Inactivation of Wild-Type p53 Protein Function by Reactive Oxygen and Nitrogen Species in Malignant Glioma Cells

Increased production of nitric oxide (NO) has been observed in multiple pathological disorders including hypoxia, inflammation, and malignancy (1). In the setting of increased NO and superoxide () production, the highly reactive anion peroxynitrite (ONOO⁻) is produced by a diffusion-limited reaction. Peroxynitrite is a potent oxidizing agent that is capable of modifying proteins by di-tyrosine formation, tyrosine nitration, and oxidation. We previously showed that inducible and constitutive isoforms of nitric oxide synthase are highly expressed in human malignant gliomas in vivo and that these tumors have evidence of tyrosine nitration, a “footprint” of peroxynitrite-mediated protein modification (2, 3).

In malignant gliomas, the p53 tumor suppressor protein is a critical regulator of tumor progression, and most malignant gliomas have evidence of functional inactivation of the p53 pathway. However, in primary glioblastoma (GBM), the most common and most malignant form of glioma, only about 30% of cases show evidence of p53 gene mutations (4). The tumors without p53 mutations often react positively with p53 antibodies, and p53 sometimes remains sequestered in the cytoplasm (5). This finding has been interpreted as indicating that besides p53 mutation, certain intracellular conditions can also be important in determining both cellular localization and function (6).

Peroxynitrite is particularly reactive with zinc-thiolate centers at the active site of zinc finger transcription factor proteins such as p53, and the p53 protein is known to be very sensitive to redox modification (7, 8, 9, 10). We demonstrated recently that concentrations of peroxynitrite consistent with those found in a hypoxic inflammatory environment such as a GBM are capable of inactivating p53-specific DNA binding of cells in culture as demonstrated by gel shift assays (3). We found that this inactivation of p53 DNA binding by peroxynitrite was associated with tyrosine nitration of wild-type p53 protein and the accumulation of nonreducible nitrated high molecular weight p53 aggregates. This finding is consistent with data indicating that zinc finger transcription factors are sensitive to inhibition by peroxynitrite (10). To determine whether exposure of tumor cells to physiologically relevant concentrations of peroxynitrite would also result in the inactivation of p53 functional activation, we designed experiments to examine these pathways in glioma cells. Here we show that direct bolus administration of peroxynitrite or the simultaneous slow release of the constituents of peroxynitrite, NO, and results in dysregulation of p53 transcriptional activity and downstream pathways.

2024 cells are a clone derived from LN-Z308 human glioblastoma cells (11) after stable transfection with an expression plasmid carrying a wild-type human p53 cDNA regulated by a doxycycline (DOX)-inducible promoter and a neomycin resistance gene (12). D54MG malignant glioma cells were kindly provided by Dr. Darell E. Bigner (Duke University Medical Center). All cells were grown in a humidified incubator at 37°C, 5% CO₂, in DMEM with 10% DOX-free fetal bovine serum (Life Technologies, Inc.), and 2024 cells were maintained in 500 μg/ml G418 and 1 μg/ml puromycin. For determinations of wild-type p53 protein expression activity, 2024 cells were cotransfected with a p53 response element-luciferase reporter construct, which is activated by wild-type p53 but not mutant p53 (13).

2024 cells were grown in DMEM with 10% fetal bovine serum to 70% confluence, and then 20 μg of the luciferase reporter plasmid containing p53 response elements were transfected into cells via electroporation as described previously (13). Cells were allowed to recover for 6–12 h, and DOX (2 μg/ml) treatment was initiated. After 6 h of DOX treatment, bolus peroxynitrite (Cayman Chemical) treatments every 90 min for 9 h (seven treatments during hours 9–15 of DOX administration) were initiated. Total peroxynitrite concentrations ranging from 0.25 to 2 mm were administered as seven equally divided boluses over the 9-h period. For each bolus, cells were taken from the incubator, the media were aspirated, and 5 ml of fresh HBSS were plated onto cells. Peroxynitrite was administered directly into sterile HBSS (Sigma), and the plate was rapidly rotated to disperse the peroxynitrite into solution. After 1 min, HBSS was aspirated from the plate, and the original media containing DOX were replaced, and cells were put back into the incubator until the next peroxynitrite bolus. Control cells were treated identically, except that no peroxynitrite was added. Cells were washed with PBS and lysed with 200 μl of lysis buffer [25 mm trisphosphate (pH 7.8), 2 mm DTT, 2 mm diaminocyclohexane tetraacetic acid, 10% glycerol, and 1% Triton X-100]. Extracts were assayed in triplicate for luciferase activity in a total volume of 130 μl (30 μl of cell extract, 20 mm tricine, 0.1 mm EDTA, 1 mm MgCO₃, 2.67 mm MgSO₄, 33.3 mm DTT, 0.27 mm CoA, 0.47 mm luciferin, and 0.53 mm ATP), and light intensity was measured using a luminometer (Promega, Madison, WI) over a 30-s time period. Peroxynitrite treatments of D54MG cells, p53 immunoprecipitations, and immunoblots for p53 and nitrotyrosine were performed as described previously (3).

Experiments to determine luciferase activity of SIN-1-treated 2024 cells containing a DOX-inducible p53 gene were performed by growing cells to 70% confluence as described, at which time cells were cotransfected with the luciferase reporter construct as described. Cells were allowed to recover for 16 h, after which time media were replaced with DMEM without serum containing DOX (2 μg/ml) or no DOX. Serum was removed so as not to diminish the effect of SIN-1 on the cells. After 6 h, fresh DMEM without serum was added to cells without DOX, and DMEM with DOX (2 μg/ml) was added to two other groups of cells, and SIN-1 (1 mm) was added into one of these groups of cells. This process was repeated every 4 h for the next 16 h, such that two groups of cells received 16 total hours of DOX treatment, and one of these groups received fresh SIN-1 treatment every 4 h during that 16-h period. At the end of the 24-h experiment, cells were harvested, and luciferase activity was determined for each group as described above.

2024 cells were treated with DOX or no DOX to induce p53 as described above. After 9 h, media were removed from the plates of both the non-DOX-treated and DOX-treated cells (2 μg/ml), and HBSS with or without SIN-1 (1 mm) was added. HBSS plus DOX (either with or without SIN-1) was replenished with fresh HBSS containing DOX with or without SIN-1 every 3 h for the last 9 h for a total of 18 h of DOX treatment. Nuclear and cytoplasmic cell lysates were then prepared as described previously (7), and proteins were separated by SDS-PAGE. Immunoblots were performed using anti-p53 polyclonal antibody (1:500; Santa Cruz Biotechnology Biotech) and anti-p21^WAF1 monoclonal antibody (1:200; BD-PharMingen) as described previously (3).

To determine whether direct administration of peroxynitrite could inhibit p53 transcriptional activation, we developed an in vitro cell system containing an inducible wild-type p53 and a p53 response element-luciferase construct that is activated by wild-type but not mutant p53 (13). We used the LN-Z308 human malignant glioma cell line (which does not express p53 protein), into which a DOX-inducible wild-type human p53 gene construct has been stably transfected (Fig. 1; Ref. 12). In these cells (clone 2024), DOX administration induces human wild-type p53 protein expression, which leads to a strong induction of luciferase activity (Fig. 2,A). When we treated 2024 cells with DOX to induce wild-type p53 and concurrently administered boluses (seven boluses over 9 h) of peroxynitrite, we detected a dramatic, dose-dependent decrease in p53-mediated luciferase activity (Fig. 2 A). Repeated bolus administrations of low concentrations of peroxynitrite were used because the half-life of peroxynitrite under physiological conditions is in the millisecond range, the impact of peroxynitrite is cumulative, and administration of a high concentration of peroxynitrite at one time point may be toxic to cells. A marked reduction of p53-mediated luciferase activity was achieved with a total concentration of <0.5 mm peroxynitrite, and at >1 mm peroxynitrite, p53-mediated luciferase activity was completely inhibited. These data indicate that peroxynitrite in the 0.25–1 mm range causes changes in cells leading to inactivation of p53 transcriptional activity.

As controls to show that this decreased luciferase activity was not due to the effects of peroxynitrite on the DOX-regulated transcription system, we repeated the experiments using a different glioma cell line (GM47.23) with a stably transfected dexamethasone-inducible p53 promoter system cotransfected with the same p53 response element-luciferase reporter as described previously (13). In this system, we obtained results identical to those shown in Fig. 2,B. As an additional control to demonstrate that luciferase activity itself was not sensitive to inhibition by peroxynitrite, we used a luciferase reporter construct derived from the p53-luciferase plasmid construct, except that the p53 reporter element was replaced by a GAL4 reporter construct. GAL4 treatment of cells caused potent induction of GAL4-luciferase activity, and 1 mm bolus administration of peroxynitrite had no inhibitory effect on luciferase activity in this system (Fig. 2 C). These control experiments indicate that the inhibitory effects of peroxynitrite on wild-type p53 activity we observed are not due to inactivation of luciferase or the DOX-inducible system by peroxynitrite.

To circumvent the need for bolus treatments and test whether lower levels of sustained peroxynitrite exposure could also inhibit p53 transcriptional activity, we treated cells with SIN-1. This compound decomposes in solution into both and NO, which together form peroxynitrite. The concentrations of and NO produced by treatment of cells with 1 mm SIN-1 are consistent with fluxes of these free radicals encountered in vivo under inflammatory conditions (14, 15). As shown in Fig. 3, only background levels of p53-mediated luciferase activity are detected in 2024 cells without DOX treatment, and a strong induction occurs after 24 h of DOX treatment (2 μg/ml). When DOX-treated cells were exposed to SIN-1 for the last 16 h of the 24-h DOX treatment, almost complete inhibition of p53-mediated luciferase activity was observed. This effect was not due to toxic effects of SIN-1 on the cells because cell viability as determined by trypan blue exclusion did not differ between the two groups of cells after treatment. These data indicate that sustained peroxynitrite production at concentrations similar to those that may occur in vivo can inhibit wild-type p53 transcriptional activity in a manner similar to direct peroxynitrite bolus treatment.

Because these data indicated that p53 transcriptional activity with an exogenous reporter gene is inhibited by peroxynitrite, we next wished to determine whether administration of peroxynitrite could also inhibit induction of endogenous p53 target genes such as p21^WAF1, a principal mediator of the p53-induced G₁ cell cycle arrest (16). For these experiments, we performed immunoblots for p53 and p21^WAF1 before and after DOX treatment of 2024 cells and with or without administration of the peroxynitrite donor SIN-1. Cell lysates were fractionated into nuclear and cytoplasmic components to determine the amounts of p53 and p21^WAF1 localized in each cellular compartment. As illustrated in the control lane in Fig. 4, which shows total untreated 2024 cell total lysate, p53 and p21^WAF1 are scarcely detected before DOX administration. After 18 h of DOX treatment, p53 and p21^WAF1 proteins are both readily detected in nuclear and cytoplasmic fractions. After 1 mm SIN-1 was administered concurrently during hours 10–18 of DOX treatment, a slight decrease of p53 and p21^WAF1 protein was detected in the nucleus and a greater decrease in p53 was observed in the cytoplasm. These data indicate that SIN-1 treatment of 2024 cells that are actively expressing wild-type p53 results in a decrease in detectable cytoplasmic p53 and p21^WAF1 expression, raising the possibility that in addition to decreasing p53 functional activity, oxidation or tyrosine nitration of p53 may cause increased p53 protein degradation.

To confirm that covalent modifications of wild-type p53 protein (including nitration or oxidation) occur after 1 mm SIN-1 exposure, we treated the D54MG cells with an equivalent range of concentrations of authentic peroxynitrite (10–100 μm), calculated based on the fact that 1 mm SIN-1 decomposes in solution at a rate of approximately 2.7 μm/min to form peroxynitrite (15). Such concentrations of peroxynitrite are thought to be similar to those achieved in the brain under conditions of inflammation and hypoxia-reperfusion (17). We treated the D54MG cells with actinomycin-D to induce p53 expression and then administered peroxynitrite to the cells in culture. We then immunoprecipitated the p53 protein with a p53-specific monoclonal antibody and performed immunoblots using a p53-specific polyclonal antibody (Fig. 5,A). The blot was then stripped and reprobed with a monoclonal antibody specific for nitrotyrosine (Fig. 5,B). These data indicated that treatment of D54MG cells in culture with as little as 10 μm authentic peroxynitrite is sufficient to cause nonreducible high molecular weight aggregates of wild-type p53 protein as well as tyrosine nitration of the p53 protein itself (Fig. 5). Such modifications of p53 protein were associated with significant loss of wild-type p53 DNA specific binding in our previous studies (3).

The data presented here indicate that both authentic peroxynitrite and SIN-1, a molecule that produces peroxynitrite by simultaneously generating NO and , can inhibit wild-type p53 protein transcriptional activity in 2024 glioblastoma cells containing a wild-type-specific p53-luciferase reporter construct. Physiologically relevant concentrations of peroxynitrite were associated with decreased p21^WAF1 protein expression and covalent modifications of p53 protein. Overall, these data indicate that concentrations of peroxynitrite similar to those that occur during inflammatory conditions and malignancy can cause posttranslational modifications of wild-type p53 protein that are associated with dysregulation of p53 transcriptional activity and downstream pathways.

In vivo, peroxynitrite is produced during inflammatory conditions by the presence of excess NO and and can posttranslationally modify and inactivate proteins, especially zinc finger transcription factors such as p53. We demonstrated previously that malignant gliomas have evidence of peroxynitrite-mediated tyrosine nitration in vivo and that physiologically relevant concentrations of peroxynitrite could inhibit the specific DNA binding ability of purified wild-type p53 as well as wild-type p53 protein in D54MG cells in culture (3). The data presented here further suggest that peroxynitrite can cause altered transcriptional activity and downstream functional activation of wild-type p53 in both 2024 cells and D54MG cells.

In support of our findings, others have also shown that treatment of tumor cells with very high concentrations of NO can result in tyrosine nitration and mutant conformation of wild-type p53 (7, 18). Because eight of the nine tyrosines in the p53 molecule are located in the critical DNA binding region, it is possible that covalent modification of these residues by p53 could result in a mutant p53 conformation or directly interfere with DNA binding residues. Peroxynitrite could also inhibit wild-type p53 protein function through other mechanisms. Oxidizing agents are known to modify both conformation and sequence-specific DNA binding of p53 in vitro, and peroxynitrite causes zinc to be released from the zinc-thiolate center of zinc finger transcription factors (9, 10). Zinc binding and redox regulation are, at least in part, distinct determinants of the binding of p53 to DNA, and investigators have shown that p53 is subject to more than one level of conformational modulation through oxidation-reduction of cysteines at or near the p53-DNA interface (9). Thus, oxidation of these critical p53 cysteines by peroxynitrite could have dramatic effects on p53 function.

In an inflammatory tumor environment, the effect of peroxynitrite on cell signaling pathways is certain to be complex, and p53 is likely just one of many targets of peroxynitrite modification. Investigators have demonstrated that constituents of the epidermal growth factor receptor, Ras, and p120 Src signaling pathways can become activated by peroxynitrite (19, 20, 21), and thus cell proliferation, growth arrest, and apoptotic pathways are likely affected by peroxynitrite through a complex interplay of multiple activating and inhibitory signaling pathways. Nevertheless, the data presented here indicate that concentrations of peroxynitrite that may occur in vivo can cause posttranslational modification of wild-type p53 and that this may be an important epigenetic mechanism through which cell cycle checkpoint controls can become dysregulated. It is conceivable that with prolonged exposure of cells to increased concentrations of NO and , as might occur in a chronic inflammatory process or malignancy, oncogenic progression could develop due to both the direct mutagenic properties of these reactive species themselves and their posttranslational inhibitory effects on p53 function.

We thank Drs. LeeAnn MacMillan-Crow and Alvaro Estevez for helpful comments and suggestions.