Mutant p53 Protects Cells from 12-O-Tetradecanoylphorbol-13-Acetate–Induced Death by Attenuating Activating Transcription Factor 3 Induction

Multiple stress responses are regulated by the p53 tumor suppressor gene (1). p53 is a transcription factor that activates specific target genes through direct binding to a p53 consensus sequence (2). p21^WAF1 and GADD45 are activated transcriptionally by p53 in response to DNA damage, and this activation is involved in the control of cell cycle checkpoints (3, 4). Alternatively, p53 can transactivate genes, such as BAX, PUMA, and Noxa, leading to induction of apoptosis (5–7).

Wild-type (WT) p53 plays a pivotal role in preventing tumor development. Indeed, in >50% of human primary tumors p53 is mutated. Notably, the predominant mode of p53 inactivation is by point mutation rather than by deletion or truncation. These data coupled with the observation that mutant p53 is generally highly overexpressed in tumors have led to the hypothesis that mutant p53 possesses gain-of-function activities. This hypothesis is supported by the results of in vivo and in vitro studies. For example, mice harboring mutant p53 display allele-specific tumor spectra, higher metastatic frequency, enhanced cell proliferation, and higher transformation potential compared with their p53-null counterparts (8, 9). In addition, endogenous mutant p53 or reconstitution of mutant p53 expression in p53-null cells augments the transformed phenotype (10–13). Although the question of how mutant p53 contributes to tumor initiation and progression has been addressed intensively, the molecular mechanism that underlies the role of mutant p53 in malignant transformation remains unclear. One putative mechanism is that mutant p53 alters the expression of specific genes and thus interferes with the onset of the apoptotic process. This is supported by recent data showing that the EGR1 and NFκB2 genes can be activated and the CD95 and MSP-1 genes repressed by mutant p53 (14–17). These findings prompted the present study to investigate whether mutant p53 influences cell fate by altering gene expression. Due to the fact that activation of the protein kinase C (PKC) pathway by phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) has been associated with neoplastic transformation, carcinogenesis, and tumor cell invasion, we decided to pursue the role of mutant p53 in this pathway. At low doses, TPA is a potent PKC activator and its major biological effects are exerted via the PKCs (18), which in turn can induce either cell survival or cell death (19, 20). In contrast, high doses of TPA lead to a down-regulation of PKC activity (21–23). Notably, in certain cell types, attenuation of PKC activity, either due to exposure to high doses of TPA or due to PKC inhibitors, results in cell death (24, 25).

The cyclic AMP-responsive element binding protein (CREB)/activating transcription factor (ATF) and activator protein-1 (AP-1) transcription factor families were shown to be induced in cells treated with low doses of TPA (26, 27). ATF3 is a member of the CREB/ATF subfamily of bZIP transcription factors (28). It has been shown to stabilize WT p53 by blocking its ubiquitination (29) and is also a p53 downstream target gene, creating a regulatory feedback loop (30). Evidence supporting interactions between p53 and ATF3 proteins has led to the proposal that ATF3 can inhibit p53 transactivation capacity (31). ATF3 is induced in response to stress agents, such as UV and ionizing radiation (IR), not only via p53 but also by p53-independent pathways (32, 33). Similar to p53, ATF3 expression is associated with cell cycle arrest and apoptosis. Overexpression of ATF3 in HeLa and HT-1080 cells induces G₁ arrest on IR treatment (31, 33), whereas in HeLa-S3 cells treated with etoposide ATF3 enhances apoptosis (34).

We report here that expression of mutant p53 results in reduced sensitivity to cell death induced by high doses of TPA. We show that this protection is mediated by attenuation of p53-independent ATF3 induction. Furthermore, we show that mutant p53 attenuates ATF3 expression by two complementary mechanisms involving the MEF2 motif on the ATF3 promoter and expression of the MEF2D gene. This study provides novel insights into the molecular mechanisms by which mutant p53 exerts gain-of-function activity.

Chemicals and reagents. TPA, doxorubicin, and DMSO were from Sigma (St. Louis, MO). Go6976 and rottlerin were from Calbiochem-Novabiochem (Bad Soden, Germany). Tumor necrosis factor α (TNFα) was from Biological Industries (Beit Haemek, Israel). Transforming growth factor β (TGFβ) was from R&D Systems (Minneapolis, MN).

Cell culture and treatments. The amphotropic and ecotropic Phoenix retrovirus-producing cells were from the American Type Culture Collection (Manassas, VA). The immortalized primary human embryonic lung fibroblasts (WI-38) were created and described previously by our laboratory (35). The ovarian cancer SKOV3 cell line stably expressing either an empty vector, p53R175H, or p53R248W was a gift from Prof. P.M. Chumakov (University of California, San Diego, CA). The fibrosarcoma HT-1080 cell line was kindly provided by Dr. M. Brandeis (Hebrew University, Jerusalem, Israel). WI-38 cells were grown in MEM supplemented with 10% FCS, 1 mmol/L sodium pyruvate, 2 mmol/L l-glutamine, and antibiotics. Phoenix, HT-1080, and SKOV3 cells were grown in DMEM supplemented with 10% FCS and antibiotics. All cells were maintained in a humidified incubator at 37°C and 5% CO₂.

TPA and rottlerin were applied in DMSO to a final concentration of 1 μg/μL and 500 μmol/L, respectively. Treatment of cells with doxorubicin (0.2 μg/mL), TNFα (10 ng/mL), TGFβ (2 ng/mL), TPA (100 ng/mL, 10 μg/mL), rottlerin (5 μmol/L), and Go6976 (10 nmol/L) was conducted at 60% confluency. Equal amounts of DMSO were added to the control plates.

Plasmids. The ATF3 expression plasmid was constructed by cloning the ATF3 open reading frame, obtained by reverse transcription-PCR (RT-PCR) with specific primers (see Supplementary Materials), into the pGEMT-easy vector (Promega, Madison, WI) and then restricted with EcoRI and inserted into the pBabe-puro expressing vector. Four ATF3 promoter-luciferase reporters were constructed by cloning the ATF3 promoter, obtained by genomic PCR amplification with specific primers (see Supplementary Materials), into the pGEMT-easy vector with subsequent subcloning into the pGL3-Basic luciferase plasmid (Promega). Seven motifs along the ATF3 promoter were mutated using the Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) following the manufacturer's instructions (see Supplementary Materials for primer sequences). pLXSN-GSE56-neo was obtained by subcloning the GSE56 BamHI-restricted fragment from pBabe-GSE56-puro into the pLXSN vector (Ossovskaya et al., 1996). The pLXSN-p53R175H-neo was obtained by subcloning the p53R175H BamHI-restricted fragment from pcDNA-p53R175H into the pLXSN vector. The expression plasmids MEF2A and MEF2D were kindly provided by Dr. E. Bengal (Technion-Israel Institute of Technology, Haifa, Israel). The p53 short hairpin RNA (shRNA) vector and its mouse NOXA shRNA control vector were kindly provided by Dr. D. Ginsburg (Bar-Ilan University, Ramat Gan, Israel).

Transfections and reporter assays. p53-null and p53R175H SKOV3 cells were plated at 2 × 10⁴ per well in a 24-well plate 24 hours before transfection. Cells were transfected with jetPEI transfection reagent (Polyplus Transfection, Illkirch, France) using 300 ng/well of the relevant ATF3 promoter-luciferase reporter construct, 600 ng/well Bluescript carrier plasmid, and 100 ng/well of pCMV-β-galactosidase expression vector for normalization of transfection efficiency. Cells were treated with 10 μg/mL TPA 48 hours after transfection. Following 24 hours of treatment, cell extracts were prepared and luciferase and β-galactosidase activities were determined using Promega materials and procedures.

Fluorescence-activated cell sorting analysis. Cells were plated in 10-cm dishes and treated with 10 μg/mL TPA for 72 hours. Cells were subsequently trypsinized and fixed in 70% ethanol/30% HBSS for 24 hours. Cells were then rehydrated for at least 30 minutes in PBS, washed, resuspended in PBS containing 50 μg/mL propidium iodide and 10 μg/mL RNase A, and subjected to fluorescence-activated cell sorting (FACS)-based cell cycle analysis.

Cell proliferation assay. Cells were seeded in 24-well culture dishes at 60% confluency. Cells were incubated with 10 μg/mL TPA or 0.2 μg/mL doxorubicin for 72 hours. Cell proliferation was determined by using a colorimetric assay with WST-1 reagent (Roche, Mannheim, Germany) following the manufacturer's instructions.

Western blot and retroviral infections. Western blot analysis and retroviral infection were done as described in ref. 35. The following primary antibodies were used: anti-p53 (DO-1; kindly provided by Dr. D. Lane, Ninewells Hospital and Medical School, Dundee, Scotland), anti-PKCα and PKCδ (kindly provided by Y. Dicken, Bar-Ilan University), anti-p21, anti-ATF3, anti-MEF2 (C-19; Santa Cruz Biotechnology, Santa Cruz, CA), anti-poly(ADP-ribose) polymerase-1 (PARP-1; C-2-10; Biomol, Plymouth Meeting, PA), anti-vinculin (Sigma), anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; MAB374; Chemicon, Temecula, CA), and anti-tubulin (T7816; Sigma).

Chromatin immunoprecipitation analysis. Chromatin immunoprecipitation (ChIP) was conducted as described in ref. 14. To detect binding of p53R175H or MEF2 to the ATF3 promoter, quantitative RT-PCR (QRT-PCR) primers amplifying the ATF3 promoter were used. To detect binding of p53R175H or MEF2 to the MEF2D promoter, QRT-PCR primers amplifying the MEF2D promoter were used (see Supplementary Materials).

Real-time RT-PCR analysis. Total RNA was extracted using the Versagene RNA cell kit (Gentra Systems, Inc., Minneapolis, MN). An aliquot of 2 μg of total RNA was reverse transcribed using Moloney murine leukemia virus reverse transcriptase (Promega) and random hexamer primers. QRT-PCR was done using the ABI 7000 machine (Applied Biosystems, Foster City, CA) with SYBR Green PCR Master Mix (Applied Biosystems). Specific primers were designed to the following genes: ATF3, p21^WAF, MEF2A, MEF2B, MEF2C, and MEF2D. cDNA levels were normalized to GAPDH levels and amplified with appropriate primers (see Supplementary Materials for all primers).

ATF3 induction following high-dose TPA is attenuated by mutant p53. Because p53 and ATF3 are cross-talking regulators of cell fate decisions, we decided to explore further the molecular relationships between these two transcription factors. Furthermore, as both p53 and CREB/ATF are involved in the PKC pathway that has been implicated in malignant progression, we focused on p53-ATF3 cross-talk following treatment with three compounds that modulate PKC activity: TGFβ, TNFα, and TPA. To that end, we established congenic cell lines differing in their p53 status. hTERT-immortalized normal human fibroblasts (WI-38) and a human fibrosarcoma cell line (HT-1080), both harboring WT p53, were infected stably with a retrovirus containing either an empty vector (neo), a dominant-negative p53 peptide (GSE56), or a tumor-associated p53 mutant R175H (p53R175H). The efficiency of GSE56 and p53R175H in blocking WT p53 activity was validated by measuring expression of the p53 target gene p21^WAF1 at both mRNA and protein levels (Supplementary Fig. S1).

Whereas 24 hours of treatment with TNFα (10 ng/mL) or TGFβ (2 ng/mL) did not change the levels of ATF3, low doses of TPA (100 ng/mL) down-regulated ATF3 expression (Supplementary Fig. S2). In contrast, high doses of TPA (10 μg/mL) induced ATF3 expression independently to the p53 status (Fig. 1A , compare lanes neo and GSE56). Notably, ATF3 induction was significantly attenuated in cells expressing the p53R175H mutant in both the WI-38 and HT-1080 backgrounds. In light of the finding that p53 status influences ATF3 expression specifically under conditions of high-dose TPA treatment, we focused on p53-ATF3 cross-talk under this condition, and therefore, from this point on in this study ‘TPA treatment’ refers only to high-dose TPA treatment, unless stated otherwise.

It was important to evaluate the response of ATF3 to TPA in cells that express mutant p53 in a p53-null background as opposed to the p53 WT background of HT-1080 and WI-38. To that end, ATF3 gene expression was examined in the genetic background of SKOV3, a p53-null ovarian cancer cell line. It was observed that SKOV3 cells stably expressing mutant p53 R175H (p53R175H) or R248W (p53R248W) exhibited attenuated ATF3 induction in response to TPA, both at the mRNA and protein levels, when compared with SKOV3 cells expressing empty vector (p53 null; Fig. 1B and C). Taken together, these results suggest that mutant p53 interferes with p53-independent cellular pathways that mediate ATF3 induction on TPA treatment. This represents a novel gain-of-function activity for mutant p53.

Next, we tested whether this gain-of-function activity is a feature of other p53 mutants. To this end, a genomic fragment of ∼1.3 kbp corresponding to the ATF3 promoter (28) was cloned upstream of a luciferase reporter. This full-length promoter (pLucATF3-1287) was cotransfected into p53-null SKOV3 cells together with different p53 mutants, and reporter activity was assayed after treatment with TPA for 24 hours. Notably, p53 mutants R175H, H179G, R248W, and D281G reduced significantly ATF3 promoter induction following TPA treatment when compared with the empty vector. In contrast, p53R273H as well as WT p53 lacked this repressor activity (Fig. 1D). A same phenomenon was observed with the partially truncated promoter pLucATF3-293 (Supplementary Fig. S3). Thus, the capacity to interfere with TPA induction of ATF3 is not restricted to specific mutant.

Mutant p53 protects cells from TPA-induced cell death. Having established that mutant p53 interferes with TPA-induced ATF3 activation, we wanted to examine the consequences of mutant p53 presence during TPA exposure for cell fate. To that end, p53-null and p53R175H SKOV3 cells were treated with TPA for 72 hours and assayed for cell proliferation using a colorimetric assay. Cell proliferation values indicate that already at 48 hours p53R175H presence confers cells with a significant proliferation advantage on TPA treatment (Fig. 2A). To delineate whether this was due to inhibition of cell death or of cell cycle arrest, DNA content was measured by flow cytometry to reveal changes in cell cycle distribution following TPA treatment. Whereas p53-null SKOV3 cells exhibited G₁ arrest following 48 hours of treatment and pronounced cell death following 72 hours, p53R175H and p53R248W SKOV3 cells displayed prolonged G₁ arrest with a low amount of cell death even after 72 hours (Fig. 2B). To further characterize the cellular death observed, the cleavage of the cell death marker PARP-1 (36) was measured. The level of full-size PARP-1 protein was reduced dramatically in p53-null cells following 72 hours of TPA treatment (Fig. 2C). In contrast, in p53R175H cells its level remained unchanged during the entire course of treatment. This protection from TPA by p53R175H SKOV3 cells was abolished when a p53 shRNA vector (shp53) was introduced to the cells as indicated by the PARP-1 analysis at 72-hour time point (Fig. 2D). In summary, these data suggest that p53-null cells cease to proliferate sooner than p53R175H cells as indicated by the lower mitochondrial activity observed at 48 hours after TPA treatment. More importantly, at 72 hours past TPA treatment, p53-null SKOV3 cells display significantly higher amount of cell death as measured by FACS and PARP-1 cleavage compared with their mutant p53-expressing counterparts. These data support the hypothesis that mutant p53 possesses a gain-of-function activity whereby it blocks TPA-induced cell death.

ATF3 sensitizes cells to TPA-induced cell death. The data thus far indicate that mutant p53 inhibits TPA-induced cell death and that this correlates with the ability of mutant p53 to attenuate ATF3 gene expression. This raises the question whether the attenuation of ATF3 induction by p53R175H is required for p53R175H to exert its protective effect. To answer this question, we examined the consequences of expressing ectopically ATF3 in the SKOV3 cell lines during TPA treatment. Both p53-null and p53R175H SKOV3 cells were infected with either an ATF3 carrying retrovirus (pBabe-ATF3) or an empty vector (pBabe-puro; Fig. 3A,, compare lanes 1 and 7 and lanes 4 and 10). Importantly, whereas in SKOV3 cells expressing p53R175H the protein level of ATF3 was transiently up-regulated before returning to basal levels after 72 hours, in SKOV3 cells expressing both p53R175H and ATF3 sustained high levels of ATF3 protein were observed (Fig. 3A , compare lanes 6 and 12).

To assess the phenotypic effect of this elevation of ATF3 protein levels, the SKOV3 cell lines with each p53 status, with or without ATF3 ectopic expression, were exposed to either TPA (10 μg/mL) or doxorubicin (0.2 μg/mL) and cell proliferation was assessed. Both treatments resulted in pronounced cell death (Fig. 3B). For the doxorubicin-treated cells, ectopic ATF3 expression did not affect significantly the amount of cell death following 72 hours of treatment in either p53-null or p53R175H cells. In contrast, when cells were treated with TPA, ATF3 ectopic expression enhanced cell death in p53R175H cells at the 72-hour time point. Moreover, Western blot analysis of extracts generated from these SKOV3 cell lines was conducted to measure the level of the full-size PARP-1. In p53-null cells, the full-size PARP-1 protein completely disappeared following 72 hours of TPA treatment both in the presence and in the absence of ectopic ATF3 expression (Fig. 3A, compare lanes 3 and 9). In contrast, in p53R175H cells the full-size PARP-1 protein was observed to disappear more completely in the cells expressing ectopic ATF3 than in those without ectopic ATF3 (Fig. 3A,, compare lanes 6 and 12). In agreement with that, p53-null SKOV3 cells harboring an ATF3 shRNA vector (shATF3) were less sensitive to TPA compared with their control vector (shmNOXA) counterparts as indicated by the PARP-1 analysis at 72-hour time point (Fig. 3C). These results implicate ATF3 as a cell death promoter under conditions of TPA treatment. Moreover, these data extend the correlation between ATF3 expression and the ability of p53R175H to influence TPA-induced cell death, lending support to the hypothesis that one mechanism by which p53R175H protects cells from TPA-induced cell death partially involves the attenuation of ATF3 expression.

p53R175H is acting downstream to the PKC proteins. It was shown that TPA exerts its cellular effects mainly via modulation of the PKC proteins (18). Whereas treatment with TPA at low doses activates most PKC isoforms, high doses of TPA result in depletion of PKC isoforms and consequently in inhibition of the PKC pathway. To discover whether the gain-of-function activity of p53R175H is mediated by PKCs, we monitored the protein levels of two major PKC isoforms that were shown to influence cell survival, PKCα and PKCδ, in SKOV3 and WI-38 cell lines treated with TPA for 72 hours. Both PKC isoforms were depleted following the treatment independently of the p53 status (Fig. 3D). These results suggest that p53R175H is acting downstream to the PKC proteins, but upstream to ATF3, in the TPA response pathway. Drugs that inhibit PKC activity have the potential to mimic the phenotype induced by high doses of TPA. We examined ATF3 gene expression after treatment with isoform-specific inhibitors of PKC activity. PKCδ is inhibited specifically by rottlerin at 5 μmol/L (37), whereas the activity of PKCα, as well as that of PKCβ1, is inhibited by Go6976 at 10 nmol/L (38). In p53-null SKOV3 cells, treatment with rottlerin induced ATF3 protein to a level similar to that induced by TPA treatment (Fig. 3D). As expected, cells expressing mutant p53 were less sensitive to rottlerin treatment compared with their p53-null counterparts, similarly to the results obtained with the TPA treatment (Supplementary Fig. S4). In contrast, treatment with Go6976 did not affect ATF3 levels. These results point to a specific role of the PKCδ isoform in ATF3 induction by TPA.

Activation of ATF3 by TPA is mediated via CREB/ATF and MEF2 motifs. To map the region responsible for ATF3 transcriptional activation on TPA treatment, a deletion analysis of the ATF3 promoter was carried out. Several 5′-truncated constructs were generated and placed upstream of the luciferase reporter gene (Fig. 4A). These were each transfected transiently into p53-null SKOV3 cells that were then treated with TPA for 24 hours before assessment of luciferase activity. pLucATF3-1287 was induced approximately 2- to 3-fold on TPA treatment (Fig. 4B). A similar response to TPA was observed with the partially truncated construct pLucATF3-293. Although further truncation, pLucATF3-104, resulted in a construct with significantly reduced basal activity, the transcriptional activity of this construct was still induced by TPA. In contrast, truncation of another 39 nucleotides resulted in a construct pLucATF3-65 that had lost its ability to respond to TPA but which still exhibited basal activity higher than that of the control empty construct (pLucEMPTY). These data suggest that the genomic region spanning from −65 to −104 relative to the transcription start site (TSS) is responsible for ATF3 induction on TPA treatment. Using the MatInspector database (39), we found that this promoter region harbors three known regulatory motifs [i.e., MEF2, CREB/ATF, and CAAT/enhancer binding protein (C/EBP)]. Directed mutagenesis of either the CREB/ATF or the MEF2 sites, but not of the C/EBP site, in the pLucATF3-293 reporter resulted in partial loss of the reactivity to TPA (Fig. 4C). Simultaneous mutations of both CREB/ATF and MEF2 motifs completely abolished the induction of pLucATF3-293 reporter activity on TPA treatment, suggesting that these two transcription factor binding motifs are required for the induction of ATF3 by TPA.

ATF3 induction by TPA is attenuated by p53R175H via the MEF2 site. Having established that ATF3 induction by TPA is mediated by the CREB/ATF and MEF2 sites, we wanted to elucidate whether either of these or any other binding sites mediate attenuation of ATF3 induction by p53R175H. Although the majority of p53 tumor-derived mutants, including R175H, are impaired in their DNA-binding domain, it has been reported that certain p53 mutants can nevertheless bind specific regulatory sites within promoters by interacting with other transcription factors, such as ETS-1 and SP1 (40–42). QRT-PCR-ChIP analysis was conducted on p53R175H-expressing SKOV3 cells following TPA treatment to discover whether p53R175H interacts with the ATF3 promoter. Indeed, p53R175H was interacted with the ATF3 promoter in nontreated cells and this interaction was enhanced following TPA treatment (compare αp53 with αHA in Fig. 5A). An in silico search for putative transcription factor binding sites within the ATF3 promoter with known involvement in ATF3 regulation revealed five motifs in addition to MEF2 and CREB/ATF: SP1, ATF4, ETS-1, C/EBP, and CDE. To investigate which of these sites mediates attenuation of ATF3 induction by p53R175H, seven ATF3 promoter constructs upstream of the luciferase reporter gene were generated, each having one of the sites mutated (Fig. 5D). Mutation in the MEF2 motif abolished most effectively the effect of p53R175H on the induction of ATF3 by TPA (Fig. 5B and C), for the fold induction of all ATF3 reporters except for the one harboring a mutant MEF2 site was lower in p53R175H cells than in p53-null cells. This suggests that p53R175H attenuates the ATF3 promoter activity primarily via the MEF2 site.

p53R175H attenuates TPA-induced MEF2D expression. The data presented above implicate the involvement of the MEF2 family of transcription factors both in TPA-induced ATF3 expression and in the attenuation of this induction by p53R175H. To study in more detail this role of MEF2 proteins, we monitored the mRNA levels of the four MEF2 members in p53-null and p53R175H SKOV3 cells following 24 hours of TPA treatment using the QRT-PCR analysis. Whereas MEF2A, MEF2C, and MEF2D were induced following TPA treatment, no significant induction of MEF2B was observed (Fig. 6A). MEF2A and MEF2C mRNA levels were induced similarly in both p53-null and p53R175H cells. In contrast, MEF2D induction was much stronger in the p53-null cells compared with their p53R175H counterparts. A differential induction in response to TPA in p53-null versus p53R175H cells was also observed for MEF2 protein levels (Fig. 6A).

To examine the possibility that MEF2D mediates TPA-induced ATF3 expression, p53-null and p53R175H SKOV3 cells were transiently cotransfected with the reporter pLucATF3-293 and either with a control vector (pBabe-puro), MEF2A (pBabe-MEF2A), or MEF2D (pBabe-MEF2D) expression constructs. Notably, ectopic expression of both MEF2A and MEF2D increased reporter gene expression in the absences of treatment, indicating that these transcription factors are activators of ATF3 (Fig. 6C). The apparently contradictory results that both overexpression of MEF2 (Fig. 6C) and mutation in the MEF2 motif (Fig. 4C) lead to up-regulation of ATF3 promoter activity can be resolved by the hypothesis that in the absence of TPA treatment a repressor protein is bound to the MEF2 site, which is released from the promoter by MEF2D on TPA treatment. When the p53-null SKOV3 cell expressing the pLucATF3-293 reporter gene and either empty vector, MEF2D, or MEF2A were treated with TPA, only for MEF2D was TPA-induced expression attenuated (Fig. 6C). This result indicates that specifically MEF2D is responsible for the transcriptional induction of ATF3 following TPA treatment. The same result was observed when using the reporter pLucATF3-293 + mCREB/ATF, which implies that the combination of mutant CREB/ATF site and MEF2D expression can mimic the induction of ATF3 as was observed in the response to TPA. Based on these experiments, we hypothesize that in p53R175H-expressing SKOV3 cells in response to TPA treatment the removal of a repressor from the MEF2 site by MEF2D is prevented, leading to attenuated induction of ATF3.

The next question pertains to how p53R175H might exert this action, of preventing the removal of a repressor bound to the MEF2 motif. The observation that MEF2D induction was much stronger in p53-null cells raised the possibility that p53R175H influences events at the MEF2 site in the ATF3 promoter not only via its interacting there but also indirectly through binding to the MEF2D promoter and inhibiting MEF2D induction. QRT-PCR-ChIP analysis was conducted on p53R175H SKOV3 cells following 24 hours of TPA treatment. Indeed, p53R175H displayed interaction with the MEF2D promoter (Fig. 6B), although this interaction was weak compared with ATF3 promoter binding (Fig. 5A). These results support the premise that an additional mechanism by which p53R175H attenuates ATF3 induction following TPA treatment might be via direct binding to the MEF2D promoter and inhibition of MEF2D expression.

Expression of several genes was shown to be modulated by mutant p53, including MDR1, EGFR, PCDNA, IL-6, BFGF, HSP70, and BAG-1 (reviewed in ref. 43). In this study, we explore how mutant p53 affects the immediate early response gene ATF3 and exerts its gain of function in the TPA-PKC pathway.

Kieser et al. (44) showed a link between mutant p53 and the PKC pathways by showing that mutant, but not WT, p53 increases the transcription of the vascular endothelial growth factor (VEGF) gene following low doses of TPA treatment. They suggested that VEGF induction following PKC activation is mediated by a member of the AP-1 family of transcription factors and that mutant p53 might enhance AP-1 activity. Indeed, c-fos, a member of the AP-1 family of transcription factors, was shown to be induced in mutant p53 cells (45).

In agreement with others, we showed here that high doses of TPA act to inhibit PKCs due to their ability to promote depletion of the PKC isoforms (21–23). Consistent with these previous studies, we observed that ATF3 induction following high doses of TPA is WT p53 independent and is correlated negatively with the levels of PKCs. However, we discovered that expression of mutant p53 disrupted this correlation and resulted in attenuated ATF3 induction. Importantly, this attenuation of ATF3 by mutant p53 was shown to correlate with protection of cells from TPA-induced death.

Having shown a novel gain-of-function activity for mutant p53, we investigated the molecular mechanism underlying this activity. First, we revealed that both CREB/ATF and MEF2 sites are responsible to TPA-induced ATF3 activation. Moreover, we show that both mutation in the MEF2 site and expression of the MEF2D protein up-regulate the basal activity of the ATF3 promoter, suggesting that a repressor protein binds to the MEF2 site in basal conditions. In contrast, mutations of both MEF2 and CREB/ATF sites do not result in increased promoter activity. These results may suggest that a potential repressor bound to MEF2 site that interferes with the binding of an activator to the CREB/ATF. Next, we found that p53R175H interacts with the ATF3 promoter and provide evidence that it attenuates ATF3 induction via this promoter binding specifically at the MEF2 site. Furthermore, we uncovered a second mechanism by which p53R175H influences ATF3 expression, for p53R175H interacts also with the MEF2D promoter and attenuates the expression of the ATF3 activator MEF2D. (Fig. 6D). Increased activation of the TPA-PKC pathway has been associated with malignant transformation in breast, lung, and gastric carcinoma cell lines. Conversely, inhibition of the TPA-PKC pathway has been shown to inhibit the invasive and metastasis potential of some malignant cells. Therefore, inhibiting the TPA-PKC pathway by antisense or inhibitors is considered likely to result in tumor regression and such inhibitors are used in cancer therapy today, such as UCN01, PKC412, bryostatin, and ISIS3521 (reviewed in ref. 46). Our discovery that mutant p53 interferes with TPA-induced cell death at a point in the pathway below the PKCs raises the worrying possibility that its expression in human tumors may desensitize them to such therapies.

Furthermore, ATF3 was shown to enhance cell death in several cell systems (34, 47, 48). Recently, it was shown that ATF3 can stabilize the tumor suppressor gene p53 and suppress Ras-stimulated tumorigenesis (29, 49). These tumor-suppressive activities of ATF3 have led to the hypothesis that down-regulated ATF3 expression may be an important feature of tumors compared with normal tissues. Indeed, Yan and Boyd (50) have shown recently that ATF3 expression is down-regulated in tumor versus normal tissues. Thus, in light of the results of the present study, it is tempting to speculate that the clinical aggressive features displayed by mutant p53-expressing tumors are at least partially exerted via the negative regulation of mutant p53 of the ATF3 gene.

Grant support: Flight Attendant Medical Research Institute Center of Excellence, European Commission's Sixth Framework Programme grant LSHC-CT-2004-503576, and Yad Abraham Center for Cancer Diagnosis and Therapy.

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.