The GTPase KRAS suppresses the p53 tumor suppressor by activating the NRF2-regulated antioxidant defense system in cancer cells

In human cancer cells that harbor mutant KRAS and WT p53 (p53), KRAS contributes to the maintenance of low p53 levels. Moreover, KRAS depletion stabilizes and reactivates p53 and thereby inhibits malignant transformation. However, the mechanism by which KRAS regulates p53 is largely unknown. Recently, we showed that KRAS depletion leads to p53 Ser-15 phosphorylation (P-p53) and increases the levels of p53 and its target p21/WT p53-activated fragment 1 (WAF1)/CIP1. Here, using several human lung cancer cell lines, siRNA-mediated gene silencing, immunoblotting, quantitative RT-PCR, promoter–reporter assays, and reactive oxygen species (ROS) assays, we demonstrate that KRAS maintains low p53 levels by activating the NRF2 (NFE2-related factor 2)–regulated antioxidant defense system. We found that KRAS depletion led to down-regulation of NRF2 and its targets NQO1 (NAD(P)H quinone dehydrogenase 1) and SLC7A11 (solute carrier family 7 member 11), decreased the GSH/GSSG ratio, and increased ROS levels. We noted that the increase in ROS is required for increased P-p53, p53, and p21Waf1/cip1 levels following KRAS depletion. Downstream of KRAS, depletion of RalB (RAS-like proto-oncogene B) and IκB kinase–related TANK-binding kinase 1 (TBK1) activated p53 in a ROS- and NRF2-dependent manner. Consistent with this, the IκB kinase inhibitor BAY11-7085 and dominant-negative mutant IκBαM inhibited NF-κB activity and increased P-p53, p53, and p21Waf1/cip1 levels in a ROS-dependent manner. In conclusion, our findings uncover an important role for the NRF2-regulated antioxidant system in KRAS-mediated p53 suppression.

The p53 protein exerts its tumor suppressor activity by triggering transient cell cycle arrest, cellular senescence, and apoptosis in response to stresses such as oncogene activation and DNA damage (1)(2)(3)(4). In the absence of stress, normal cells maintain low p53 levels by inducing its degradation. MDM2, an E3 ubiquitin-protein ligase that directly binds p53 and targets it for protein degradation by the proteasome, plays a major role in maintaining low levels of p53 in normal cells (5,6). The binding of MDM2 to p53 is inhibited by the phosphorylation of p53 on amino acids involved in the MDM2/p53-binding interface. For example, during DNA damage, ataxia telangiectasia mutated (ATM) 2 kinase is activated to phosphorylate p53 on Ser-15, which leads to its stabilization by inhibiting binding of MDM2 to p53, its ubiquitination, and its degradation (7)(8)(9). As such, the stabilized p53 allows the cells to repair the DNA damage by inducing cell cycle arrest at the G 1 phase (10), which is mediated at least in part by transcriptional activation of the cyclindependent kinase inhibitor p21 Waf1/cip1 . Thus, one proposed mechanism by which p53 prevents cancer is by arresting or eliminating cells with DNA damage and hence limiting the formation of oncogenic mutations.
Most human cancers protect themselves from the tumor suppressor activity of p53 by inactivating it, leading to genetic instability and increased mutation frequency. Approximately half of human cancers harbor mutations in the TP53 gene, which results in loss of WT p53 function as well as oncogenic gain of function (3,4). In the remaining cancers that harbor WT TP53, several oncogenic pathways lead to the inactivation of the p53 protein. In human cancer cells that harbor mutant (mt) KRas and WT p53, we have shown that KRas and RalB, but not Akt1/2 and c-Raf, are required for maintaining low levels of p53 and that depletion of KRas and RalB stabilizes and reactivates p53 to inhibit malignant transformation (11). Importantly, depletion of KRas and RalB led to ATM kinase activation and phosphorylation of p53 on Ser-15 and the subsequent increase in p53 half-life, which in turn led to a p53-dependent up-regulation of its transcriptional target p21 Waf1/cip1 (11). However, the mechanism by which depletion of KRas and RalB leads to activation of ATM and the subsequent activation of p53 is not known.

Depletion of KRas and RalB promotes p53 phosphorylation on serine 15 and p53 activation by a ROS-dependent mechanism
Because ATM can be activated directly by reactive oxygen species (ROS) (12), we reasoned that depletion of KRas and RalB may lead to the generation of ROS and subsequent phosphorylation of p53. To test this hypothesis, we depleted KRas and RalB from A549 human lung cancer cells, which harbor mt KRas and WT p53, and processed the cells for ROS-level and Western blotting analyses as described under "Experimental procedures." Depletion of KRas and RalB resulted in increased cellular ROS levels ( Fig. 1A) and increased levels of the phosphorylation of p53 on Ser-15 (Fig. 1C). This in turn led to increased levels of total p53 and its transcriptional target p21 Waf1/cip1 (Fig. 1C). Consistent with the ROS-level increase, Fig. 1B shows that depletion of KRas and RalB decreased the ratio of the reduced form of GSH to its oxidized form (GSSG) (GSH/GSSG). These effects were Ras isoform-specific in that KRas, but not HRas and NRas, depletion increased p53 levels in A549 cells (Fig. S1A). Similarly, KRas and RalB, but not HRas and NRas depletion, increased p53 levels in H460, another human lung cancer cell line with mt KRas and WT p53 (Fig. S1B). We next determined whether the increase in the phosphorylation of p53 and its up-regulation, following KRas and RalB depletion, require ROS, by treating the cells with the ROS scavenger and potent antioxidant N-acetylcysteine (NAC). Fig. 1D shows that NAC attenuated the increased phosphorylation of p53 on Ser-15 and total p53 and p21 Waf1/cip1 levels following depletion of KRas and RalB in A549 cells. In two other human lung cancer cell lines with mt KRas and WT p53 (H460 and H1944), NAC treatment also prevented increased phosphorylation of p53 on Ser-15 and increased total p53 and p21 Waf1/cip1 levels following depletion of KRas and RalB (Fig. S2, A and B). These results indicate that depletion of KRas and RalB leads to phosphorylation and up-regulation of p53 by a mechanism requiring ROS. To provide further support for the involvement of ROS in the up-regulation of p53, we treated A549 cells with erastin, an inhibitor of SLC7A11, the cystine-glutamate antiporter that decreases levels of ROS by increasing cellular levels of cysteine and subsequently the antioxidant GSH (13). Fig. 1 (E and F) shows that treatment with erastin increased cellular ROS levels and promoted p53 phosphorylation and increased p53 and p21 Waf1/cip1 levels.

Depletion of KRas and RalB down-regulates the expression of the master regulator of the antioxidant defense system Nrf2 and its transcriptional targets
The fact that depletion of KRas and RalB resulted in increased ROS levels suggested the involvement of the antioxidant defense system in p53 induction. To explore this mechanism further, we first determined whether KRas and RalB are required for the expression of the master regulator of the anti- KRas suppresses p53 by Nrf2 activation oxidant defense system, the basic leucine zipper transcription factor Nrf2 (nuclear factor (erythroid-derived 2)-like 2) (14 -16). Nrf2 protects cells from oxidative damage by inducing the expression of several antioxidant genes that maintain a redox balance by reducing intracellular ROS levels. These genes include, for example, NQO1 (NAD(P)H quinone oxidoreductase), an enzyme responsible for reduction/detoxification of reactive quinones that cause redox cycling and oxidative stress (17), and the above-mentioned cystine-glutamate antiporter SLC7A11 (13). As shown in Fig We next determined whether the decrease in Nrf2 expression mediates p53 up-regulation following KRas and RalB depletion. We reasoned that if this is the case, then a compound able to increase Nrf2 levels could rescue from KRas and RalB siRNA up-regulation of p53. To this end, we used the clinical agent bardoxolone-methyl (CDDO-ME), a compound known to greatly increase the cellular levels of Nrf2 by binding to and preventing the Nrf2 negative regulator, Keap1, from inducing the ubiquitination and proteasomal degradation of Nrf2. Fig. S3 shows that in control vehicle-treated A549 cells, the levels of

KRas suppresses p53 by Nrf2 activation
P-p53, p53, and p21 were increased following depletion of KRas and RalB relative to nontargeting siRNA. In contrast, in CDDO-ME-treated cells, Nrf2 levels were greatly increased, and the ability of KRas and RalB siRNAs to increase the levels of P-p53, p53, and p21 was blocked (Fig. S3), suggesting that low levels of Nrf2 are required for p53 up-regulation following KRas and RalB depletion.

IB kinase-related TBK1 depletion leads to activation of p53 in a ROS-dependent manner
To determine the mechanism by which depletion of KRas and RalB down-regulates Nrf2 and activates p53, we focused on the IB kinase (IKK)-related TBK1 (Tank-binding kinase 1), because TBK1 has been shown to bind in a complex with RalB and its effector Sec5, and RalB/TBK1 activation has been shown to contribute to cancer cell survival and oncogenesis (18). Therefore, we reasoned that if TBK1 mediates the ability of RalB to regulate Nrf2 levels and p53 up-regulation, then TBK1 itself may be required for Nrf2 expression and suppression of p53 up-regulation. To test this hypothesis, we depleted TBK1 from A549 cells and determined whether it is required for the expression of Nrf2 and suppression of p53. Fig. 3A shows that depletion of TBK1 resulted in decreased Nrf2 expression and increased p53 phosphorylation and total p53 and p21 Waf1/cip1 levels. The decrease in Nrf2 expression was at the transcriptional level, as documented by the reduction of the mRNA lev-els of Nrf2 (Fig. 3B) and its target genes NQO1 (Fig. 3C) and SLC7A11 (Fig. 3D). Consistent with the decreased expression of Nrf2, depletion of TBK1 also resulted in increased intracellular ROS levels (Fig. 3E). The ROS scavenger NAC prevented TBK1 siRNA from activating p53 (Fig. 3F), suggesting that ROS is required for p53 activation following TBK1 depletion. Taken together, the results shown in Figs. 1-3 suggest that KRas, RalB, and TBK1 suppress p53 by activation of the antioxidant defense system. TBK1 and IKK⑀ are serine/threonine protein kinases related to the IKK family of kinases and play a key role in coordinating the activation of the NF-B pathway. TBK1, just like IKK⑀, can phosphorylate the NF-B inhibitor IB␣ and promote its proteasomal degradation, thereby allowing activation and nuclear translocation of the NF-B complex (19,20). To provide further support for the involvement of TBK1 and the NF-B pathway in the regulation of p53 through the antioxidant defense system, we next investigated the effects of inhibiting the NF-B pathway on the Nrf2 anti-oxidant system, ROS levels, and p53 activation. To this end, A549 cells were co-transfected with the dominant-negative IB␣ mt IB␣M (21) and a NF-Bresponsive promoter-reporter construct; the cells were then treated with vehicle or NAC. Fig. 4A shows that expression of the IB␣M results in decreased NF-B promoter activity, confirming that IB␣M inhibited the NF-B pathway in A549 cells.

KRas suppresses p53 by Nrf2 activation
Furthermore, in the absence of NAC, IB␣M expression resulted in decreased Nrf2, Nqo1, and SLC7A11 mRNA (Fig.  4B); decreased Nrf2 protein levels (Fig. 4D); increased ROS levels (Fig. 4C); and increased levels of p53 phosphorylation, p53, and p21 Waf1/cip1 (Fig. 4D). The induced ROS and activation of p53 by IB␣M was abrogated by NAC (Fig. 4, C and D), suggesting that the ability of IB␣M to activate p53 depends on ROS. Further support for the involvement of the NF-B pathway in p53 activation through the antioxidant defense system was provided by experiments using BAY 11-7085, an irreversible inhibitor of IB␣ phosphorylation that prevents activation of the NF-B pathway (22). Treatment of A549 cells with BAY 11-7085 resulted in decreased NF-B promoter activity (Fig.  4E), confirming that it inhibited the NF-B pathway; decreased Nrf2, Nqo1, and SLC7A11 mRNA (Fig. 4F); decreased Nrf2 protein levels (Fig. 4H); increased ROS levels (Fig. 4G); and increased levels of p53 phosphorylation, p53, and p21 Waf1/cip1 in a ROS-dependent manner (Fig. 4H). Taken together, these data suggest the involvement of the NF-B pathway in the regulation of p53 through the antioxidant defense system.

Discussion
The genome guardian tumor suppressor p53 is one of the pivotal gate keepers that prevent normal cells from becoming cancerous; therefore, it is not surprising that approximately half of cancers contain p53-inactivating mutations and the other half harbor oncogenic events that inactivate WT p53 (3,4,23). Consistent with this, we have shown that in human tumors with WT p53, mt KRas is required for maintaining low p53 levels and that depletion of KRas increases p53 stability by activating ATM to phosphorylate p53 on Ser-15, leading to a 6-fold increase in p53 half-life (11). However, the mechanism by which KRas regulates ATM phosphorylation of p53 Ser-15 to maintain low p53 levels remained unknown. The fact that ROS can directly activate ATM by reacting with ATM Cys-2991 and inducing a disulfide cross-linked, oxidized, and active ATM dimer (12) prompted us to test the hypothesis that KRas regulation of p53 is mediated by ROS. In this study, we discovered that KRas regulation of p53 Ser-15 phosphorylation depends on ROS cellular levels and is mediated by the antioxidant program (Fig. 5).
In contrast to normal cells that maintain redox homeostasis through an antioxidant defense system tightly regulated by Nrf2, cancer cells protect themselves from oxidative stress by aberrantly up-regulating Nrf2 (24 -27). In this study, we demonstrated that KRas depletion decreased expression of Nrf2 and its transcriptional targets the ROS detoxifying/ antioxidant enzyme NQO1 and the cysteine-glutamate antiporter SLC7A11 (28,29). Consistent with this, depletion of

KRas suppresses p53 by Nrf2 activation
KRas decreased the GSH/GSSG ratio and increased ROS levels.
Although it has been shown that KRas G12D activates Nrf2 and lowers ROS levels (30), whether ROS mediates the ability of mt KRas to maintain low p53 levels is not known. Here, we showed that the ROS scavenger NAC rescued from p53 activation induced by depletion of KRas, suggesting that ROS is required for p53 Ser-15 phosphorylation and up-regulation. Furthermore, CDDO-ME, a compound known to increase the levels of Nrf2 by binding to Keap1 and inhibiting the proteasomal degradation of Nrf2, rescued from KRas siRNA up-regulation of p53, suggesting that low levels of Nrf2 are required for p53 up-regulation. Our results suggest that depletion of KRas leads to down-regulation of Nrf2 and its antioxidant targets, decreased GSH/GSSG ratio, and increased ROS levels, which in turn can activate ATM to phosphorylate Ser-15 p53, preventing its degradation and hence increasing its stability. These results are significant because they suggest that one mechanism by which tumors that harbor mt KRas and WT p53 overcome the ability of WT p53 to induce oxidative damage and tumor cell death is through Nrf2 up-regulation and low ROS levels and therefore low levels of Ser-15 phosphorylation leading to p53 degradation. We found the requirement of KRas for the maintenance of low p53 levels is Ras isoform-specific because depletion of HRas and NRas resulted in little up-regulation of p53. This is consistent with previous reports showing that overexpressing mt KRas, but not HRas or NRas, reduced p53 levels (31). Furthermore, we have shown that downstream of KRas, RalB, but not Akt1/2 and c-Raf, is required for maintaining low levels of p53 (11). However, whether RalB regulates p53 stability through the antioxidant defense system is not known. In this study, we found that depletion of RalB leads to down-regulation of Nrf2, decreased GSH/GSSG ratio, and increased ROS and that the resulting increased p53 Ser-15 phosphorylation and p53 upregulation require increased intracellular levels of ROS.
RalB is a member of an IB kinase-related TBK1 complex that contributes to cancer cell survival (18), which led us to investigate whether TBK1 also regulates p53 stability. Our results show that depletion of TBK1 increases p53 Ser-15 phosphorylation and up-regulates p53 and that this is mediated by ROS. Further support for the regulation of p53 by the TBK1/ NF-B pathway was provided by the demonstration that the NF-B pathway inhibitors, BAY 11-7085 and the dominantnegative IB␣M, increased p53 phosphorylation and p53 levels in a ROS-dependent manner. Although the involvement of the TBK1/NF-B pathway in the regulation of p53 and its dependence on the anti-oxidant program is novel, the decrease in Nrf2 levels by BAY 11-7085 and IB␣M is consistent with reports showing that NF-B binds B-binding sites on the Nrf2 promoter and increases its mRNA expression in leukemia cells (32). Furthermore, the same study showed that NF-B inhibitors decreased Nrf2 levels and significantly reduced clonogenicity of leukemia patient cells and improved their chemotherapeutic responsiveness (32). Although our studies do not demonstrate a causal link between Nrf2 and the NF-B pathway-dependent regulation of p53, together with these previous results (32), they suggest that one potential way that tumors maintain low levels of ROS and p53 is through TBK1/NF-Bmediated Nrf2 redox detoxification.
Although our results show that downstream of KRas, the Ral/TBK1/NF-B pathway contributes to the tumor's ability to maintain low ROS levels, other pathways may also be involved in KRas maintaining low ROS levels. For example, in MEFs expressing G12D KRas, the MEK inhibitor AZD6244 decreased

KRas suppresses p53 by Nrf2 activation
Nrf2 and its target genes and increased ROS levels, whereas endogenous expression of activated B-Raf increased P-ERK and Nrf2 and decreased ROS (30). The demonstration that depletion of KRas, RalB, and TBK1, ectopic expression of IB␣M, and treatment with BAY 11-7085 all lead to decreased Nrf2 and increased ROS opens up new therapeutic avenues of targeting these pathways to up-regulate WT p53, which in turn can induce tumor cell death. This is significant because tumor cells and cancer stem cells must maintain low levels of intracellular ROS for their survival and to preserve self-renewal capacities (24,26,33,34). Finally, although studies in this manuscript do not demonstrate a tumorigenic potential of Nrf2-mediated alterations of the p53 pathway, our previous studies showed that depletion of K-Ras and RalB activates p53 and inhibits anchorage-independent growth and invasion and interferes with cell cycle progression in a p53-dependent manner (11).
Taken together, our results have provided mechanistic insights to the central question of how mt KRas regulates WT p53 by uncovering the critical role of the Nrf2 antioxidant defense system in KRas regulation of p53 phosphorylation and stability (Fig. 5). This is a novel mechanism that is relevant to KRas-driven cancer development. Nrf2, a key regulator for the maintenance of redox homeostasis, has been shown to contribute to several hallmarks of cancer, including uncontrolled proliferation, self-renewal, metabolic reprogramming, drug resistance, invasion, and metastasis (14 -16). In the present study, we provided evidence for a new oncogenic role of Nrf2 in cancer development as a mediator of oncogenic KRas suppression of p53.

siRNA transfection
Cancer cells (0.2 ϫ 10 6 /well) were plated in 6-well plates and transiently transfected the next day with 20 M of siRNA using Lipofectamine RNAiMAX transfection reagent (re-agent 13778-150; Invitrogen) according to the manufacturer's instructions. Transfected cells were collected 48 h after transfection for ROS and GSH assays and RT-PCR and Western blotting analyses.

DNA transfection
The cells were plated overnight to reach 70 -80% confluency and were then transfected with plasmid DNA of pCMX IB␣ mt (IB␣M, Addgene, 12329) using Lipofectamine 2000 (11668-027, Invitrogen) transfection reagent according to the manufacturer's instructions. Transfected cells were harvested 48 h after transfection for ROS, luciferase reporter assay, and assessment of mRNA and protein expression levels.

Western blotting
The cells were harvested, washed, and processed for Western blotting as described by us (11).

Measurement of intracellular ROS levels
Intracellular ROS levels were estimated using the cell-permeable oxidation-sensitive probe H2DCF-DA. The cells were seeded into dark, clear-bottomed, 96-well plates at 1000 cells/ well. The cells were transfected the next day with siRNA; 48 h after transfection, the cells were washed with Hanks' balanced salt solution (HBSS; 14025-076, Invitrogen). The cells were then incubated with 20 M of H2DCF-DA in HBSS in the dark at 37°C for 45 min. After incubation, H2DCF-DA was removed, and the cells were washed with HBSS. The ROS levels were determined by measurement of fluorescence of dichlorofluorescein (excitation at 480 nm and emission at 530 nm) and normalized to cellular protein content, which was determined by the sulforhodamine B (SRB, 230162, Sigma) assay (35).

Microtiter plate assay for the measurement of GSH/GSSG ratio
The cells were harvested using 0.05% Trypsin-EDTA (1ϫ) and washed three times with ice-cold PBS. GSH and GSSG levels were measured by enzymatic assays using the Enzy-Chrom GSH/GSSG assay kit (EGTT-100; BioAssay System, Hayward, CA), following the manufacturer's instructions. Detection of reaction products was monitored at 0 and 10 min using a microplate reader at 412 nm. GSH and GSSG levels were calculated from standard curves of GSH and GSSG.

KRas suppresses p53 by Nrf2 activation
TGATGG-3Ј (reverse) for GAPDH. All primers were synthesized by Integrated DNA Technologies. The RT-PCRs were performed in 20-l volumes that included 2 l of RT products (1:100 diluted) as template, 10 l of SYBR green PCR master mix, and 20 pmol of individual primer. The PCR amplifications (40 cycles at 95°C for 15 s and at 55°C for 60 s) were performed using the QuantStudio 3 RT-PCR system (Applied Biosystems). For each target gene, the average Ct values calculated from triplicate PCRs were normalized to the average Ct values for GAPDH. These normalized values were then used to calculate a value expressing the extent of knockdown relative to the nonspecific control siRNA according to the formula 2 Ϫ(mean⌬⌬Ct) .

Luciferase reporter assay
The cells (3000/well) were seeded in 96-well plates and cotransfected the next day with DNA: 300 ng of pNFB MetLuc reporter (Clontech) and 20 ng of ␤-gal according to the manufacturer's instructions.

Statistical analyses
The graphics were prepared using GraphPad Prism software. In this study, t tests were used to determine the statistical significance of the results. p values Ͻ 0.05 were chosen as a threshold for statistical significance.