Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Acute oxidant lung injury is characterized by hypertrophy, inflammation, edema, and cell death, resulting in respiratory distress and increased morbidity and mortality (1). Lungs are particularly at risk from oxidant injury because of their large surface area, which is constantly exposed to both molecular oxygen as well as various oxidative pollutants, such as ozone (2). Moreover, supplemental oxygen is a common clinical intervention for newborns, children, and adults with respiratory distress (3). Severe hyperoxic lung injury is associated with death of epithelial cells, predominantly bronchiolar ciliated and alveolar type I cells, and alveolar endothelial cells (1, 4). Oxygen toxicity occurs through the conversion of molecular oxygen to more cytotoxic-reduced species of superoxide anion, hydrogen peroxide, and hydroxyl radicals (5). Reactive oxygen species exert their cytotoxicity through production of DNA strand breaks, peroxidation of lipids, and enzymatic alterations of proteins and amino acids (6). It is likely that cells growth-arrest in order to repair cellular damage, or undergo apoptosis when the damage is too extensive to be repaired (7).

The cyclin-dependent kinase inhibitor p21^Cip1/WAF1 (p21) regulates the cellular response to DNA damage by promoting growth arrest, facilitating DNA repair, and regulating apoptosis. Initially, p21 was identified as a protein that was transcriptionally induced by the tumor suppressor p53 (8). Subsequent studies found that DNA damage from radiation or cis-platinum increased p21 expression through accumulation of p53 (9, 10). p21 inhibits cell-cycle progression in G1/S by binding and inhibiting the activities of the G1 cyclins D and E and the S-phase cyclin A (11, 12); p21 may also inhibit DNA replication or regulate DNA repair through its ability to bind and inhibit the activity of proliferating cell nuclear antigen (PCNA) (13). p21 promotes DNA repair following ultraviolet (UV) irradiation or adriamycin-induced damage (14, 15). Although several studies have shown that p21 regulates apoptosis in some cell lines, it is unclear whether this activity is indirectly due to its growth inhibitory activities (15). In addition to being regulated by p53, which accumulates following DNA damage, p21 transcription is increased by the growth inhibitory cytokine transforming growth factor- (TGF-) (18, 19). p21 messenger RNA (mRNA) stability is also post-transcriptionally regulated by oxidative stress through activation of the mitogen-activated protein (MAP) kinase pathway (20). Thus, p21 expression is regulated through multiple pathways and can participate in regulating cell growth, DNA repair, and apoptosis.

Hyperoxia inhibits pulmonary cell proliferation in vitro and in vivo (21, 22) and can result in cell death. Marked cell proliferation occurs during recovery to replace cells that were severely injured or killed during the oxidant stress (22, 23). Studies using an SV40 transformed rat alveolar type II cell line found that hyperoxia inhibited their growth; growth resumed upon recovery in room air (21). Hyperoxia inhibited cell growth in G1 through production of p21 that inhibited cyclin E activity (24). Less is known about molecular signals that regulate hyperoxia-mediated changes in cell proliferation, DNA integrity, and cell survival in vivo. We have recently documented DNA damage and accumulation of p53 and TGF- in bronchiolar epithelial cells and in alveolar cells of lungs acutely exposed to > 95% oxygen (25, 26). Moreover, other studies have suggested that hyperoxia kills pulmonary cells in vivo through an apoptotic pathway (27). Collectively, these observations demonstrate that hyperoxia inhibits cell proliferation, promotes cell injury and death, and results in increased expression of molecules that have been shown previously to participate in regulating the cellular response to injury.

Because numerous studies have clearly demonstrated that p21 plays a central role in regulating how cells respond to injury, the present study was designed to investigate the hypothesis that the expression of p21 increases in lungs of mice exposed to hyperoxia. The finding that hyperoxia does indeed increase p21 expression in vivo is significant because it suggests a molecular mechanism by which oxidative stress regulates pulmonary cell growth arrest, DNA repair, and/or apoptosis.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue and Reagents

Normal adult (8-wk) pathogen-free male C57Bl/6J mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and housed for 1 wk under standard barrier conditions prior to exposure to oxygen. Mice were exposed to room air (control) or > 95% oxygen by placing the cages inside a Plexiglas chamber through which prewarmed, humidified, and filtered oxygen was delivered through a 0.22-µm filter. Oxygen concentrations were monitored each day with a miniOX I analyzer from Catalyst Research Corp. (Owings Mills, MD). Animals were allowed water and food ad libitum and were killed after 24, 48, and 72 h of exposure with intraperitoneal (i.p.) pentobarbital (65 mg/kg injected intraperitoneally). Recovery experiments were initiated by exposing the mice to hyperoxia for 64 h, at which time the cages were returned to room air. This exposure time was chosen because preliminary experiments found that longer exposures resulted in high mortality during the recovery period. Mice were killed every 24 h over a 4-d recovery period. All exposures and handling of the mice were approved by the University of Rochester's University Committee on Animal Resources (Rochester, NY).

The lungs were exposed and the right lobes were ligated and removed for isolation of total RNA. The left lobe was inflation-fixed through the trachea with 100 mM cacodylic acid, pH 7.4, with 2% glutaraldehyde at 10 cm pressure for 15 min. Lungs were further fixed for 12 h in the same buffer, dehydrated through graded alcohol, and embedded in paraffin and 4-µm sections were prepared. Protein homogenates were prepared from some lungs that were first perfused through the right cardiac ventricle with 10 ml Dulbecco's phosphate-buffered saline (PBS) containing 1% glucose, 0.25 mg/ml gentamicin, and 0.2 mM ethyleneglycol-bis-(-aminoethyl ether)-N,N'-tetraacetic acid (EGTA), followed by lavaging twice with 1 ml of the same buffer lacking EGTA. Three to four mice were analyzed per exposure time for each parameter studied.

Immunohistochemistry, Western Blots, and Antibodies

Sections were deparaffinized and hydrated prior to blocking of endogenous peroxidase with hydrogen peroxide/ methanol. Nonspecific antibody reactions were first blocked with goat serum before addition of the primary antibody. Polyclonal rabbit antimouse p21 and goat antimouse p27 were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Primary antibody controls used an equivalent concentration of nonimmune rabbit IgG and resulted in sections without staining (data not shown). Slides were washed extensively in PBS prior to incubating with the appropriate biotinylated secondary antibody with avidin- enzyme complex (Vector Laboratories, Burlingame, CA) as previously described (25). Sections were reacted with 3,3'-diaminobenzidine and counterstained with methyl green.

Perfused and lavaged lungs were homogenized at 4°C in 50 mM Tris (pH 7.4), 150 mM sodium chloride, 2 mM ethylenediaminetetraacetic acid (EDTA), 25 mM sodium fluoride, 25 mM -glycerol phosphate, 0.1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 0.2% triton X-100, 0.3% NP-40, 10 µg/ml leupeptin, 10 µg/ml pepstatin, and 10 µg/ml aprotinin. Cell lysates were centrifuged 13,000 rpm at 4°C and the supernatants aliquotted. Approximately 20% of the lysate was stored at 80°C for determination of protein concentration, and the remainder was immediately boiled in 2× Laemmli buffer (1× is 62.5 mM Tris [pH 6.8], 2% sodium dodecyl sulfate [SDS], 10% glycerol, 0.025% bromophenol blue, and 5% -mercaptoethanol). Protein concentration was determined using a modified Lowry assay (Bio-Rad, Hercules, CA) with bovine serum albumin (BSA) as a standard. Proteins were electrophoretically separated by size on 15% polyacrylamide-SDS gels and transferred to nitrocellulose. Membranes were blocked in PBS containing 5% nonfat dry milk overnight at 4°C before incubating with primary antibody (1 µg/ml) for 1 h. Nonspecific interactions were removed by washing the membranes in PBS containing 0.05% Tween-20 before incubating the blots in peroxidase-conjugated secondary antibody at 1:5,000 (Jackson ImmunoResearch Labs, West Grove, PA). Blots were extensively washed again and conjugates were visualized by chemiluminescence (Amersham, Arlington Heights, IL) and exposure to Kodak Bio-Max film. Relative changes in protein expression were determined by scanning densitometry of blots using Metamorph Image Analysis (Universal Imaging Corp., West Chester, PA).

RNA Extraction and Analysis

Lungs were homogenized in 4 M guanidine isothiocyanate, 0.5% N-laurylsarcosine, 20 mM sodium citrate, and 0.1 M 2-mercaptoethanol using a Techmar homogenizer (Techmar Company, Cincinnati, OH). RNA was extracted using acid phenol and phase lock columns (5 Prime-3 Prime, Boulder, CO) and resuspended in diethylpyrocarbonate-treated water. The amount of RNA in an aqueous solution was determined by absorbance at 260 nm. RNA was electrophoretically separated on a 1.0% agarose-formaldehyde gel and transferred to Nytran. Blots were prehybridized and hybridized at 65°C in 1% BSA, 7% SDS, 0.5 M sodium phosphate, and 1 mM EDTA. Radioactive complementary DNA (cDNA) probes were prepared by random primer labeling (GIBCO/BRL, Grand Island, NY) and [³²P] dCTP (3,000 Ci/mmol; New England Nuclear, Boston, MA). Hybridized blots were washed briefly in 1% BSA, 40 mM sodium phosphate, and 2 mM EDTA twice at room temperature prior to stringent washing in the same buffer at 65°C. Washed blots were visualized on Kodak X-OMAT AR film (Eastman Kodak, Rochester, NY). RNA blots were hybridized with a 454-base pair (bp) cDNA containing the second exon of the mouse p21 gene (28), and were normalized to expression of mRNA for the ribosomal subunit L32 as previously described (25).

In Situ Hybridization

Sections were hybridized with RNA probes derived by transcription from the same p21 cDNA template used in the Northern blot analyses. Radiolabeled sense and antisense probes were synthesized using Sp6 and T7 promoters with ³³P to a specific activity of 1.44 × 10⁹ dpm/µg. The probes were hydrolyzed to a length of approximately 200 bp by alkaline hydrolysis.

Hybridizations were performed as described previously with slight modifications (29). Briefly, 4-µm sections were placed on N-tris[hydroxymethyl]methyl-2-aminoethane sulfonic acid (TES)-treated slides, dried, and deparaffinized. The sections were incubated with Proteinase K and equilibrated with 100 mM triethanolamine-HCl (pH 8.0) before incubation with 0.25% acetic anhydride. The slides were washed in 2× standard saline citrate (SSC), dehydrated, and dried. Prehybridization was at 53°C for 3 h. The slides were then rinsed in 2× SSC and dehydrated. The sections were hybridized for 16 h at 60°C with 30 ng/kb/ml probe. Sections were washed, digested with ribonucleases (RNases) A and T1, and washed in RNase buffer. The slides were then washed in 0.1× SSC at 66°C and in 0.1× SSC at room temperature. The slides were dehydrated, dipped in a 1:1 dilution of NTB-2 emulsion (Eastman Kodak), exposed at 4°C for 2 wk, developed, fixed, and stained with hematoxylin and eosin.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Hyperoxia Increases Expression of p21 mRNA

Northern blot analyses were performed on total lung RNA harvested from lungs of mice exposed to room air or > 95% oxygen for 24, 48, or 72 h. A 2.0-kb p21 mRNA transcript was detected faintly in lungs of room air-exposed mice (Figure 1). The expression of p21 mRNA from room air-exposed lungs was detected only after long exposures of the film or by PhosphorImage Analysis. The abundance of p21 mRNA increased slightly by 24 h of exposure and increased markedly after 48 and 72 h of exposure. To confirm that the blots were equally loaded, they were reprobed for expression of the ribosomal mRNA L32. Densitometric scanning of blots containing RNA from three mice per time point revealed that the expression of p21 mRNA relative to L32 mRNA increased nearly 60-fold after 72 h of hyperoxia (Figure 1b).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Northern blot analysis of p21 expression in hyperoxic mouse lung. (a) Mice were exposed to room air (0) or > 95% oxygen for 24, 48, or 72 h. Total lung RNA was isolated, separated by electrophoresis, and hybridized with [32P] cDNA for p21 and L32. Each lane shows one lung at a given time of exposure. (b) Intensity of hybridization was scanned by densitometry and normalized to expression of L32, and the data are graphically represented compared with room air-exposed (control) levels (n = 3). Hyperoxia significantly increased p21 mRNA expression by analysis of variance (ANOVA) (P < 0.05) at all times.

In situ hybridization was used to identify the cell types that expressed p21 mRNA. Sense and antisense probes were generated from the same cDNA probe that was used for the Northern blot analysis, and were hybridized to sections prepared from room air- and hyperoxia-exposed lungs. Room air-exposed lungs had minimal grains that were uniformly detected in all cell types (Figure 2a). p21 mRNA abundance initially increased in some bronchiolar epithelial cells located near the airway/alveolar junction after 24 h of hyperoxia (data not shown). Abundant p21 mRNA was observed in bronchiolar epithelial cells and to a lesser extent in alveolar cells by 48 h of hyperoxia (Figure 2b). Intense hybridization was observed in both bronchiolar epithelium and in alveolar cells of lungs exposed to hyperoxia for 72 h (Figure 2c). At higher magnification of the alveolus, p21 grains were observed in both cuboidal epithelial cells, characteristic of alveolar type II cells, and in thin-walled cells, which may be alveolar type I epithelial cells, microvascular endothelial cells, and/or interstitial fibroblasts (data not shown). Although some resident alveolar cells had more signal than others, the precise identification of the p21-expressing cells could not be ascertained with light microscopy. Hyperoxia did not cause an appreciable change in p21 abundance in fibroblasts underlying the airway, smooth-muscle cells, perivascular regions of larger blood vessels, or airway cells of the lobar bronchus. Sections of lung hybridized with the sense probe contained minimal grains (data not shown).

View larger version (109K): [in this window] [in a new window]   Figure 2.   In situ hybridization of p21 expression in hyperoxic mouse lung. Mice were exposed to room air (a) or > 95% oxygen for 48 (b) or 72 (c) h. Sections of lung were hybridized with a [33P]-labeled antisense p21 riboprobe. Sections were counterstained with hematoxylin and eosin. Filled arrow depicts bronchiolar epithelial cells, and open arrow depicts alveolar cells. (c) Bar = 100 µm.

Hyperoxia Increases Protein Expression of p21^Cip1/WAF1 But Not p27^Kip1

Western blot analyses were performed on homogenates prepared from room air- and oxygen-exposed lungs to determine whether hyperoxia increased p21 protein abundance. A protein of 21 kD was initially detected in homogenates prepared from lungs exposed to hyperoxia for 24 h (Figure 3a). The abundance of p21 increased markedly by 48 h and remained elevated at 72 h of exposure. Scanning densitometry determined that p21 increased approximately 60-fold over the first 48 h of exposure (Figure 3b). p21 is structurally related to the cyclin-dependent kinase inhibitor p27^Kip1 (p27). Because the expression of p27 is not altered during the cell cycle (30), we evaluated its expression as a control to demonstrate that the Western blots were equally loaded. Expression of p27 was readily detected by Western blotting of homogenates prepared from room air-exposed mice and was not altered during 3 d of oxygen exposure (Figure 3a). Thus, hyperoxia specifically increased p21 protein levels without altering the expression of the related family member p27. Furthermore, the uniform expression of p27 during the exposure period provides evidence that the Western blots were equally loaded.

View larger version (12K): [in this window] [in a new window]   Figure 3.   Expression of p21 and p27 protein in hyperoxic mouse lung. (a) Mice were exposed to room air (0) or hyperoxia for 24, 48, or 72 h, and total lung protein (50 µg) was immunoblotted for p21 or p27. Relative molecular weights in kilodaltons are listed to the left of the figure. Each lane shows one lung at a given time of exposure. (b) Blots were scanned by densitometry, and p21 expression is graphically represented compared with room air-exposed levels (n = 4). Hyperoxia significantly increased p21 expression by ANOVA (P < 0.05) at all times.

Changes in cellular expression of p21 protein was determined by immunohistochemistry. Faint cytoplasmic staining was observed in the bronchiolar epithelium from room air-exposed mice (Figure 4a). A modest increase in p21 staining was observed in some bronchiolar epithelial cells near the bronchiolar/alveolar junction after 24 h of hyperoxia (data not shown). Intense p21 immunostaining was observed throughout the bronchiolar epithelium by 48 (Figure 4c) and 72 (Figure 4e) h of exposure. Higher magnification revealed that whereas expression of p21 was generally expressed uniformly in bronchiolar epithelium of lungs exposed for 48 h, it was not uniformly expressed after 72 h of hyperoxia. Examination of the parenchyma from room air-exposed lungs found low, but uniform, p21 immunostaining in alveolar cells (Figure 4b). This level of p21 staining was not significantly altered by 24 h of hyperoxia (data not shown). However, p21 staining increased markedly in alveolar cells of lungs exposed to hyperoxia for 48 (Figure 4d) and 72 (Figure 4f) h. Although p21 staining was detected in cuboidal alveolar epithelial cells, characteristic of type II cells, the antibody also immunoreacted with the flattened cells between the cuboidal cells. Because the thin cells may consist of type I epithelial cells, endothelial cells, and interstitial fibroblasts, it is presently unclear whether hyperoxia increased p21 expression in all or only some of these cell types. Moreover, although abundant p21 staining was observed in alveolar cells of lungs exposed for 48 h, it was not observed in all resident cuboidal alveolar cells by 72 h of exposure (see open arrow in Figure 4f). Hyperoxia did not increase p21 staining in endothelial cells of large blood vessels, smooth-muscle cells, or airway cells of the lobar bronchi (data not shown). Sections reacted with an equivalent concentration of nonimmune rabbit IgG as a negative control had minimal staining (data not shown).

View larger version (155K): [in this window] [in a new window]   Figure 4.   Immunohistochemical detection of p21 in hyperoxic mouse lung. Mice were exposed to room air (a and b), or > 95% oxygen for 48 (c and d) or 72 (e and f  ) h. Their lungs were immunostained for p21, detected as an orange/brown stain. Sections were counterstained with methyl green. Filled arrow depicts p21-positive cell, and open arrow depicts p21-negative cell. Bar: panels a, c, and e = 50 µm; panels b, d, and f = 20 µm.

p21 Expression Decreases during Recovery from Hyperoxia

To determine whether repair from hyperoxic injury is associated with changes in p21 expression, mice were exposed to hyperoxia for 64 h and returned to room air for up to 4 d. The expression of p21 mRNA was markedly increased after 64 h of exposure, modestly decreased by 1 d recovery, and substantially decreased over the next 2 d (Figure 5a). Scanning densitometry of Northern blots containing RNA from three mice at each time point revealed that p21 expression had returned to control levels by the third day of recovery relative to the expression of L32 (data not shown). In situ hybridization of sections from these mice showed abundant p21 expression in the bronchiolar epithelium and throughout the parenchyma after 64 h of hyperoxia (Figure 5b). The expression of p21 mRNA decreased in the airways of 1-d-recovered mice with a less demonstrable change in alveolar regions (Figure 5c). By Day 2 of recovery, spotty p21 expression was observed in alveolar cells with little detectable p21 expression in the bronchiolar epithelium (Figure 5d). p21 mRNA expression was only detected above background in an occasional cell by 3 d recovery (see airway epithelial cell in Figure 5e). There was no detectable difference in p21 expression between 3- and 4-d recovered lungs (data not shown). Recovery from hyperoxia is often associated with inflammation caused by recruitment of macrophages and neutrophils. Higher-power magnification of inflamed regions of lungs from recovered mice on Days 3 and 4 showed that macrophages had minimal p21 grains and neutrophils did not have any grains (data not shown).

View larger version (146K): [in this window] [in a new window]   Figure 5.   Recovery from hyperoxia is associated with decreased p21 mRNA. (a) Mice were exposed to room air (C) or > 95% oxygen for 64 h (64hO2) and allowed to recover in room air for 1 (1DR), 2 (2DR), 3 (3DR), or 4 (4DR) d. Total lung RNA was isolated, separated by electrophoresis, and hybridized with [32P] cDNA for p21. Each lane shows one lung at a given time of exposure and/or recovery. Sections from 64 h-exposed (b) and 1-, 2-, and 3-d-recovered lungs (c, d, and e, respectively) were hybridized with [33P]-labeled antisense p21 probe. Sections were stained with hematoxylin and eosin. Filled arrow depicts bronchiolar epithelial cells, and open arrow depicts alveolar cells. (e) Bar = 100 µm.

Immunohistochemistry was performed to determine whether recovery was associated with decreased p21 protein. Abundant p21 staining was observed in both the bronchiolar epithelium and the parenchyma of 1-d-recovered mice (Figure 6a). The intensity of p21 staining decreased in both bronchiolar epithelium and alveolar cells after 2 d of recovery (Figure 6b). p21 staining continued to decrease on the third day (Figure 6c) and was nearly undetectable in the majority of bronchiolar and alveolar cells by the fourth day of recovery (Figure 6d).

View larger version (89K): [in this window] [in a new window]   Figure 6.   Recovery from hyperoxia is associated with decreased p21 protein. Mice were exposed to hyperoxia for 64 h and allowed to recover in room air for 1 (a), 2 (b), 3 (c), and 4 (d) d, and their lungs were immunostained for p21, detected as an orange/brown stain. Sections were counterstained with methyl green. (d) Bar = 100 µm.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study demonstrates that p21 mRNA and protein abundance increased in lungs of mice exposed to > 95% oxygen. The cell types with the largest increase in p21 expression were the bronchiolar epithelial cells and resident alveolar cells. p21 expression was observed in cuboidal alveolar cells, suggestive of type II cells, and in flattened cells, suggestive of alveolar type I epithelial, microvascular endothelial, and/or interstitial fibroblast cells. Changes in p21 expression were not observed in cells of the large lobar bronchi, fibroblasts underlying the bronchiolar epithelium, endothelial cells of larger vessels, or neutrophils. Although the precise identity of the p21-producing cells in the lung remains to be determined, the present findings demonstrate that p21 expression increases in regions of the lung that are known to be significantly injured by hyperoxia (1, 4). Recovery from hyperoxia was associated with a decrease in p21 expression that was observed first in the bronchiolar epithelium and later in alveolar cells.

The observation that p21 is increased during hyperoxic lung injury is significant because p21 plays a central role in regulating the cellular response to agents that cause DNA damage. In this capacity, p21 inhibits G1-to-S-phase progression by binding and inactivating cyclins D, E, and A (11, 12). D-type (D1, D2, D3) cyclins and cyclin E bind various cyclin-dependent kinases (CDKs) that regulate entry into S phase by phosphorylating the retinoblastoma gene product Rb (31). Hypophosphorylated Rb maintains the G1 phase by blocking the activity of the transcription factor E2F. Phosphorylation of Rb by active G1 cyclin/ CDKs releases E2F, which regulates transcription of genes that control S-phase progression. Thus, binding of p21 to G1 cyclins inhibits their ability to phosphorylate Rb, resulting in G1 growth arrest. p21 may also play an important role in regulating senescence of fibroblasts (32). We and others have found that hyperoxia inhibits cell proliferation in vivo, which is followed by cell proliferation after 2 d of recovery (22, 23, and data not shown). Although it is unclear whether p21 regulates lung proliferation, loss of p21 expression during recovery is temporally associated with proliferation of bronchiolar epithelial cells and parenchymal cells (data not shown). Perhaps more compelling is the observation that SV40-transformed rat type II epithelial cells exposed to hyperoxia show increased expression and binding of p21 to cyclin E/CDK 2 complexes and result in G1 growth arrest (24). Collectively, these findings support the concept that p21 mediates changes in proliferation induced by hyperoxia.

p21 has also been implicated in regulating DNA repair processes through its ability to bind PCNA. PCNA is an auxiliary factor for DNA replication and repair and is found at sites of DNA damage (13, 14, 33). The human colon carcinoma cell line HCT116, which lacks p21, is less efficient at repairing UV-radiation- or cis-platinum-induced DNA damage than cells transfected with p21 (14). Similarly, p21 under the control of an inducible promoter system enhanced repair of UV-damaged DNA in p21-deficient colorectal carcinoma cells (15). Expression constructs lacking the carboxy-terminal PCNA binding domain of p21 failed to rescue DNA repair activities (13). One mechanism by which p21 may regulate DNA repair is through inhibiting binding of DNA-(cytosine-5) methyltransferase (MCMT) to PCNA (34). MCMT methylates newly replicated DNA, which is required for histone association and imprinting of DNA for further replications. This observation suggests that p21 regulates mismatch and nucleotide excision repair of damaged DNA by inhibiting damaged DNA from becoming hypermethylated. Although it is clear that hyperoxia damages DNA through the production of reactive oxygen species, the source of the free radicals is presently unknown. Because hyperoxic injury is often associated with an influx of macrophages and neutrophils, it is likely they may contribute to the accumulation of hydroxyl radicals that injure cells (35). The observation that p21 is markedly elevated after 48 h of hyperoxia, when cell injury has been documented, supports the concept that it may play a role in regulating repair of DNA damaged by reactive oxygen species.

Although the present study demonstrates that hyperoxia increased p21 expression, further studies are needed to clarify how hyperoxia regulates p21 expression. Cell-culture experiments have shown increased transcription of the p21 gene by TGF- and p53. TGF- induces p21 in a p53-independent manner (18, 19), and p53 induces p21 in response to DNA damage (8, 10). Analysis of the p21 promoter revealed that TGF- and p53 increase p21 transcription through distinct cis-acting regions (19). p21 mRNA stability is also regulated post-transcriptionally by redox status through activation of the MAP kinase pathway (20). One study demonstrated that the half-life of p21 protein in the human osteosarcoma cell line U2OS was 30 to 60 min (36). Because p21 is ubiquitinated, it remains to be determined whether this post-translational modification plays any role in regulating the half-life of p21. The present study found that hyperoxia increased p21 mRNA and protein expression in cell types that have previously been shown to express both TGF- and p53 in response to hyperoxia (25, 26). Although it is still unclear how hyperoxia regulates TGF- and p53 expression, the present study supports the concept that TGF-, p53, and/or simple changes in oxidant status may be responsible for regulating the observed changes in p21 expression.

Molecular signals that regulate cell proliferation, survival, and repair from oxidant injury remain to be identified. The findings in this study demonstrate that hyperoxia increases expression of p21, which plays a central role in regulating how cells respond to injury. The observation that hyperoxia causes DNA damage and accumulation of p53, TGF-, and now p21 supports the concept that pulmonary cells actively respond to oxidant damage. Further studies are needed to clarify the role of p21 in orchestrating the cellular response to oxidant injury.