1. The function of multiple IkB : NF-kB complexes in the resistance of cancer
cells to Taxol-induced apoptosis
Qiang G Dong1
, Guido M Sclabas1,5
, Shuichi Fujioka1
, Christian Schmidt1
, Bailu Peng1
,
TianAi Wu1
, Ming-Sound Tsao6
, Douglas B Evans1
, James L Abbruzzese2
, Timothy J McDonnell3
and Paul J Chiao*,1,4
1
Department of Surgical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, TX 77030, USA;
2
Department of GI Medical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, TX 77030, USA;
3
Department of Molecular Pathology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, TX 77030, USA;
4
Department of Molecular & Cellular Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, TX
77030, USA; 5
Department of Visceral and Transplantation Surgery, Inselspital, University of Bern, 3010 Bern, Switzerland;
6
Division of Cellular and Molecular Biology, Ontario Cancer Institute, Princess Margaret Hospital, University of Toronto,
Toronto, Ontario, Canada M5G 2M9
The Rel/NF-kB transcription factors play a key role in
the regulation of apoptosis and in tumorigenesis by
controlling the expressions of specific genes. To determine
the role of the constitutive activity of RelA in tumorigen-
esis, we generated pancreatic tumor cell lines that express
a dominant negative mutant of IkBa (IkBaM). In this
report, we show that the inhibition of constitutive NF-kB
activity, either by ectopic expression of IkBaM or by
treating the cells with a proteasome inhibitor PS-341
which blocks intracellular degradation of IkBa proteins,
downregulates the expression of bcl-xl. We identified two
putative NF-kB binding sites (kB/A and B) in the bcl-xl
promoter and found that these two sites interact with
different NF-kB proteins. p65/p50 heterodimer interacts
with kB/A site whereas p50/p50 homodimer interacts
with kB/B. The bcl-xl promoter reporter gene assays
reveal that NF-kB dependent transcriptional activation is
mainly mediated by kB/A site, indicating that bcl-xl is
one of the downstream target genes regulated by RelA/
p50. Both IkBaM and PS-341 completely abolish NF-kB
DNA binding activity; however, PS-341, but not ectopic
expression of IkBaM, sensitized cells to apoptosis induced
by Taxol. This is due to the Taxol-mediated reactivation
of RelA through phosphorylation and degradation of IkBb
and the re-expression of NF-kB regulated bcl-xl gene in
these cancer cells as ectopic expression of the bcl-xl gene
confers resistance to Taxol-induced apoptosis in PS-341
sensitized cells. These results demonstrate the important
function of various NF-kB/IkB complexes in regulating
anti-apoptotic genes in response to apoptotic stimuli, and
they raise the possibility that NF-kB : IkBa and NF-
kB : IkBb complexes are regulated by different upstream
activators, and that NF-kB plays a key role in pancreatic
tumorigenesis.
Oncogene (2002) 21, 6510 – 6519. doi:10.1038/sj.onc.
1205848
Keywords: apoptosis; gene expression; NF-kB; IkB;
bcl-xl
Introduction
The c-rel member of the Rel/NF-kB pleiotropic
transcription factor family was first identified as a
cellular homologue of the v-rel oncogene, suggesting
that other Rel/NF-kB members are oncogenes (Verma
et al., 1995; Baldwin, 1996). The Rel/NF-kB family,
which consists of RelA, Rel (v-rel), RelB, p50 (p105),
and p52 (p100), and can form heterodimers and
homodimers among themselves, controls the expression
of numerous genes (Verma et al., 1995; Baldwin, 1996).
In most cell types, Rel/NF-kB proteins are sequestered
in the cytoplasm in an inactive form through their
noncovalent association with the inhibitor IkB (Sen
and Baltimore, 1986). This association masks the
nuclear localization signal of Rel/NF-kB, thereby
preventing Rel/NF-kB nuclear translocation and
DNA binding activity (Baeuerle and Baltimore, 1989).
Stimulation of these cells leads to phosphorylation of
IkBa, which triggers rapid degradation of the inhibitor,
and consequently, Rel/NF-kB proteins are released and
translocated into the nucleus, where they activate
transcription of target genes (Verma et al., 1995;
Baldwin, 1996). One of the key target genes regulated
by RelA is its inhibitor IkBa. We previously cloned
IkBa cDNA and promoter and described a feedback
inhibition pathway for control of IkBa gene transcrip-
tion and downregulation of transient activation of Rel/
NF-kB activity (Chiao et al., 1994). Activation of NF-
kB is achieved through the signal-induced proteolytic
degradation of IkBs, which is mediated by the 26S
proteosome (Chen et al., 1995; Brown et al., 1995;
Mercurio et al., 1997). Stimulation of cells by various
inducers that leads to phosphorylation of IkBa at
serine residues 32 and 36 by IkB kinases (IKKs)
triggers the rapid degradation of the inhibitor
Received 13 September 2001; revised 1 July 2002; accepted 5 July
2002
*Correspondence: PJ Chiao, E-mail: pjchiao@notes.mdacc.tmc.edu
Oncogene (2002) 21, 6510 – 6519
ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00
www.nature.com/onc

2. (Woronicz et al., 1997; DiDonato et al., 1997; Zandi et
al., 1997). Which kinases regulate IKK activity remains
unclear; however, it appears likely that members of the
MAP kinase kinase kinase (MAPKKK) family play a
key role in control of IKK kinase complex. These
findings have provided a useful tool for detailed
functional analysis of components of an IkB kinase
complex and a better understanding of the signal
transduction cascade leading to activation of Rel/NF-
kB.
Several reports suggest that members of the Rel/NF-
kB and IkB families are involved in the development of
cancer (Gilmore et al., 1996). For instance, v-rel,
carried by a highly oncogenic retrovirus, causes
aggressive tumors in young birds and is able to
transform avian lymphoid cells and fibroblasts (Moore
and Bose, 1988; Sylla and Temin, 1986). The mutated
c-rel oncogene transforms cells (Moore and Bose,
1988). The Tax protein from the leukemogenic virus
HTLV-1 is a potent activator of Rel/NF-kB, and the
growth of Tax-induced tumors in mice is inhibited by
antisense RelA constructs (Kitajima et al., 1992). Tax-
mediated activation of NF-kB resulted from direct
interactions of Tax and MEKK1, a component of an
IkB kinase complex leading to enhanced IkB kinase
phosphorylation of IkBa (Yin et al., 1998). In addition,
the genes encoding c-Rel, Bcl-3, p105 (p50), and p100
(p52) are located at sites of recurrent genomic
rearrangements in cancer (Gilmore et al., 1996).
Recently, we showed that RelA is constitutively
activated in human pancreatic adenocarcinomas and
in a number of human pancreatic tumor cell lines
(Wang et al., 1999a).
A major cause of death of mice that lack the RelA
subunit of Rel/NF-kB transcription factors in embryo-
nic development is a massive apoptosis of fetal liver
cells; this suggests a role for RelA in protecting cells
from pro-apoptotic stimuli (Beg et al., 1995). This
observation is supported by several recent reports
showing that tumor necrosis factor a (TNFa) readily
induced apoptosis in macrophages and a fibroblast cell
line established from RelA knockout mice but not in
wild-type cells (Beg et al., 1995). This phenotype in
RelA knockout cells could be inhibited by transfected
RelA (Beg et al., 1995). Similarly, radiation-, daunor-
ubicin- or TNFa-induced apoptosis is potentiated in
HT1080 human fibrosarcoma, human embryonic
fibroblast, and Jurkat cell lines transfected with
dominant negative IkBa (Wang et al., 1996; van
Antwerp et al., 1996). In our studies, expression of
the dominant-negative IkBa mutant proteins did not
enhance TNFa-induced apoptosis in cells that ectopi-
cally express high levels of Bcl-2 proteins, which did
not inhibit RelA activation or RelA-dependent trans-
activation; this suggests that Bcl-2 functions
downstream of the RelA anti-apoptotic signaling
pathway (Herrmann et al., 1997, 1998). We also
showed that RelA and the RelA target genes uPA
were constitutively activated in about 70% of human
pancreatic cancers and nine of 11 human pancreatic
cancer cell lines but not in normal pancreatic tissues
and nontumorigenic pancreatic epithelial cells (Wang et
al., 1999a,b). Pancreatic adenocarcinoma is the fifth
leading cause of adult cancer mortality in the US with
1 – 2% survival rates in 5 years (DiGiuseppe et al.,
1996). Early detection of pancreatic cancer has not yet
been developed; however, at the time of diagnosis,
most pancreatic cancer patients present with locally
advanced disease or metastasis. Resistance to apoptosis
in human pancreatic cancer is thought to render most
chemotherapeutic agents and radiation treatments
ineffective (DiGiuseppe et al., 1996). Thus, dysregula-
tion of Rel/NF-kB activity in tumor cells may be
critical in the inhibition of pro-apoptotic stimuli. We
therefore undertook a study of the regulation of bcl-xl
gene by NF-kB and the role of bcl-xl in control of
apoptosis in pancreatic cancer cells.
Results
Constitutive RelA activity induced overexpression of
Bcl-xl in human pancreatic cancer cell lines
We previously reported that RelA and the RelA target
genes uPA were constitutively activated in about 70%
of human pancreatic cancers and nine of 11 human
pancreatic cancer cell lines but not in normal
pancreatic tissues and nontumorigenic pancreatic
epithelial cells (Wang et al., 1999a,b). Interestingly,
the levels of bcl-xl protein are increased in pancreatic
cancer cell lines exhibiting constitutive RelA activity
(Figure 1a,b). As shown in Figure 1c, the constitutive
RelA activity is found in the nuclear extracts from
AsPc-1, MDAPanc-28, Capan-1 and Panc-1 cell lines,
as we previously reported (Wang et al., 1999a,b), but
not in that of the E6E7 immortalized HPDE cells.
Figure 1d shows that TNF-a-inducible RelA/p50 NF-
kB DNA binding activity in the E6E7 immortalized
HPDE and Jurkat cell lines as controls for the
constitutive activation of RelA in these pancreatic
cancer cell lines.
In order to determine whether the constitutive RelA
activity induced overexpression of bcl-xl, we
constructed a retroviral-expressing Flag-tagged phos-
phorylation mutant of IkBa (S32, 36A) (IkBaM) to
specifically inhibit RelA activity in the pancreatic
cancer cells. Human pancreatic tumor cell lines AsPc-
1, MDAPanc-28, Capan-1 and Panc-1 cells were
infected with Flag-tagged IkBaM and control retro-
virus and selected for the resistance to puromycin
(800 ng/ml). The pooled AsPc-1 and MDAPanc-28
puromycin-resistant cells were used for our study. The
expression of Flag-tagged IkBaM was detected in these
cells infected by Flag-tagged IkBaM retrovirus but not
in those infected by control retrovirus (Figure 2a). The
constitutive RelA activity was inhibited by IkBaM
(Figure 2b). As shown on Western blot analysis
(Figure 2c,d), the expression of bcl-xl was also
inhibited in cells that expressed Flag-tagged IkBaM
but not in the control cells. The results suggest that
constitutive RelA results in upregulation of bcl-xl in
these pancreatic cancer cell lines. To verify this finding,
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6511
Oncogene

3. we performed Northern blot analysis using AsPc-1/
CTL/puro, AsPc-1/Flag-tagged IkBaM, MDAPanc-28/
CTL/puro, MDAPanc-28/Flag-tagged IkBaM cells
(Figure 2e,f). Taken together, these results provide
consistent evidence that bcl-xl could be induced by
constitutive RelA in these pancreatic cancer cells.
The promoter of bcl-xl contains multiple kB enhancers
and is regulated by RelA/p50 heterodimers
We next identified two kB enhancers in the 800-bp
promoter sequence of the human bcl-xl gene. The
diagram shows that two kB enhancer sites
(GGGGACTGCCcagggaGTGACTTTCC) are located
341- and 365-nt upstream of the bcl-xl translation start
site and between 759 and 784 upstream of the
transcription start site of the bcl-xl gene (Figure 3a).
Sequence analysis revealed that the presence of the
TATA box is located 25-bp upstream of the transcrip-
tion initiation site and other previously identified cis-
regulatory elements (Figure 3a). To determine with
which members of Rel/NF-kB these two kB sites (kB/A
and kB/B) in the bcl-xl promoter will interact, we
synthesized oligonucleotides probe A, which only
contains the kB/A site, and oligonucleotides probe B,
which only contains the kB/B site and performed
electrophoretic mobility shift assays to analyse the kB
DNA-binding activity using these probes. As shown in
Figure 3b, probe kB/A interacted with the RelA/p50
heterodimers as in the control using HIV/kB probe,
whereas probe kB/B interacted with the p50/p50
homodimers. To confirm these differential binding
activities, competition and supershift assays were
carried out using wild-type and mutant kB/A and
kB/B oligonuceotides and specific antibodies to RelA
and p50. The analyses show that the kB/A site interacts
with RelA/p50 heterodimers only and the kB/B site
binds to p50/p50 homodimers (Figure 3c,d). To
determine whether these two kB sites (kB/A and kB/
B) identified in the bcl-xl promoter are functional, 800-
bp and 520-bp genomic fragments of the bcl-xl gene
promoter were cloned into pBLCAT3 vector (Figure
3a). By using PCR-mediated site-directed mutagenesis,
mutant kB/A and kB/B sites were generated in the 520-
bp fragment of the bcl-xl gene promoter (Figure 3a).
High levels of CAT activities were detected in COS
cells cotransfected with the 800-bp or 520-bp bcl-xl
gene promoter-CAT construct and the RelA/NF-kB
expression vector (Figure 3e, lanes 2 and 4). Little
CAT activity increase was detected in the cells
cotransfected with the bcl-xl gene promoter-CAT
construct with the mutant kB/A site and the RelA/
NF-kB expression vector (Figure 3e, lane 6), but
increasing transactivation (twofold) was observed when
Figure 1 Constitutive RelA activity correlated with overexpres-
sion of bcl-xl in human pancreatic cancer cell lines. (a,b and c)
Expression of bcl-xl, b-actin and electrophoretic mobility shift
assay for NF-kB DNA binding activity in nontumorigenic human
pancreatic ductal epithelial cell line (HPDE/E6E7), AsPc-1, Panc-
28, Capan-1 and Panc-1 human pancreatic cancer cell lines, as in-
dicated. (d) Electrophoretic mobility shift assay for NF-kB DNA
binding activity in HPDE/E6E7 and Jurkat cells stimulated with
or without TNF-a (10 ng/ml) as indicated
Figure 2 Inhibition of constitutive RelA activity and expression
of bcl-xl by IkBaM. (a – f), lane 1, Panc-28/Puro; lane 2, Panc-28/
IkBaM; lane 3, AsPc-1/Puro; lane 4, AsPc-1/IkBaM. (a) expres-
sion of Flag-IkBaM. (b) Electrophoretic mobility shift assay for
NF-kB DNA binding activity. (c and d) Western blot analyses
for expression of bcl-xl and b-actin. (e and f) Northern blot ana-
lyses for bcl-xl and GADPH expression in the pancreatic cancer
cell lines, as indicated
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6512
Oncogene

4. the RelA expression vector was cotransfected with a
bcl-xl gene promoter-CAT construct with a mutant kB/
B site and the RelA/NF-kB expression vector (Figure
3e, lane 8). These results show that the kB/A site,
which binds to RelA/p50, is the key kB enhancer for
regulating the expression of the bcl-xl gene, suggesting
Figure 3 The promoter of bcl-xl contains multiple kB enhancers that are regulated by RelA/p50 heterodimers and p50/p50 homo-
dimers. (a) Schematic diagram of bcl-xl promoter and bcl-xl promoter reporter gene constructs with wild-type and mutant kB/A and
kB/B sequences indicated. (b) Electrophoretic mobility shift assay (EMSA) for NF-kB DNA binding activity with the kB probes
using 10 mg of AsPc-1 cell nuclear extracts; lane 1, HIV-kB probe; lane 2 kB/A probe, lane 3, kB/B probe. (c) EMSA using
10 mg of AsPc-1 nuclear extracts with 32
P-labeled kB/A probe for the competition assay with wild-type and mutant kB/A oligonu-
cleotides, and supershift with anti-RelA antibody with or without RelA peptide as indicated. (d) EMSA using 10 mg of AsPc-1 nu-
clear extracts with 32
P-labeled kB/B probe for competition assay with wild-type and mutant kB/B oligonucleotides, and supershift
with anti-p50 antibody with or without p50 peptide as indicated. (e) Reporter gene assays for analysis of the bcl-xl promoter.
Various bcl-xl CAT reporter gene constructs were transfected in COS cells with or without RelA expression plasmid
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6513
Oncogene

5. that both kB/A and kB/B sites play important roles in
regulating bcl-xl gene expression.
Taxol induced expression of bcl-xl via IkB kinase
activated multiple NF-kB : IkB complexes
Since overexpression of the bcl-xl gene is regulated by
RelA, it was of interest to determine whether down
regulation of bcl-xl expression by inhibiting RelA
DNA binding activity in human pancreatic cancer cells
will confer the susceptibility to apoptosis induced by
Taxol, which is one of the apoptosis-inducing
chemotherapeutic agents used in the current treatment
of pancreatic cancer. We therefore used various
concentrations of Taxol to induce cell death in AsPc-
1/IkBaM, MDAPanc-28/IkBaM and Panc-1/IkBaM
cells at different time periods. MTT assays and FACS
analysis (fluorescence activated cell sorting) showed
that Taxol did not induce apoptosis in these cells (data
not shown). To determine whether alternative signaling
pathways are activated by Taxol to overcome the
inhibition of RelA by dominant negative IkBa
(IkBaM), further analyses were performed using
AsPc-1/puro and AsPc-1/IkBaM stimulated with Taxol
(40 mM) for 2 h. As shown in Figure 4a, the expression
of a Flag-tagged IkBaM was confirmed. Interestingly,
while Taxol enhanced the constitutive RelA/p50-DNA
binding activity in AsPc-1/puro cells (Figure 4b, lanes 1
and 2, threefold induction by phosphoimage analysis),
it reactivated RelA/p50 activity in AcPc-1/IkBaM cells
(Figure 4b, lanes 3 and 4, threefold induction by
phosphoimage analysis). Oct-1 DNA binding activity
as loading control is shown in Figure 4c. These results
suggest that RelA/p50 activity was activated from
other IkB/RelA/p50 complexes by Taxol. To determine
this possibility, immunoblot analyses for IkBa and
IkBb were performed. The phosphorylated IkBa was
detected in AsPc-1/puro cells with and without
stimulation of Taxol (Figure 4d), which is consistent
with constitutive RelA activation in these cells.
However, no phosphorylation of IkBa was observed
in the AsPc-1/IkBaM cells, and the presence of the
slower migrating Flag-tagged phosphorylation mutant
of IkBa (S32, 36A) was detected (Figure 4e, lanes 3
and 4). The Taxol stimulation induced degradation of
IkBb in AsPc-1/IkBaM, but not in AsPc-1/puro cells
(Figure 4f), suggesting that the reappearing RelA
activity observed in lane 4 of Figure 4b was induced
by Taxol through the activation of NF-kB complexes
containing IkBb. Furthermore, the inhibition of bcl-xl
expression by IkBaM was overcome by Taxol in the
AsPc-1/IkBaM cells (Figure 4g). Since IKK1/a and
IKK2/b play a key role in regulation of NF-kB
activation, we tested whether the IKKs are essential
in Taxol-induced NF-kB activation. NF-kB DNA
binding activity is induced by Taxol in IKK WT
MEF cells but not in the IKK1/a7/7
and IKK2/b7/7
MEF cells, suggesting that Taxol-induced RelA
activation requires IKK activation (Figure 5a). As a
control, NF-kB DNA binding activity induced by
TNF-a was also analysed in these cells (Figure 5b).
This is consistent with previous reports that TNF-a
and Taxol activated IKK2/b (Chen et al., 1995; Brown
et al., 1995; Mercurio et al., 1997; Woronicz et al.,
1997; DiDonato et al., 1997; Zandi et al., 1997;
Blagosklonny et al., 1996; Heilker et al., 1999a). Thus,
the reactivation of bcl-xl expression by Taxol empha-
sizes the role of RelA-regulated bcl-xl expression in
controlling apoptosis of human pancreatic cancer cells.
Figure 4 Induction of bcl-xl expression by Taxol via activation
of multiple NF-kB : IkB complexes. Both AsPc-1/puro control
and AsPc-1/IkBaM cells were stimulated with or without 40 mM
Taxol for 2 h as indicated. Cytoplasmic and nuclear extracts
and RNA were isolated for analyses. Western blot analysis using
affinity purified, specific antibodies against Flag-tag (a), phospho-
IkBa (d), IkBa (e), IkBb (f), Bcl-xl (g) and b-actin (h). (b) EMSA
for analysis for NF-kB DNA-binding activity in the nuclear ex-
tracts of AsPc-1/puro and AsPc-1/IkBaM cells with or without
Taxol stimulation. (c) Oct-1 DNA-binding activity for loading
control
Figure 5 Taxol-induced NF-kB activation is dependent on IkB
kinase. The nuclear extracts were isolated for EMSA from IKK
WT, IKK1/a7/7
and IKK2/b7/7
MEF cells unstimulated in
both (a and b) lanes 1, 3 and 5, and stimulated with Taxol
(10 mM) for 2 h in (a), or TNF-a (10 ng/ml) for 30 min in (b)
lanes 2, 4 and 6
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6514
Oncogene

6. Proteosome inhibitor PS-341 suppressed constitutive
RelA/p50 activity and bcl-xl expression
The proteosome inhibitor PS-341 was tested for time
and dose dependent inhibition of constitutive RelA
activity (Figure 6a). The inhibition of constitutive RelA
activity was observed at 100 nM of PS-341 at 4 h (Figure
6a, lane 2), and RelA activity was completely inhibited
in the cells treated with 10 nM PS-341 for 24 h (Figure
6a, lane 8). The Oct-1 DNA-binding activity was not
altered in the presence of various concentrations of PS-
341 over 4 to 48 h (Figure 6b). Accompanied by the
decreasing RelA DNA binding activity in the nucleus,
the level of bcl-xl protein is also reduced by PS-341
(Figure 6c,d lanes 3 – 6 and 9 – 11). Moreover, the
expression of Bcl-xl and A20 mRNA is induced by
Taxol and decreased by inhibition of RelA by PS-341
(Figure 7a). These results suggest that inhibition of
RelA constitutive activity by PS-341 in pancreas cancer
cells may sensitize them to Taxol-induced cell death.
Taxol induced apoptosis in PS-341 sensitized pancreatic
cancer cell lines with constitutive RelA activity
To determine whether inhibition of RelA constitutive
activity in pancreas cancer cells by PS-341 sensitizes
them to Taxol-induced cell death, AsPc-1 cells were
treated with 10 nM PS-341 and 50 mM Taxol from 24 to
72 h and the numbers of viable cells were determined
by MTT assays. As shown in Figure 8a, there was a
fivefold survival reduction in the cells treated with PS-
341 and Taxol compared with those treated with PS-
341 or Taxol alone. The findings were similar for
MDAPanc-28, Capan-1 and Panc-1 cells (data not
shown). FACS analysis showed that Taxol-induced
apoptosis in PS-341 sensitized AsPc-1 cells (Figure 8b).
These results suggest that inhibition of RelA-regulated
bcl-xl expression may render AsPc-1 pancreatic cancer
cells predisposed to Taxol-induced apoptosis.
Ectopic expression of bcl-xl confers resistance to
apoptosis in the pancreatic cancer cells treated with
PS-341 and taxol
To determine the role of bcl-xl in PS-341-mediated
sensitization to Taxol-induced pancreatic cancer cell
Figure 6 Suppression of constitutive RelA/p50 activity and bcl-xl
expression by proteosome inhibitor PS-341. Lanes 1 – 6 PS-341
(100 nM) at 0, 4, 8, 16, 24, 48, respectively. Lanes 7 – 11, 24 h
PS-341 treatment at 0, 10, 50, 100, 500 nM, respectively. (a) PS-
341 time course and dose-dependent inhibition of RelA/p50
NF-kB-DNA binding activity. (b) Oct-1 DNA-binding activity
for loading control. (c) Western blot analysis of PS-341 time
course and dose-dependent inhibition of bcl-xl expression. (d) b-
actin for loading control
Figure 7 Inhibition of Taxol-induced bcl-xl and A20 expression
by proteosome inhibitor PS-341. AsPc-1 cells were untreated
and treated with Taxol (50 mM), PS-341 (100 nM) or both, as in-
dicated. Total RNA was isolated, and Northern blot analysis was
performed using bcl-xl cDNA, A20 cDNA and GAPDH probes
a
b
Figure 8 Analysis of Taxol- and PS-341-induced apoptosis in
pancreatic cancer cells. (a) Viable cells were determined by
MTT assays 24 h after treatment of AsPc-1 cells with Taxol
(50 mM), PS-341 (10 nM) or both, as indicated. (b) The sub-G1
faction was assessed by FACS analysis after propidium iodide
staining in AsPc-1 cells treated with (a), Taxol (50 mM); (b),
PS – 341 (10 nM) or (c), both
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6515
Oncogene

7. apoptosis, we tested whether ectopic expression of bcl-
xl cDNA would induce resistance to Taxol-mediated
apoptosis in AsPc-1 cells. These cells were transfected
with 0.5 mg of GFP expression vectors for labeling the
cells and LacZ plasmid for transfection efficiency
control in the presence and absence of 5 mg of bcl-xl
expression plasmid. After 48 h, these cells were treated
with PS-341 (10 nM), Taxol (50 mM), or both for 24 h,
and FACS analysis was performed to determine the
number of GFP positive cells (Figure 9). The analysis
showed that the number of GFP-positive cells in PS-
341- and Taxol-treated AsPc-1 cells with ectopic
expression of bcl-xl gene was almost the same as the
control AsPc-1 cells and AsPc-1 cells treated with PS-
341 or Taxol, whereas the number of GFP-positive
cells in PS-341- and Taxol-treated AsPc-1 cells without
a transfected bcl-xl gene was significantly reduced
(Figure 9). These combined results suggest that over-
expression of bcl-xl confers resistance to Taxol-induced
apoptosis in PS-341-sensitized pancreatic cancer cells.
Discussion
The transcription factor NF-kB plays a critical role in
controlling expression of its downstream target genes in
immune and inflammatory responses, development, and
apoptosis (Verma et al., 1995; Baldwin, 1996; Gilmore
et al., 1996). Important advances have been made in
understanding the signal transduction cascades that
control NF-kB activation in response to proinflamma-
tory cytokines such as TNF-a and IL-1. Several lines of
evidence indicate that the transcription factor NF-kB is
constitutively activated in cancer cells, suggesting that
NF-kB plays an important role in tumorigenesis
(Verma et al., 1995; Baldwin, 1996); however, the
downstream target genes regulated by NF-kB in the
tumorigenic process remain to be identified. The
current study sought to determine and characterize
the expression of bcl-xl regulated by RelA activity and
the role of bcl-xl in regulation of apoptosis induced by
Taxol in human pancreatic cancer cells, which are
resistant to apoptosis induced by most chemotherapeu-
tic agents and radiation (DiGiuseppe et al., 1996). This
ineffectiveness in the treatment of pancreatic adeno-
carcinoma in part results in this disease being the fifth
leading cause of adult cancer mortality in the US.
We have demonstrated that bcl-xl plays a key role in
the control of apoptosis and transcription of bcl-xl is
induced by constitutive RelA/NF-kB activity in human
pancreatic cancer cells. The induction of bcl-xl is likely
mediated directly by the kB sites present in the
upstream promoter element of the bcl-xl gene.
Furthermore, IkBaM-inhibited transcription of bcl-xl
is reinduced by p65(RelA) from other IkB/NF-kB
complexes such as IkBb/p65(RelA) complex through
Taxol-mediated phosphorylation and degradation of
IkBb in the pancreatic cancer cells expressing IkBaM.
These results demonstrate the importance of various
NF-kB/IkB complexes in regulation of antiapoptotic
genes in cancer cells.
We previously showed that RelA and its downstream
target gene uPA are constitutively activated in human
pancreatic adenocarcinomas and in a number of human
pancreatic tumor cell lines (Wang et al., 1999a,b). To
determine the role of constitutive RelA in pancreatic
cancer cells, we generated several cell lines expressing
IkBaM by retroviral infections. We have shown that the
constitutive NF-kB activity is efficiently abrogated
either by ectopic expression of IkBaM or by treating
the cells with PS-341, an inhibitor of proteosome
responsible for the intracellular degradation of IkB
proteins. Blocking of constitutive NF-kB activity by
either IkBaM or PS-341 inhibits the expression of bcl-xl
(Figure 2). We identified the two putative NF-kB
binding sites (kB/A and B) in the bcl-xl promoter. Bcl-
xl promoter reporter gene assays indicate that the NF-
kB-dependent transcriptional activation is mainly
mediated by the kB/A site, indicating that bcl-xl is
one of the downstream target genes regulated by RelA/
p50. Several investigators have recently reported the
regulation of bcl-xl gene expression by RelA/p50
(Glasgow et al., 2000; Chen et al., 2000; Stroka et al.,
1999; Bui et al., 2001; Tsukahara et al., 1999). In this
report, we have shown that the promoter of bcl-xl gene
contains two different kB sites, kB/A and kB/B, and,
hence, is a target of various NF-kB complexes. The
RelA/p50 heterodimers interact with the kB/A site
whereas the p50/p50 homodimers interacts with the kB/
B site. It has been shown that the p65(RelA)/p50
heterodimers play a key role in activation of transcrip-
tion of NF-kB downstream target genes; however, p50/
p50 homodimers appear to function as transcriptional
Figure 9 Expression of exogenous bcl-xl protects against PS-341-
and Taxol-mediated apoptosis. AsPc-1 cells were grown in 6-well
dishes and transfected with GFP and LacZ expression plasmids in
the presence and absence of bcl-xl expression plasmids. Twenty-
four hours after transfection, these AsPc-1 cells were treated with
Taxol (50 mM), PS-341 (10 nM), or both for 24 h, and the number
of GFP-positive cells were determined by FACS analysis and
transfection efficiencies were normalized by LacZ activity
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6516
Oncogene

8. activators or repressors, depending on the cell type
(Ishikawa et al., 1996, 1998). Our results from reporter
gene assay (Figure 3f) suggest that both kB/A and kB/B
sites in bcl-xl promoter play an important role in the
control of the expression of this key apoptotic mediator
and p50/p50 homodimers may positively be involved in
the bcl-xl expression. PS-341-mediated time dependent
inhibition of bcl-xl expression suggested that the half-
life of bcl-xl protein is greater than 4 h, which is
consistent with a previous report (Pardo et al., 2002).
Ectopic expression of bcl-xl confers resistance to Taxol-
induced apoptosis in PS-341 sensitized cells. These
results clearly demonstrate the importance of various
NF-kB/IkB complexes in regulation of anti-apoptotic
genes.
IkBaM and PS-341 completely abolish the RelA
constitutive activity in the pancreatic cancer cells
(Figure 2). However, PS-341, but not ectopic expression
of IkBaM, sensitized cells to apoptosis induced by
Taxol (Figures 8 and 9). Several reports showed that
Taxol activated NF-kB via IKK2/b, and IkBa, IkBb
and IkBe complexes are the substrates of IKK2/b (Lee
et al., 1997; Jordan et al., 1996; Blagosklonny et al.,
1996; Heilker et al., 1999a,b). The reactivation of RelA
induced by Taxol in pancreatic cancer cells that express
IkBaM occurred through phosphorylation and degra-
dation of IkBb, resulting in the re-expression of the
NF-kB regulated bcl-xl gene in these cancer cells
(Figures 4 and 6). These results are consistent with
our finding that TNF-a increased constitutive RelA
DNA binding activity in pancreatic cancer cells and
induced RelA DNA-binding activity in the pancreatic
cancer cells that express IkBaM (data not shown). The
expression of both NF-kB downstream target gene bcl-
xl and A20 was significantly reduced by a NF-kB
inhibitor, PS-341, in the presence and absence of Taxol
stimulation (Figure 7, lane 3 and 4), further supporting
that PS-341 is a potent inhibitor for NF-kB activation
by blocking the degradation of IkBs. Our data obtained
using stable clones pooled from the infection of control
and IkBaM retrovirus suggest that inhibition of NF-kB
activity by IkBaM may not be sufficient to block the
bcl-xl expression after Taxol stimulation, because other
IkB/NF-kB complexes such as IkBb/NF-kB complex
can be activated to induce bcl-xl expression. Therefore,
sensitization of pancreatic cancer cell lines to apoptosis
induced by anti-cancer agents such as Taxol (Figures
7 – 9) requires complete inhibition of NF-kB activation
by blocking degradation of both IkBa and IkBb
proteins. This is consistent with the important role of
NF-kB in control of apoptosis. Our results suggest that
inhibiting the degradation of IkBs is critical in order to
suppress NF-kB activation induced by various antic-
ancer agents.
In summary, the current study provides direct
evidence that bcl-xl is one of the downstream target
genes regulated by RelA in pancreatic cancer cells and
the bcl-xl plays a key role in control of Taxol-induced
apoptosis in pancreatic cancer cells. Multiple p65RelA/
IkB complexes are activated in response to the
apoptotic signal induced by Taxol and result in the
re-expression of the NF-kB regulated bcl-xl gene in
pancreatic cancer cells. Furthermore, we identified the
two putative NF-kB binding sites (kB/A and B) in the
bcl-xl promoter, and these two sites interact with
different NF-kB heterodimers and homodimers in
regulating bcl-xl gene expression. Taken together, our
results suggest the importance of various NF-kB/IkB
complexes in regulating anti-apoptotic genes.
Materials and methods
Cell culture
The human pancreatic adenocarcinoma cell lines AsPc-1,
Capan-1, and Panc-1 were obtained from the American Type
Cell Culture (Rockville, MD, USA). MDAPanc-28 cells were
obtained from Dr Marsha L Frazier (M.D. Anderson Cancer
Center). IKK+/+
, IKK1/a7/7
and IKK2/b7/7
MEF cells
were provided by Dr Inder M Verma (Salk Institute, La
Jolla, CA, USA). All cells were cultured in Dulbecco’s
modified Eagle’s medium (DMEM) (Gibco BRL, Gaithers-
burg, MD, USA) containing 10% fetal bovine serum, 100 U/
ml penicillin (Gibco BRL) and 10 mg/ml streptomycin (Gibco
BRL) in a 378C incubator at 5% CO2 atmosphere.
Retroviral infections of human pancreatic cancer cell lines
The CMV-Flag- IkBaM/puror
retroviral vector was generated
by replacing the XhoI – HindIII fragment containing the TRE
minCMV promoter and rTTA sequence of the pRetro-On/
puror
with XhoI – BamHI fragment of CMV-Flag-mutant
IkBa (IkBaM). The CMV-IkBaM was provided by Dr Inder
M Verma (Salk Institute, La Jolla, CA, USA). The CMV-
Flag-IkBaM/puror
and pRetro-On/puror
control retroviruses
were generated and infections were performed as described
previously (Naviaux et al., 1996). Forty-eight hours after
infection, pancreatic cancer cell lines were seeded in a 100-
mM dish at a density of 56105
cells in the culture medium
containing 500 mg puromycin (Clontech, Palo Alto, CA,
USA). Both control and IkBaM puromycin-resistant
pancreatic cancer cells were pooled for Flag-tagged IkBaM
expression analysis.
Western blot analysis
Cytoplasmic extracts were prepared as described by Andrews
and Faller (1991). Samples were denatured and subjected to
10% sodium dodecyl sulfate-polyacrylamide gel electrophor-
esis (SDS – PAGE). The resolved proteins were transferred to
an Immobilon-P membrane (Millipore, Bedford, MA, USA).
The membrane was blocked with 5% nonfat milk in
phosphate-buffered saline (PBS) containing 0.2% Tween-20
and incubated with affinity-purified mouse antibody against
Flag epitope, b-actin, rabbit antibody against IkBa, IkBb,
RelA (65), or bcl-xl (Santa Cruz Biotechnology Inc., Santa
Cruz, CA, USA). Membranes were washed in PBS containing
0.2% Tween-20 and probed with horseradish peroxidase-
coupled secondary goat anti-rabbit or mouse IgG antibodies
(Amersham, Arlington Heights, IL, USA). Proteins were
detected using a chemoluminescence kit (Amersham).
Northern blot analysis
Total RNA was isolated from the pancreatic cancer cell lines
by the acid guanidium thiocyanate phenol chloroform
extraction, as described by Chomczynski and Sacchi (1987).
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6517
Oncogene

9. Ten micrograms of total RNA was subjected to electrophor-
esis in 1% denaturing agarose gels, transferred to nylon
membranes (Strategene, La Jolla, CA, USA) and ultraviolet
radiation-crosslinked and baked. The membranes were
hybridized with 32
P-dCTP-labeled 0.57-kb EcoRI – BamHI
bcl-xl cDNA or 1.2-kb PstI glyceraldehyde-3-phosphate
dehydrogenase (GAPDH) cDNA probes. The Northern blots
analyses were performed, essentially, as described before
(Wang et al., 1999b).
Nuclear extracts and electrophoretic mobility shift assay
(EMSA)
EMSAs were performed as described (Wang et al., 1999a;
Andrews and Faller, 1991). Briefly, nuclear extracts were
prepared using the method of Andrews and Faller (1991).
After preparation, 10 mg of nuclear extract was incubated
with 1 mg of poly (dI-dC) (Pharmacia, Piscataway, NJ, USA)
in 10 ml of binding buffer (75 mM NaCl, 15 mM Tris-HCl,
pH 7.5, 1.5 mM EDTA, 1.5 mM DTT, 25% glycerol and
20 mg/ml BSA) for 30 min at 48C. 32
P-labeled double-
stranded oligonucleotides containing the kB site (underlined)
found in the HIV-1 LTR (5’-CTCAACAGAGGG-
GACTTTCCGAGAGGCCAT) were incubated as probes.
Unlabeled double-stranded oligonucleotides (50-fold molar
excess) containing the mutant HIV-kB (5’-CTCAACA-
GAGTTGACTTTTCGAGAGGCCAT-3’) and wild-type
HIV-kB were used for competition studies (Wang et al.,
1999a). For antibody supershifts, 2 ml of the polyclonal
antibodies against p65 and p50 (Santa Cruz Biotechnology)
were preincubated for 30 min at room temperature before the
probe was added. The probe was allowed to bind for 20 min
on ice. Reaction mixtures were analysed on 4% polyacryla-
mide gels containing 0.256TBE (22.5 mM Tris, 22.5 mM
borate, 0.5 mM EDTA, pH 8.0) buffer. Fold induction was
determined by phosphoimage analysis and the kB binding
intensities were normalized to those of Oct-1.
Site-directed mutagenesis
By using PCR-mediated site-directed mutagenesis, the wild-
type kB elements (5’-GGGGACTGCC-3’) (kB/A) and (5’-
GTGACTTTCC-3’) (kB/B) in bcl-xl promoter-CAT reporter
gene was substitutited with mutant kB sites (5’-
TGAGACTGCC-3’) and (5’-GTGACTTTTT-3’), respec-
tively, where underlined bases are the mutated sequences.
Polymerase chain reaction (PCR)-mediated site-directed
mutagenesis were performed using a double-stranded, site-
directed mutagenesis kit from Stratagene. Sequences of the
mutated kB sites were confirmed by DNA sequencing.
Transient transfections and reporter gene assay
An 1100-bp genomic fragment in the 5’ region of the bcl-xl
gene (710, 71090) was cloned using the Promoter Finder
Kit (Clontech), and sequenced. The 1100-bp (710, 71090)
and 560 (710, 7570) bcl-xl-promoter fragments were then
ligated upstream of a chloramphenicol acetyl transferase
reporter gene into the pblCAT-3 plasmid. Transient transfec-
tion was performed as described previously (Chiao et al.,
1994). Briefly, 1 – 2 mg of the 1.1-kb bcl-xl promoter-CAT,
0.5-kb bcl-xl promoter-CAT with wild-type kB sites, 0.5-kb
bcl-xl promoter-CAT with mutated kB/A or mutated kB/B
site, and 5 or 10 mg of RelA/NF-kB expression plasmids
(CMV-p65RelA) (Chiao et al., 1994) were cotransfected into
COS cells with 1 – 5 mg of b-actin LacZ, or b-actin Firefly-
luciferase and TK-Rellina luciferase reporter genes. Forty-
eight hours after transfection, cells were harvested, and the
CAT reaction was performed as previously described (Chiao
et al., 1994). Briefly, the relative transfection efficiencies were
determined by using the cotransfected LacZ expression
plasmid (1 mg, b-actin-LacZ), and subsequent b-galactosidase
activities in the cell extracts were used to normalize the
transfection efficiencies. Luciferase assays were performed
using Dual-LuciferaseTM
Reporter Assay System (Promega,
Madison, WI, USA) and according to the instructions from
the manufacturer. CAT activity (percent conversion to
acetylated chloramphenicol) was determined by phospho-
image analysis.
Flow cytometry analysis
Cells expressing green fluorescent protein (GFP) were
analysed by flow cytometry in FL3 channel. Quantification
of the GFP expression levels in FACS analysis was
determined by the geo mean on the CellQuest Software
(Palo Alto CA, USA), which represents the mean value of the
area covered by curve. All flow cytometry analyses were
repeated three times. Apoptosis in pancreatic cancer cell lines
was detected using flow cytometry as previously described
(Sambrook et al., 1989). Briefly, the control and treated cells
were collected at various time points following treatment,
washed with PBS, and fixed in 800 ml of 80% ethanol (EtOH)
overnight. Thirty minutes before flow cytometry, 300 ml of
propidium iodide solution (0.1% sodium citrate, 0.1%
Triton-X 100, 20 mg RNase/ml, 50 mg propidium iodide)
was added to the fixed cells and vortexed for resuspension.
FACS was performed according to standard protocol.
Acknowledgements
We are grateful to Dr Inder M Verma for generously
providing the IKK+/+
, IKK1/a7/7
and IKK2/b7/7
MEF
cells, Dr David McConkey for kindly providing PS-341.
We also thank Carol Kohn and Pat Thomas for editorial
assistance. This work was supported by grants from
National Cancer Institute CA73675, CA78778, CA75517,
the Lockton Fund for Pancreatic Cancer Research and
Institution Research Grant from M.D. Anderson Cancer
Center.
References
Andrews NC and Faller DV. (1991). Nucleic Acids Res., 19,
2499.
Baeuerle PA and Baltimore D. (1989). Genes Dev., 3, 1689 –
1698.
Baldwin Jr AS. (1996). Annu. Rev. Immunol., 14, 649 – 683.
Beg AA, Sha WC, Bronson RT, Ghosh G and Baltimore D.
(1995). Nature, 376, 167 – 170.
Blagosklonny MV, Schulte T, Nguyen P, Trepel J and
Neckers LM. (1996). Cancer Res., 56, 1851 – 1854.
Brown K, Gerstberger S, Carlson L, Franzoso G and
Siebenlist U. (1995). Science, 267, 1485 – 1488.
Bui NT, Livolsi A, Peyron JF and Prehn JHM. (2001). J. Cell
Biol., 152, 753 – 764.
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6518
Oncogene

10. Chen ZJ, Hagler J, Palombella VJ, Melandri F, Scherer D,
Ballard D and Maniatis T. (1995). Genes Dev., 9, 1586 –
1597.
Chen C, Edelstein LC and Gelinas C. (2000). Mol. Cell. Biol.,
20, 2687 – 2695.
Chiao PJ, Miyamoto S and Verma IM. (1994). Proc. Natl.
Acad. Sci. USA, 91, 28 – 32.
Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162,
156 – 159.
DiDonato JA, Hayakawa M, Rothwarf DM, Zandi E and
Karin M. (1997). Nature, 388, 548 – 554.
DiGiuseppe JA, Yeo CJ and Hruban RH. (1996). Adv. Anat.
Pathol., 3, 139 – 155.
Gilmore TD, Koedood M, Piffat KA and White D. (1996).
Oncogene, 13, 1367 – 1378.
Glasgow JN, Wood T and Perez-Polo JR. (2000). J.
Neurochem., 75, 1377 – 1389.
Heilker R, Freuler F, Vanek M, Pulfer R, Kobel T, Peter J,
Zerwes HG, Hofstetter H and Eder J. (1999a). Biochem-
istry, 38, 6231 – 6238.
Heilker R, Freuler F, Pulfer R, Di Padova F and Eder J.
(1999b). Eur. J. Biochem., 259, 253 – 261.
Herrmann JL, Beham A, Sarkiss M, Chiao PJ, Rands MT,
Bruckheimer EM, Brisbay S and McDonnell TJ. (1997).
Exp. Cell Res., 234, 442 – 451.
Herrmann JL, Briones Jr F, Brisbay S, Logothetis CJ and
McDonnell TJ. (1998). Oncogene, 17, 2889 – 2899.
Ishikawa H, Ryseck RP and Bravo R. (1996). Oncogene, 13,
255 – 263.
Ishikawa H, Claudio E, Dambach D, Raventos-Suarez C,
Ryan C and Bravo R. (1998). J. Exp. Med., 187, 985 – 996.
Jordan MA, Wendell K, Gardiner S, Derry WB, Copp H and
Wilson L. (1996). Cancer Res., 56, 816 – 825.
Kitajima I, Shinohara T, Bilakovics JDAB, Xu X and
Nerenberg M. (1992). Science, 258, 1792 – 1795.
Lee FS, Hagler J, Chen ZJ and Maniatis T. (1997). Cell, 88,
213 – 222.
Mercurio F, Zhu H, Murray BW, Shevchenko A, Bennett
BL, Li J, Young DB, Barbosa M, Mann M, Manning A
and Rao A. (1997). Science, 278, 860 – 866.
Moore BE and Bose HRJ. (1988). Virology, 162, 377 – 387.
Naviaux RK, Costanzi E, Haas M and Verma IM. (1996).
Virology, 70, 5701 – 5705.
Pardo OE, Arcaro A, Salerno G, Raguz S, Downward J and
Seckl MJJ. (2002). J. Biol. Chem., 277, 12040 – 12046.
Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular
Cloning. Cold Spring Harbor Laboratory Press: Cold
Spring Harbor, New York.
Sen R and Baltimore D. (1986). Cell, 47, 921 – 928.
Stroka DM, Badrichani AZ, Bach FH and Ferran C. (1999).
Blood, 93, 3803 – 3810.
Sylla BS and Temin HM. (1986). Mol. Cell. Biol., 6, 4709 –
4716.
Tsukahara T, Kannagi M, Ohashi T, Kato H, Arai M, Nunez
G, Iwanaga Y, Yamamoto N, Ohtani K, Nakamura M and
Fujii M. (1999). J. Virol., 73, 7981 – 7987.
Van Antwerp D, Martin SJ, Kafri T, Green DR and Verma
IM. (1996). Science, 274, 787 – 789.
Verma IM, Stevenson JK, Schwarz EM, Antwerp DV and
Miyamoto S. (1995). Genes Dev., 9, 2723 – 2735.
Wang CY, Mayo MW and Baldwin Jr AS. (1996). Science,
274, 784 – 787.
Wang W, Larry L, Evans DB, Abbruzzese J and Chiao PJ.
(1999a). Clin. Cancer Res., 5, 119 – 127.
Wang W, Abbruzzese J, Evans DB and Chiao PJ. (1999b).
Oncogene, 18, 4554 – 4563.
Woronicz JD, Gao X, Cao Z, Rothe M and Goeddel DV.
(1997). Science, 278, 866 – 869.
Yin MJ, Christerson LB, Yamamoto Y, Kwak YT, Xu S,
Mercurio F, Barbosa M, Cobb MH and Gaynor RB.
(1998). Cell, 93, 875 – 884.
Zandi E, Rothwarf DM, Delhase M, Hayakawa M and
Karin M. (1997). Cell, 91, 243 – 252.
Multiple NF-kB : IkB complexes regulate bcl-xl expression
QG Dong et al
6519
Oncogene