IκBα Overexpression in Human Breast Carcinoma MCF7 Cells Inhibits Nuclear Factor-κB Activation but Not Tumor Necrosis Factor-α-induced Apoptosis

Nuclear factor-κB (NF-κB) is one of major component induced by tumor necrosis factor-α (TNF), and its role in the signaling of TNF-induced cell death remains controversial. In order to delineate whether the involvement of NF-κB 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 IκBα that is resistant to both phosphorylation and proteolytic degradation and that behaves as a potent dominant negative IκBα protein. Our results demonstrated that the mutated IκBα was stably expressed in the transfected MCF7 cells and blocked the TNF-induced NF-κB nuclear translocation. Indeed, TNF treatment of these cells induced the proteolysis of only the endogenous IκBα but not the mutated IκBα. The nuclear NF-κB released from the endogenous IκBα within 30 min of TNF treatment was rapidly inhibited by the mutated IκBα. There was no significant difference either in cell viability or in the kinetics of cell death between control cells and the mutated IκBα transfected cells. Furthermore, electron microscopic analysis showed that the cell death induced by TNF in both control and mutated IκBα transfected cells was apoptotic. The inhibition of NF-κB translocation in mutated IκBα-transfected cells persisted throughout the same time course that apoptosis was occurring. Our data provide direct evidence that the inhibition of NF-κB did not alter TNF-induced apoptosis in MCF7 cells and support the view that TNF-mediated apoptosis is NF-κB independent.


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 IB␣ that is resistant to both phosphorylation and proteolytic degradation and that behaves as a potent dominant negative IB␣ protein. Our results demonstrated that the mutated IB␣ 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 IB␣ but not the mutated IB␣. The nuclear NF-B released from the endogenous IB␣ within 30 min of TNF treatment was rapidly inhibited by the mutated IB␣. There was no significant difference either in cell viability or in the kinetics of cell death between control cells and the mutated IB␣ transfected cells. Furthermore, electron microscopic analysis showed that the cell death induced by TNF in both control and mutated IB␣ transfected cells was apoptotic. The inhibition of NF-B translocation in mutated IB␣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 IB, which traps NF-B. Following degradation of IB, the heterodimer translocates to the nucleus, where it participates in transcriptional regulation of numerous genes (3)(4)(5). Several proteins, collectively termed IB, share the property of retaining NF-B dimers and preventing their translocation to the nucleus (6). To date, the most extensively studied IB protein is IB␣ (37 kDa) encoded by the human MAD-3 gene or its homologues in different species (7). The mechanisms that lead to the degradation of IB proteins are poorly understood, but involve changes in the phosphorylation state of IB (8,9). Two serines in the N-terminal domain of IB␣, Ser-32 and Ser-36, were shown to be critical for IB␣ stability. Substitution of Ser-32 and Ser-36 by alanine residue rendered IB␣ undegradable by cellular activators (10 -12). Among the many proteins exhibiting IB function, IB␣ 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/IB␣ 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 -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 80amino 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 TNFsensitive MCF7 cells. In the present report, we describe the consequences of the stable expression of the mutated IB␣ on NF-B activation and TNF-mediated cell killing. * This work was supported in part by grants (to S. C.) from the INSERM and from the Association pour la Recherche sur le Cancer (ARC 6627) and from EEC BIOTECH (BIO2-CT92-0130) (to M. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Recipient of grants from the Institut de Formation Supérieure BioMédicale (IFSBM) and the Association pour la Recherche sur le Cancer (ARC).

Transfection of MCF7 Cells with Mutated MAD-3 cDNA and Cell
Culture-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 g/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 g/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% CO 2 .
Determination of Cell Viability-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 (%) ϭ 100 ϫ (A 1 /A 0 ), cell lysis (%) ϭ 1 Ϫ cell viability (%), where A 1 and A 0 were the absorbance obtained from TNF-treated and untreated cells, respectively. The mean value of quadruplicate was used for analysis. Highly purified (Ͼ99%) recombinant TNF (specific activity 6.63 ϫ 10 6 units/mg of protein) was kindly provided by A. G. Knoll (Luwigshafen, Germany).
Nuclear Extracts and Electrophoretic Mobility Shift Assays (EMSAs)-Transfected MCF7 cells (15 ϫ 10 6 ) 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, 5Ј-end-labeled with [␥-32 P]ATP using the T4 kinase (31).
Southern Blot and Northern Blot Analysis-Genomic DNA was extracted from transfected MCF7 cells and digested by HindIII/XabI enzymes before electrophoresed (10 g) in a 0.8% agarose gel and transferred to nylon membrane hybond-N (Amersham Corp.). Total RNA (15 g) 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 [␣-32 P]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.
Western Blotting-Determination of IB␣ content in MCF7 cell clones was performed by Western blotting of cytosolic protein extracts using a specific monoclonal antibody for IB␣, MAD10B antibody (32). The MAD10B antibody recognizes both wild-type and mutated IB␣. 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 g) 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. IB␣ 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).
Morphological Examination by Electron Microscopy-Control and TNF-treated cells (1 ϫ 10 7 ) fixed with 2% glutaraldehyde in phosphatebuffered 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.

RESULTS AND DISCUSSION
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 (Ͼ75%) 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 IB␣ gene, while the mutated IB␣ transfected clones (MAD-1001, -1706, -1904, -1906) contained an additional band representative of the mutated exogenous IB␣ gene. EMSA analysis (Fig. 2B) demonstrated that the introduction of exogenous IB␣ 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 IB␣. 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 IB␣.
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 IB␣ itself (35,36). IB␣ 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 IB␣ and the endogenous IB␣ in the transfected cell lines, we tested the cytosol of control cells (pcN-183) and mutant IB␣-transfected cells (MAD-1906) for IB␣ 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 IB␣-specific monoclonal antibody evidenced a major band with a electrophoretic mobility of 36 kDa, corresponding to wild-type IB␣ (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 IB␣ correlated with nuclear translocation of NF-B in these cells (Fig. 3A, lane 2). After 4 h of TNF treatment, newly synthesized IB␣ was detected in cytosols, witnessing the early NF-B activation. The de novo synthesized IB␣ 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 IB␣ or an IB␣ unrelated protein that was not evidenced in other cell types. Unexpectedly, a 24-h treatment with TNF showed again a reduced amount of IB␣ (Fig. 3B,  lane 4). Thus, in the control cells the IB␣ 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 IB␣ from the control T cell line and the 36-kDa band from pcN-183 cells, corresponding therefore to the endogenous IB␣. The amount of the reduced mobility band was equal to the amount of the wild-type IB␣. In the T cell line used here as control, transfection of the mutated MAD-3 cDNA also gen-  3 and 7), and 24 h (lanes 4 and 8). Nuclear proteins (15 g) extracted from untreated cells or TNF-treated cells were subjected to electrophoretic mobility shift assay as described under "Experimental Procedures." B, Western blot analysis of IB␣ content in transfected cells. Proteins (50 g) from the cytosolic fraction of cells used in A were subjected to 10% SDS-polyacrylamide gel electrophoresis and transferred on to polyvinylidene difluoride membrane. Wild-type (wt) and mutated IB␣ proteins were visualized with IB␣-specific antibody as described under "Experimental Procedures." Human T cell line (lane 9) was used as wild-type IB␣ control.

FIG. 2.
A, Southern blot analysis of IB␣-transfected MCF7 cells. MCF7 cells were transfected by control pcDNA3 vector (pcN-) or mutated IB␣ (MAD-) and the stable transfectants were obtained. Genomic DNA was extracted from transfected clones, digested by HindIII/XbaI, and subjected to a 0.8% agarose gel as described under "Experimental Procedures." Digested genomic DNA (10 g/lane) was then transferred to nylon membrane and hybridized with 32 P-labeled specific IB␣ cDNA probe. B, EMSA study of transfected cells. Transfected MCF7 cells (10 ϫ 10 6 ) were incubated for 90 min in the presence or absence of TNF (50 ng/ml). Nuclear proteins (15 g) extracted from untreated cells (Ϫ) or TNF-treated cells (ϩ) were subjected to electrophoretic mobility shift assay as described under "Experimental Procedures." erated 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 IB␣ and a total and persistent disappearance of the endogenous IB-␣. Thus, the mutated IB-␣ 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 IB␣ in these cells. However, the lack of resynthesis of the endogenous IB␣ at the 4-h or later time points suggests that this faint NF-B nuclear translocation was not sufficient to enhance IB␣ 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 IB␣ 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 IB␣ within 30 min of TNF treatment was rapidly inhibited by the mutated IB␣. The lack of further endogenous IB␣ synthesis may be the consequence of the inhibition of NF-B, since the IB␣ 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 IB␣ but not the mutated IB␣. Additionally, the endogenous IB␣ served as a marker of NF-B activity. The results shown in Fig. 3B demonstrated that in the MAD-1906 cells, the endogenous IB␣ was not re-synthesized in response to TNF, in contrast to what occurred in control cells. We conclude from these results that the mutated IB␣ 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, andMAD-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, IB␣ 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 andMAD-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 coordi-nately 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.   1 and 5) or in the presence of TNF (50 ng/ml) for 2 h (lanes 2 and 6), 48 h (lanes 3 and 7), and 72 h (lanes 4 and 8). Nuclear proteins (10 g) extracted from untreated cells or TNF-treated cells were subjected to electrophoretic mobility shift assay as described under "Experimental Procedures."