(Received for publication, August 5, 1996, and in revised form, October 17, 1996)
From the INSERM CJF 9411 "Cytokines et
Immunité Antitumorale," Institut Gustave Roussy, 94805 Villejuif, and the ¶ Laboratoire d'Immunologie Cellulaire et
Tissulaire, CNRS URA 625, Hôpital
Pitié-Salpêtrière, 75013 Paris, France
Nuclear factor-B (NF-
B) is one of major
component induced by tumor necrosis factor-
(TNF), and its role in
the signaling of TNF-induced cell death remains controversial. In order
to delineate whether the involvement of NF-
B activation is required
for triggering of the apoptotic signal of TNF, we inhibited the nuclear
translocation of this transcription factor in TNF-sensitive MCF7 cells
by introducing a human MAD-3 mutant cDNA coding for a mutated
I
B
that is resistant to both phosphorylation and proteolytic
degradation and that behaves as a potent dominant negative I
B
protein. Our results demonstrated that the mutated I
B
was stably
expressed in the transfected MCF7 cells and blocked the TNF-induced
NF-
B nuclear translocation. Indeed, TNF treatment of these cells
induced the proteolysis of only the endogenous I
B
but not the
mutated I
B
. The nuclear NF-
B released from the endogenous
I
B
within 30 min of TNF treatment was rapidly inhibited by the
mutated I
B
. There was no significant difference either in cell
viability or in the kinetics of cell death between control cells and
the mutated I
B
transfected cells. Furthermore, electron
microscopic analysis showed that the cell death induced by TNF in both
control and mutated I
B
transfected cells was apoptotic. The
inhibition of NF-
B translocation in mutated I
B
-transfected
cells persisted throughout the same time course that apoptosis
was occurring. Our data provide direct evidence that the inhibition of
NF-
B did not alter TNF-induced apoptosis in MCF7 cells and support
the view that TNF-mediated apoptosis is NF-
B independent.
Cytokine-dependent activation of transcription factors
such as NF-B is one of the mechanisms by which signals are
transmitted from the extracellular surface to the nucleus to enhance
the transcription of specific genes (1, 2). The activation of
cytoplasmic NF-
B heterodimer consisting of p50 and p65 polypeptides
has been shown to require the degradation of a cytoplasmic inhibitor
I
B, which traps NF-
B. Following degradation of I
B, the
heterodimer translocates to the nucleus, where it participates in
transcriptional regulation of numerous genes (3, 4, 5). Several proteins, collectively termed I
B, share the property of retaining NF-
B dimers and preventing their translocation to the nucleus (6). To date,
the most extensively studied I
B protein is I
B
(37 kDa) encoded
by the human MAD-3 gene or its homologues in different species (7). The
mechanisms that lead to the degradation of I
B proteins are poorly
understood, but involve changes in the phosphorylation state of I
B
(8, 9). Two serines in the N-terminal domain of I
B
, Ser-32 and
Ser-36, were shown to be critical for I
B
stability. Substitution
of Ser-32 and Ser-36 by alanine residue rendered I
B
undegradable
by cellular activators (10, 11, 12). Among the many proteins exhibiting
I
B function, I
B
is the only inhibitor that in response to cell
stimulation dissociates from the NF-
B heterodimer complex, with
kinetics matching NF-
B translocation to the nucleus (13, 14). It was therefore suggested that the inducible activation of NF-
B is mainly
regulated by NF-
B/I
B
dissociation (6, 9, 15).
Tumor necrosis factor- (TNF),1
originally described for its antitumor activity, is one of the
cytokines known to activate NF-
B within minutes, leading to the
transcriptional activation of various important cellular and viral
genes (16, 17). The activation of NF-
B is considered integral to the
transfer of the TNF signal to the nucleus (18). Both TNF receptors (p55 TNF-R1 and p75 TNF-R2) independently mediate NF-
B activation by TNF
(19, 20, 21, 22). The nature of signaling mechanisms mediating the effects of
TNF on NF-
B activation remains poorly defined. It has been shown
that TNF first activates phosphatidylcholine-specific phospholipase C
and leads to the sequential activation of an acidic sphingomyelinase
and the production of ceramide, which in turn causes the activation of
NF-
B (23, 24). Mutagenesis studies have identified an 80-amino acid
region within the cytoplasmic domain of p55 TNF-R1 that is required for
initiation of both apoptosis and NF-
B activation (25). However,
several recent studies debated the involvement of NF-
B activation in
TNF-induced apoptosis. The report of Dbaibo et al. (26)
suggested that ceramide mediated the effects of TNF on growth
inhibition of Jurkat lymphoblastic leukemia cells, but was unable to
activate NF-
B. In addition, TNF was reported to be capable of
activating NF-
B in different cell models resistant to its cytotoxic
action (27, 28). In order to directly examine whether the NF-
B
activation is an essential requirement for triggering the apoptotic
signal of TNF, we chose an approach based on the inhibition of the
translocation of this transcription factor by introducing a
dominant-negative human MAD-3 mutant construct into the TNF-sensitive
MCF7 cells. In the present report, we describe the consequences of the
stable expression of the mutated I
B
on NF-
B activation and
TNF-mediated cell killing.
The MAD-3 double point mutant (positions 32 and 36)
construct was described by Traeckner et al. (11) and was a
kind gift by Patrick A. Baeuerle, Tularik, Inc., San Francisco. The
empty vector used for the generation of control cells was the
pcDNA3 purchased from Invitrogen. The transfection of human breast
carcinoma cell line MCF7 with the expression constructs was performed
by the calcium phosphate precipitation method (29). 1000 cells were
plated per 10-cm tissue culture plates. After 10-14 days selection in
growth medium containing 200 µg/ml G418 (Sigma), four to five resistant colonies were isolated from each plate and
examined for IB
expression by Southern blot analysis, and the
positive clones were maintained in culture medium with 100 µg/ml G418
for more than 2 months. All cell lines were routinely cultured in RPMI
1640 medium containing 5% fetal calf serum, 1% penicillin-streptomycin, 1% L-glutamine at 37 °C in a
humidified atmosphere with 5% CO2.
Cells viability was
determined using crystal-violet staining method as described previously
(28). Absorbance (A), which was proportional to cell
viability, was measured at 540 nm. TNF-mediated cell lysis was assessed
by comparing the viability of untreated cells with that of treated
cells using the following calculation: cell viability (%) = 100 × (A1/A0), cell lysis
(%) = 1 cell viability (%), where A1
and A0 were the absorbance obtained from TNF-treated and untreated cells, respectively. The mean value of
quadruplicate was used for analysis. Highly purified (>99%) recombinant TNF (specific activity 6.63 × 106
units/mg of protein) was kindly provided by A. G. Knoll (Luwigshafen, Germany).
Transfected MCF7 cells (15 × 106)
were incubated in the presence or absence of 50 ng/ml TNF. The cells
were then trypsinized and washed with phosphate-buffered saline.
Nuclear extracts were prepared according to the procedure of Dignam
et al. (30). Gel mobility shift assays were performed with a
synthetic double-stranded 31-mer oligonucleotide containing the B
sequences of the human immunodeficiency virus long terminal repeat,
5
-end-labeled with [
-32P]ATP using the T4 kinase
(31).
Genomic DNA was
extracted from transfected MCF7 cells and digested by
HindIII/XabI enzymes before electrophoresed (10 µg) in a 0.8% agarose gel and transferred to nylon membrane hybond-N (Amersham Corp.). Total RNA (15 µg) was electrophoresed in a 1.2% agarose gel and transferred to nitrocellulose membrane hybond-C (Amersham). The membranes were hybridized overnight at 42 °C with the probe labeled with [-32P]dCTP using a megaprime
DNA labeling system (Amersham). The hybridized membranes were washed
and exposed to Hyperfilm-MP (Amersham). The blot of RNA was stripped by
boiling in 0.1% SDS and probed again with
-actin probe to confirm
equal loading of RNA samples.
Determination of IB
content in MCF7
cell clones was performed by Western blotting of cytosolic protein
extracts using a specific monoclonal antibody for I
B
, MAD10B
antibody (32). The MAD10B antibody recognizes both wild-type and
mutated I
B
. The cytosolic fractions of MCF7 cells used for EMSA
analysis (as described previously) were denatured by boiling in SDS and
2-mercaptoethanol. Equal amounts of protein extracts (50 µg) were
subjected to 10% polyacrylamide gel electrophoresis in denaturing
conditions (33). Fractionated proteins were transferred onto
polyvinylidene difluoride membranes using the Hoeffer semi-phor system.
The efficiency of the electrotransfer was assessed by Ponceau Red
staining of the polyvinylidene difluoride membranes. I
B
protein
was revealed with MAD10B hybridoma supernatant diluted 400-fold. The
antigen-antibody complex was visualized by enhanced chemiluminescence
method (ECL, Amersham) using the horseradish peroxidase-coupled
anti-mouse antibody (Bio-Rad).
Control and TNF-treated cells (1 × 107) fixed with 2% glutaraldehyde in phosphate-buffered saline were pelleted at low speed. The pellet was washed in Sorensen buffer (67 mM phosphate buffer, pH 7.4), post-fixed in 2% osmium tetroxide, dehydrated with graded ethanol and propylene oxide, and included in "Epon" resin by usual techniques. Sections of cells were stained with uranyl acetate and lead citrate and observed with a Zeiss EM 902 electron microscope. Enhanced contrast was obtained by selecting elastic electrons using the slit of a spectrophotometer.
The TNF-sensitive human breast carcinoma MCF7 cell line was used
in this study to examine the effect of the inhibition of NF-B
activation on its susceptibility to the cytotoxic action of TNF. As
shown in Fig. 1A, MCF7 cells were highly
sensitive to the cytotoxicity of TNF. Following 72 h of exposure
to TNF, optimal lysis (>75%) of MCF7 cells was obtained at 50-100
ng/ml of TNF. The results of EMSA indicate that in the absence of TNF, MCF7 cells showed no constitutive activation of NF-
B (Fig. 1B, lane 1). After 90-min incubation, TNF induced in these cells a significant activation of NF-
B (Fig. 1B, lane 2). The
specific binding of NF-
B to DNA could be abrogated with an excess of
unlabeled probe (Fig. 1B, lane 3). The fast migrating
B-binding protein detected in both TNF-treated and untreated cells
was not selective for
B sequence (Fig. 1B, lanes 1 and
2), since its binding was not abrogated by an excess of
unlabeled probe (Fig. 1B, lane 3).
In order to inhibit TNF-induced NF-B translocation to the nucleus,
we transfected MCF7 cells with the mutated MAD-3 cDNA which was
unsusceptible to phosphorylation at positions 32 and 36 and which was
found to resist degradation in transient transfections (11). The stable
transfected clones of the control vector pcDNA3 (pcN-) and of the
mutated MAD-3 gene (MAD-) were first screened by Southern blot
analysis. As shown in Fig. 2A, the control
clones (pcN-112 and pcN-183) contained only the endogenous wild-type I
B
gene, while the mutated I
B
transfected clones (MAD-1001, -1706, -1904, -1906) contained an additional band representative of the
mutated exogenous I
B
gene. EMSA analysis (Fig. 2B)
demonstrated that the introduction of exogenous I
B
mutant led to
a significant suppression of TNF-induced NF-
B activation in the four
representative MAD-3-transfected clones as compared with the level of
NF-
B translocation after 90-min treatment with TNF in the control
pcN-112 and pcN-183 cells.
To examine the stability and the efficiency of the NF-B inhibition
in MAD-3 mutant transfectants during a long term incubation with TNF,
kinetic analysis of NF-
B translocation was performed. The treatment
of control pcN-183 cells with TNF for 30 min (Fig. 3A, lane
2) resulted in a significant NF-
B translocation
that further persisted and accumulated until after at least 24-h
treatment (Fig. 3A, lanes 3 and 4). In contrast,
in MAD-1906 cells, after 30-min incubation with TNF (Fig. 3A,
lane 6), only marginally activated NF-
B was observed in the
nuclear extract, that probably corresponded to the NF-
B released
from rapidly degraded endogenous I
B
. No further activation of
NF-
B could be detected after 4 h (Fig. 3A, lane 7)
or 24 h (Fig. 3A, lane 8) treatment with TNF, thus
suggesting that the NF-
B was inhibited by a stabilized association with the mutated I
B
.
It has been shown in various cell lines that the endogenous IB
is
rapidly degraded as a consequence of cell stimulation by TNF or phorbol
esters (3, 5, 13, 34). As a result, NF-
B translocates to the
nucleus, where it participates in the initiation of the transcription
of numerous genes. One of the target genes of NF-
B is I
B
itself (35, 36). I
B
degradation is followed by its de
novo synthesis as a consequence of the early NF-
B activation
(4). In an attempt to compare the stability of the transgenic I
B
and the endogenous I
B
in the transfected cell lines, we tested
the cytosol of control cells (pcN-183) and mutant I
B
-transfected
cells (MAD-1906) for I
B
expression by Western blotting. The
cytosols were obtained from the same cells as those whose nuclear
extracts were used in the EMSA experiment in Fig. 3A. In
control pcN-183 cells, in the absence of TNF, the I
B
-specific
monoclonal antibody evidenced a major band with a electrophoretic
mobility of 36 kDa, corresponding to wild-type I
B
(Fig.
3B, lane 1). The treatment of pcN-183 cells with
TNF for 30 min resulted in a dramatic reduction of the 36-kDa band. This reduction of I
B
correlated with nuclear translocation of NF-
B in these cells (Fig. 3A, lane 2). After 4 h of
TNF treatment, newly synthesized I
B
was detected in cytosols,
witnessing the early NF-
B activation. The de novo
synthesized I
B
did not inhibit NF-
B nuclear translocation
(Fig. 3A, lane 3). An additional faint band of 38 kDa was
also detected by the antibody in these cells at the 4-h point (Fig.
3B, lane 3) but not in a control T cell line (Fig. 3B,
lane 9). This 38-kDa band could correspond to a transient
phosphorylated I
B
or an I
B
unrelated protein that was not
evidenced in other cell types. Unexpectedly, a 24-h treatment with TNF
showed again a reduced amount of I
B
(Fig. 3B, lane 4).
Thus, in the control cells the I
B
underwent at least two cycles
of proteolysis/resynthesis in 24 h of TNF treatment. The second
proteolytic step could explain the increased amount of NF-
B after
24 h of TNF activation as compared with 30-min activation (Fig.
3A, lanes 2 and 4). In the MAD-1906 cells, two
bands (36 and 38 kDa) were detected by the antibody in the absence of
TNF (Fig. 3B, lane 5). The 36-kDa band comigrated with the
I
B
from the control T cell line and the 36-kDa band from pcN-183
cells, corresponding therefore to the endogenous I
B
. The amount
of the reduced mobility band was equal to the amount of the wild-type I
B
. In the T cell line used here as control, transfection of the
mutated MAD-3 cDNA also generated a reduced electrophoretic mobility product.2 Thus the slower
migrating band probably corresponded to the product of the mutated
MAD-3 cDNA. As already mentioned, treatment of MAD-1906 cells with
TNF resulted in a small increase in the amount of the mutated I
B
and a total and persistent disappearance of the endogenous I
B-
.
Thus, the mutated I
B-
was not degraded in response to TNF.
Treatment of MAD-1906 cells with TNF for 30 min induced a faint NF-
B
translocation (Fig. 3A, lane 6). This was probably due to
the degradation of the endogenous I
B
in these cells. However, the
lack of re-synthesis of the endogenous I
B
at the 4-h or later
time points suggests that this faint NF-
B nuclear translocation was
not sufficient to enhance I
B
transcription. At the 4- and 24-h
time points, NF-
B was no longer detectable in nuclei of MAD-1906
cells (Fig. 3A, lanes 7 and 8). Concomitantly,
the amount of mutated I
B
was increased and persisted in the
cytosolic fraction of these cells (Fig. 3B, lanes 7 and 8). These observations suggest that the nuclear NF-
B
released from the endogenous I
B
within 30 min of TNF treatment
was rapidly inhibited by the mutated I
B
. The lack of further
endogenous I
B
synthesis may be the consequence of the inhibition
of NF-
B, since the I
B
itself is one of the target genes of
NF-
B.
Together, these results demonstrated that the mutated IB
was
stably expressed in the transfected MCF7 cells and that TNF treatment
of the MAD-1906 cells induced the proteolysis of only the endogenous
I
B
but not the mutated I
B
. Additionally, the endogenous
I
B
served as a marker of NF-
B activity. The results shown in
Fig. 3B demonstrated that in the MAD-1906 cells, the endogenous I
B
was not re-synthesized in response to TNF, in contrast to what occurred in control cells. We conclude from these results that the mutated I
B
inhibited efficiently NF-
B nuclear translocation and activation in the MAD-1906 cells.
To further test the functional effect of the inhibition of NF-B
translocation in the MAD-3-transfected clones, we studied the
expression of one of the TNF-inducible genes, mitochondrial manganous
superoxide dismutase (37), in these cells. The mitochondrial manganous
superoxide dismutase gene presents potential
B site(s) in its
promoter region and the induction of its expression is closely
associated with NF-
B activation by TNF (38, 39). The results of
Northern blot analysis (Fig. 4) showed that TNF significantly induced the expression of mitochondrial manganous superoxide dismutase mRNA in control clones (pcN-112 and pcN-183). In contrast, no induction of this gene was observed in the four mutant
MAD-3-transfected clones (MAD-1001, MAD-1706, MAD-1904, and MAD-1906).
This correlated with the inhibition of the NF-
B activation in these
MAD-3-transfected cells. Therefore, at least two known NF-
B target
genes, I
B
and mitochondrial manganous superoxide dismutase, were
negatively regulated in mutated MAD-3-transfected clones, indicating a
functional inhibition of NF-
B in these cells.
A kinetic study was then performed to determine the sensitivity of the
mutated MAD-3-transfected cells to the cytotoxic effect of TNF. When
the transfected cells were incubated with 50 ng/ml of TNF during 6-72
h (Fig. 5), there was no significant difference between
control (pcN-183) and the MAD-3-transfected (MAD-1904 and MAD-1906)
cells, neither in the cell viability nor in the kinetics of cell death.
After 48-h incubation with TNF, we even observed a slightly more
elevated cell lysis in the two MAD-3 transfected clones as compared
with the control pcN-183 cells. Furthermore, in order to examine the
sensitivity of these transfected cells to short term treatment of TNF
and the nature of cell death, the cells were treated with TNF for
24 h, and the electron micrograph analysis was performed. The
results (Fig. 6) showed that the cell death induced by
TNF in both control and MAD-3-transfected clones was apoptotic with the
dense and vacualized cytoplasm and the condensation of the chromatin
along the nuclear membrane. To verify if inhibition of NF-B
persisted throughout the whole time course of the apoptotic process, we
tested the nuclear extract of control (pcN-183) and MAD-3-transfected
(MAD-1906) cells for NF-
B binding activity by EMSA between 24 and
72 h of TNF treatment. The treatment of control pcN-183 cells with
TNF for 48 h (Fig. 7, lane 3) resulted in an accumulated NF-
B translocation as compared with 2-h short time
treatment with TNF (Fig. 7, lane 2). The activation of
NF-
B persisted until 72 h (Fig. 7, lane 4) but at a
lower level due to the important cell lysis at this time point. In
contrast, no activation of NF-
B could be detected in MAD-1906 cells
after 2-h treatment with TNF (Fig. 7, lane 6), neither at
48 h nor at 72 h (Fig. 7, lanes 7 and
8). The possibility of a re-activation of NF-
B in
MAD-3-transfected cells during a prolonged TNF incubation (24 h -72 h)
can therefore be ruled out. The nuclear translocation of NF-
B is
blocked in these cells throughout the same time course that apoptosis
is occurring. These data clearly indicated that the inhibition of
NF-
B activation had no effect on TNF-mediated apoptotic cell
death.
It is admitted that TNF signaling involves multiple second messenger
pathways that function independently or coordinately to transduce
distinct biological responses of TNF. Our results directly demonstrated
that NF-B activation is not required for induction of apoptosis by
TNF. This is in agreement with the reports of Hsu et al.
(40) indicating that TNF-R1-associated death domain protein (TRADD)
directly interacts with one of the TNF-R2-associated factors (TRAF2)
and the Fas-associated factor (FADD) to induce NF-
B activation and
apoptosis, respectively. A TRAF2 mutant acts as a dominant-negative
inhibitor of TNF-mediated NF-
B activation, but does not affect
TNF-induced apoptosis. Conversely, a FADD mutant is a dominant-negative
inhibitor of TNF-induced apoptosis, but does not inhibit NF-
B
activation. Thus, it is suggested that TNF-R1 may utilize distinct
TRADD-dependent mechanisms to activate signaling pathways
for NF-
B activation and apoptosis. TNF has been shown to mediate its
action through activation of several other transcriptional factors,
including c-Jun/AP-1, c-Fos, c-Myc, IRF-1, and early growth response
gene (Egr-1) (for review, see Ref. 17). However, Egr-1, c-Fos, c-Jun,
and c-Myc have been implicated in cell proliferation. Therefore, the
nuclear factors distinct from NF-
B that are capable of mediating
TNF-induced apoptotic signal still remain to be identified.
We thank R. Stancou for her technical
assistance and acknowledge D. Coulaud (from the Laboratoire de
Microscopie Cellulaire et Moléculaire, CNRS, URA 147) for helping
us in performing the electron microscopy experiments and Dr. X.-J. Ma
(Wistar Institute) for critical reading of the manuscript. We are
grateful to Prof. P. A. Baeuerle (Tularik, Inc. San Francisco) for
providing mutated IB
expression construct.