification of the samples was performed with a Molecular Dynamics PhosphorImager and adjusted to the quantity of 18S rRNA detected with methylene blue staining. Values represent means of three animals ± SD, and are expressed as fold increases over controls (C) set as 1. There was no change at 24 h and 48 h of hyperoxia, whereas there was a significant increase in bax and bcl-x mRNA levels at 72 h and 90 h compared with control values (P < 0.001).

Role of p53 in Hyperoxia-induced Alveolar Damage

To explore the role of p53, we exposed p53^/ null mice and their p53^+/+ littermates to hyperoxia for 90 h. We compared the w:d lung-weight ratio and (w/d)/BW (normalized to initial body weight) in p53^/ and p53^+/+ mice, because as a measure of lung injury, lung weight reflects vascular leakage caused by disruption of the alveolar- capillary barrier. There were minor albeit statistically significant differences between p53^/ mice and their littermates in terms of (w/d)/BW in air-breathing mice (P = 0.05), showing slightly less lung tissue in p53^/ mice. However, the percentage of lung weight increase over the baseline after hyperoxia was identical in both types of mice (Table 1). Likewise, there was no difference in the number of TUNEL-positive cells in the knockout and control groups at 90 h of hyperoxia exposure (not shown). These data indicate that the presence of p53 is not necessary for hyperoxia-induced lung injury or apoptosis.

Expression and Role of Pulmonary Fas and Fas Ligand

In many cell types, ligation of Fas can lead to apoptosis (14, 18, 19). In our study, only a small increase, although significant (P = 0.05), in Fas mRNA level was observed (1.5-fold) at 72 h and 90 h of hyperoxia (Figure 6). Fas ligand mRNA was not detected by Northern blot analysis either in control or in hyperoxia-exposed lungs (not shown). These observations suggest that the Fas cascade may not participate in hyperoxia-induced lung injury. To test this hypothesis, lpr (Fas null) mice were exposed to hyperoxia. If Fas promotes cell death in hyperoxic lung injury, lpr mice would be predicted to exhibit less injury. However, the extent of lung injury was even worse in lpr mice than in controls (Table 1). Taken together, these data suggest that Fas activation is not an important contributor along the pathways to cell death induced by oxygen.

View larger version (34K): [in this window] [in a new window]   Figure 6.   Quantification of Fas mRNA in murine lung during hyperoxia. Mice were exposed to hyperoxia for the indicated times, and total lung RNA was isolated as described. Quantification was achieved as described in Figure 5. Values represent the mean of three animals ± SD and are expressed as fold increases over controls (C) set as 1. Fas mRNA was present in control lungs, and its level did not change until 48 h of hyperoxia; at 72 h and 90 h there was a slight but significant increase (P = 0.05) in the Fas mRNA level.

Proteolytic ICE-like, CPP32-like, and Cathepsin D Activities

The execution phase of apoptosis is known to occur via members of the caspase superfamily of proteases. To determine whether caspases might be activated in hyperoxic lung, we measured the enzymatic activities in lung lysates of three different proteases (ICE-like, CPP32-like, and cathepsin D-like) (6, 20). CPP32-like activity and ICE-like activity were either at or below the detectable level and did not change during hyperoxia (Table 2). Cathepsin D activity was easily detected in control lungs and did not increase during hyperoxia. The liver extract from a mouse injected with 10 µg of anti-Fas was used as positive control for each measurement.

To test further the possibility that caspases participate in hyperoxic lung injury, mice were injected and then infused continuously with the tripeptide caspase inhibitor Z-VAD. A bolus injection of Z-VAD, given at 1 h before Fas activation with an mAb, has been reported to block hepatocyte death completely (14). In our experiments, lung epithelial cell death occurred over several days. Therefore, we used a bolus injection of a higher dose than used in the liver experiment, followed by a continuous infusion (0.03 mg/h). Under these conditions, Z-VAD had no effect on oxygen-induced lung damage at 90 h of exposure. Wet lung weights were 0.21 ± 0.04 g for the DMSO-infused group (n = 7) and 0.22 ± 0.04 g for the Z-VAD-infused group (n = 7). The appearence of nucleosomal ladders in DNA isolated from injured lungs was also unchanged (not shown), indicating that apoptosis was not altered by this treatment.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In virtually all animal models of hyperoxia, injury and death of endothelial and alveolar epithelial cells result in a massive leak of fluid into the alveoli, leading to poor gas exchange and respiratory failure. We found in the present study that hyperoxic lung-cell death showed features of apoptosis and necrosis, and that both probably occurred simultaneously. The occurrence of necrosis is supported by the fact that cell damage was in many cases associated with loss of integrity of the plasma membrane, as well as with cytoplasmic swelling, evidenced by EM in both endothelial and epithelial cells. We also found evidence for apoptosis. Supporting observations for this were, first, that the nuclei of damaged endothelial or epithelial cells exhibited chromatin condensation typical of apoptotic cells, and not a random nuclear destruction. Second, nucleosomal DNA fragmentationa hallmark of apoptosiswas also evident. However, the ladders were sometimes superimposed upon diffuse, nonnucleosomal smears consistent with simultaneous necrosis. Third, there was a marked increase in the expression of several of the proteins and genes known to be associated with apoptosis. Fourth, the TUNEL assay, which is generally considered an indicator of apoptosis, showed significant numbers of TUNEL-positive cells. However, there remains some controversy over the ability of the TUNEL assay to distinguish between internucleosomal fragmentation and random DNA fragmentation (15, 21). It should also be noted that although the maximum extent of TUNEL-positive cells was about 16%, Mantell and associates reported a much higher TUNEL-positive index in similar experiments performed with hyperoxia-exposed mice (22). This quantitative discrepancy might reflect the relative difference in manipulating very sensitive assays. Importantly, both studies show that the extent of cell death increases as the injury becomes more severe.

These observed alterations seem incompatible with an exclusive PCD process, and our results are compatible with what is currently known about oxygen toxicity. High oxygen concentrations are toxic for most cellular constituents and induce irreversible damage. Exposure to hyperoxia is often considered similar to oxidative stress (23). In vitro, exposure to hyperoxia increases levels of reactive oxygen species (Ros), which can, among their deleterious effects, provoke DNA-strand scission (apoptosis), lipid peroxidation (necrosis) of cellular membranes, and activation of various genes whose products are involved in inflammation and cell death (24, 25). Apoptosis and necrosis are commonly conceived as different and mutually exclusive processes (2). For example, in cultured alveolar epithelial cells, hyperoxia leads to necrosis and not apoptosis (9). The possibility remains that apoptosis and necrosis might occur in an exclusive fashion in different lung cells simultaneously, or concomitantly in the same cell.

The biochemical pathways leading to apoptotic cell death are not yet completely elucidated. p53 can induce apoptosis in some types of injury and its expression was strongly upregulated during hyperoxia in our study. p53 is induced in vivo and in vitro after exposure to ionizing radiation, a cellular stress known to provoke massive DNA damage and to have some similarity with other oxidative stresses (4). There is also evidence that ROS can influence transcriptional activators of p53 (26). p53 is generally more strongly expressed in epithelial cells than in mesenchymal cells, and recently, O'Reilly and colleagues described the presence of p53 in bronchial cells of mouse lung and in bronchial and parenchymal cells of hyperoxia-exposed lungs (27). In contrast to radiation, in which the absence of p53 increases resistance (4), we were surprisingly unable to demonstrate an attenuation of oxygen-induced lung damage in p53-deficient mice. Therefore, despite strong induction of p53, no causal relation between p53 upregulation and hyperoxia-induced lung damage could be established. These apparently conflicting results could have several explanations. Perhaps p53 expression is an epiphenomenon that also accompanies necrosis, and is associated with a loss of DNA integrity. Another possibility is that p53 contributes only to epithelial PCD, and attenuation of this effect is not enough to provide significant clinical benefit.

Expression of p53, one of the regulators of apoptosis, is an upstream event to bax gene activation (16), and in accord with this possibility, bax genes are induced in close correlation to p53. Concomitantly, the bcl-x gene was also activated in hyperoxia-exposed lungs. However, we could not discern with Western blotting which form of the protein was synthesized. Indeed, there are two known forms of bcl-x, with bcl-xs promoting and bcl-xl repressing apoptosis in vitro (17).

The interaction with Fas on the cell surface is known to trigger the apoptotic cascade in various cells, epithelial or mesenchymal (14, 18, 19). We observed a strong basal expression of Fas mRNA, which was only slightly upregulated during hyperoxia, and to a lesser extent than in a model of bleomycin-induced fibrosis (28). DNA-strand breaks and Fas upregulation were seen immediately after intratracheal instillation of bleomycin (2 h to 12 h), and also in the late phase (after 6 d), whereas FasL, synthesized by lymphocytes, could be detected in the late phase only (7 d), suggesting that FasL is abundant when lymphocytes are present within the lung. In our model, we observed pulmonary sequestration of platelets and neutrophils, but not of lymphocytes (29). In any case, lpr mice do not manifest increased resistance to hyperoxia, indicating that Fas is not necessary for acute oxygen-induced alveolar damage.

It is possible that various PCD pathways converge into the same executioner machinery, which includes CPP32-like and ICE-like proteases (3, 20). We evaluated the role of the caspase system in two ways. First, we explored the activity of these proteases and did not observe a substantial increase in their activity in correlation with alveolar cell death. By contrast, Mantell and colleagues detected the presence of an ICE cleavage-protein by Western blot analysis of hyperoxic lung. However, the cleavage product was detected only in the nuclear fraction and not in the lung extract (22). Our functional assays showed no significant caspase activity within the lung, especially when compared with the activity detected in the liver of mice injected with anti-Fas mAb. Second, alveolar damage and alveolar cell death were not attenuated by Z-VAD treatment, which is documented to inhibit the activity of the ICE-like proteases. Because this type of treatment was effective in preventing Fas-induced hepatocyte PCD, which is associated with increased caspase activity, and not hyperoxia-induced lung damage, it is possible that hepatocytes and alveolar epithelial cells have different susceptibilities to Fas ligation despite the expression of Fas in these cell types (14, 19). Another explanation would be that hyperoxic lung injury does not involve the caspases commonly implicated in apoptosis (18). Indeed, it has been described that in an Hep3 cell line, apoptosis could occur without the activation of these proteases, suggesting that other, yet unknown proteases are involved in this process (30).

In summary, our results suggest that during hyperoxia, cell death occurs via at least two different and/or additive mechanisms, necrosis and apoptosis. Because interfering with one of the pathways involved in PCD did not provide significant protection against hyperoxic damage, PCD may be only a minor contributor to alveolar damage and cell death from hyperoxia.