A R T I C L E
Bnip3L is induced by p53 under hypoxia, and its knockdown
promotes tumor growth
Peiwen Fei, Wenge Wang, Seok-hyun Kim, Shulin Wang, Timothy F. Burns, Joanna K. Sax,
Monica Buzzai, David T. Dicker, W. Gillies McKenna, Eric J. Bernhard, and Wafik S. El-Deiry*
University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104
p53-dependent apoptosis is a major determinant of its tumor suppressor activity and can be triggered by hypoxia. No p53
target is known to be induced by p53 or to mediate p53-dependent apoptosis during hypoxia. We report that p53 can
directly upregulate expression of Bnip3L, a cell death inducer. During hypoxia, Bnip3L is highly induced in wild-type p53-
expressing cells, in part due to increased recruitment of p53 and CBP to Bnip3L. Apoptosis is reduced in hypoxia-exposed
cells with functional p53 following Bnip3L knockdown. In vivo, Bnip3L knockdown promotes tumorigenicity of wild-type
versus mutant p53-expressing tumors. Thus, Bnip3L, capable of attenuating tumorigenicity, mediates p53-dependent
apoptosis under hypoxia, which provides a novel understanding of p53 in tumor suppression.
apoptotic activity of p53 is essential for tumor suppression.
During tumor growth, hypoxia is a major stress that cells en-
counter (Achison and Hupp, 2003; Vaupel et al., 1991). To date,
a number of p53 targets have been identified and proposed to
mediate p53-dependent cell death in response to genotoxic
stresses (Fei et al., 2002; Nakano and Vousden, 2001; Oda et
al., 2000a, 2000b; Villunger et al., 2003; Jeffers et al., 2003).
However, none of them hasbeen shown to mediate p53-depen-
dent cell death under hypoxia. In fact, p53 has been shown to
be incapable of upregulating its known transcriptional targets
during hypoxia, due to a failure to recruit coactivator proteins
such as CBP or p300 to their promoters (Koumenis et al., 2001).
We report here that Bnip3L is a proapoptotic transcriptional
target of p53. Its induction by p53 under hypoxia appears to
occur through the recruitment of coactivator CBP to Bnip3L,
leading to a greater magnitude of Bnip3L induction in wild-
type as compared to p53-deficient cells. Silencing of Bnip3L
significantly blocks the apoptotic advantage in wild-type p53-
expressing cells under hypoxia. Bnip3L knockdown promotes
tumorigenicity in human tumor xenograft models if tumor cells
contain wild-type but not mutant p53. In particular, nontumori-
genic U2OS osteosarcoma cells were converted into a tumori-
genic state following injection of nude mice with U2OS cells
harboring stable Bnip3L knockdown. Our results suggest that
Bnip3L can suppress tumor xenograft growth in vivo and is
a mediator of p53-dependent apoptosis under hypoxia. The
regulation of Bnip3L by p53 provides a novel mechanism by
which p53 acts as a tumor suppressor in vivo.
Solid tumors are poorly oxygenated as compared with normal
tissues and possess regions of hypoxia (Vaupel et al., 1991).
Apoptosis induced by hypoxia is a mechanism for elimination
of stressed cells (Shimizu et al., 1996). In response to hypoxia,
hypoxia-inducible factor 1 (HIF-1) activates genes involved in
menza, 2000), which contribute to adaptive survival. HIF-1 also
activates proapoptotic members of the Bcl-2 family, including
Bnip3L (Denko et al., 2003; Piret et al., 2002), which induces
cell death, and plays an essential role in cardiac cell death
during hypoxia (Kubasiak et al., 2002; Yussman et al., 2002).
HIF-1-dependent (An et al., 1998; Carmeliet et al., 1998; Piret
et al., 2002; Hansson et al., 2002) and/or -independent mecha-
nisms (Wenger et al., 1998; Pan et al., 2004). Wild-type p53-
or survival during hypoxia associated with tumor progression
(Schmaltz et al., 1998).
p53 is frequently mutated in human cancer and is consid-
ered a “guardian of the genome” in preventing cancer (Chan et
al., 2000; Lane, 1992; Levine, 1997). Abundant evidence (Sy-
monds et al., 1994; Attardi and Jacks, 1999; Aurelio et al., 2000;
Bardeesyet al.,1995;Eischenet al.,2001;Meijerink etal.,1998;
Schmitt et al., 2002; Soengas et al., 2001) indicates that the
S I G N I F I C A N C E
The p53 gene is important in tumor suppression during hypoxia, but none of its known proapoptotic targets is transactivated by p53
under hypoxic conditions. We show here that p53 can transactivate proapoptotic Bnip3L during hypoxia and that CBP and p53 are
recruited to Bnip3L in vivo. Although Bnip3L is a HIF target, its silencing only significantly protects cells in culture from apoptosis
induced by hypoxia if cells contain wild-type p53. Silencing Bnip3L promotes tumor growth in vivo through reduced sensitivity to
hypoxia and increased proliferation. This work provides insights into hypoxic death of tumors and possible strategies for therapeutic
restoration of tumor sensitivity during hypoxia.
CANCER CELL : DECEMBER 2004 · VOL. 6 · COPYRIGHT © 2004 CELL PRESS 597
A R T I C L E
Figure 1. Bnip3L is a p53-regulated gene
A and B: Bnip3L induction at the mRNA level (A) or protein level (B) depends on wild-type p53 status following 5-FU and cisplatin (Cis) exposure of U2OS
cells. The arrows indicate the two human Bnip3L transcripts (3.9 and 1.6 kb). Ethidium staining shows the 18S rRNA in the lower panel as a loading control.
The quantification of Bnip3L induction by a phosphorimager is provided.
C: Bnip3L mRNA induction in MEFs carrying temperature-sensitive p53.
D and E: In vivo dependence of Bnip3L expression on wild-type p53 or ? irradiation-induced DNA damage in mice by in situ hybridization analysis. Spl,
spleen; Ile, ileum; DC, descending colon; Duo, duodenum; Jej, jejunum.
expression was observed in wild-type p53-expressing U2OS
(3-fold), as compared to untreated U2OS cells or treatment of
U2OS cells carrying human papillomavirus E6 that targets p53
for degradation (Figure 1A). A similar pattern of Bnip3L protein
induction was observed in chemotherapy-treated U2OS-Neo
cells but not U2OS-E6 cells (Figure 1B). We further confirmed
the regulation of Bnip3L by p53 in VM10 murine embryonic
fibroblasts carrying temperature-sensitive p53 (Sax et al., 2002)
and in p53?/?versus p53?/?mice. Bnip3L mRNA levels in-
(32?C), and a further increase occurred upon adriamycin expo-
sure (Figure 1C). Using in situ hybridization, we found that
Bnip3L expression (Figure 1D) increased in irradiated spleen,
ileum, and transverse colon of wild-type but not p53-deficient
mice. We also noted a higher basal expression of Bnip3L in
Bnip3L is a p53-regulated gene
Among the known activities of p53, sequence-specific DNA
binding and transactivation explain the majority of its effects
(El-Deiry, 1998). We hypothesized that the apoptotic effect of
p53 under hypoxia may be mediated through transcriptional
control, which prompted us to analyze candidate apoptotic tar-
get(s) that might mediate p53-dependent cell death during hyp-
oxia. Using a microarray screening approach (Sax et al., 2002),
we found that Bnip3L is upregulated 2.8-fold by wild-type p53.
Bnip3L is a known apoptotic mediator (Piret et al., 2002) that
is ubiquitously expressed as 1.6 and 3.9 kb mRNA transcripts
(Yasuda et al., 1999). To confirm that Bnip3L is a p53-regulated
gene, we examined Bnip3L induction in human tumor cells with
tic agents (Figures 1A and 1B). A higher level of Bnip3L mRNA
CANCER CELL : DECEMBER 2004
A R T I C L E
Figure 2. Bnip3L is a direct target of p53
A: The human Bnip3L gene, located on chr. 8, contains six exons (green rectangles) and five introns. Putative p53 binding sites (A, B, C, E, E?, F, and F?;
oval shapes) with 85% or greater homology to the p53 binding consensus sequence are shown along with their sequences (nucleotides in uppercase
indicate homology, whereas lowercase indicates a mismatch with the consensus). Fragments D and F contain two putative p53 binding sites each (E? and
E or F1 and F2, respectively). The red color indicates the fragments (B, D, and E) that could mediate p53-dependent transactivation. The first nucleotide
is 8727 bp upstream of the first exon, based on the gene prediction method “BestRefseq” using accession number NT-023666. The relative locations for
exons and putative p53 sites are labeled.
B: P53 binds to human Bnip3L in vivo. U2OS cells were incubated with (?) or without (?) 2 ?g/ml 5-FU for 18 hr, and this was followed by a ChIP assay.
Anti-Rb antibody was used as a negative control. DR5, Puma, and Bax were used as positive controls.
C and D: p53 transactivates reporters containing Bnip3L fragments. Reporter constructs (labeled in the figure as A, B, C, D, E, or F) and the original pGL-3
plasmid vector (V) were mixed with a renilla luciferase-expressing plasmid at a ratio of 10:1 as a transfection efficiency control. Plasmids were transfected
into U2OS cells by Lipofectamine 2000 (Invitrogen). In D, reporter constructs with fragments B, D, or E or the pGL-3 vector (V) were transfected with pcDNA-
wtp53 in Saos-2 cells. The reporter construct containing fragment D was cotransfected with mtp53a (R248G) or mtp53b (R275H) in Saos-2 cells. Transfected
U2OS cells were plated in triplicate for 24 hr after transfection and treated with 2 ?g/ml 5-FU for 20 hr. The relative fold-luciferase activity was standardized
to the renilla luciferase activity. The DNA fragments B, D, and E increased reporter activity by 2-, 5-, and 3-fold in U2OS cells, and up to 6-, 10-, and 6-fold
high at 10 hr after transfection in Saos-2 cells, respectively, as compared to the basal reporter luciferase expression.
1E). Taken together, Bnip3L expression is induced by p53 fol-
lowing DNA damage in vitro and in vivo.
U2OS cells (Figure 2B). In contrast, no DNA was recovered from
precipitates of untreated cells or cells immunoprecipitated with
an anti-RB control antibody. Thus, following 5-FU exposure,
p53 localizes to human Bnip3L in vivo. To confirm that these
p53 binding DNA segments can confer p53-dependent tran-
scriptional activity, the six DNA fragments containing the above
sites (Figure 2A) were cloned individually upstream of the mini-
mal SV40 promoter in a reporter plasmid. We found that the
ure 2C) and 10-fold in Saos-2 cells (Figure 2D) cotransfected
with a wild-type p53-expressing plasmid. The D-reporter was
not induced by either of two different mutant p53-expressing
plasmids (plasmid “a” contains mutant p53 with an R248G sub-
stitution, whereas plasmid “b” contains mutant p53 with an
R275H substitution). The B- and E-reporters were also induced
Bnip3L is a direct target of p53
To gain insight into whether p53 might regulate Bnip3L directly,
we searched the NCBI database and found that human Bnip3L
contains seven putative p53 DNA binding sites with 85% or
greater identity to the p53 consensus DNA binding sequence
(El-Deiry et al., 1992) (Figure 2A). To determine whether p53
can bind to candidate p53 sites in human Bnip3L, a chromatin
immunoprecipitation (ChIP) assay was performed using lysates
from U2OS cells with or without 5-FU treatment. Precipitation
with an anti-p53 antibody of 200–300 bp DNA fragments corre-
sponding to sites B, D, or E was observed in 5-FU-treated
CANCER CELL : DECEMBER 2004599
A R T I C L E
by p53 to similar levels (over 2-fold in U2OS cells and 6-fold in
2C and 2D). The expression levels of the transfected p53 pro-
teins in Saos-2 cells and the original images of the luciferase
assays are provided in Supplemental Figure S1 at http://www.
cancercell.org/cgi/content/full/6/6/597/DC1/. These results
are consistent with the ChIP assay results (Figure 2B). Site A-
and C-reporters were weakly transactivated by p53 (Figure 2C),
possibly because they contain a spacer region between the
p53 DNA binding half-site decamers (Tokino et al., 1994). It
is possible that site E? may have contributed to induction of
luciferase activity in the D-reporter versus the E-reporter (Fig-
ures 2C and 2D). In summary, Bnip3L appears to be a target of
p53, which can be directly regulated by p53 through sequence-
specific DNA binding and transcriptional activation.
p53 targets in recruiting p53 protein to its genomic locus under
The work of Giaccia and colleagues indicated that p53 tar-
gets induced by DNA damage are not upregulated during hyp-
oxia, because p53 fails to recruit CBP or p300 to their DNA
binding sites (Koumenis et al., 2001). Here, we show that CBP
is efficiently recruited to the p53 DNA binding regions of Bnip3L
(regions B, D, and E), but not to Bax, Puma, or DR5 (Figure 4B).
This provides a mechanistic insight into the difference in the
regulation of Bnip3L versus other known p53 targets under
hypoxic conditions. Because HIF-1 is known to regulate Bnip3L
(Piret et al., 2002) and HIF-1 function relies on the coactivators
CBP and p300 (Arany et al., 1996), we hypothesized that HIF-1
may recruit CBP to the Bnip3L locus regardless of p53. We
found that five of six DNA fragments containing putative p53
DNA binding sites within Bnip3L were coimmunoprecipitated
with anti-CBP antibody under hypoxic conditions either in wild-
type U2OS-Neo or p53-deficient U2OS-E6 cells (Figure 4B).
However, not all of the fragments immunoprecipitated by the
anti-CBP antibody were bound by p53 (Figures 2B and 4A).
BecauseCBP wasrecruitedto Bnip3L inhypoxia-exposedp53-
deficient U2OS-E6 cells, the results argue that CBP recruitment
is independent of p53 but may facilitate coactivation in cells
with wild-type p53. Thus, p53 can enhance induction of Bnip3L,
in concert with HIF-1 and/or other factors under hypoxic condi-
To further investigate the kinetics and magnitude of Bnip3L
induction in wild-type versus p53-deficient cells under hypoxic
conditions, we performed a detailed time course to unravel the
relationships between Bnip3L mRNA, Bnip3L protein, and p53
protein (Figure 4C). We found that Bnip3L mRNA induction oc-
curs to higher levels in wild-type p53-expressing U2OS-Neo
cells (16-fold) as compared to the observed 5-fold induction in
the U2OS-E6 cells (Figure 4C). The induction of Bnip3L mRNA
and protein peaked at 24 hr after p53 stabilization and was
detected by 18 hr in the hypoxia-exposed U2OS-Neo cells (Fig-
ure 4C). It appears that increases in Bnip3L protein levels occur
to a much greater extent in wild-type p53-containing cells as
compared to the p53-deficient cells exposed to severe hypoxia
(Figure 4C). Thus, a clear difference in the magnitude of Bnip3L
mRNA and protein induction depends on p53 status in hypoxia-
exposed cells. Bnip3L mRNA was induced 16-fold at 24 hr in
U2OS-Neo cells but only 5-fold in the U2OS-E6 cells at 24 hr.
Despite the 5-fold increase in Bnip3L mRNA expression by 24
hr after hypoxia, no increase in Bnip3L protein expression was
observed in the U2OS-E6 cells exposed to severe hypoxia for
30 hr. This is in contrast to the 5-fold increase in Bnip3L protein
expression detected by 24 hr in hypoxia-exposed U2OS-Neo
cells (Figure 4C). Thus, one aspect of the importance of the
p53-dependent regulation of Bnip3L involves the magnitude of
Bnip3L mRNA and protein induction under conditions of hyp-
Bnip3L, unlike other proapoptotic targets of p53,
remains inducible by p53 under hypoxia
In response to hypoxia, Bnip3L and its close family member
Bnip3 have been identified as apoptotic mediators (Denko et
al., 2003; Sowter et al., 2001). Wild-type p53 accumulates in
tumor cells under hypoxic conditions (Figure 3A) and can exert
However, none of the known apoptotic targets of p53 appear
to be induced by p53 or to mediate p53-dependent apoptosis
under hypoxic conditions. Thus, we explored the possibility that
Bnip3L may be regulated by p53 during hypoxia. We found that
Bnip3L mRNA is induced to significantly higher levels during
hypoxia in wild-type p53-expressing U2OS-Neo, PA1-Neo, or
parental U2OS cells as compared to the respective p53-defi-
cient U2OS-E6, PA1-E6, or Saos-2 cells (Figures 3B–3D). Simi-
larly, Bnip3L protein is elevated to much higher levels in wild-
type p53-expressing cells as compared to p53-deficient control
cells during hypoxia (Figures 3E and 3F). To investigate the
uniqueness of Bnip3L regulation by p53 under hypoxia, expres-
sionof the Bnip3L-related Bnip3,orthe otherp53 targets Puma,
DR5, Bax, or p21WAF1wasanalyzed. Unlike Bnip3L, the induction
with or without functional p53 (Figure 3G). No transcriptional
induction of Puma, DR5, Bax, or p21 was observed under hyp-
after hypoxia (Figure 3G), which is in agreement with recent
findings that Bax and Bid are repressed under hypoxia (Erler et
transcriptionally upregulated targets of p53, is upregulated by
p53 under hypoxia. Moreover, the regulation by p53 during
hypoxia is specific to Bnip3L and not to the closely related
The contrast in induction patterns under hypoxia between
Bnip3L and other targets of p53 prompted us to determine
tin may vary among its target genes. ChIP assays revealed that
genomic regions of Bnip3L bound by p53 in response to 5-FU
treatment (Figure 2B) were the same fragments bound by p53
under hypoxia (Figure 4A). In contrast, p53 did not localize to
the p53 binding regions of Bax, Puma, or DR5 under hypoxia
(Figure 4A), although p53 was observed to bind to these genes
following exposure to genotoxic stresses (Figure 2B; Kaeser
and Iggo, 2002). Thus, Bnip3L is directly bound by p53 under
hypoxia and appears to be unique among known proapoptotic
Bnip3L mediates p53-dependent
apoptosis during hypoxia
In response to hypoxia, p53 contributes to a higher apoptotic
activity in wild-type p53-containing cells (Schmaltz et al., 1998)
(Figures 5A and 5B). Bnip3L was highly induced in cells with
wild-type p53 (Figures 3 and 4C) as compared to the E6-
expressing cells under hypoxia. The kinetics of cell death in
hypoxia-exposed U2OS-Neo and U2OS-E6 cells (Figure 5B)
CANCER CELL : DECEMBER 2004
A R T I C L E
Figure 3. Bnip3L is inducible by p53 under hypoxia
A: P53 protein level is elevated in response to severe hypoxia.
B–F: A higher Bnip3L mRNA (B–D) or protein (E and F) induction occurs during severe hypoxia in wild-type as compared to p53-deficient cells.
G: Wild-type p53 status does not predict expression levels of other proapoptotic targets, or the close family member of Bnip3L, Bnip3, under hypoxic
conditions. “?” indicates normoxia in experiments; “?” refers to 0.1% O2.
correlated well with the kinetics of Bnip3L mRNA and protein
induction (Figure 4C). To further address the relevance of the
observed greater magnitude of induction of Bnip3L when cells
contain wild-type p53, we generated Tet-inducible clones of
Bnip3L in a wild-type p53-deficient background (Figure 5C).
Apoptosis was observed in the Tet-inducible clones of Bnip3L
following Bnip3L mRNA and protein induction in a manner inde-
pendent of p53 regulation (Figure 5C). These results demon-
strate a dosage effect of Bnip3L mRNA and protein expression
induction is sufficient to induce apoptosis when conditionally
when induced by p53 under hypoxia).
We investigated the role of Bnip3L in p53-dependent cell
death under hypoxia through siRNA-mediated knockdown.
siRNA oligonucleotidesdirected against Bnip3L inPA1-neo and
U2OS-Neo cells reduced Bnip3L mRNA by 80% (Figure 5D).
We found a 60% reduction in Bnip3L protein in the transfected
ulationof Bnip3L expression (Figures5D and5E) duringhypoxia
was correlated with significant blockade of apoptotic death in
wild-type p53-containing PA1-Neo and U2OS-Neo cells (Fig-
CANCER CELL : DECEMBER 2004601
A R T I C L E
Figure 4. Bnip3L remains as a direct target of p53 under hypoxia
A: P53 binds to the human Bnip3L locus but not to Bax, Puma, or DR5 during hypoxia in vivo. A ChIP assay was performed with anti-Rb antibody as a
negative control. For 22 hr, 0.1% O2was used as the hypoxia condition.
B: CBP is recruited to the p53 binding region of Bnip3L but not to the p53 binding regions of Bax, Puma, or DR5. A ChIP assay was performed with anti-Rb
antibody as a negative control.
C: Kinetics and magnitude of induction of Bnip3L mRNA, Bnip3L protein, and p53 protein under severe hypoxia in U2OS cells. The quantitative information
for Bnip3L mRNA expression, analyzed using a phosphorimager, is provided below the blots.
ures 5F and 5G, upper panels). In contrast, the p53-deficient
E6-expressing cell lines were not affected (PA1-E6; Figure 5F)
or became only slightly more resistant to cell death after Bnip3L
knockdown (U2OS-E6; Figure 5G). As additional evidence for
these effects, we tested different Bnip3L RNAi sequences (Sup-
plemental Figure S2 at http://www.cancercell.org/cgi/content/
full/6/6/597/DC1/). Bnip3L knockdown protected PA1-Neo and
U2OS-Neo cells from the apoptotic effects of severe hypoxia
(Supplemental Figure S2). The results in Figure 5 and Supple-
mental Figure S2 suggest that Bnip3L contributes to apoptotic
death in wild-type p53-expressing cells exposed to hypoxia.
The induction of Bnip3L by hypoxia in E6-expressing cells ap-
pears to have little effect in inducing cell death as noted by
the lack of substantial change in percent of cells undergoing
apoptosis following Bnip3L knockdown. The kinetics of the ob-
served death and lack of significant effect following Bnip3L
knockdown correlate well with the observation that Bnip3L pro-
tein expression in E6-expressing cells was not detectably ele-
vated under the conditions studied here (Figure 4C). These re-
sults support the conclusion that Bnip3L appears to mediate in
part p53-dependent apoptosis under hypoxia. These studies
growth advantage to tumors under hypoxic conditions.
Knockdown of Bnip3L promotes human tumor
xenograft growth in vivo
opment (Hammond et al., 2002; Vaupel et al., 1991). Wild-type
p53-expressing tumor cells, which possess a higher apoptotic
et al., 1996). The high apoptotic activity of wild-type p53-
expressing cells under hypoxia correlates well with the high
frequency of p53 mutations in human tumors, which provide a
selective growth advantage under hypoxic stress during tumor
evolution and progression. Because Bnip3L is a proapoptotic
target of p53 functioning under hypoxic conditions, a role in
suppressing tumor formation may provide a mechanism by
which p53 mediates tumor suppression, particularly under hyp-
oxic conditions. Thus, in order to elucidate the importance of
Bnip3L in tumor formation, we generated stable cell lines U-11
andU-17 (derivedfromU2OS)that constitutivelyexpresssiRNA
directed against human Bnip3L, or pooled U2OS cells infected
with Bnip3L RNAi-expressing retrovirus (U2OS p3LKD cells)
(Supplemental Figure S3A at http://www.cancercell.org/cgi/
content/full/6/6/597/DC1/). At 10 days after injection of 10 mil-
cells U2OS-s; control and Bnip3L knockdown injections on op-
CANCER CELL : DECEMBER 2004
A R T I C L E
Figure 5. Bnip3L is a mediator of p53-dependent cell death under hypoxia
A and B: Higher apoptotic activity is observed in wild-type p53-expressing PA1-Neo and U2OS-Neo cells as compared to PA1-E6 and U2OS-E6 cells exposed
to severe hypoxia.
C: Induction of exogenous Bnip3L mRNA and protein expression promotes apoptosis in mutant p53-expressing SW480 tumor cells following use of the Tet-on
inducible system to restore regulated expression of Bnip3L. The level of Bnip3L mRNA progressively increased according to the dose of doxycycline. Bnip3L
protein was clearly increased at the highest dose of doxycycline (see 54 hr time point). The percentage of cell death correlated with the increased level
of Bnip3L protein. No increase in cell death was observed in parental SW480 cells treated with 1 ?g/ml doxycycline for 54 hr (data not shown). The
quantitative information for Bnip3L mRNA expression is provided below the blots.
D and E: Both levels of Bnip3L mRNA (D) and protein (E) were analyzed to determine the efficiency of RNAi by real-time quantitative RT-PCR (D) or Western
blotting under hypoxia (E). A similar result to that in Figure 5E was noted in Western blotting the RNAi-transfected PA1 cells (data not shown).
F and G: Knockdown of Bnip3L diminishes the apoptotic death in both PA1-Neo and U2OS-Neo cells. “3Li” refers to Bnip3L RNAi-transfected cells.
posite flanks as indicated in Supplemental Figure S3B), the U-
11, U-17, and pooled U2OS p3LKD xenografted cells formed
tumors. In contrast, the U2OS-s xenograft disappeared by 5
days after injection, consistent with prior reports that U2OS
cells do not form tumors in nude mice (American Type Culture
a constitutively expressing firefly luciferase gene cloned in a
retrovirus vector was introduced into the U2OS-s, U-11, U-17,
and p3LKD cells to monitor the dynamic course of tumor forma-
luciferase-expressing U-11, U-17, or p3LKD and U2OS-s, were
roughly equal, with perhaps more signal from the parental
U2OS-s injection sites at the beginning of the experiment (day
0). On subsequent days, the bioluminescence emitted from the
U2OS-s cells decreased as compared to the U-11, U-17, or
p3LKD xenografts, and gradually the control bioluminescence
from the U2OS-s xenografts disappeared (compare days 0, 3,
and 6 in Figure 6). In contrast, the bioluminescence emitted
by the U-11, U-17, or p3LKD xenografts initially increased in
intensity and persisted through day 21 (Figure 6) and subse-
quently decreased (data not shown). A higher intensity of biolu-
minescence was also observed in Bnip3L knockdown PA1 xe-
nografts (Supplemental Figure S3C) as compared to the control
PA1 xenograft (PA1-s) over time, and a slight growth advantage
was also observed in the mutant p53-expressing SW480 xeno-
grafts with Bnip3L knockdown (Supplemental Figure S4). These
data suggest that Bnip3L suppresses tumor growth as a tumor
suppressor, particularly in tumors that contain wild-type p53.
We speculate that the ultimate reduction of bioluminescence in
the Bnip3L-silenced U2OS derivatives (data not shown beyond
21 days) may reflect the action of other p53 targets and/or
other death inducing protein(s), which may contribute to tumor
suppression. Moreover, host immunity resulting from NK cell
activity and/or other factors may contribute to the elimination
of U2OS-derivative xenografts and the eventual suppression of
the Bnip3L-silenced tumors.
Bnip3L is a mediator of p53-dependent radiosensitivity
Wild-type p53 plays a critical role in radiosensitivity (Fei et al.,
2002; Fei and El-Deiry, 2003). Tumors harboring mutations in
p53 generally carry a poor prognosis following radiotherapy, as
has been observed, for example, in breast cancer (Marchetti et
al., 2003). However, the efficacy of ionizing radiation directly
relies on adequate oxygen tensions (Weinmann et al., 2003).
To probe a potential role of Bnip3L in p53-dependent radiosen-
sitivity, mice were irradiated with 5 Gy on day 4 after injection
CANCER CELL : DECEMBER 2004603
A R T I C L E
Figure 6. Knockdown of Bnip3L promotes tumorigenicity
U-11, U-17, pooled U2OS 3L knockdown (p3LKD; Bnip3L silencing was mediated through retrovirus infection of RNAi-expressing cDNA), or U2OS-s cells were
infected with firefly luciferase-expressing retroviruses. Ten million cells of each were injected into mice. On day 0, xenografts of U2OS-s/U-11 (upper), U2OS-
s/U-17 (middle), and U2OSs/p3LKD (bottom) emitted similar levels of bioluminescence, or there was a slightly higher emission from the U2OS-s xenografts.
Bioluminescence was imaged over time and quantified. The bioluminescence from U2OS-s xenografts disappeared by day 6, whereas U-11, U-17, and
p3LKD xenografts continued to emit light detectable up to day 21 shown here. A typical mouse image (the same mouse is shown over the time course)
is presented from a total of four mice for U-11/U-17 xenografts or from a total of eight mice for p3LKD xenografts. Quantification with error bars of the
photon counts per unit time per cm2from the wild-type (U2OS-s) as compared to the Bnip3L-silenced tumor xenografts appears to the right.
of 10 million U2OS-s versus either U-11 or p3LKD cells (Figure
7A and Supplemental Figure S5A at http://www.cancercell.org/
cgi/content/full/6/6/597/DC1/). On day 4, the U2OS-s xeno-
grafts were still alive and were excised at 6 hr after irradiation and
compared with the nonirradiated controls (Figure 7A and Supple-
mental Figure S5). Active caspase-3 appeared to be significantly
increased in irradiated U2OS-s xenografts as compared to the
nonirradiated U2OS-s and either the irradiated or untreated U-11
or p3LKD xenografts (Figures 7B and 7C; Supplemental Figures
S5B and S5C). There was more active caspase-3 in irradiated
U-11 or p3LKD versus unirradiated U-11/p3LKD xenografts, re-
spectively, consistent with the hypothesis that Bnip3L may be
irradiation. Nonetheless, silencing of Bnip3L appears to reduce
apoptotic activity initiated in irradiated wild-type p53-containing
xenografts, suggesting an important role for Bnip3L among the
apoptotic targets of p53.
(Figure 7D and Supplemental Figure S5D at http://www.
for the Ki67 antigen confirmed the proliferative advantage in the
Figure7D andSupplementalFigure S5D),andstaining foractive
caspase 3 confirmed a lower level of apoptosis in U-11 versus
U2OS-s and p3LKD versus U2OS-s (Figure 7D and Supplemen-
(Figure 7D) or pimonidazole adducts (Supplemental Figure S5D)
showed comparable staining between U2OS-s, U-11, or U2OS
p3LKD tumors, suggesting similar degrees of hypoxia. These
results suggest that silencing of Bnip3L does not reduce the
hypoxic gene induction response leading to increased Glut1
expression; rather, despite the hypoxia, there is reduced apo-
ptosis and increased proliferation in the U-11 or p3LKD versus
U2OS-s tumors. One interpretation is that silencing of Bnip3L
results in reduced sensitivity to hypoxia and increased cell sur-
vival that is permissive for tumor proliferation and the observed
growth advantage of U-11 or p3LKD tumors under hypoxia.
Decreased sensitivity to hypoxia in vivo contributes
to tumor growth following Bnip3L
silencing in U2OS xenografts
We further examined the U2OS-s control, U-11, and pooled
U2OS (p3LKD) tumors with Bnip3L silencing to gain insights
into whether the growth advantage and greater proliferation in
U-11 or p3LKD may be due to reduced sensitivity to hypoxia.
U2OS-s, U-11, or p3LKD xenografts were excised from tumor-
bearing mice on day 4,prior to regression of the U2OS-s tumors
A high frequency of p53 mutations is associated with a growth
advantage in developing tumors (Vogelstein et al., 2000). Nu-
and Jacks, 1999; Aurelio et al., 2000; Bardeesy et al., 1995;
Eischen et al., 2001; Meijerink et al., 1998; Schmitt et al., 2002;
CANCER CELL : DECEMBER 2004
A R T I C L E
Figure 7. Knockdown of Bnip3L is accompanied by reduced apoptosis in irradiated or unirradiated xenografts, and the reduced sensitivity to hypoxia-
mediated cell death is accompanied by increased proliferation of tumors
A–C: Two mice were injected with 10 million cells of U2OS-s or U-11 cells. On day 4 (images in A), mice were either untreated or treated with 5 Gy ionizing
radiation, and the xenografts were excised 6 hr later. The image shown was obtained from one of the two mice (the same mouse) at the two different
time points shown. The active caspase-3 signal (brown color in cytoplasm) (B) was highest in the irradiated (?IR) U2OS-s xenograft, moderate in the
nonirradiated (?IR) U2OS-s xenograft, low in the ?IR U-11 xenograft, and lowest in the unirradiated U-11 xenograft. The relative amount of active caspase-3
expression is shown (C) by using NIH ImageJ software.
D: Day 4 xenografts were harvested, and the nonirradiated U2OS-s or U-11 xenografts revealed similar levels of Glut1. In Bnip3L-silenced xenografts, active
caspase-3 levels were lower, whereas Ki67 expression was higher as compared to the control xenograft.
Soengas et al., 2001) that the apoptotic effect of p53 plays a
critical role in tumor suppression. To date, several proapoptotic
targets of p53 have been suggested to mediate p53-dependent
apoptosis following exposure to genotoxic stresses. However,
none of the known p53-activated proapoptotic genes has been
shown to mediate p53-dependent apoptosis in response to
hypoxia, as it occurs during tumor development and progres-
sion. Here, we demonstrate that Bnip3L is a recognized target
of p53 and is a strong candidate mediator of p53-dependent
tumorigenicity such that silencing of Bnip3L allows conversion
of nontumorigenic U2OS-s cells to tumorigenic xenografts. Si-
lencing Bnip3L reduces sensitivity to apoptosis under hypoxic
conditions, thereby permitting a higher degree of proliferation
(Figure 7D and Supplemental Figure S6A). These studies sug-
gest that loss of Bnip3L regulation by p53 under hypoxia may
be a pivotal event in tumor development, which contributes to
(Graeber et al., 1996) grow into tumors. Taken together, the
work shown here provides a mechanism regarding how p53
tions. Moreover, this work provides a mechanistic insight into
the selection pressure that occurs under hypoxia and an apo-
ptotic mediator that needs to be suppressed to permit tumor
P53 induces Bnip3L under hypoxia
Bnip3L, unlike its family member Bnip3, not only responds to
HIF-1 (Sowter et al., 2001; Bruick, 2000) but also responds to
wild-type p53 (Figures 1–3). Other known proapoptotic targets
of p53 are either uninduced (Koumenis et al., 2001), repressed
(Figure 3G; Erler et al., 2004), or induced without a preference
to p53 status in hypoxic cells (Kim et al., 2004). To date, Bnip3L
is the only apoptotic target of p53 identified that appears to be
upregulated by wild-type p53 in response to hypoxia. Here, we
show that p53 localizes to the p53 DNA binding sites of Bnip3L
but not to Bax, Puma, or DR5 during hypoxia (Figure 4A), al-
though these targets can be induced and p53 localizes to their
genomic loci following DNA damage. A failure of p53 to trans-
activate its known targets under hypoxia has been previously
reported and was attributed to a lack of recruitment of CBP or
the lack of induction of Bax, Puma, or DR5 during hypoxic
exposure of wild-type p53-expressing human tumor cell lines
(Figure 3G). We confirmed that CBP is not recruited to the
p53 DNA binding regions of Bax, Puma, or DR5 (Figure 4B)
consistent with the Koumenis et al. (2001) model. However,
CBP was efficiently localized to the p53 DNA binding regions
of Bnip3L (Figure 4B). Two regions that bound CBP, including
p53 DNA binding site-containing fragments A and C, were not
shown to be bound by p53 (Figures 2B and 4A), which suggests
CANCER CELL : DECEMBER 2004605
A R T I C L E
of its location near the first coding exon of Bnip3L (Figure 2A),
its abilityto bindp53 (Figure 2B)including underhypoxia (Figure
4A), and its potential to mediate p53-dependent transactivation
of promoter-reporter plasmids (Figures 2C and 2D). Moreover,
a candidate HIF site (Supplemental Figure S6B at http://www.
base pairs of the p53 site B. Thus, it is possible that CBP is
recruited to the genomic region of site B through HIF, and then
p53 can be recruited to the region through its DNA binding
response element. We note that the mouse Bnip3L locus con-
tains candidate p53 binding sites, with 80% homology to the
consensus sequence, both upstream of the first coding exon
and downstream of the last coding exon (Supplemental Figure
S6C). We also note that there are two candidate p53 binding
sites with 75% homology to the p53 consensus in intron 1 of
mouse Bnip3L (Supplemental Figure S6C) and that there is a
75% homology site in intron 1 of human Bnip3L (data not
shown). Thus, regulation of Bnip3L by p53 appears to be con-
Figure S6C) and evidence for p53-dependent Bnip3L induction
will examine the contribution of specific p53 sites to Bnip3L
regulation under various conditions including DNA damage and
hypoxia, with detailed analysis of the spatial and temporal inter-
actions between HIF-1, CBP, and p53 proteins at the genomic
regulatory region of Bnip3L. To date, an analysis documenting
the importance of any specific genomic DNA response element
for p53 has not been performed for any gene, including those
that contain multiple p53 sites. Such studies on the mecha-
nism(s) by which p53, CBP, and HIF coordinately regulate
Bnip3L may provide a greater understanding of Bnip3L regula-
tion during p53-dependent tumor suppression under hypoxia
and may help to find ways to maintain or restore apoptotic
target gene expression and prevent tumor progression.
Figure 8. Model of Bnip3L regulation by p53 under hypoxia
A: Under normoxia, p53 recruits coactivators (Co-a) CBP or p300 to its con-
sensus binding sequence (CBS) and transactivates its target genes in re-
sponse to DNA damage. During hypoxia, p53 fails (B) to recruit coactivators
to its binding sequence and does not transactivate its known proapoptotic
targets or p21. Bnip3L, through HIF-1 (C), recruits coactivator CBP close to
p53 binding regions under hypoxia. p53 protein interacts with CBP while
interacting with its binding sites. A cooperative interaction involving HIF-1,
wild-type p53. In such cells, Bnip3L appears to be a major mediator of
apoptosis when the cells are exposed to severe hypoxia.
that other cis elements within Bnip3L may play an essential role
in the recruitment of CBP to Bnip3L under hypoxia. We propose
a model (Figure 8) aiming to explain how p53 may regulate
Bnip3L under hypoxia. Under hypoxic conditions, p53 fails to
interact with CBP (Koumenis et al., 2001) and interacts with
its p53 DNA binding sites weakly and transiently (Figure 8B).
However, p53 may interact with CBP and bind firmly to putative
p53 DNA binding sites when the local level of CBP is stimulated
by HIF-1, which primes the putative p53 DNA binding sites in
the vicinity of HIF-1 response elements (HRE). We speculate
thatCBP promotesp53recruitment byaltering chromatinstruc-
ture to allow access to DNA binding sites for p53 or by physical
proximity to allow interaction between CBP and p53 (Figure
8C), and this will not occur in p53-deficient cells, although CBP
may be present (Figure 4B). In our studies, p53 protein is absent
in E6-expressing cells under hypoxia, although HIF-1 protein is
present (Supplemental Figure S6A at http://www.cancercell.
org/cgi/content/full/6/6/597/DC1/). Thus, in cells lacking p53,
our model predicts a failure to optimally induce Bnip3L to suffi-
cient levels, thereby leading toresistance to the apoptotic effect
of severe hypoxia.
We note that the genomic locus of Bnip3L contains p53
DNA binding sites upstream of the first coding exon and down-
stream of the last coding exon (Figure 2A). It is not uncommon
for p53-regulated genes to contain multiple p53 response ele-
ments or to have sites located within downstream introns. It is
not clear if sites downstream of coding exons are significantly
different in function from sites located within introns in terms
taining site B is a good candidate mediator of the observed
p53-dependent Bnip3L transactivation under hypoxia because
Bnip3L is an apoptotic mediator
of p53 under hypoxic conditions
We have shown a clear difference in the magnitude of Bnip3L
mRNA and protein induction depending on p53 status in hyp-
oxia-exposed cells (Figure 4C). p53 protein stabilization oc-
curred before the induction of Bnip3L mRNA and protein, and
a much greater extent of induction of Bnip3L was found in wild-
type p53-containing cells as compared to p53-deficient cells
following exposure to hypoxia. Thus, p53 appears to play a
critical role in the magnitude of Bnip3L induction under hypoxia.
We believe that the magnitude of Bnip3L induction in wild-
type p53-expressing hypoxic cells is a major determinant of
subsequent cell death. To further explore the possible impor-
tance of the greater magnitude of induction of Bnip3L in wild-
type p53-containing cells exposed to hypoxia, we established
Tet-inducible clones of Bnip3L in mutant p53-expressing
SW480 cells. This design was intended to (1) uncouple p53
regulation from the level of Bnip3L induction and (2) determine
whether Bnip3L induction to high levels (not normally achieved
in the absence of wild-type p53) would be sufficient to induce
apoptosis in p53-deficient cells. The Tet-inducible Bnip3L
SW480 cells demonstrated that there is a dosage effect of
Bnip3L mRNA and protein expression with regard to the ob-
in a p53-deficient background required high levels of Bnip3L
CANCER CELL : DECEMBER 2004
A R T I C L E
mRNA and protein, and correlated well with what we described
regarding Bnip3L regulation under hypoxia in wild-type versus
tionshowedonlymodestincreasesin Bnip3L mRNAexpression
that we believe may not result in efficient cell death induction
under hypoxia (Figure 4C). Moreover, the studies using Tet-
inducible Bnip3L in p53-deficient SW480 cells confirm that
Bnip3L is sufficient to induce apoptosis when conditionally ex-
pressed and therefore may be sufficient when induced by p53
under hypoxia. This is important because p53 may induce other
targets that may be involved in cell death under hypoxia, but
induction of Bnip3L is sufficient once a certain level of induction
To further support ourconclusion that p53-dependent regu-
lation of Bnip3L under hypoxia is relevant to the observed apo-
ptosis and tumor suppression, we investigated the effects of
Bnip3L silencing in p53-deficient cells. The p53-deficient cells
were already more resistant to hypoxia-induced apoptosis (Fig-
ures 5A and 5B), and the magnitude of Bnip3L mRNA and
protein induction by hypoxia was low (Figure 4C). Silencing of
Bnip3L in p53-deficient U2OS-E6 cells only slightly reduced the
observed apoptosis following exposure to hypoxia (Figure 5G)
as compared to U2OS-Neo cells, where a dramatic protection
was observed (Figure 5G). Taken together with the results of
reinforce two conclusions: (1) p53-deficient human tumor cells
die inefficiently when exposed to severe hypoxia, despite the
fact that Bnip3L and other hypoxia-inducible targets are in-
duced, and (2) blockade of Bnip3L has much more significant
ing cells because of its greater magnitude of induction. We
believe the greater magnitude of Bnip3L induction represents
conclude that hypoxic death is inefficient in p53-deficient cells,
possibly because sufficient levels of Bnip3L are not achieved
(Figure 4C). When such levels are artificially created in cells
lacking wild-type p53, cell death is observed (Figure 5C). We
containing cells results from Bnip3L regulation by p53, but we
cannot rule out other mediators of p53, direct effects of p53,
and/or other factors that may also contribute to the cell death
triggered by hypoxia.
Bnip3L appears to play a crucial role in tumor suppression
as documented by our observations that silencing of Bnip3L
lowers sensitivity to cell death induced by hypoxia (Figure 7D
content/full/6/6/597/DC1/) in vivo. Moreover, Bnip3L knock-
down converts nontumorigenic U2OS cells into tumorigenic xe-
nografts. This was demonstrated by two approaches: (1) stable
clonal cells and (2) retrovirus-infected pooled cells to eliminate
potential artifacts due to selection during cloning procedures
(Figure 6). These results are consistent with the concept that a
lower apoptotic potential promotes tumor growth (Graeber et
al., 1996), and this may be compromised when p53 is mutated.
silencing occurred in U2OS xenografts, despite the fact that
these cells contain wild-type p53 and maintain a modest apo-
ptotic response to ionizing irradiation (Figure 7B and Supple-
mental Figure 5B). We propose that p53-dependent upregula-
tion of Bnip3L during hypoxia provides a mechanism by which
the p53 tumor suppressor functions in vivo in inhibiting tumor
Altered Bnip3L regulation by p53 in cancer
Deficient Bnip3L upregulation may occur in several ways in
tumors. These include the loss of functional p53 through p53
mutation, or loss of Bnip3L function. Some mutations of Bnip3L
have been reported in breast cancer and ovarian cancer; its
expression is low or absent in both lung cancer cell lines and
human lung cancers (Lai et al., 2003; Sun et al., 2004). This
is consistent with the present demonstration that silencing of
Bnip3L promotes tumorigenicity (Figure 6 and Supplemental
Figures S3 and S4 at http://www.cancercell.org/cgi/content/
full/6/6/597/DC1/). In addition, there is a possibility that Bnip3L
regulation by p53 could be impaired during tumor formation
even in some tumors that retain wild-type p53, possibly through
changes in selectivity of target gene activation that promote
tumor growth. For example, we observed no Bnip3L upregula-
tion by p53 in HCT116 or H460 tumor cell lines following expo-
sure to DNA damaging agents or hypoxia (data not shown). A
recent report demonstrated, among other changes, downregu-
lation of Bnip3L in HCT116 cells selected for resistance to 5-FU
(De Angelis et al., 2004). Another study reported hypermethyla-
tion of Bnip3 in pancreatic cancer (Okami et al., 2004). Thus, a
number of genetic or epigenetic changes, which may occur
tion by p53 when both are present, thereby contributing to
decreased sensitivity to hypoxia and tumor progression.
Bnip3L and Bnip3 are potent apoptotic mediators during
hypoxia (Kubasiak et al., 2002; Piret et al., 2002; Yussman et
al., 2002); however, Bnip3 is not upregulated by p53. Interest-
ingly, we found that the kinetics of Bnip3 mRNA and protein
induction were slower than those of Bnip3L mRNA and protein
induction in hypoxia-exposed SW480 cells (Supplemental
Figure S4C at http://www.cancercell.org/cgi/content/full/6/6/
597/DC1/), and in vivo knockdown of Bnip3 in the SW480 cells
had no apparent effect on xenograft growth (Supplemental Fig-
ures S4A and S4B). Taken together, without upregulation by
p53, Bnip3L or Bnip3 may play a role in hypoxia-triggered cell
death with a suggestion that Bnip3L may be induced earlier
and contribute slightly more to the cell death. However, these
effects are weak in terms of their tumor suppressive potential
in cells lacking wild-type p53, and such p53-deficient human
cells appear to have little difficulty in surviving and forming
tumors in vivo, regardless of the silencing of either Bnip3 or
Bnip3L. Bnip3L appears to play a crucial role in tumor suppres-
sion that is compromised when p53 is mutated or Bnip3L is
silenced, and its induction provides a mechanism by which p53
functions in vivo in inhibiting tumor growth. In tumor progres-
sion, cells with lower expression of Bnip3L have a lower apo-
ptotic potential, and this may facilitate their growth and expan-
sion despite hypoxia (Figures 6 and Supplemental Figures S3
and S4). In tumor therapy, a lower expression of Bnip3L is
expected to contribute to radioresistance (Figure 7B and Sup-
plemental Figure S5B). We suggest that Bnip3L regulation by
p53 may contribute to hypoxia-induced cell death and tumor
ment therefore offers a strategy for promoting increased sensi-
tivity to hypoxia and increased apoptosis in response to therapy
of hypoxic tumors.
CANCER CELL : DECEMBER 2004607
A R T I C L E
nase gene using a transcriptional targeting approach. Cancer Gene Ther. 7,
Reporter plasmids, transfection, and reporter imaging assays
Six genomic DNA fragments containing sites A, B, C, D, E, and F (in Figure
2A) were cloned into the pGL3 promoter-reporter (Promega). Transfection
was performed using Lipofectamine 2000 (Invitrogen). The bioluminescent
images were obtained using a Xenogen IVIS as previously described (Wang
and El-Deiry, 2003, 2004).
Arany, Z., Huang, L.E., Eckner, R., Bhattacharya, S., Jiang, C., Goldberg,
M.A., Bunn, H.F., and Livingston, D.M. (1996). An essential role for p300/
CBP in the cellular response to hypoxia. Proc. Natl. Acad. Sci. USA 93,
Attardi, L.D., and Jacks, T. (1999). The role of p53 in tumour suppression:
lessons from mouse models. Cell. Mol. Life Sci. 55, 48–63.
Bnip3L siRNA oligonucleotide transfection
siRNA oligonucleotidesdirected against human Bnip3L included 5?-AA-CAG
UUCCUGGGUGGAGCUA-3? for Supplemental Figure S3 (http://www.
cancercell.org/cgi/content/full/6/6/597/DC1/) and 5?-AA-CACGUACCAUCC
UCAUCCU-3? for Figure 4, and scrambled oligonucleotides were purchased
from Dharmacon Research (Lafayette, CO). The transfection was performed
as suggested by Dharmacon.
have selective dominant-negative effects on apoptosis but not growth arrest
in human cancer cell lines. Mol. Cell. Biol. 20, 770–778.
Bahnson, A.B., Dunigan, J.T., Baysal, B.E., Mohney, T., Atchison, R.W.,
Nimgaonkar, M.T., Ball, E.D., and Barranger, J.A. (1995). Centrifugal en-
hancement of retroviral mediated gene transfer. J. Virol. Methods 54,
Bardeesy, N., Beckwith, J.B., and Pelletier, J. (1995). Clonal expansion and
attenuated apoptosis in Wilms’ tumors are associated with p53 gene muta-
tions. Cancer Res. 55, 215–219.
Nude mice (4–6 weeks old) from Charles River laboratory were injected with
300 ?l Matrigel (Invitrogen) flurry (prepared at a 1:1 ratio with 1? PBS)
containing10million cellswithorwithout constitutivelyexpressedluciferase.
The luciferase activity was measured by intraperitoneal injection of 50 mg/
mouse D-luciferin into anesthetized mice, followed by detection of live im-
ages using the Xenogen IVIS as described (Wang and El-Deiry, 2003, 2004).
Bruick, R.K. (2000). Expression of the gene encoding the proapoptotic Nip3
protein is induced by hypoxia. Proc. Natl. Acad. Sci. USA 97, 9082–9087.
Brummelkamp, T.R., Bernards, R., and Agami, R. (2002). A system for stable
expression of short interfering RNAs in mammalian cells. Science 296,
Establishment of stable siRNA-expressing cell lines
and luciferase-expressing retrovirus
mental Data at http://www.cancercell.org/cgi/content/full/6/6/597/DC1/)
was cloned into pSuper or Super-retro plasmid, and firefly luciferase cDNA
was cloned into pBabe-puro plasmid. The retroviruses made from Phoenix
cells werespinoculated into cellsas described(Bunnell et al.,1995; Bahnson
et al., 1995; Burns and El-Deiry, 2001). The cell line selection was performed
as described previously (Brummelkamp et al., 2002).
Bunnell, B.A., Muul, L.M., Donahue, R.E., Blaese, R.M., and Morgan, R.A.
(1995). High-efficiency retroviral-mediated gene transfer into human and
nonhuman primate peripheral blood lymphocytes. Proc. Natl. Acad. Sci.
USA 92, 7739–7743.
Burns, T.F., and El-Deiry, W.S. (2001). Identification of inhibitors of TRAIL-
induceddeath (ITIDs)inthe TRAIL-sensitivecoloncarcinomacell lineSW480
using a genetic approach. J. Biol. Chem. 276, 37879–37886.
Carmeliet, P., Dor, Y., Herbert, J.M., Fukumura, D., Brusselmans, K., Dewer-
chin, M., Neeman, M., Bono, F., Abramovitch, R., Maxwell, P., et al. (1998).
Role of HIF-1? in hypoxia-mediated apoptosis, cell proliferation and tumour
angiogenesis. Nature 394, 485–490.
Inducible Bnip3L expression using the Tet-on system
These procedures were performed according to the protocol provided by
the manufacturer of the Tet-on system (Clontech). The Bnip3L cDNA was
subcloned into the pRevTRE vector at the BamH1 and the HindIII sites.
Chan, T.A., Hwang, P.M., Hermeking, H., Kinzler, K.W., and Vogelstein, B.
(2000). Cooperative effects of genes controlling the G(2)/M checkpoint.
Genes Dev. 14, 1584–1588.
A more detailed description of the Experimental Procedures is provided
in the Supplemental Data at http://www.cancercell.org/cgi/content/full/6/6/
De Angelis, P.M., Fjell, B., Kravik, K.L., Huag, T., Tunheim, S.H., Reichelt,
W., Beigi, M., Clausen, O.P., Galteland, E., and Stokke, T. (2004). Molecular
characterizations of derivatives of HCT116 colorectal cancer cells that are
resistant to the chemotherapeutic agent 5-fluorouracil. Int. J. Oncol. 24,
Denko, N.C., Fontana, L.A., Hudson, K.M., Sutphin, P.D., Raychaudhuri, S.,
Altman, R., and Giaccia, A.J. (2003). Investigating hypoxic tumor physiology
through gene expression patterns. Oncogene 22, 5907–5914.
P.F. received the Scholar In Training Award at the 95th AACR meeting in
Orlando, FL (March 2004). This work was supported by NIH Grants
(CA75454, CA75138, and CA105008) and by funds from the Howard Hughes
Medical Institute to W.S.E.-D.
Eischen, C.M., Roussel, M.F., Korsmeyer, S.J., and Cleveland, J.L. (2001).
Bax loss impairs Myc-induced apoptosis and circumvents the selection of
p53 mutations during Myc-mediated lymphomagenesis. Mol. Cell. Biol. 21,
Received: February 14, 2004
Revised: June 6, 2004
Accepted: October 21, 2004
Published: December 20, 2004
El-Deiry, W.S. (1998). Regulation of p53 downstream genes. Semin. Cancer
Biol. 8, 345–357.
El-Deiry, W.S., Kern, S.E., Pietenpol, J.A., Kinzler, K.W., and Vogelstein, B.
(1992). Definition of a consensus binding site for p53. Nat. Genet. 1, 45–49.
Erler, J.T., Cawthorne, C.J., Williams, K.J., Koritzinsky, M., Wouters, B.G.,
Wilson, C., Miller, C., Demonacos, C., Stratford, I.J., and Dive, C. (2004).
Hypoxia-mediated down-regulation of Bid and Bax in tumors occurs via
hypoxia-inducible factor 1-dependent and -independent mechanisms and
contributes to drug resistance. Mol. Cell. Biol. 24, 2875–2889.
Achison, M., and Hupp, T.R. (2003). Hypoxia attenuates the p53 response
to cellular damage. Oncogene 22, 3431–3440.
An, W.G., Kanekal, M., Simon, M.C., Maltepe, E., Blagosklonny, M.V., and
Neckers, L.M. (1998). Stabilization of wild-type p53 by hypoxia-inducible
factor 1?. Nature 392, 405–408.
Fei, P., and El-Deiry, W.S. (2003). P53 and radiation responses. Oncogene
Anderson, L.M., Krotz, S., Weitzman, S.A., and Thimmapaya, B. (2000).
Breast cancer-specific expression of the Candida albicans cytosine deami-
Fei, P., Bernhard, E.J., and El-Deiry, W.S. (2002). Tissue-specific induction
of p53 targets in vivo. Cancer Res. 62, 7316–7327.
CANCER CELL : DECEMBER 2004
A R T I C L E
Graeber, T.G., Peterson, J.F., Tsai, M., Monica, K., Fornace, A.J., Jr., and
Giaccia, A.J. (1994). Hypoxia induces accumulation of p53 protein, but acti-
vation of a G1-phase checkpoint by low-oxygen conditions is independent
of p53 status. Mol. Cell. Biol. 14, 6264–6277.
Piret, J.P., Mottet, D., Raes, M., and Michiels, C. (2002). Is HIF-1? a pro-
or an anti-apoptotic protein? Biochem. Pharmacol. 64, 889–892.
Sax, J.K., Fei, P., Murphy, M.E., Bernhard, E., Korsmeyer, S.J., and El-Deiry,
W.S. (2002). BID regulation by p53 contributes to chemosensitivity. Nat. Cell
Biol. 4, 842–849.
Graeber, T.G., Osmanian, C., Jacks, T., Housman, D.E., Koch, C.J., Lowe,
S.W., and Giaccia, A.J. (1996). Hypoxia-mediated selection of cells with
diminished apoptotic potential in solid tumours. Nature 379, 88–91.
Schmaltz, C., Hardenbergh, P.H., Wells, A., and Fisher, D.E. (1998). Regula-
tion of proliferation-survival decisions during tumor cell hypoxia. Mol. Cell.
Biol. 18, 2845–2854.
Hammond, E.M., Denko, N.C., Dorie, M.J., Abraham, R.T., and Giaccia, A.J.
(2002). Hypoxia links ATR and p53 through replication arrest. Mol. Cell. Biol.
Schmitt, C.A., Fridman, J.S., Yang, M., Baranov, E., Hoffman, R.M., and
Lowe, S.W. (2002). Dissecting p53 tumor suppressor functions in vivo. Can-
cer Cell 1, 289–298.
Hansson, L.O., Friedler, A., Freund, S., Rudiger, S., and Fersht, A.R. (2002).
Two sequence motifs from HIF-1? bind to the DNA-binding site of p53.
Proc. Natl. Acad. Sci. USA 99, 10305–10309.
Semenza, G.L. (2000). HIF-1 and human disease: one highly involved factor.
Genes Dev. 14, 1983–1991.
Jeffers,J.R., Parganas,E.,Lee,Y., Yang,C.,Wang,L., Brennan,J.,Maclean,
K.H., Han, J., Chittenden, T., Ihle, J., et al. (2003). Puma is an essential
mediator of p53-dependent and -independent apoptotic pathways. Cancer
Cell 4, 321–328.
Shimizu, S., Eguchi, Y., Kamiike, W., Itoh, Y., Hasegawa, J., Yamabe, K.,
Otsuki, Y., Matsuda, H., and Tsujimoto, Y. (1996). Induction of apoptosis as
well as necrosis by hypoxia and predominant prevention of apoptosis by
Bcl-2 and Bcl-XL. Cancer Res. 56, 2161–2166.
fails to support the latency model for regulation of p53 DNA binding activity
in vivo. Proc. Natl. Acad. Sci. USA 99, 95–100.
Soengas, M.S., Capodieci, P., Polsky, D., Mora, J., Esteller, M., Opitz-Araya,
X., McCombie, R., Herman, J.G., Gerald, W.L., Lazebnik, Y.A., et al. (2001).
Inactivation of the apoptosis effector Apaf-1 in malignant melanoma. Nature
Kim, J.Y., Ahn, H.J., Ryu, J.H., Suk, K., and Park, J.H. (2004). BH3-only
protein Noxa is a mediator of hypoxic cell death induced by hypoxia-induc-
ible factor 1-?. J. Exp. Med. 199, 113–124.
Sowter, H.M., Ratcliffe, P.J., Watson, P., Greenberg, A.H., and Harris, A.L.
(2001). HIF-1-dependent regulation of hypoxic induction of the cell death
factors BNIP3 and NIX in human tumors. Cancer Res. 61, 6669–6673.
Koumenis, C., Alarcon, R., Hammond, E., Sutphin, P., Hoffman, W., Murphy,
sis from p53-dependent transactivation. Mol. Cell. Biol. 21, 1297–1310.
of BNIP3L in lung cancer]. Ai Zheng 23, 8–14.
Symonds, H., Krall, L., Remington, L., Saenz-Robles, M., Lowe, S., Jacks,
T., and Van Dyke, T. (1994). p53-dependent apoptosis suppresses tumor
growth and progression in vivo. Cell 78, 703–711.
Kubasiak, L.A., Hernandez, O.M., Bishopric, N.H., and Webster, K.A. (2002).
Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2
family protein BNIP3. Proc. Natl. Acad. Sci. USA 99, 12825–12830.
Tokino, T., Thiagalingam, S., El-Deiry, W.S., Waldman, T., Kinzler, K.W., and
Vogelstein, B. (1994). p53 tagged sites from human genomic DNA. Hum.
Mol. Genet. 3, 1537–1542.
Lai, J., Flanagan, J., Phillips, W.A., Chenevix-Trench, G., and Arnold, J.
(2003). Analysis of the candidate 8p21 tumour suppressor, BNIP3L, in breast
and ovarian cancer. Br. J. Cancer 88, 270–276.
Vaupel, P., Schlenger, K., Knoop, C., and Hockel, M. (1991). Oxygenation
of human tumors: evaluation of tissue oxygen distribution in breast cancers
by computerized O2tension measurements. Cancer Res. 51, 3316–3322.
Lane, D.P. (1992). p53, guardian of the genome. Nature 358, 15–16.
Levine, A.J. (1997). p53, the cellular gatekeeper for growth and division. Cell
Villunger, A., Michalak, E.M., Coultas, L., Mullauer, F., Bock, G., Ausser-
lechner, M.J., Adams, J.M., and Strasser, A. (2003). p53- and drug-induced
apoptotic responses mediated by BH3-only proteins puma and noxa. Sci-
ence 302, 1036–1038.
Marchetti, P., Cannita, K., Ricevuto, E., De Galitiis, F., Di Rocco, Z.C., Tessi-
tore, A., Bisegna, R., Porzio, G., De Rubeis, G.P., Ventura, T., et al. (2003).
Prognostic value of p53 molecular status in high-risk primary breast cancer.
Ann. Oncol. 14, 704–708.
Vogelstein, B., Lane, D., and Levine, A.J. (2000). Surfing the p53 network.
Nature 408, 307–310.
Wang, W., and El-Deiry, W.S. (2003). Bioluminescent molecular imaging of
endogenous and exogenous p53-mediated transcription in vitro and in vivo
using an HCT116 human colon carcinoma xenograft model. Cancer Biol.
Ther. 2, 196–202.
Meijerink, J.P., Mensink, E.J., Wang, K., Sedlak, T.W., Sloetjes, A.W., de
Witte, T., Waksman, G., and Korsmeyer, S.J. (1998). Hematopoietic malig-
nancies demonstrate loss-of-function mutations of BAX. Blood 91, 2991–
Wang, S., and El-Deiry, W.S. (2004). Inducible silencing of KILLER/DR5
in vivo promotes bioluminescent colon tumor xenograft growth and confers
resistance to chemotherapeutic agent 5-fluorouracil. Cancer Res. 64, 6666–
Nakano, K., and Vousden, K.H. (2001). PUMA, a novel proapoptotic gene,
is induced by p53. Mol. Cell 7, 683–694.
Oda, E., Ohki, R., Murasawa, H., Nemoto, J., Shibue, T., Yamashita, T.,
Tokino, T., Taniguchi, T., and Tanaka, N. (2000a). Noxa, a BH3-only member
of the Bcl-2 family and candidate mediator of p53-induced apoptosis. Sci-
ence 288, 1053–1058.
Weinmann, M., Welz, S., and Bamberg, M. (2003). Hypoxic radiosensitizers
and hypoxic cytotoxins in radiation oncology. Curr. Med. Chem. Anti-Canc.
Agents 3, 364–374.
Oda, K., Arakawa, H., Tanaka, T., Matsuda, K., Tanikawa, C., Mori, T.,
Nishimori, H., Tamai, K., Tokino, T., Nakamura, Y., and Taya, Y. (2000b).
p53AIP1, a potential mediator of p53-dependent apoptosis, and its regula-
tion by Ser-46-phosphorylated p53. Cell 102, 849–862.
Wenger, R.H., Camenisch, G., Desbaillets, I., Chilov, D., and Gassmann,
M. (1998). Up-regulation of hypoxia-inducible factor-1? is not sufficient for
hypoxic/anoxic p53 induction. Cancer Res. 58, 5678–5680.
Yasuda,M., Han,J.W.,Dionne, C.A.,Boyd,J.M., andChinnadurai, G.(1999).
BNIP3?: a human homolog of mitochondrial proapoptotic protein BNIP3.
Cancer Res. 59, 533–537.
Okami, J., Simeone, D.M., and Logsdon, C.D. (2004). Silencing of the hyp-
oxia-inducible protein Bnip3 in pancreatic cancer. Cancer Res. 64, 5338–
Yussman, M.G., Toyokawa, T., Odley, A., Lynch, R.A., Wu, G., Colbert, M.C.,
Aronow, B.J., Lorenz, J.N., and Dorn, G.W., II (2002). Mitochondrial death
protein Nix is induced in cardiac hypertrophy and triggers apoptotic cardio-
myopathy. Nat. Med. 8, 725–730.
Pan, Y., Oprysko, P.R., Arsham, A.M., Koch, C.J., and Simon, M.C. (2004).
p53 cannot be induced by hypoxia alone but responds to the hypoxic micro-
environment. Oncogene 23, 4975–4983.
CANCER CELL : DECEMBER 2004 609