Introduction

To respond and adapt to changes in its external environment, such as exposure to growth factors, cytokines or stress-inducing stimuli such as ionising radiation, the cell must initiate a program of regulated gene expression. This response depends on many factors such as the cell type and the number and intensity of the new stimuli. To a large extent these changes are mediated by nuclear DNA-binding proteins, which function in a combinatorial manner to activate or repress the transcription of target genes. Many cellular stimuli result in activation of both the tumour suppressor p53 and NF-κB.

TNFα initiates a broad range of cellular responses including growth stimulation of various non-transformed cell lines (e.g. fibroblasts), and, in contrast, G1-growth arrest of many other transformed cell lines (e.g. the human breast carcinoma MCF-7) or apoptosis.1,2,3,4 At a molecular level, TNFα binds the TNF receptor and induces its trimerization. The ligand-bound activated TNF receptor associates with the cytoplasmic adaptor protein TRADD, which functions as a platform to recruit several signalling molecules (reviewed in5). The active TNF receptor complexes can interact with caspase proteases via TRADD and FADD to induce apoptosis.6,7 TNF receptor complexes also interact, via TRADD, with TRAF2 and RIP proteins, which are involved in the activation of JNK/AP-1 and NF-κB.6,7,8,9 Different studies have shown that TNF-activated NF-κB plays an anti-apoptotic role against TNF cytotoxicity.10,11 In fact, NF-κB regulates a set of anti-apoptotic genes including members of the Bcl-2 family, cellular inhibitors of apoptosis protein (c-IAP2 and c-IAP1), A20 and superoxide dismutase 2 (SOD2) (reviewed in12). Interestingly, NF-κB also regulates the expression of the tumour suppressor gene p53,13 although the role of this regulation is not well understood.14,15

A large number of evidence places the p53 suppressor gene at a nodal point in the regulation of diverse stress responses. Indeed, a number of cellular stresses including DNA damage, heat shock, hypoxia, hyperoxia, metabolic changes, oncogene activation, and others, converge on p53 protein and cause its activation and stabilization.16 p53 is a transcription factor which exerts its function mainly through binding to a specific DNA responsive element. Activation of p53 by post-translational events (reviewed in17) increases the affinity of the protein for its DNA consensus sites, leading to an increased expression of a set of target genes. Accumulation of the protein generally results from an extended half-life brought about by post-translational modifications that could be distinct from those involved in p53 activation.18 But accumulation of p53 could also arise from an up-regulation of its gene transcription rate.13

A number of p53-regulated genes are associated with induction of apoptosis or cell cycle arrest.19 One of the most important is the p21/Waf1 gene, an inhibitor of several cyclin-dependent kinases, and a target for diverse signals that induce growth arrest and differentiation. While p53 is not required for induction of p21/Waf1 transcription during development and in most tissues of the adult mouse, the p53-dependent regulation of p21/Waf1 is critical for the response to DNA damage (reviewed in20). Activation of p53 following exposure to DNA damage caused by UV- or γ-irradiation or chemotherapeutic agents leads to p53-dependent transcription of p21/Waf1. Cells without functional p53 cannot induce p21/Waf1.21 However, there is accumulating evidence that the expression of the p21/Waf1 gene can be induced upon stimulation through a pathway, which does not require p53 (reviewed in20). Kobayashi et al. showed an increased rate and accumulation of p21/Waf1 in TNFα-treated cells expressing a transcription-inactive p53 mutant.22 In contrast, Jeoung et al reported that in MCF-7 cells induction of p21/Waf1 protein is closely related to p53 induction. Based on their results, they suggested that induction of p21/Waf1 protein by TNFα is mediated by p53.4 By analyzing the response to TNFα of cell lines derived from MCF-7 and expressing either the HPV E6 or DD protein, two p53 inhibitors, or IκBα, an inhibitor of NF-κB, we provide evidence that p21/Waf1 is not regulated by p53 but by NF-κB in TNFα-treated MCF-7 cells. This observation leads us to demonstrate that TNFα induces accumulation but not activation of p53.

Results

Induction of p21/Waf1 gene expression in MCF-7 following TNFα treatment is independent of p53 expression

In agreement with published data,23,24 results reported in Figure 1A show that treatment of MCF-7 cells with TNFα (10 ng/ml) resulted in a transient arrest of cells at the G0/G1 phase followed by apoptosis. The cyclin-dependent kinase inhibitor p21/Waf1 is one regulator of the cell cycle that can cause cell growth arrest.25 Under our experimental conditions, p21/Waf1 mRNA was significantly increased by TNFα, 4 h after treatment (Figure 1B). These results corroborate those obtained by Jiang and Porter.24 Previous results have indicated that p53 accumulation preceded activation of p21/Waf1 in response to TNFα4 suggesting a causal relationship between p53 activation, p21/Waf1 stimulation and G1 arrest. To test the possible implication of p53 in p21/Waf1 accumulation in TNFα-treated MCF-7 cells, the induction of p21/Waf1 gene expression by TNFα was tested in MCF-7/E6, a MCF-7 derived cell line stably transfected with a cDNA encoding HPV16/E6 protein.26 Expression of E6 protein in MCF-7/E6 cells has been shown to impair the p53-dependent p21/Waf1 accumulation induced by γ-radiation.26 The viral protein E6 promotes p53 degradation by the ubiquitination pathway.27 This degradation accounts for the lower steady-state level of p53 protein in MCF-7/E6 as compared to MCF-7/neo (Figure 1C, t0). This difference was enhanced following TNFα treatment, which significantly increased the p53 protein level in MCF-7 cells but not in MCF-7/E6 cells (Figure 1C, t4 and 16). Unexpectedly, p21/Waf1 mRNA (Figure 1B and C) and protein (Figure 1D) levels were increased in TNFα-treated MCF-7/E6 as strongly as in MCF-7/neo, suggesting that TNFα induces p21/Waf1 accumulation via a p53-independent pathway. To confirm these results, we established the MCF-7/DD-TO cell line, a MCF-7 derived cell line engineered to exhibit a tetracycline inducible expression of p53DD, a dominant negative truncated p53 mutant. Two independent clones were analyzed in parallel. As already reported,28 the p53DD expression stabilizes MCF-7-endogenous wt-p53 (Figure 2A, lanes ‘+tet’), but inhibits its transcriptional activity as shown by transient expression of luciferase reporter gene under the control of a p53 responsive element (Figure 2B). However, the p53DD expression (Figure 2C, lanes ‘+tet’) had no effect on the increase of p21/Waf1 mRNA level brought by TNFα treatment. These results further support the assumption that TNFα stimulates the p21/Waf1 gene expression by a p53-independent pathway.

Figure 1
figure 1

Effect of TNFα on MCF-7 cell cycle and p21/WAF1 gene expression. (A) TNFα induces a transient G1-growth arrest in MCF-7 cells. Cells were mock-treated (Control) or treated with TNFα (50 ng/ml) for 30 and 56 h. The DNA content was analyzed by flow cytometry. (B) and (C) TNF-induced p21/WAF1 mRNA accumulation is independent of p53. MCF-7/neo and MCF-7/E6 cells were mock-treated or treated with TNFα for 4 h or 16 h. Expression of p21/WAF1 was analyzed by Northern blot (B) and by quantitative RT–PCR (C). (D) Expression of HPV/E6 in MCF-7/E6 cells prevents the accumulation of p53 protein in TNFα-treated cells but is without effect on the increased expression of p21/Waf1. Following incubation with TNFα for the indicated times, volumes of cellular extract corresponding to 10 μg of total proteins were loaded on a 10% (for p53 detection) or 12% (for p21/Waf1 detection) SDS-polyacrylamide gel and analyzed by Western blot using the antibodies DO-7 to reveal p53 and anti p21/Waf1 (Ab1 from Calbiochem) to reveal p21/Waf1

Full size image
Figure 2
figure 2

Inactivation of p53 by the truncated p53DD protein cells does not impair TNFα-induced p21/WAF1 gene expression. MCF-7/DD-TO cell line expresses the p53-DD cDNA under the control of a Tet-on promoter. Clones 2 and 40 are two clones isolated independently. Tetracycline (1 μg/ml) was added 24 h after cell plating (tet+). Non tetracycline exposed cells (tet-) were treated in parallel. (A) Tetracycline induces the expression of p53DD protein. p53DD expression was evaluated by Western blot using PAb122, 24 h after the addition of tetracycline. (B) Induction of p53DD expression impairs the p53 transactivation function. Six hours after the addition of tetracycline, the cells were transiently transfected with the luciferase reporter gene put under the control of the p53 responsive element of the human Waf1 promoter (pE1B-hWaf1). Parallel cultures were transfected with a control plasmid that includes the luciferase reporter gene but not the p53 responsive element. Luciferase activity was measured after an additional incubation of 24 h with (+tet) or without (−tet) tetracycline. Results are expressed relative to those obtained with the control plasmid. (C) Inactivation of p53 by p53DD does not affect TNFα-induced p21/WAF1 expression. Tetracycline (tet+) was added, or not (tet-) 4 h before the addition of TNFα. Total RNA was extracted at the indicated time following TNFα addition. p21/WAF1 mRNA levels were evaluated by quantitative RT–PCR. Results are expressed as fold activation relative to non-TNFα-treated cells in the absence and in the presence of tetracycline, respectively

Full size image

p21/Waf1 expression is regulated by NF-κB in response to TNFα

In response to diverse signals that induce growth arrest and differentiation, p21/Waf1 is regulated at a transcription level by a number of distinct transcription factors, besides p53 (reviewed in20). Recently it was shown that, in TNFα-treated Ewing tumour cells, the stimulation of p21/Waf1 expression is dependent on NF-κB activation.29 We then asked whether, in TNFα-treated MCF-7 cells, the NF-κB activation could account for the p53-independent stimulation of p21/Waf1.

Under our experimental conditions, NF-κB was already strongly induced in the 15 min following TNFα addition to MCF-7 (Figure 3A) and this activation was prolonged for at least 16 h (Figure 6B, middle panel). We then analyzed, in both MCF-7 and MAD-1904, the effect of 4- and 16-h TNFα-treatment on p21/Waf1 mRNA expression. MAD-1904 is a stable transformant of MCF-7 expressing the IκBα(A32/36) mutant, a superrepressor form of IκBα30 which behaves like a constitutive repressor of NF-κB activity.31 SOD2 gene that is known to be regulated by TNFα through an intronic NF-κB element,32 was used as a positive control.33 As expected, TNFα treatment significantly increased the level of SOD2 mRNA in MCF-7 but not in MAD-1904 (Figure 3B). Similarly, p21/Waf1 expression was severely impaired in MAD-1904 cells, strongly suggesting that NF-κB activation accounts for the up-regulation of p21/Waf1 gene expression in MCF-7 cells in response to TNFα-treatment. To reinforce this assumption further, NF-κB was inactivated using a chemical inhibitor, sodium salicylate (NaSal).34 A 30-min pre-incubation of MCF-7 cells with 5 mM or 20 mM NaSal was followed by a 16-h incubation with TNFα. As shown in Figure 3C, 20 mM NaSal abrogated the TNFα-induced SOD2 expression and both the endogenous and TNFα-induced expression of p21/Waf1, showing that NF-κB does account for the TNFα-dependent up-regulation of p21/Waf1 in MCF-7 cells. These results confirm and extend to MCF-7 those obtained by Javelaud et al.,29 showing that in MCF-7, like in Ewing tumour cells, the regulation of p21/Waf1 transcription depends on NF-κB. A functional NF-κB binding site within the p21/Waf1 gene remains to be identified to prove definitely that the p21/Waf1 gene is directly transactivated by NF-κB.

Figure 3
figure 3

TNFα induces p21/WAF1 expression via NF-κB. (A) Time-course analysis of NF-κB activation following TNFα treatment. MCF-7 cells were incubated with TNFα for the indicated times. DNA-binding activity of NF-κB was then analyzed by EMSA, using the NF-κB consensus DNA-binding site from Promega. (B) The induction of p21/WAF1 and SOD2 gene expression following TNFα treatment is impaired in MAD1904 cells. Total RNA was extracted from MCF-7/neo and MAD1904 cells either mock-treated or treated with TNFα for 4 h or 16 h before RNA extraction. Twenty μg of RNA were electrophoresed in 1% agarose-6% formadehyde gel and transferred to nylon membrane (Hybond N, Amersham) by electrophoresis. The blot was successively hybridized with p21/WAF1, SOD2, and GAPDH cDNA probes. (C) Sodium salicylate (NaSal) abrogated the TNFα-mediated stimulation of p21/WAF1 gene expression. MCF-7 cells were mock-treated or treated with the indicated amount of NaSal. After 30 min of incubation, cells were mock-treated or treated with 10 ng/ml TNFα for 16 h. p21/WAF1, SOD2 and GAPDH expression was analyzed as described in B

Full size image
Figure 6
figure 6

DNA-binding activity of p53 is not stimulated by TNFα. (A) p53 expressing in MCF-7 is able to bind its responsive element. MCF-7 cells were incubated for 16 h with adriamycin (10 μg/ml). DNA binding activity of p53 was analyzed by EMSA. Nuclear extract (10 μg of total protein) was incubated with 32P-labelled p53-consensus binding site (lane 2) and with PAb1801 antibody (lane 3), a mixture of PAb421 and PAb1801 antibodies (lane 4), PAb421 antibody alone (lanes 5, 6 and 7). The specificity of the DNA binding was checked by adding 50× molar excess of either specific (S, lane 6) or non-specific (NS, lane 7) cold oligonucleotides. The vertical bar on the right of the figure indicates the migration position of either supershifted (DNA bound by PAb1801/p53 or PAb421/p53 molecules) or super-supershifted (DNA bound by PAb1801/PAb421/p53 molecules) specific complexes. (B) No stimulation of p53 DNA-binding activity in three TNFα-treated cell lines expressing wt-p53. MCF-7 (lanes 1–5), Hep G2 (6–10) and U-2 OS (lanes 11–15) cells were incubated with 10 ng/ml of TNFα for the indicated times or with 10 μg/ml of adriamycin (ADR) for 16 h. DNA-binding activity of cellular extract was analyzed by EMSA with p53 (upper panel), NF-κB (middle panel) and Sp1 (lower panel) specific DNA-binding sites. The vertical bar on the left of the figure indicates the migration position of specific shifted complexes. (C) Analysis of p53-specific DNA binding activity with equivalent amount of whole p53 expressed in cells treated with either TNFα or adriamycin. MCF-7 were non-treated (lane 1) or treated with TNFα (lane 2) or adriamycin (lane 3) as described in (B). DNA binding activity of p53 was analyzed on a PhastGel gradient 4–15% (Amersham Pharmacia biotech). Binding assays were performed with 10 μg of total proteins of each extract (C1) or with a volume of cellular extract adjusted to have the same amount of p53 protein (C2). The p53 concentration was estimated by ELISA. Binding assays were performed with a volume of cellular extract corresponding to 10 (C2, lane 1), 5 (C2, lane 2) or 1.25 (C2, lane 3) μg of total proteins

Full size image

TNFα-dependent G1-arrest is independent of NF-κB activation

To examine further the implication of NF-κB-mediated activation of p21/Waf1 gene expression in TNFα-induced G1 arrest, the percentage of cells in G1 was evaluated comparatively by flow cytometry in both MCF-7/neo and MAD-1904 cells. Cells were treated in parallel with a mitotic inhibitor, the nocodazole.35 This allows identification of the true G1-arrest cell population (cells unable to proceed through S-phase) as illustrated in Figure 4A by comparing panels A1 versus A4, and B1 versus B4. Addition of nocodazole to the culture medium considerably increases the G2/M peak. But, in MCF-7 TNFα-treated cells there is only a modest increase in G2/M peak in the presence of nocodazole (Figure 4A, compare panels A3 and A4). This result clearly illustrates the TNFα-induced G1 arrest. This G1 arrest is equally observed with MAD-1904 (Figure 4A, compare panels A6 and A8), suggesting that TNFα-specific G1 arrest is not dependent on NF-κB-mediated induction of p21/Waf1 gene expression. This conclusion was further supported by analyzing the levels of p21/Waf1 protein comparatively in MCF-7/neo and MAD-1904. As expected, the extent of p21/Waf1 protein accumulation in TNFα-treated MCF-7/neo cells is not affected by nocodazole treatment (Figure 4B, compare lanes 3 and 4). Moreover, and in agreement with the inactivation of NF-κB in MAD-1904 cells, the p21/Waf1 protein level was very low in this cell line (Figure 4B, lanes 5 to 8) as compared to MCF-7/neo cells (Figure 4B, lanes 1 to 4).

Figure 4
figure 4

TNFα-induced G1 arrest is independent of NF-κB. (A) MCF-7/neo (upper panels, A1 to A4) and MAD1904 (lower panels, A5 to A8), cells were mock-treated (panels A1, A2 and A5, A6) or treated with 10 ng/ml of TNFα (panels A3, A4 and A7, A8). For the cells treated with nocodazole (panels A2, A4 and A6, A8), 0.6 μg of this drug was added 16 h before collecting the cells. The cultures were collected 40 h after the addition of TNFα. DNA content was analyzed by flow cytometry. (B) p21/WAF1 protein levels were evaluated by ELISA in cultures treated in parallel. Results are expressed as Arbitrary Units (AU)

Full size image

p53 is not activated by TNFα

The fact that p21/Waf1 gene expression is not stimulated by p53 in response to TNFα in MCF-7 strongly suggests that the increase in p53 protein level is not concomitant with its activation. To further support this assumption, the effect of TNFα on p53-transcriptional activity was checked by analyzing the expression of Ddb2, a recently identified p53 target gene that encodes for the p48 subunit of a human damage-specific DNA binding protein.36 As expected, Ddb2 expression is stimulated in response to γ irradiation in MCF-7/neo but is not in MCF-7/E6 cells. Yet, the expression of this gene is not stimulated in response to TNFα treatment (Figure 5), showing again that TNFα-mediated induction of p53 gene expression is not accompanied by its transcriptional activation. To better understand the lack of p53-target gene stimulation following the induction of wt-p53 accumulation in TNFα-treated MCF-7 cells, the affinity of p53 for its DNA responsive element was measured by electromobility shift assays (EMSA). Nuclear extracts of MCF-7 cells, mock-treated or treated with 10 ng/ml of TNFα for 1.5, 3 and 16 h, were incubated with double-strand DNA sequences containing a binding site either for p53, NF-κB or Sp1 transcription factors. The p53 binding site used consists in a perfect p53 DNA responsive element, with two adjacent consensus decamers (Pu Pu Pu C A/T T/A Py Py Py). Adriamycin was used as a positive control for p53 activation. As expected, an intense band corresponding to the complex [DNA/p53/PAb421] was obtained with adriamycin-treated MCF-7 cellular extracts (Figure 6B, upper panel, lane 5). In contrast, there was no apparent stimulation of p53-specific DNA-binding activity following TNFα treatment, irrespective of the treatment duration (Figure 6B, upper panel, compare lane 1 with lanes 2, 3, and 4). On the contrary, and in agreement with published data,37 both adriamycin and TNFα activated the specific DNA-binding activity of NF-κB (Figure 6B, middle panel, compare lane 1 to lanes 2, 3, 4 and 5). The Sp1 binding activity that was stimulated neither by adriamycin nor by TNFα treatment was used as a negative control (Figure 6B, bottom panel). To insure that the difference in p53-specific DNA binding activity of cellular extracts treated either with adriamycin or TNFα did not simply reflect a lower level of p53 accumulation in TNFα- versus adriamycin-treated cells, p53 concentration was evaluated by an ELISA assay. Indeed, the p53 concentration was increased by a factor 8 in adriamycin treated-cells and by only a factor 2 in TNFα-treated cells (data not shown and Figure 8). DNA band-shift assay was then repeated by adjusting the volume of adriamycin- and TNFα-treated cellular extracts to have an equivalent amount of p53 in each DNA-binding reaction mixture. Results presented in Figure 6, show a significant activation of p53-specific DNA binding activity in adriamycin-treated cellular extract even with a p53 concentration equivalent to that of untreated or TNFα-treated cellular extracts (panel C2, compare lanes 1 and 2 to lane 3). From these results, we conclude that TNFα is not able to activate the MCF-7-endogenous wt-p53. This observation was extended to two other cell lines derived from tumours of different origins, but expressing wild-type p53. Similarly to the results obtained with MCF-7, TNFα did not activate specific DNA binding activity of the endogenous wt-p53 expressed either in Hep G2, a cell line derived from a hepatocellular carcinoma (Figure 6B, upper panel, compare lane 6 with lanes 7, 8, and 9), or in U-2 OS, a cell line derived from an osteosarcoma (Figure 6B, upper panel, compare lane 11 with lanes 12, 13, and 14). These results demonstrated that the specific DNA-binding activity of wt-p53 is not stimulated by TNFα, at least in the three cell lines tested.

Figure 5
figure 5

The p53-target gene Ddb2 is not stimulated by TNFα. MCF-7/neo and MCF-7/E6 cells were incubated with 10 ng/ml of TNFα for the indicated times or irradiated at 6 Gy and incubated for an additional 16 h period. Expression of Ddb2 was analyzed by quantitative RT–PCR. Results were expressed as fold-stimulation relative to non-treated cells (CTRL)

Full size image
Figure 8
figure 8

TNFα induces p53 accumulation at both mRNA and protein levels. MCF-7 cells were incubated for different times with 10 ng/ml of TNFα. p53 protein level was evaluated by ELISA (solid circle) and mRNA was quantified by RT –PCR (open square) as described in Materials and Methods

Full size image

p53 accumulates within the nuclei of TNFα-treated cells

Nuclear localization of p53 has been shown to be essential for the activity of p53 protein.38 In normal, non-stressed cells, p53 sub-cellular localization is variable and depends on the phase of the cell cycle. p53 is predominantly nuclear from approximately mid G1 to G1/S transition, then becomes increasingly cytoplasmic as the cell progresses through the cell cycle.39 Similar regulation of p53 intracellular localization through the cell cycle progression has been described for MCF-7.40 We then checked if p53 nuclear exclusion could account for the accumulation of an inactive form of p53 in cells treated with TNFα. TNF-treated or untreated MCF-7 cells were subjected to cell fractionation, and p53 protein contents in both nuclear and cytoplasmic fractions were evaluated by Western Blot. Results are presented in Figure 7A. p53 protein was not detectable in mock-treated cells, nor in the cytoplasmic or nuclear fractions. As expected, p53 accumulated in the nuclei of γ-irradiated cells. Comparable results were obtained in response to TNFα treatment. A p53 band was easily detectable in the nuclear but not in the cytoplasmic fraction of cells incubated in the presence of TNFα for 4 and 16 h. We conclude that nuclear exclusion cannot account for the fact that p53 is transcriptionally inactive in TNFα-treated cells.

Figure 7
figure 7

Intracellular localization and phosphorylation status of p53 in TNFα-treated cells. (A) p53 accumulates in the nucleus of MCF-7 cells in response to TNFα. MCF-7 cells were mock-treated (lanes 1, 2), exposed to γ-radiation at a dose of 6 Gy for 4 h (lanes 3, 4) or treated with TNFα for 4 h (lanes 5, 6) or 16 h (lanes 7, 8). Following treatment, cells were fractionated. Relative levels of p53 within the cytoplasm (Cy) and nuclear (N) fractions were evaluated by Western blot using anti-p53 DO-7 antibody. (B) Time course of p53 expression and phosphorylation at Ser-15, Ser-20 and Ser-392 in MCF-7 cells following adriamycin (B1) or TNFα (B2) treatment. MCF-7 cells were treated with adriamycin (10 μg/ml) or with TNFα (10 ng/ml) for the indicated time. A volume of cellular extract corresponding to 60 μg (B1 and B2, lanes 1 to 8), 30 μg (B1, lane 7b) or 90 μg (B2, lane 7b) of total proteins was separated on a 10% SDS–PAGE and analyzed by Western blot using specific antibodies as indicated

Full size image

TNFα treatment does not induce p53 phosphorylation at Ser-15, -20 and -392, in MCF-7 cells

The C- and N-terminal domains are exposed to covalent post-translational modifications (phosphorylation, acetylation, sumoylation) that modulate p53 activity (reviewed in17,41). In particular, depending on the stress, phosphorylation at Ser-15 and -20 in the N-terminal region and -392 in the C-terminal region participate to p53 activation and/or stabilization.42 The fact that TNFα treatment led to an accumulation of a transcriptional inactive form of p53 prompts us to analyze the phosphorylation state of p53 in TNFα-treated cells. Adriamycin treatment that activates p53 was used as a positive control. p53 phosphorylation was revealed by Western blot using antibodies directed against phosphorylated Ser-15, -20 or -392. As expected, the adriamycin treatment induced p53 phosphorylation at these three sites (Figure 7, panel B1). Phosphorylation at Ser-20 and Ser-392 occurred earlier (1–2 h after adding adriamycin) than phosphorylation at Ser-15, which was detectable only 8 h after the addition of adriamycin to the culture medium. In contrast, no phosphorylated form of p53 was revealed in TNFα-treated MCF-7 cells (Figure 7, panel B2). As already mentioned, the p53 concentration is lower in TNFα- than in adriamycin-treated cells (compare the intensity of the bands revealed by DO-7 in panel B1 and B2). To ensure that no detection of p53 phosphorylated forms in TNFα-treated cellular extracts was not a bias imputable to the limit of antibody sensitivity, Western blot were performed by loading an equivalent amount of whole p53 protein. Results are presented in Figure 7 for a 16 h treatment with either adriamycin (panel B1, lane 7b) or TNFα (B2, lane 7b). Although the DO-7-band intensity was comparable in both cases, again there was no specific band revealed by antibody directed against the phosphorylated forms of p53 in TNFα-treated cell extracts. We conclude that, in contrast to adriamycin treatment, TNFα leads to accumulation of a p53 protein non-phosphorylated at position 15, 20 and 392.

Concomitant increase of p53 protein and mRNA levels in MCF-7, in response to TNFα

The promoter sequences of the p53 gene contains binding elements for transcription factors that may regulate expression in response to defined stimuli, including motifs for AP1, NF-κB and Myc/Max/USF.43 The fact that TNFα activated NF-κB makes p53 gene a good candidate to be transcriptionally regulated by TNFα. We then quantified, at different times following the addition of TNFα to the culture medium, the relative levels of p53 mRNA by real-time detection of PCR products. The level of p53 protein was estimated in parallel cultures by ELISA assay as described in experimental procedures. Results presented in Figure 8 show that TNFα had doubled the level of p53 mRNA 16 h after the addition of TNFα. Interestingly, the time-courses of p53 mRNA and protein accumulation are comparable in terms of kinetics and magnitude. This observation suggests that TNFα-induced accumulation of p53 protein results from a higher rate of transcription rather than from a stabilization of the protein.

All taken together, these results demonstrate that TNFα induces the accumulation of an inactive form of p53 by regulating the p53 transcription level.

Discussion

In this paper, we show that TNFα induces the expression of both p53 and p21/Waf1 in MCF-7 cell line. We provide evidence that NF-κB but not p53 is involved in the regulation of p21/Waf1 by showing that NF-κB but not p53 inactivation impairs TNFα-induced stimulation of p21/Waf1 expression. This led us to demonstrate that TNFα induces the accumulation of an inactive form of p53 unable to bind specifically to DNA and to activate the expression of its target genes. Consistent with this assumption, we report that Ser-15, Ser-20, and Ser-392 are not phosphorylated in response to TNFα. It is well documented that phosphorylation at Ser-15 and Ser-392 up-regulates the p53's transcriptional activity.44 The notion that p53 accumulation could be dissociated from its activation has already been proposed. Chernov et al.18 showed that PKC inhibitors induce an increase in p53 lifetime without inducing p53 activation. Tryptic digestion of stabilized protein shows that although PKC inhibitor decreases the overall level of p53 phosphorylation, this treatment leads to the appearance of a new set of phosphopeptides, leading the authors to propose that accumulation and activation of p53 could result from distinct phosphorylation events.

In our cellular model, both p53 mRNA and protein levels increased in parallel, strongly suggesting that the rate of p53 gene transcription could account for the accumulation of p53 protein. Indeed the p53 promoter contains a functional NF-κB site43 supporting the hypothesis that the increase level of p53 mRNA could arise from the TNFα-mediated NF-κB activation. In this line of evidence, it was previously reported that p53 gene expression is activated through NF-κB after Benzo(a)pyrene treatment.15

The idea that p53 accumulation results from a higher transcription rate rather than from the protein stabilization is emphasised by the fact that TNFα did not induce p53 phosphorylation at Ser-20. Indeed, phosphorylation at Ser-20 has been shown to inhibit p53 interaction with the ubiquitin-ligase MDM2, preventing p53 degradation by the proteasome pathway and thereby promoting p53 stabilization.50

Some papers describe the existence of an emerging PKR/p53 pathway involved in TNF-induced apoptosis (reviewed in45). It has been demonstrated that the TNFα-inducible kinase PKR can phosphorylate p53 in vitro at serine 392.46 Phosphorylation of serine 392 has been shown, in vitro, to convert p53 to an active form,47 suggesting that TNFα could activate p53 through PKR. Our results argue against this interpretation since we did not detect p53 modification at serine 392 in TNFα-treated cells. Alternatively, PKR, which has been reported to activate NF-κB (reviewed in48), could participate in the increase of p53 gene expression. In support of this notion, it was demonstrated that p53 mRNA expression level is increased in PKR-overexpressing cells.49

On the other hand, NF-κB inactivation, which diminishes the expression of p21/Waf1, has no effect on TNFα-dependent G1-arrest. This fact would imply that p21/Waf1 is not involved in the G1 arrest induced by TNFα. We can not exclude, however, the possibility that a very low remaining expression of p21/Waf1 could account for the G1-arrest identified in TNFα-treated MAD-1904 cells. In different cellular models, other investigators have found that p21/Waf1 is not required to induce G1-arrest in response to TNFα. Shiohara et al.51 have shown that overexpression of p21/Waf1 antisense in ME-180 cells does not abrogate TNFα-dependent G1-growth arrest. Additionally, Nalca and Rangnekar52 have demonstrated that interleukin-1 (IL-1), another cytokine promoting cytostatic effects, induces p53 and p21/Waf1 expression. However, and in agreement with our results with TNFα, cells in which p53 and p21/Waf1 have been switched off are still arrested in the G1-phase following IL-1 incubation.52 These results raise the question of the pathway by which cytokines such as TNFα or IL-1, could induce G1-arrest. In preliminary experiments, we showed that the product of the retinoblastoma gene (Rb) is under-phosphorylated in response to TNFα (data not shown). Rb has been implicated in the regulation of the cell cycle through its interaction with the transcription factor E2F. Dephosphorylated Rb interacts with E2F, leading to functional inhibition of this protein that is essential for cell-cycle progression (reviewed in53). In the hypothesis of an inhibition of Rb phosphorylation by TNFα, G1 arrest would depend on the maintenance of hypo-phosphorylated forms of Rb rather than on p21/Waf1 overexpression or p53 activation.

Although TNFα-induced p21/Waf1 accumulation seems not to be involved in G1 arrest, it might play an antiapoptotic role against TNFα-induced apoptosis. Indeed, it has been shown that expression of an antisense of p21/Waf1 sensitizes MCF-7 cells to TNFα-dependent apoptosis.24 Consistent with this, cytoplasmic accumulation of p21/Waf1 observed during monocytic differentiation is concomitant with a resistance to various apoptotic stimuli including TNFα.54 Therefore, it will be interesting to analyze the MCF-7 sub-cellular localization of p21/Waf1 in response to TNFα.

In conclusion, we show that TNFα stimulates the expression of both p21/Waf1 and p53, in MCF-7 cells, and that the TNFα-related stimulation of p21/Waf1 is dependent on NF-κB activation. The accumulation of p53 that arises from an increase of p53 gene transcription, could also depend on TNFα-mediated activation of NF-κB. Although TNFα leads to the accumulation of a transcriptional inactive form of p53, it cannot be excluded that p53 accumulation could be involved in TNFα-dependent induction of apoptosis. Indeed, it has been reported that p53 can induce apoptosis independently of its transactivation properties (reviewed in55). Moreover, some published results suggested an implication of p53 in TNFα-dependent apoptosis.35,56 We are presently analyzing the possible relationship between p53 accumulation and TNFα-induced apoptosis.

Materials and Methods

Cells and treatment

The cells U-2 OS (derived from a human osteosarcoma), MCF-7 (derived from a human breast carcinoma) and Hep G2 (derived from a human hepatocellular carcinoma) express wt-p53. MCF-7/MAD1904 cells, expressing IκBα, the dominant negative of NF-κB35 and MCF-7/E6, expressing HPV16/E6 protein26 were a gift from S Chouaib. MCF-7/DD-TO cell lines were derived from MCF-7 using the tetracycline-regulated mammalian expression (T-RexTM) system from Invitrogen. T-RexTM System uses regulatory elements from the E. coli Tn10-encoded tetracycline (Tet) resistance operon. In this system, expression of the gene of interest is repressed in the absence and induced in the presence of tetracycline. MCF-7 cells were first stably transfected with pcDNA6/TR, which encodes the Tet repressor (TetR) under the control of the human CMV promoter. The TetR-expressing cell line was then stably transfected with pDDm-TO, a tetracycline inducible expression vector obtained by inserting the sequence that encodes the mouse p53DD truncated protein, into the Invitrogen vector, pcDNA4/TO/B. The p53DD truncated protein includes the mouse p53-amino acid residues 1–14 and 302–390.57 Transfections were performed according to the manufacturer's protocol.

All cells were maintained at 37°C in DMEM containing 10% foetal calf serum (FCS). MCF-7/MAD1904 and MCF-7/E6 cell lines were cultured in the presence of 0.2 mg/ml of geneticin (Gibco BRL). The p53DD protein expression was induced by adding tetracycline to the medium (1 μg/ml). All the treatments were performed 48 h after plating (3.106 cells per 10 cm Petri dish), except otherwise indicated. TNFα (Sigma) was added to a final concentration of 10 ng/ml. Sodium salicylate was added at the indicated concentration for 30 min before adding TNFα (10 ng/ml), and incubated for an additional 16 h. Adriamycin (10 μg/ml, Sigma) was added for 16 h. The cells were exposed to γ irradiation at a dose of 6 Gy and a dose rate of 2.0 Gy/min, using a 137Cs source (IBL 637 apparatus, Cisbio International).

Northern blot

Northern-blot analysis was performed as described.28 Preparation of the probes has been described.58

Western blot

Western blots were performed as described.58 Whole p53 protein was detected using DO-7 or PAb122 supernatant hybridomas, the phosphorylated forms using anti-phosphop53 Ser-15, -20, and -392 from Cell Signaling technology. The p21/Waf1 protein was revealed with the anti-p21/Waf1 (Ab1) from Calbiochem.

Cell fractionation

Cells were scraped, washed and lysed in buffer A (20 mM HEPES, pH 7.6, 10 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 1 mM DTT, 20% glycerol, 4 mM Pefabloc) containing 0.1% NP-40. The lysate was centrifuged for 5 min at 500×g. Supernatant was collected and centrifuged at 100 000×g for 1 h at 4°C. Soluble material which represents the cytosol fraction, was conserved at −80°C, while the pellet was resuspended in ice-cold sucrose buffer I (320 mM sucrose, 3 mM CaCl2, 2 mM MgCl2, 0.1 mM EDTA, 10 mM Tris-base, pH 8.0, 1 mM DTT, 0.1% NP-40). The resuspended pellet was loaded on a cushion of sucrose buffer II (2 mM sucrose, 5 mM MgCl2, 0.1 mM EDTA, 10 mM Tris-base, pH 8.0, 1 mM DTT) and centrifuged at 30 000×g at 4°C for 45 min. The pellet (nucleus) was washed twice in buffer A and nuclear proteins were extracted using the same buffer containing additionally 0.5 M NaCl. Then, nuclear and cytosol fraction (30 μg of total protein) were separated on a 10% gel SDS–PAGE and p53 was detected by Western blot.

EMSA

Nuclear extracts were prepared as described.58 For p53 DNA binding analysis, nuclear extract (10 μg of total protein) was incubated with 0.4 ng of 32P-labelled self-renatured palindromic probe [TCGACGGACATGCCCGGGCATGTCC] (the underlined nucleotides correspond to the consensus sequence). Complexes were resolved on a 4% native acrylamide gel or on a PhastGel gradient 4–15% (Amersham Pharmacia biotech) as described.58 For Sp1 and NF-κB DNA binding analysis, nuclear extract (5 μg of total protein) was incubated with Sp1 and NF-κB-32P-labelled consensus probes, and complexes were separated on a 4% or 6% non-denaturing acrylamide gel, respectively. Sp1 and NF-κB probes were purchased from Promega.

ELISA

Cells were scraped, washed twice with ice-cold PBS and lysed for 30 min at 4°C in buffer containing 50 mM Tris-base (pH 8.0), 150 mM NaCl, 5 mM EDTA, 1% NP40 and 1 mM Pefabloc (Boehringer Mannheim). Solubilized extracts were collected after centrifugation for 15 min at 10 000×g. Total protein concentration was determined by the Bradford assay (Biorad Protein Assay). The p53 and p21/Waf1 protein concentrations in extract were estimated using the ‘p53 pan ELISA’ kit (Roche Molecular Biochemicals) and the ‘Waf1 ELISA’ kit (Oncogene Research Product), respectively.

Cell cytometry analysis

MCF-7 cells were collected, washed with PBS and then fixed in 70% ethanol. After 16 h at −20°C, cells were washed with PBS and then incubated for 30 min at 37°C in 0.5 ml of PBS containing 50 μg/ml propidium iodide and 50 μg/ml RNAse A. Cell DNA content was evaluated by flow cytometry using a FacScan (Becton Dickinson).

Luciferase assay

Luciferase assays were performed as previously reported using the pE1B-hWaf1 plasmid, which contains the human Waf1-p53 responsive element cloned upstream the luciferase gene.59

Real-time quantitation mRNA

p53 mRNA was quantified using the TaqMan apparatus (Perkin Elmer Biosystems), while p21/Waf1 and Ddb2 mRNAs were quantified using Lightcycler apparatus (Roche Molecular Biochemicals).

For p53 quantification, cDNA synthesis was performed in a volume of 20 μl using TaqMan reverse transcription reagents (Perkin Elmer Biosystems, Foster City, USA), as previously described.60 For the β-actin gene, the sense primer [ACCGAGGCCCCCCTG] and the probe [VIC-CCCAAGGCCAACCGCGAGAAACAGCCTGGATAGCAACGTACA] were located in exon 3 and the antisense primer [ACAGCCTGGATAGCAACGTACA] in exon 4. For the p53 gene, the sense primer [CCCAGCCAAAGAAGAAACCA) was in exon 9, the probe [6FAM-CCCTTCAGATCCGTGGGCGTGA-TAMRA] at the junction between exon 9 and 10 and the antisense primer in exon 10 [CTCGGAACATCTCGAAGCG]. PCR was performed as previously described.60 The use of VIC and 6-FAM reporter dye on the β-actin and p53 probes, respectively, allowed quantification from a multiplex PCR reactions. Results are normalized to β-actin.

For p21/Waf1 and Ddb2 mRNA quantification, cDNA synthesis was performed using oligo dT (Sigma) and AMV reverse transcriptase as described by the manufacturer (Promega). Real-time quantitative PCR was done using the ‘FastStart DNA Master SYBR Green’ kit (Roche Molecular Biochemical). PCR were performed with the oligonucleotide pairs [GGACCTGTCACTGTCTTGTA]/[GGCTTCCTCTTGGAGAAGAT] for p21/Waf1; [GAACAACTAGGCTGCAAGAC]/[ATTCGGCTACTAGCAGACAC] for Ddb2 and [AGCTCACTGGCATGGCCTTC]/[ACGCCTGCTTCACCACCTTC] for GAPDH. Results are normalized to GAPDH.