THE JOURNAL OF BIOLOGICAL CHEMISTRY
29184 –29191, 2003
© 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
Necrotic Cell Death in Response to Oxidant Stress Involves the
Activation of the Apoptogenic Caspase-8/Bid Pathway*
Received for publication, February 14, 2003, and in revised form, May 14, 2003
Published, JBC Papers in Press, May 15, 2003, DOI 10.1074/jbc.M301624200
Xue Wang‡, Stefan W. Ryter‡, Chunsun Dai§, Zi-Lue Tang‡, Simon C. Watkins¶, Xiao-Ming Yin§,
Ruiping Song‡, and Augustine M. K. Choi‡储
From the ‡Division of Pulmonary, Allergy, and Critical Care Medicine, Department of Medicine, the §Department of
Pathology, and the ¶Center for Biologic Imaging, Department of Cell Biology and Physiology, School of Medicine,
University of Pittsburgh, Pittsburgh, Pennsylvania 15213
Human epithelial (A549) cells exposed to hyperoxia
die by cellular necrosis. In the current study, we dem-
onstrated the involvement of apoptogenic factors in ep-
ithelial cell necrosis in response to hyperoxia, including
the formation of the Fas-related death-inducing signal-
ing complex and initiation of mitochondria-dependent
apoptotic pathways. We showed increased activation of
both Bid and Bax in A549 cells subjected to hyperoxia.
Bax activation involved a Bid-assisted conformational
change. We discovered that the response to hyperoxia in
vivo predominantly involved the activation of the Bid/
caspase-8 pathway without apparent increases in Bax
expression. Disruption of the Bid pathway by gene de-
letion protected against cell death in vivo and in vitro.
Likewise, inhibition of caspase-8 by Flip also protected
against cell death. Taken together, we have demon-
strated the involvement of apoptogenic factors in epi-
thelial cell responses to hyperoxia, despite a final out-
come of cellular necrosis. We have, for the first time,
identified a predominant role for the caspase-8/Bid
pathway in signaling associated with hyperoxic lung
injury and cell death in vivo and in vitro.
The clinical treatment of respiratory failure often requires
supplemental oxygen therapy. Unfortunately, exposure to an
elevated oxygen tension (hyperoxia) may cause acute and
chronic lung injury. Prolonged hyperoxia triggers an extensive
inflammatory response in the lung that degrades the alveolar-
capillary barrier, leading to impaired gas exchange and pulmo-
nary edema (1). The pathological changes in hyperoxia-injured
lungs coincide with the injury or death of pulmonary capillary
endothelial cells and alveolar epithelial cells, which maintain
the integrity of the alveolar-capillary barrier and serve as a
first-line defense against oxidative injury. Compromised epi-
thelial cell function may permit fluid and macromolecules to
leak into the airspace, resulting in clinical respiratory failure
and death (2). Epithelial cell homeostasis requires stringent
control over proliferative and apoptotic pathways.
Hyperoxia-induced cell death may involve apoptosis and ne-
crosis, two distinct mechanisms of cell death with different
* This work was supported by National Institutes of Health Grants
R01-HL60234, R01-AI42365, and R01-HL55330 (to A. M. K. C.). The
costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked “adver-
tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate
this fact.
储 To whom correspondence should be addressed: Dept. of Medicine, Divi-
sion of Pulmonary, Allergy, and Critical Care Medicine, 3459 Fifth Ave.,
Montefiore University Hospital, 628 NW, Pittsburgh, PA 15213. Tel.: 412-
692-2210; Fax: 412-692-2260; E-mail: Choiam@msx.upmc.edu.
This is an Open Access article under the CC BY license.
29184
Vol. 278, No. 31, Issue of August 1, pp.
Printed in U.S.A.
biochemical, morphological, and functional characteristics (3).
In necrosis, an extensive cell lysis results from acute, acciden-
tal, or non-physiological injury (4). In contrast, apoptosis rep-
resents a regulated form of cell death that participates in tissue
development and homeostasis, requiring the action of pro-
teases and nucleases within an intact plasma membrane (5).
The delineation of apoptosis and necrosis as mutually exclusive
processes, however, has not been convincingly demonstrated in
vivo or in vitro.
Previous studies describe apoptosis as a major histological
feature of hyperoxia-induced lung injury in vivo (6 – 8).
Petrache et al. (9) demonstrated induction of apoptosis in mu-
rine macrophage cell lines in response to in vitro hyperoxia. In
contrast, hyperoxia primarily caused necrosis in A549 and in
Type-I epithelial and murine lung bronchial cells (10 –12). Fur-
thermore, pre-treatment with hyperoxia inhibited oxidant-in-
duced apoptosis in A549 cells (13). Thus, the mechanisms un-
derlying hyperoxic lung injury and cell death may vary in a
tissue-specific manner and may involve apoptosis, necrosis, or
a mixed cell-death phenotype.
Members of the death receptor superfamily, including Fas
(14) and tumor necrosis factor-␣ (TNF-␣)1 (15), as well as Bcl
family members (i.e. Bax and Bcl-XL) have been implicated in
hyperoxic lung injury. Adult mice exposed to hyperoxia (⬎95%
O2) for 72 h displayed increased whole-lung Bax and Bcl-XL
mRNA levels; unaltered Bak, Bad, or Bcl-2 mRNA levels; and
decreased Bcl-w and Bfl-1 mRNA levels (14, 16). Increases in
Bcl-XL protein, but not Bax protein, have been reported in
response to hyperoxia in the mouse lung (16, 17). Hyperoxia
may induce other cell death-related molecules, such as p53 and
p21Cip1/WAF1/Sdi1 (p21) (14, 16, 18 –20). Expression of the p53
protein responds to DNA damage and, in turn, regulates genes
involved in growth control, DNA repair, and apoptosis. By
increasing the expression of pro-apoptotic Bcl-2 family mem-
bers such as Bax, p53 may promote cell death. A major regu-
latory target of p53, the cyclin-dependent kinase inhibitor
p21, may protect against oxidative lung injury by inhibiting
cell proliferation and DNA replication, and promoting DNA
repair (11).
Mice with different genetic backgrounds differ in oxygen
sensitivity. In vivo studies using genetically modified mouse
strains have revealed the relative roles of apoptosis-related
proteins in oxygen sensitivity of the lung. Fas⫺/⫺ (lpr) mice (14)
and TNFRI/II⫺/⫺ mice (21) did not display resistance to hyper-
1
The abbreviations used are: TNF, tumor necrosis factor; TUNEL,
terminal deoxynucleotidyl transferase-mediated dUTP nick end label-
ing; PBS, phosphate-buffered saline; LDH, lactate dehydrogenase;
FasL, Fas ligand; DISC, death-inducing signal complex; TNFR-I, TNF
receptor type I.
This paper is available on line at http://www.jbc.org
 Pathways to Cell Death in Hyperoxia
oxia. Murine embryonic fibroblasts derived from Bax⫺/⫺ and
Bak⫺/⫺ mice resisted hyperoxia-induced cytotoxicity (22). In
contrast, p21⫺/⫺ mice were more sensitive than wild-type mice,
whereas p53⫺/⫺ mice did not display any modulation of oxygen
sensitivity (11, 14, 16). Hyperoxia induced the expression of
p21 and of growth arrest and DNA damage-inducible genes in
p53⫺/⫺ mice (11, 23). These observations suggest that down-
stream regulatory targets of p53 such as p21 and Bax modulate
cellular sensitivity to hyperoxia, although p53 itself is appar-
ently not essential.
Despite these observations, the one or more signal transduc-
tion pathways involved in hyperoxic lung injury and cell death
remain incompletely understood. In the current study, we dem-
onstrate the involvement of apoptogenic factors in epithelial
cell responses to hyperoxia, despite a final outcome of cellular
necrosis. We have, for the first time, identified a predominant
role for the caspase-8/Bid pathway in signaling associated with
hyperoxic lung injury and cell death in vivo and in vitro. A
second pathway involving Bax gains importance under condi-
tions that inhibit or abolish the caspase-8/Bid pathway. We
therefore provide an explanation for the lack of Bax modulation
in the lungs of wild-type mice subjected to a lethal dose of
hyperoxia.
EXPERIMENTAL PROCEDURES
Chemicals and Reagents—Digitonin, disuccinimidyl suberate, etopo-
side, and other chemicals were from Sigma Chemical Co. (St. Louis,
MO). Rabbit or goat polyclonal antibodies against Bax, Bcl-XL, Bid,
caspase-8, caspase-9, cytochrome c, Fas, TNFR-1, and protein A-aga-
rose were from Santa Cruz Biotechnology (Santa Cruz, CA) and were
used for Western blotting. Anti-Bax 6A7 antibody was purchased from
BD Pharmingen (San Diego, CA) and used for immunoprecipitation
experiments. The Flip, Bcl-XL, and LacZ inserted in an adenovirus
expression system were from the Molecular Medicine Institute Pro-
grams of Excellence in Gene Therapy Vector Core Facility, University of
Pittsburgh.
Hyperoxia Exposure and Animal Experimentations—Normal, lpr
(Fas null, Fas⫺/⫺), and gld (Fas ligand null, FasL⫺/⫺) adult (7 weeks)
pathogen-free male C3H/HeJ mice and appropriate controls were ob-
tained from The Jackson Laboratory (Bar Harbor, ME). C57BL Bid
wild-type (Bid⫹/⫹) and Bid-null (Bid⫺/⫺) adult (7 weeks) mice were bred
at the University of Pittsburgh. Mice were kept in room air (control) or
exposed to ⬎98% oxygen by placing the cages inside a plexiglass cham-
ber. Animals were injected with 1 ⫻ 108 of adenovirus (inserted with
appropriate gene) per mouse through the trachea into the lung 2 days
prior to oxygen exposure. Food and water were provided normally.
Animals were observed closely for overall mortality, and the time of
death was recorded. Immediately after death, mice were necropsied,
and their lungs were harvested for histochemical and biochemical
analyses.
Cell Culture and Treatments—Human lung adenocarcinoma A549
cells (ATCC CCL185) were grown in F12K medium (Invitrogen) sup-
plemented with 10% fetal bovine serum. Cells were maintained at 37 °C
in 95% room air, 5% CO2 in a humidified chamber. Subconfluent cul-
tures were used in all experiments, with the cells adhered 24 h prior to
the experimental treatment. Cells were cultured in sealed chambers
flushed with 95% O2 and 5% CO2. Control cells were cultured in 95%
room air, 5% CO2. For cells expressing the appropriate gene, 1 ⫻ 105 of
adenovirus per milliliter was added to cells grown to about 30% conflu-
ence into the medium 2 days before exposure to O2. Increased glucose
consumption has been observed in a number of cell types exposed to
hyperoxia in culture. In 10 ml of medium, glucose was depleted and
cellular ATP decreased after 36 h (24). We refreshed the medium and
gases daily to control for the equilibrium between apoptosis and necro-
sis that is influenced by maintaining the ATP levels (5). At each time
point, cell viability was determined by Trypan blue dye exclusion anal-
ysis (Invitrogen). Apoptosis was determined by morphological changes
and TUNEL assays.
Preparation of Primary Fibroblast Cultures from Mouse Lung—Lung
tissue was excised from either wild-type or Bid⫺/⫺ mice. The lung tissue
was soaked in a 1% antibiotic-antimycotic PBS (Invitrogen, Carlsbad,
CA) twice for 20 min. The tissue was minced into 1-mm pieces, which
were plated onto p100 dishes (15–20 pieces per plate). Fetal bovine
serum was added dropwise over each tissue piece, and then the plates
29185
were incubated for 4 h at 37 °C. Then 2 ml of Dulbecco’s modified
Eagle’s medium containing 10% fetal bovine serum and antibiotics was
added to each plate. The plates were restored to the incubator and
monitored every day until the fibroblasts reached confluence.
Cross-linking of Fas—After treatment of hyperoxia, A549 cells were
grown to 90% confluence and were then washed three times with cold
PBS. Freshly prepared disuccinimyl suberate in Me2SO was then added
to a final concentration of 0.25 mM to the cells in 1 ml of PBS. The dishes
were incubated for 30 min on a rocker platform. After cross-linking,
each dish was washed two times with PBS. The cells were then har-
vested for Western blotting analysis.
Annexin V/Propidium Iodide Staining—Using the annexin V-FITC
kit from BD Pharmingen we followed the manufacturer’s protocol.
Briefly, after hyperoxia treatment for 72 h, A549 cells were washed
with cold PBS and resuspended with binding buffer (10 mM HEPES/
NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl2) before transferring 1 ⫻ 105
cells to a 5-ml tube. Then 5 ␮l of annexin V and 5 ␮l of propidium iodide
were added, and the cells were incubated for 15 min in the dark.
Binding buffer (400 ␮l) was then added to each tube and analyzed by
flow cytometry.
DNA Ladder—A549 cells treated with hyperoxia were trypsinized
and washed with PBS, and then the cells were spun down and resus-
pended in lysis buffer containing 50 mM Tris-HCl, pH 7.8, 10 mM
EDTA-2Na⫹, and 0.5% SDS. RNase A was then added at a concentra-
tion of 0.5 mg/ml and incubated at 37 °C for 60 min. Next, the protein
was degraded using 0.5 mg/ml proteinase K at 50 °C for 60 min. DNA
fragmentation was visualized with ethidium bromide staining after
electrophoresis on a 1.2% agarose gel.
Electron Microscopy Assays—Cell cultures were fixed in 2.5% glut-
araldehyde in 0.1 M sodium cacodylate buffer (pH 7.2) for 1 h at 4 °C.
The cells were then postfixed in 1% osmium tetroxide, dehydrated in a
graded series of ethanol, and embedded in LX112 (Ladd Corp.,
Burlington, VT). Thin sections (60 nm) were cut, stained with uranyl
acetate and lead citrate, and examined on a Zeiss EM 10 transmission
electron microscope.
Cytosol Isolation—At different times after exposure to hyperoxia,
A549 cells were harvested in 0.05% of digitonin in extraction buffer
containing 50 mM Hepes, pH 7.5, 50 mM KCl, 5 mM EGTA, and 2 mM
MgCl2 with protease inhibitors. The cell extracts were spun at 14,000 ⫻
g for 20 min, and the supernatants were removed and used for Western
blotting.
Lactate Dehydrogenase Release Assay—LDH release was measured
using a commercially available assay (Cytotoxicity Detection Kit, Roche
Applied Science, Indianapolis, IN). Lung fibroblasts derived from the
wild-type or Bid⫺/⫺ strains were treated with hyperoxia. After gentle
agitation, 200 ␮l of medium was removed at different time points to be
used for the assay. The samples were incubated (30 min) with buffer
containing NAD⫹, lactate, and tetrazolium. LDH converts lactate to
pyruvate, thus generating NADH. The NADH then reduces tetrazolium
(yellow) to formazan (red), which was detected by absorbance (490 nm).
Statistical Analysis—Fisher’s exact test was used for analysis of the
survival rate. The unpaired Student’s t test was used for other data
analyses as indicated, and a value of p ⬍ 0.05 was considered
significant.
RESULTS
Hyperoxia Induced Necrosis in A549 Epithelial Cells Involves
the Release of Apoptogenic Factors from the Mitochondria—
Morphological signs of injury appear in alveolar endothelial
and epithelial cells within 48 –72 h of exposure to lethal levels
of oxygen (2). Continued exposure for more than 72 h can result
in epithelial cell death, and consequently, loss of alveolar in-
tegrity, airway fluid accumulation, and mortality. Hyperoxia
primarily caused necrosis in A549 alveolar epithelial cells, as
determined by fluorescence DNA labeling, in situ end labeling
of DNA, and electron microscopy (10).
We exposed human epithelial cells (A549) to in vitro hyper-
oxia and then monitored the cells for survival. Flow cytometry
with PI/Annexin-V double staining and DNA-laddering analy-
ses demonstrated that hyperoxia caused A549 cell death pri-
marily by necrosis (Fig. 1, B and C). Etoposide-induced apo-
ptosis of Jurkat cells was used as positive controls (Fig. 1A).
Quantitative data from our flow cytometry analyses (PI/An-
nexin-V double staining) demonstrate that ⬎96% of the dead
cells were necrotic, whereas 3% were apoptotic after 72 h of
29186 Pathways to Cell Death in Hyperoxia

FIG. 1. Hyperoxia-induced necrosis in A549 cells. A, etoposide (100 ng/ml)-induced apoptosis of Jurkat cells for 6 h used as a positive control
for apoptosis. B and C, A549 cells were cultured under 95% room air/5% CO2 (Room air) or 95% O2/5% CO2 (hyperoxia) for the time indicated and
analyzed by PI/Annexin V double staining and flow cytometry (B) or DNA ladder analysis (C).

hyperoxia exposure. Furthermore, we also performed electron
microscopy analysis of A549 cells at 72 h of hyperoxia, confirm-
ing the classic feature of necrosis as shown in Fig. 2. The
epithelial cell growth medium was refreshed daily to maintain
high levels of intracellular ATP (24). Hyperoxia induced the
necrotic death of A549 cells even in the presence of a high
glucose concentration in the medium. Under these experimen-
tal conditions, ⬍50% of these cells survived after 72 h of hy-
peroxia, by trypan blue analysis (Fig. 3A).
We hypothesized that apoptogenic factors may participate in
hyperoxia-induced A549 cell necrosis. In many cell types, apo-
ptosis may proceed through receptor-dependent pathways ini-
tiated by Fas ligand (FasL) or tumor necrosis factor-␣ (TNF-␣).
FasL-mediated apoptosis begins with the formation of a death-
inducing signal complex (DISC), that involves the death recep-
tor Fas, Fas-associated death domain, and caspase-8. The con-
sumption and recruitment of Fas leads to apparent decreases
in Fas protein levels. Activated caspase-8 may in turn either
activate effector caspases (i.e. caspase-3) or cleave Bid. The
activated form of Bid triggers apoptosis by mitochondrial trans-
location, which stimulates cytochrome c release (25). On the
other hand, receptor-independent apoptosis may involve the
mitochondrial-translocation of pro-apoptotic Bcl-2 family mem-
bers such as Bax. In the mitochondria, Bax oligomers form
membrane channels by opening the permeability transition
pore, leading to cytochrome c release and caspase activation
(26). Once released in the cytosol, cytochrome c binds to Apaf-1,
stimulating its ATP-dependent oligomerization, which in turn
leads to the recruitment of procaspase-9 and its activation to
caspase-9.
The role of apoptotic pathways in hyperoxia-induced A549
epithelial cell necrosis was evaluated in vitro, by monitoring
the expression of death receptors (Fas and TNFR-I) and Bcl-2
family proteins. Following treatment of A549 cells with hyper-
oxia (24 –72 h), Fas protein levels decreased, whereas TNFR-I
protein levels remained unchanged (Fig. 3B). Anti-Fas immu-
noprecipitation revealed an interaction of caspase-8 with Fas
(Fig. 3B), demonstrating DISC formation, the initiation of the
Fas-induced apoptotic pathway. We show here for the first
time, that hyperoxia up-regulated the activation of both Bax
and Bid in A549 cells (Fig. 3C) with release of cytochrome c into
the cytosol (Fig. 3D). After 24 h exposure to hyperoxia, active
caspase-9 appeared in A549 cells (Fig. 3D).
We next examined the expression of the anti-apoptotic Bcl-2
family member Bcl-XL as a function of hyperoxic exposure in
A549 cells. Bcl-XL inhibits mitochondrial cytochrome c release
and subsequent caspase activation (27–29). After 48 h in hy-
peroxia, the expression of Bcl-XL increased in A549 cells (Fig.
3E), whereas Bcl-2 expression did not change (Fig. 3E). The
high expression of Bcl-XL apparently did not protect A549 cells
from cell death.
Overexpression of FLIP Increased Survival under Hyperoxia
in Vitro and in Vivo—Because hyperoxia treatment led to
caspase-8 and Bid activation, and increased Bcl-XL expression
in A549 cells, we hypothesized that modulation of these apo-
ptosis-related proteins may influence oxygen sensitivity in
 Pathways to Cell Death in Hyperoxia
FIG. 2. Confirmation of necrosis in A549 cells induced by hy-
peroxia. A549 cells were cultured under 95% room air/5% CO2 (control)
(A) and 95% O2/5% CO2 for 72 h (hyperoxia) (B) and analyzed by
electron microscopy.
vitro. A549 cells were infected with adenoviral constructs in-
serted with Bcl-XL, Flip, or LacZ (Fig. 4A), exposed to in vitro
hyperoxia for 72 h, and monitored for necrotic cell death. Over-
expression of the caspase-8 inhibitor, FLIP (but not Bcl-XL),
significantly protected the cells from death during hyperoxia,
relative to the LacZ control (p ⬍ 0.05) (Fig. 4B).
We hypothesized that inhibition of caspase-8 would also
modulate oxygen sensitivity in vivo. We infected C3H/HeJ mice
with adenoviral constructs inserted with Flip, Bcl-XL, or LacZ.
After 48 h, the mice were exposed to ⬎95% O2 and monitored
for overall survival. Interestingly, the FLIP group was signifi-
cantly more resistant to oxygen toxicity than the LacZ-infected
group (Fig. 4C). On the other hand, there was no significant
difference in hyperoxia resistance between the Bcl-XL-express-
ing group and the LacZ control group (Fig. 4C). These results
emphasize an important role for the caspase-8/Bid pathway in
sensitivity to hyperoxia. These results also suggest that in-
creases in Bcl-XL protein expression do not protect against
hyperoxia-induced lung injury or cell death.
Overexpression of FLIP in A549 Epithelial Cells Inhibited
Bid Expression, and Bax Conformational Change Induced by
Hyperoxia—Bid assists Bax in a conformational change that
unmasks the Bax NH2-terminal domain and coincides with
mitochondrial cytochrome c release (30). Thus, the Bax mito-
chondrial membrane insertion triggered by Bid may represent
a key step in the pathways leading to apoptosis (31). Therefore
we investigated the hypothesis that Bax conformational change
may respond to the levels of active Bid in A549 cells subjected
to hyperoxia. A549 cells were infected with adenoviral con-
structs containing Bcl-XL, Flip, or LacZ insertions and then
grown under hyperoxia (72 h). Cell lysates were immunopre-
cipitated with the anti-Bax monoclonal antibody (6A7) that
specifically recognizes the conformational change in Bax pro-
tein (32). FLIP overexpression decreased Bid protein levels and
inhibited the conformational change of Bax, compared with
LacZ controls (Fig. 5). In contrast, Bcl-XL overexpression only
29187
slightly decreased the expression of Bid protein and Bax con-
formational change relative to LacZ controls (Fig. 5). These
results support the hypothesis that Bid assists in a Bax con-
formational change that increases the efficiency of Bax in pro-
moting cell death.
lpr (Fas⫺/⫺) and gld (FasL⫺/⫺) Mice Did Not Display In-
creased Resistance to Hyperoxia but Displayed Increased Bax
Protein Expression Levels—We hypothesized that a Fas-in-
duced apoptosis-like process may precede necrotic cell death in
vivo. lpr mice (Fas⫺/⫺), however, did not display increased
resistance to hyperoxia (14). We challenged lpr (Fas⫺/⫺) and
gld (FasL⫺/⫺) mice with hyperoxia and monitored survival. As
shown in Fig. 6A, both Fas⫺/⫺ and FasL⫺/⫺ mice exhibited
similar survival curves in response to hyperoxia relative to the
C3H/HeJ wild-type mice, without significant statistical differ-
ence (p ⬎ 0.05) (Fig. 6A). We therefore investigated whether a
Fas-independent pathway, involving the expression of the
Bcl-2 family member Bax, might participate in the response to
hyperoxia. Under ambient conditions, Bax protein was ele-
vated in the lungs of the Fas⫺/⫺ mice relative to the wild-type,
with higher expression in Fas⫺/⫺ mice than in FasL⫺/⫺ mice
(Fig. 6B). Further increases in Bax protein were detected in the
lungs of Fas⫺/⫺ and FasL⫺/⫺ mice at the time of death by
hyperoxia, relative to hyperoxia-treated wild-type mice. Con-
versely, the activated form of Bid (p15) appeared only in the
lungs of wild-type mice subjected to hyperoxia (Fig. 6B). We
also found equivalent Bcl-XL expression and caspase-9 activa-
tion in the lungs of all three mice strains after hyperoxia (Fig.
6, B and C).
These results suggest that the activation of Bid appears to
represent the dominant pathway leading to cell death in pul-
monary cells during the hyperoxia response in vivo. Bid may
assist in the conformational change of pre-existing Bax-protein,
which is not directly increased by hyperoxia in vivo. On the
other hand, the deletion of the death receptor Fas promotes
increased Bax expression during in vivo hyperoxia, which
likely represents a compensatory activation of the Bax-induced
pathway.
Bid⫺/⫺ Mice Resisted Hyperoxia-induced Pulmonary Cell
Death—We then hypothesized that deletion of apoptosis-re-
lated factors downstream of Fas (i.e. Bid) would modulate ox-
ygen sensitivity in vivo. Bid knock-out (Bid⫺/⫺) mice and cor-
responding wild-type (Bid⫹/⫹) C57BL mice were exposed to
⬎95% O2 and monitored for overall survival and pulmonary
cell death. Interestingly, the Bid⫺/⫺ group was significantly
more resistant to oxygen toxicity than the corresponding wild-
type littermates (Fig. 7A). We also observed increased resist-
ance to hyperoxia in cells isolated from Bid⫺/⫺ null mice when
compared with cells from negative littermate controls (Fig. 7B).
These results again emphasize an important role for the
caspase-8/Bid pathway in sensitivity to hyperoxia.
DISCUSSION
The therapeutic application of supplemental oxygen to ele-
vate arterial pO2 and reduce tissue hypoxia can be offset by its
cytotoxic properties, which increase the risk of morbidity and
mortality. Over the last 30 years, morphological studies in
animal models have described hyperoxia-induced lung injury.
The first signs of damage include focal cytoplasm swelling of
microvascular endothelial cells, interstitial edema, and endo-
thelial cell fragmentation. After 72 h of exposure, type-I pul-
monary epithelial cells begin to die by necrosis (33–35). Many
studies have also demonstrated that lung cells exposed to hy-
peroxia exhibit features of apoptosis, including chromatin con-
densation, DNA fragmentation, changes in the expression of
Bcl-2 family genes, and increases in TUNEL-positive cells (12,
14, 16, 19, 20). We have also observed increases in TUNEL-
29188 Pathways to Cell Death in Hyperoxia

FIG. 3. Human lung epithelial cells, A549, are killed by hyperoxia, in association with the activation of apoptogenic factors. A, A549
cells were cultured in quadruplet under 95% room air/5% CO2 (0) and 95% O2/5% CO2 for the indicated hours. The extent of death was determined
by Trypan blue dye exclusion, and the data represent the average of three independent experiments (means ⫾ S.E.). The value for 72 h in hyperoxia
is significantly different (**, p ⬍ 0.01) from the control (0) or shorter hyperoxic exposures, using the unpaired Student’s t test. B–E, cell lysates were
then subjected to Western blot analyses as indicated to detect TNFR-1, Fas with/without cross-linking and caspase-8, (the latter from samples of
immunoprecipitation with anti-Fas) (B), Bax and Bid (C), cytochrome c and caspase-9 (D), and Bcl-2 and Bcl-XL (E).

positive staining in mouse lungs after hyperoxia treatment
relative to air-treated controls but without apparent differ-
ences in the response between the various genetically modified
strains or during overexpression of FLIP or Bcl-XL, as used in
this study (data not shown). Intravenous infusion of the
caspase inhibitor N-benzylcarbonyl-Val-Ala-Asp-fluoromethyl-
ketone into mice, however, did not protect pulmonary cells from
death, nor affect lung weights and DNA laddering (14). These
features of hyperoxic lung injury suggest that apoptosis and
necrosis may occur exclusively in the different lung cell types,
and in some cases, may occur concomitantly in the same cell
type.
The intracellular ATP concentration may represent a down-
stream signal capable of directing cells toward either type of
cell death, according to the hypothesis that high energy levels
are required for the execution of the apoptotic program,
whereas they are dissipated during necrosis (36). A number of
additional cellular factors, independent of caspase activation,
may influence the decision between apoptosis and necrosis,
including the duration of mitochondrial membrane pore open-
ing, oxidative stress, and Bcl-2 expression.
Caspases play an indispensable role in regulating many of
the morphological and biochemical features of apoptosis. Two
main pathways of procaspase activation have been proposed: (i)
extrinsic activation of initiator pro-caspases, triggered by a
death receptor (i.e. Fas and TNF-R1)-initiated receptosome
complex and (ii) intrinsic activation of initiator procaspases
initiated by formation of an apoptosome complex, triggered by
environmental insults, senescence, and developmental pro-
grams, that involves the release of cytochrome c from the
mitochondrial intermembrane space to the cytosol (5). The
Bcl-2 family of proteins controls mitochondrial integrity,
whereby the balance of pro-apoptotic (Bax) and anti-apoptotic
(Bcl-2 and Bcl-XL) proteins determines the relative sensitivity
of cells to apoptotic stimuli (37). Bcl-XL protects the mitochon-
dria by inhibiting the release of cytochrome c and subsequent
procaspase activation, and therefore modulates both extrinsic
and intrinsic apoptotic cell death. The existence of necrotic cell
death pathways regulated by an intrinsic death program dis-
tinct from that of apoptosis has also been proposed (38).
Previous studies have suggested that human A549 epithelial
cells subjected to hyperoxia died primarily by necrosis (10). Our
results (Fig. 1) support that hyperoxia induces necrotic cell
death in A549 cells. The outcome of necrosis did not strictly
depend on depletion of cellular ATP levels, because we re-
freshed the medium daily to maintain high levels of glucose
and intracellular ATP (24). The induction of necrotic cell death
in A549 cells activated an apoptotic process that involved the
formation of the Fas-related DISC. We demonstrated the in-
creased expression, activation, and mitochondrial transloca-
tion of both Bid and Bax in A549 cells subjected to hyperoxia.
Bid activation by caspase-8 involved cleavage to the p15 form,
whereas Bax activation involved a Bid-assisted conformational
change. Activation of both Bax and Bid stimulated the release
 Pathways to Cell Death in Hyperoxia 29189

FIG. 4. FLIP (but not Bcl-XL) increased survival under hyperoxia in vitro and in vivo. A549 cells infected with an adenovirus-inserted
gene, LacZ, Bcl-XL, or Flip, were cultured in quadruplet under the same conditions and experimental procedures. A, cell lysates were subjected
to Western blot analysis as indicated to detect the expression of Bcl-XL and FLIP. B, the value of survival of Flip-infected cells after 72 h under
hyperoxia is significantly different (*, p ⬍ 0.05) from control cells (LacZ-infected), but no differences occurred between Bcl-XL-infected and control
cells. C, mice were infected with an adenovirus-inserted gene, LacZ, Bcl-XL, or Flip. 48 h post-infection the mice were placed in a chamber with
⬎98% O2 and monitored for the time of death (n ⫽ 10 in each group). The mice infected with Flip displayed a significantly delayed mortality
compared with mice with LacZ and Bcl-XL (p ⫽ 0.02). There was no difference between the two groups of mice infected with LacZ and Bcl-XL.

protein has not changed in normal mouse lung in response to

FIG. 5. Flip (but not Bcl-XL) expression during hyperoxia de-
creased the level of Bid protein and blocked the change in Bax
protein conformation. A549 cells infected with an adenovirus-in-
serted gene, LacZ, Bcl-XL, or Flip, were cultured in quadruplet under
the same conditions and experimental procedures. Cell lysates were
subjected to Western blot analysis as indicated to detect the expression
of Bid and Bax, including immunoprecipitation with anti-Bax 6A7
antibody as indicated.
of mitochondrial cytochrome c and cleavage of caspase-9 (Fig.
8). The inhibition of caspase-8 protected epithelial cells from
hyperoxia-induced necrotic death. We conclude that the early
events in hyperoxia-induced epithelial cell death involve the
initiation of both receptor-mediated and mitochondria-depend-
ent apoptotic pathways, despite a final outcome of cellular
necrosis.
In mouse models, previous studies show that hyperoxia in-
duces high levels of mRNA transcription of Bcl-XL and Bax, but
not Bax protein, in mouse lungs. Here, we confirm that Bax
hyperoxia (Fig. 6B). We demonstrated for the first time that
Bid activation occurs in normal mouse lung in response to
hyperoxia. We therefore propose that a Bid-dependent pathway
dominates over a Bax-dependent pathway in wild-type mice
subjected to hyperoxia, with Bid assisting in the conforma-
tional change of pre-existing Bax protein (Fig. 8). This model is
consistent with the lack of apparent modulation of Bax protein
in vivo during hyperoxia (Refs. 16 and 17 and Fig. 6B).
In our studies we also found that the Bid⫺/⫺ genotype sig-
nificantly conferred resistance to hyperoxia-induced mortality
in vivo and in vitro. Likewise, inhibition of caspase-8 by over-
expression of Flip also resulted in a marked hyperoxia-resist-
ant phenotype. Although caspase-8 activates Bid with the
highest efficiency, other activating species may include
granzyme-B, and caspase-3 (39 – 41). The existence of alterna-
tive routes to Bid activation may explain the apparent lack of a
hyperoxia resistance phenotype in Fas⫺/⫺ mice.
In the present study, we also found that the anti-apoptotic
molecule, Bcl-XL, could not protect against hyperoxic lung in-
jury and cell death in vivo and in vitro (Fig. 4, B and C). Indeed
it has been reported that Bcl-XL does not prevent Bax-induced
mitochondrial damage (42). Bax induces cell death via apopto-
sis in cells with a low level of Bcl-XL, whereas induces necrosis
in cells with high expression of Bcl-XL (42). Meilhac et al. (43)
described a similar balance between apoptosis and necrosis
responding to Bcl-2 expression levels.
This evidence suggests that the intracellular events leading
to apoptosis and necrosis can occur sequentially. In the initial
step, there is a collapse of mitochondrial membrane potential
and loss of cellular glutathione. At this stage, pharmacologic
29190 Pathways to Cell Death in Hyperoxia

FIG. 6. FasLⴚ/ⴚ (gld) and Fasⴚ/ⴚ (lpr) mice could not manifest an increased resistance to hyperoxia compared with normal mice
(wt) but displayed changes in the expression of Bax and Bid. A, mice (either gld or lpr) were put in a chamber with ⬎98% O2 and monitored
for the time of death. No statistically significant differences in mortality occurred among the three groups (n ⫽ 12 in each group). The lysates from
four lungs of each group, in which the mice died at the same time in different groups, were subjected to Western blot analysis to detect Bax, Bid,
and Bcl-XL (B) or caspase-9 (C).

FIG. 8. Model of hyperoxia-induced cell death pathways in the

lung.

the morphological outcome can resemble apoptosis, necrosis, or

ⴚ/ⴚ
FIG. 7. Bid mice and fibroblasts with Bid null were resistant
to hyperoxia. A, mice (either wild-type or Bid⫺/⫺) were put in a
chamber with ⬎98% O2 and monitored for the time of death (n ⫽ 12 in
each group). There was a significant difference in mortality between the
two groups (p ⬍ 0.05). B, fibroblasts with/without Bid expression were
cultured in quadruplet under 95% room air/5% CO2 (0) and 95% O2/5%
CO2 for the indicated hours. LDH assay was used to test the death.
intervention in the form of caspase inhibitors can prevent cell
death. In the second phase, a disruption of the plasma mem-
brane leads to necrosis (44). Indeed, the term “programmed cell
death” now refers to any kind of cell death mediated by an
intracellular death program, irrespective of the trigger, where
a mixed phenotype (45). Based on our observations, we reason
that, in some cell types or in some conditions, the process of
necrosis shares a similar regulatory mechanism with apopto-
sis, which may include the activation of caspases. The final
manifestation of death is distinct from apoptosis, in that it
escapes inhibition by Bcl-XL.
In summary, hyperoxia induces lung injury and leads to
necrotic cell death in lung epithelial cells, mainly through a two
stage process: (i) the initiation of apoptosis-related molecules
leading to loss of normal mitochondrial function, which in-
volves both death receptor-initiated (Bid-dependent) and mito-
chondria-dependent pathways leading to release of cytochrome
c and (ii) a necrotic cell death. The activation of Bid appears to
 Pathways to Cell Death in Hyperoxia
represent the dominant pathway in pulmonary cells during the
hyperoxia response in vivo. Bid assists Bax to change its con-
formation into the active form (Fig. 8). These data, taken to-
gether, suggest that cell death in the mouse lung and in cul-
tured human A549 epithelial cells involves a cellular death
pathway sharing features of both apoptosis and necrosis.
REFERENCES
1. Horowitz, S., and Davis, J. M. (1997) in Lung Growth and Development,
(McDonald, J. A., ed) pp 577– 610, Decker Press, New York
2. Mantell, L. L., and Lee, P. J. (2000) Mol. Genet. Metab. 71, 359 –370
3. Majno, G., and Joris, I. (1995) Am. J. Pathol. 146, 3–15
4. Alison, M. R., and Sarraf, C. E. (1994) Gut 35, 577–581
5. Kroemer, G., Dallaporta, B., and Resche-Rigon, M. (1998) Annu. Rev. Physiol.
60, 619 – 642
6. Mantell, L. L., Horowitz, S., Davis, J. M., and Kazzaz, J. A. (1999) Ann. N. Y.
Acad. Sci. 887, 171–180
7. Otterbein, L. E., Chin, B. Y., Mantell, L. L., Stansberry, L., Horowitz, S., and
Choi, A. M. (1998) Am. J. Physiol. 275, L14 –L20
8. Mantell, L. L, Shaffer, T. H., Horowitz, S., Foust, R., 3rd, Wolfson, M. R., Cox,
C., Khullar, P., Zakeri, Z., Lin, L., Kazzaz, J. A., Palaia, T., Scott, W., and
Davis, J. M. (2002) Am. J. Physiol. 283, L31–L41
9. Petrache, I., Choi, M. E., Otterbein, L. E., Chin, B. Y., Mantell, L. L., Horowitz,
S., and Choi, A. M. (1999) Am. J. Physiol. 277, L589 –L595
10. Kazzaz, J. A., Xu, J., Palaia, T. A., Mantell, L., Fein, A. M., and Horowitz, S.
(1996) J. Biol. Chem. 271, 15182–15186
11. O’Reilly, M. A., Staversky, R. J., Watkins, R. H., Reed, C. K., de Mesy Jensen,
K. L., Finkelstein, J. N., and Keng, P. C. (2001) Am. J. Respir. Cell Mol.
Biol. 24, 703–710
12. McGrath-Morrow, S. A., and Stahl, J. (2001) Am. J. Respir. Cell Mol. Biol. 25,
150 –155
13. Franek, W. R., Horowitz, S., Stansberry, L., Kazzaz, J. A., Koo, H. C., Li, Y.,
Arita, Y., Davis, J. M., Mantell, A. S., Scott, W., and Mantell, L. L. (2001)
J. Biol. Chem. 276, 569 –575
14. Barazzone, C., Horowitz, S., Donati, Y. R., Rodriguez, I., and Piguet, P. F.
(1998) Am. J. Respir. Cell Mol. Biol. 19, 573–581
15. White, C. W., and Ghezzi, P. (1989) Biotherapy 1, 361–367
16. O’Reilly, M. A., Staversky, R. J., Huyck, H. L., Watkins, R. H., LoMonaco,
M. B., D’Angio, C. T., Baggs, R. B., Maniscalco, W. M., and Pryhuber, G. S.
(2000) Lab. Invest. 80, 1845–1854
17. Ward, N. S., Waxman, A. B., Homer, R. J., Mantell, L. L., Einarsson, O., Du,
Y., and Elias, J. A. (2000) Am. J. Respir. Cell Mol. Biol. 22, 535–542
18. McGrath, S. A. (1998) Am. J. Respir. Cell Mol. Biol. 18, 179 –187
19. O’Reilly, M. A., Staversky, R. J., Stripp, B. R., and Finkelstein, J. N. (1998)
Am. J. Respir. Cell Mol. Biol. 18, 43–50
29191
20. O’Reilly, M. A., Staversky, R. J., Watkins, R. H., and Maniscalco, W. M. (1998)
Am. J. Respir. Cell Mol. Biol. 19, 777–785
21. Pryhuber, G. S., O’Brien, D. P., Baggs, R., Phipps, R., Huyck, H., Sanz, I., and
Nahm, M. H. (2000) Am. J. Physiol. Lung Cell Mol. Physiol. 278,
L1082–L1090
22. Budinger, G. R., Tso, M., McClintock, D. S., Dean, D. A., Sznajder, J. I., and
Chandel, N. S. (2002) J. Biol. Chem. 277, 15654 –15660
23. O’Reilly, M. A., Staversky, R. J., Watkins, R. H., Maniscalco, W. M., and Keng,
P. C. (2001) Am. J. Physiol. 278, L552–L559
24. Allen, C. B., and White, C. W. (1998) Am. J. Physiol. 274, L159 –L164
25. Yin, X. M. (2000) Cell Res. 10, 161–167
26. Brady, H. J., and Gil-Gomez, G. (1998) Int. J. Biochem. Cell Biol. 30, 647– 650
27. Finucane, D. M., Bossy-Wetzel, E., Waterhouse, N. J., Cotter, T. G., and Green,
D. R. (1999) J. Biol. Chem. 274, 2225–2233
28. Susin, S. A., Lorenzo, H. K., Zamzami, N., Marzo, I., Snow, B. E., Brothers,
G. M., Mangion, J., Jacotot, E., Costantini, P., Loeffler, M., Larochette, N.,
Goodlett, D. R., Aebersold, R., Siderovski, D. P., Penninger, J. M., and
Kroemer, G. (1999) Nature 397, 441– 446
29. Wu, M., Xu, L., Li, X., Zhai, Z., and Shu, H. (2002) J. Biol. Chem. 277,
25617–25623
30. Desagher, S., Osen-Sand, A., Nichols, A., Eskes, R., Montessuit, S., Lauper, S.,
Maundrell, K., Antonsson, B., and Martinou, J. C. (1999) J. Cell Biol. 144,
891–901
31. Eskes, R., Desagher, S., Antonsson, B., and Martinou, J. C. (2000) Mol. Cell.
Biol. 20, 929 –935
32. Murphy, K. M., Streips, U. N., and Lock, R. B. (2000) J. Biol. Chem. 257,
17225–17228
33. Raffray, M., and Cohen, G. M. (1997) Pharmacol. Ther. 75, 153–177
34. O’Reilly, M. A. (2001) Am. J. Physiol. 281, L291–L305
35. Horowitz, S. (1999) Chest 116, Suppl. 1, 64S– 67S
36. Leist, M., Single, B., Castoldi, A. F., Kuhnle, S., and Nicotera, P. (1997) J. Exp.
Med. 185, 1481–1486
37. Oltvai, Z. N., and Korsmeyer, S. J. (1994) Cell 79, 189 –192
38. Kitanaka, C., and Kuchino, Y. (1999) Cell Death Differ. 6, 508 –515
39. Heibein, J. A., Goping, I. S., Barry, M., Pinkoski, M. J., Shore, G. C., Green,
D. R., and Bleakley, R. C. (2000) J. Exp. Med. 192, 1391–1402
40. Li, H., Zhu, H., Xu, C. J., and Yuan, J. (1998) Cell 94, 491–501
41. Luo, X., Budihardjo, I., Zou, H., Slaughter, C., and Wang, X. (1998) Cell 94,
481– 490
42. Hu, Y., Benedict, M. A., Wu, D., Inohara, N., and Nunez, G. (1998) Proc. Natl.
Acad. Sci. U. S. A. 95, 4386 – 4391
43. Meilhac, O., Escargueil-Blanc, I., Thiers, J. C., Salvayre, R., and Negre-
Salvayre, A. (1999) FASEB J. 13, 485– 494
44. Fujita, M., Kuwano, K., Kunitake, R., Hagimoto, N., Miyazaki, H., Kaneko, Y.,
Kawasaki, M., Maeyama, T., and Hara, N. (1998) Int. Arch. Allergy Immu-
nol. 117, 202–208
45. Denecker, G., Vercammen, D., Declercq, W., and Vandenabeele, P. (2001) Cell
Mol. Life Sci. 58, 356 –370