Nuclear Factor- (cid:107) B Is Activated by Hyperoxia but Does Not Protect from Cell Death*

Oxidative insults that are lethal to epithelial cells kill either via apoptosis or necrosis. Nuclear factor- (cid:107) B (NF-(cid:107) B) is a redox-sensitive transcription factor that is activated by oxidative insult, and NF- (cid:107) B activation can protect cells from apoptosis. To test if NF- (cid:107) B can protect from necrotic cell death caused by high levels of molec- ular O 2 (hyperoxia), we exposed human alveolar epithe- lial (A549) cells to hyperoxia. NF- (cid:107) B was shown to be activated and was translocated to the nucleus within minutes. Nuclear translocation persisted over the course of several days, and the levels of NF- (cid:107) B protein and mRNA increased as well. In hyperoxia, NF- (cid:107) B regu-lation was independent of mitogen-activated protein ki- nase (MAPK). In sharp contrast, there was neither nuclear translocation of NF- (cid:107) B nor any increase in expression after exposure to H 2 O 2 at a concentration where this oxidant induces both MAPK and widespread apoptosis. Despite the activation and increased expres- sion of NF- (cid:107) B in hyperoxia, this oxidant remained lethal to the cells. These observations confirm the notion that apoptosis occurs in the absence of NF- (cid:107) B activation but indicate that protection from cell death by NF- (cid:107) B is probably limited to apoptosis. 30 min. The coverslips were washed in excess water and mounted to microscope slides, and the results were examined by immunofluorescence microscopy using the UV-2A filter (Nikon Inc., Melville, NY) for rhodamine. Western Blots— Cell lysates were prepared according to the procedures recommended by New England BioLabs (Beverly, MA). Protein from each sample was loaded onto a 10 or 12% SDS-polyacrylamide gel (Bio-Rad, Hercules, CA). After electrophoresis, proteins were trans- ferred to Immobilon-P membranes (Millipore, Bedford, MA). For MAPK detection, blots were incubated with an antibody that recognizes only the activated (phosphorylated) MAPK protein (New England BioLabs). Antibody was detected by the enhanced chemiluminescent (ECL) rea- gent, CDP-Star (Boehringer Mannheim), and exposing the filter to

Oxidative stress resulting from the toxic effects of reactive oxygen intermediates (ROI) 1 plays an important role in the pathogenesis of many disease states including carcinogenesis, atherosclerosis, and inflammatory disorders (1,2). Oxidative injury can also lead to cell death, and ROI can have a role in apoptotic cell death induced by nonoxidative insults (3)(4)(5)(6)(7)(8). Direct oxidative injury often occurs as a consequence of ventilatory O 2 therapy, which is used in the treatment of critically ill patients who cannot breath efficiently. This treatment typically requires supraphysiologic concentrations of O 2 (hyperoxia), which results in an elevated level of ROI in many cell types of lung, the organ that receives the highest level of O 2 exposure (9,10). The lung is estimated to be composed of as many as 60 cell types, which complicates the study of pathways to cell death by hyperoxia. On the other hand, cultured lung epithelial cells provide simpler models for understanding certain aspects of pulmonary biology. We have recently reported that A549 cells (derived from human type II alveolar epithelial cells) and HeLa cells succumb to hyperoxia not via apoptosis but by necrosis. In contrast, lethal doses of the oxidants H 2 O 2 or superoxide kill these epithelial cells via apoptosis (11). These observations indicate that the pathways to cell death may differ, depending on the oxidant and dose used.
Very recent reports from several laboratories show that apoptotic cell death can be prevented by the expression of nuclear factor-B (NF-B) (12-15), a multisubunit transcription factor that rapidly activates the expression of genes involved in inflammation, infection, and stress (16). Taken together, these recent reports suggest that the induction of NF-B may be part of a survival mechanism used to escape cell death (17). In this report, we examine the expression of NF-B in cells exposed to lethal concentrations of hyperoxia or H 2 O 2 and show that, despite the induction of NF-B by molecular O 2 , the cells do not escape death.

EXPERIMENTAL PROCEDURES
Cell Culture-Human lung adenocarcinoma A549 cells (ATCC CCL 185) were cultured and maintained in 95% room air, 5% CO 2 as described previously (11). Subconfluent cultures were used in all experiments. For hyperoxia treatment, cells were maintained in sealed humidified chambers flushed with 95% O 2 , 5% CO 2 . Media and gases were refreshed daily. Control cells were cultured in 95% room air, 5% CO 2 . For the H 2 O 2 experiment, cells were cultured in 95% room air, 5% CO 2 , 5 mM H 2 O 2 . These cells are known to be relatively resistant to H 2 O 2 , and millimolar concentrations of H 2 O 2 are necessary to induce apoptosis in the culture medium used (11).
Immunofluorescence-All procedures were carried out at room temperature. Cells grown on coverslips were washed once with 1 ϫ phosphate-buffered saline (PBS, Life Technologies, Inc., Gaithersburg, MD) and fixed for 10 min in 10% formalin buffered in 1 ϫ PBS. Coverslips were then rinsed with 3 changes of PBS and incubated with 1% BSA solution (Panvera Corp., Madison, WI) for 10 min. Cells were incubated with a 2.5 g/ml (in 1% w/v BSA) anti NF-B p65 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for 30 min and washed with 3 changes of 1 ϫ TBS (0.1 M Tris-Cl, pH 7.4, 0.15 M NaCl). Secondary antibody, sheep anti-rabbit IgG-rhodamine (Boehringer Mannheim, Indianapolis, IN) was diluted in 1% BSA to a concentration of 27 g/ml and incubated with coverslips for 30 min. The coverslips were washed in excess water and mounted to microscope slides, and the results were examined by immunofluorescence microscopy using the UV-2A filter (Nikon Inc., Melville, NY) for rhodamine.
Western Blots-Cell lysates were prepared according to the procedures recommended by New England BioLabs (Beverly, MA). Protein from each sample was loaded onto a 10 or 12% SDS-polyacrylamide gel (Bio-Rad, Hercules, CA). After electrophoresis, proteins were transferred to Immobilon-P membranes (Millipore, Bedford, MA). For MAPK detection, blots were incubated with an antibody that recognizes only the activated (phosphorylated) MAPK protein (New England BioLabs). Antibody was detected by the enhanced chemiluminescent (ECL) reagent, CDP-Star (Boehringer Mannheim), and exposing the filter to shorowitz@winthrop.org. 1 The abbreviations used are: ROI, reactive oxygen intermediates; NF-B, nuclear factor-B; MAPK, mitogen-activated protein kinase; PBS, phosphate-buffered saline; BSA, bovine serum albumin; MOPS, 4-morpholinepropanesulfonic acid.
x-ray film. For NF-B detection, the filter was incubated with rabbit anti-NF-B p65 antibody as recommended by the supplier (Santa Cruz Biotechnology).
Northern Blot Analysis-Total RNA was purified from cells by using RNA STAT-60 (TEL-TEST "B", Inc., Friendswood, TX) according to the supplier suggestions. RNA from each sample was loaded onto a 1% agarose gel containing 2.2 M formaldehyde. The gel was electrophoresed in 1 ϫ MOPS (Sigma) buffer and transferred to a 0.2 m nylon membrane (Schleicher & Schuell, Keene, NH). The blots were exposed to UV using a DNA Transfer Lamp (Fotodyne, New Berlin, WI) and prehybridized at 50°C for at least 4 h in solution containing 50% formamide, 5 ϫ SSC, 1% blocking solution (Boehringer Mannheim), 0.1% SDS, 0.1% sarcosine, and 100 g/ml yeast tRNA. The filter was then hybridized with digoxigenin-labeled dUTP (Boehringer Mannheim) probes overnight at 50°C and washed as recommended. The hybridized signal was detected by using anti-digoxigenin antibody and developed in ECL solution (Boehringer Mannheim) according to the supplier recommendations. The probes for NF-B (full-length human p65 cDNA) and glyceraldehyde-3-phosphate dehydrogenase mRNA (full-length human cDNA) detection were labeled with random primer labeling kit (Boehringer Mannheim).

RESULTS
Because the lung is the primary target of O 2 toxicity, we utilized lung epithelial cells as a model to study signal transduction by injurious levels of hyperoxia and oxidants. In particular, A549 cells, which are derived from an adenocarcinoma of alveolar type II cells, have been extensively studied with respect to their responses to oxidants and other airborne insults (6,18). In earlier studies, we have shown that cultured A549 cells exposed continuously to 95% O 2 suffer cell death via necrosis and not by apoptosis. In contrast, lethal concentrations of H 2 O 2 or superoxide cause apoptosis (11).
Because recent reports suggest that NF-B can prevent cell death, we investigated the role of NF-B in these two distinct modes of cell death. A549 cells exposed to 95% O 2 showed evidence of cell swelling by 24 h, and the culture gradually died off over the course of one week (11). It is known that following release from the inhibitory binding protein IB, NF-B translocates from cytosol to the nucleus, where it regulates transcription (19). We therefore studied NF-B activation during hyperoxia by examining its nuclear translocation. Fig. 1 shows that control cells grown in room air had weak NF-B immunofluorescence. The signal was evident primarily in the cytoplasm although there was limited fluorescence in the nuclei of some cells. By 30 min of hyperoxia, nuclear fluorescence was more prominent, and it increased over the course of 1 day. By 24 h of hyperoxia, the cells already showed signs of swelling, and fluorescence was more intense both in the nuclei as well as in the cytoplasm of many cells. In contrast to hyperoxia, there was no nuclear translocation of NF-B when cells were exposed to concentrations of H 2 O 2 that caused apoptosis (Fig. 1, panels E-G). By 4 h of H 2 O 2 treatment, the majority (ϳ80%) of cells had undergone apoptosis (data not shown). Fig. 1 suggests that an increase in NF-B protein levels occurred during hyperoxia in addition to nuclear translocation. To examine this further, Western blots were performed. Fig. 2 shows that NF-B levels were increased as soon as 30 min after exposure to 95% O 2 , and peak levels were achieved by 24 h. Levels remained elevated for 3 days (Fig. 2). In sharp contrast, apoptotic concentrations of H 2 O 2 caused no increased NF-B protein, and there was even a slight decrease after 2 h (Fig. 2).
To determine if elevated NF-B protein levels were correlated with increased mRNA abundance, Northern blot analyses were performed. Fig. 3 shows that steady-state levels of NF-B mRNA were elevated as soon as 30 min after O 2 exposure. Message levels increased over the course of 1 day and remained elevated for 3 days. In contrast, the abundance of the message encoding the glyceraldehyde-3-phosphate dehydrogenase was unchanged (data not shown). Unlike hyperoxia, exposure of the culture to an apoptotic dose of H 2 O 2 failed to induce NF-B mRNA in these epithelial cells. Rather, a modest decrease was observed after 2 h (Fig. 3).
The differential activation and expression of NF-B during different modes of cell death imply that different signals are transduced by hyperoxia and H 2 O 2 . We therefore investigated whether p42 and p44 MAPKs (mitogen-activated protein kinases) were activated. p42 and p44 MAPKs both function in a protein kinase cascade that plays a critical role in the regula- FIG. 1. Localization of NF-B in control, hyperoxia, and H 2 O 2 exposed A549 cells. A549 cells were grown in 95% room air, 5% CO 2 (A); exposed to 95% O 2 , 5% CO 2 for 30 min (B), for 2 h (C), and for 24 h (D); or exposed to H 2 O 2 for 30 min (E), 1 h (F), and 2 h (G). tion of cell growth and differentiation (20 -23). MAPKs are activated by variety of extracellular signals, and p42 and p44 MAPKs phosphorylate many key regulatory proteins and transcription factors that regulate cell proliferation. In addition, recent work showed that TNF, which activates NF-B, also activates the MAPK cascade (24). To study MAPK, we used antibodies that are specific to the phosphorylated or activated form of the proteins. Fig. 4 shows that there were no changes detected in the levels of phosphorylated p42 or p44 at any time during exposure to hyperoxia (phosphorylated p42 is predominant). In contrast, as soon as 10 min after incubation in an apoptosis-inducing concentration of H 2 O 2 , there was a significant increase in phosphorylated p42. Phosphorylated p44 was also increased, although lower in abundance. Interestingly, this activation was transient, and the phosphorylated protein levels decreased rapidly. DISCUSSION Multiple lines of evidence indicate that apoptosis results from severe oxidative insults, and it has even been suggested that apoptosis triggered by non-oxidative insults requires lipid peroxidation in the dying cells (25). However, cellular oxidation also can activate NF-B, which occurs through the redox-sensitive disassociation between NF-B and its inhibitory protein IB (26). Several groups have very recently reported that activation of NF-B can prevent cell death induced by TNF, X-irradiation, and chemotherapeutic agents (12)(13)(14)17). In the absence of NF-B expression or activation, these insults induce apoptosis in a variety of cell types. Taken together, these observations suggest that some cells might utilize a strategy to prevent death resulting from oxidant injury by activating NF-B, which presumably regulates downstream genes that salvage the cell. We tested this notion in two models of oxidantinduced cell death by exposing cells to lethal concentrations either of hyperoxia or H 2 O 2. Surprisingly, we found that despite the rapid activation of NF-B, hyperoxia remains lethal, indicating that NF-B is not sufficient to protect the cells from this oxidative insult. Moreover, when apoptosis was triggered in the cells by another oxidative insult, H 2 O 2 , NF-B was not induced.
The mode of cell death induced by these two oxidative insults is different, hyperoxia caused necrosis while H 2 O 2 caused apoptosis (11). Thus, although recent reports indicate that apoptotic cell death is avoidable by NF-B activation, they did not address whether NF-B could provide protection from other modes of cell death. Similarly, the time course of those studies was relatively short, and the data did not bear on slower cell death that occurs over the course of several days. Our observation that oxidative necrotic cell death can ensue despite NF-B activation supports the notion that strategies aimed at blocking NF-B activation might accelerate cell death (via apoptosis) but implies that with sufficient time the overall amount of cell death might be no different. Also, because necrosis in vivo is associated with inflammation and apoptosis typically is not, when radiation or chemotherapy are administered to patients, it might actually be preferable to encourage or at least permit some degree of inflammation to hasten the removal of dead tumor cells.
Interestingly, not only was NF-B activated, but its expression was also induced at the protein and mRNA levels, and increasing amounts of NF-B were found in cells exposed to prolonged hyperoxia. This suggests that NF-B synthesis was also persistent. In attempting to identify the pathway leading to NF-B activation and induction by hyperoxia, we found that the MAP kinase cascade was not induced, which is distinct from other observations (24).
Non-apoptotic epithelial cell death, which is morphologically similar to hyperoxia, occurs at lower (and perhaps more physiologically relevant) levels of oxidants than are required for the induction of apoptosis (11). Interestingly, it is also known that non-apoptotic cell death in the roundworm Caenorhabditis elegans, can occur in mutants of a gene family that include deg-1, mec-4, and mec-10. In these cases, cell death morphologically resembles hyperoxic or low oxidant cell necrosis and is characterized by swelling and lysis of a specific group of neurons (27). Likewise, germ cells in mes-3 Ϫ worms undergo a necrosis, not apoptosis (28). Although mec-4, encodes a subunit of a mechanosensory ion channel, the function of these genes are not yet fully understood. Perhaps they are steps along a pathway leading to necrosis. Taken together with our observations, it may prove valuable to explore pathways to alternate (non-apoptotic) modes of cell death. Cell lysates were isolated from control (C) A549 cells and cells exposed to hyperoxia or 5 mM H 2 O 2 for the times indicated. Thirty g of protein were electophoresed, blotted, and incubated with an antibody specific for the phosphorylated forms of p42/p44 MAPK.