SIRT2 Deacetylates and Inhibits the Peroxidase Activity of Peroxiredoxin-1 to Sensitize Breast Cancer Cells to Oxidant Stress-Inducing Agents

SIRT2 is a predominantly cytoplasmic member of the class III histone deacetylases (HDAC), or sirtuins, which function as NAD⁺-dependent lysine deacetylases (1–3). Similar to HDAC6, SIRT2 colocalizes with microtubules in the cytosol and deacetylates lysine 40 on α-Tubulin (2). During G₂–M phase of the cell cycle, SIRT2 is nuclear, where it deacetylates histone protein residues H4K16, H3K56, and H3K18 as well as the BUB1-related kinase (BUBR1), thereby controlling the activity of the anaphase-promoting complex/cyclosome (APC/C) and normal mitotic progression (4–6). Because of this activity, SIRT2 prevents chromosomal instability during mitosis (4, 5). SIRT2 has also been recognized as a tumor suppressor, because its loss in mice is associated with mammary tumors and hepatocellular carcinomas (5). Among other notable substrates deacetylated by SIRT2 are metabolic enzymes, including glucose 6 phosphate dehydrogenase (G6PD), ATP citrate lyase (ACLY), and phosphoglycerate mutase (PGAM; refs. 7–9). Oxidative stress has been shown to induce SIRT2-mediated deacetylation of the transcription factor FOXO3a (10). This induces the transcriptional activity of FOXO3a, causing increased expression of its target genes, including BIM (11). In NIH3T3 cells, ectopic overexpression of SIRT2 was also shown to induce BIM and promote cell death following exposure to hydrogen peroxide (H₂O₂; ref. 10). Recently, SIRT2 was shown to interact with receptor-interacting protein 3 (RIP3) and deacetylate RIP1, leading to the formation of a stable RIP1/RIP3 complex and promotion of TNFα-induced necroptosis (12).

H₂O₂ is a toxic byproduct of normal cellular processes in aeorobic organisms and is detoxified by antioxidant enzymes including catalase, glutathione peroxidases, and peroxiredoxins (Prdx; refs. 13, 14). Prdxs are a family of ubiquitously expressed, 22 to 27 kDa, thiol-dependent peroxidases, with a conserved cysteine residue (15, 16). Prdx-1 is a two-cysteine residue member of the PRDX family of proteins (15, 16). Prdx-1 exists as a homodimer and reduces H₂O₂, utilizing thioredoxin (TRX) as the electron donor for the antioxidation (14, 16). Prdx-1 is expressed at high levels in the cytosol of transformed cells and is further induced by oxidative stress, for example, due to exposure to H₂O₂, which oxidizes the conserved cysteine of Prdx-1 to sulfenic acid (15, 16). Besides its cytoprotective antioxidant function, Prdx-1 plays a role in cellular processes involving redox signaling and reactive oxygen species (ROS; refs. 17, 18). In the current studies, we determined that SIRT2 binds and acts as a deacetylase for Prdx-1; whereas knockdown (KD) of SIRT2 induces acetylation, ectopic overexpression of SIRT2 deacetylates, and inhibits the ROS-neutralizing, antioxidant activity of Prdx-1. This sensitized breast cancer cells to DNA damage and apoptosis induced by H₂O₂ through a FOXO3a-BIM–mediated cell death mechanism. Consistent with this, SIRT2 overexpression also increased cell death induced by ROS-inducing agents, including arsenic trioxide (AT) and menadione.

AT and menadione were purchased from Sigma Aldrich. All antibodies were obtained from commercial sources. Detailed descriptions of the antibodies are provided in the Supplementary Methods.

The breast cancer MCF7 and MDA-MB-231, as well as HEK293 cells were obtained from ATCC. Cells were thawed, passaged, and refrozen in aliquots. Cells were used within 6 months of thawing or obtaining from ATCC. The ATCC utilizes short tandem repeat (STR) profiling for characterization and authentication of cell lines. MCF7 and HEK293 cells were cultured in DMEM with 10% FBS and 1% penicillin/streptomycin and passaged two to three times per week. MDA-MB-231 cells were cultured in RPMI1640 media containing 10% FBS and 1% penicillin/streptomycin and passaged two to three times per week (19, 20). Logarithmically growing cells were exposed to the designated concentrations and exposure interval of the drugs. Following these treatments, cells were washed free of the drug(s) using 1× PBS, and pelleted prior to performing the studies described later.

Following treatments, cells were washed with 1× PBS, then trypsinized and pelleted. Total cell lysates were combined with class-specific IgG or 2 μg of anti-Prdx-1 or anti-HA (HA-SIRT2) or anti-FLAG M2 antibody and incubated with rotation overnight at 4°C. The following day, protein A beads were added and the lysate bead mixture was rotated for 90 minutes 4°C. The beads were washed with 1× PBS three times and sample buffer was added. The beads were boiled and the samples were loaded for SDS-PAGE. Immunoblot analyses were conducted for Prdx-1, SIRT2, or acetyl lysine, as described previously (19, 20).

Seventy-five micrograms of total cell lysates were used for SDS-PAGE and immunoblot analyses, as described previously (21).

MCF7 or MDA-MB231 cells were labeled with immunofluorescence-tagged antibodies, as described previously (22).

Untreated or drug-treated cells were stained with propidium iodide (PI), and the percentage of PI-positive cells was determined by flow cytometry, as described previously (21, 22).

The Prdx activity assay was carried out according to manufacturer's instructions (Redoxica) and as previously described (23). Briefly, cells were collected by trypsinization. The cells were washed with 1× PBS and then sonicated in the activity assay buffer. The total reaction volume of 150 μL contained 50 mmol/L HEPES-NaOH buffer, Escherichia coli thioredoxin, mammalian thioredoxin reductase, and NADPH. The reaction was initiated by the addition of 2 μL of 10 mmol/L H₂O₂. NADH oxidation was monitored for 10 minutes at 340 nm.

MCF7 and HEK293 cells were transiently transfected according to the instructions of the manufacturer using Fugene 6 with plasmids containing scrambled oligonucleotide (control shRNA) or shRNA to SIRT2 containing a 21-nucleotide sequence, corresponding to SIRT2 mRNA—5′-GAAACATCCGGAACCCTTC-3′, as described previously (24). For interaction studies, HEK293 cells were transfected with pcDNA (control vector) with or plasmids for HA-SIRT2 or FLAG-Prdx-1.

DNA damage and repair at an individual cell level was determined by the comet assay as previously described (19, 20). A detailed method is provided in the Supplementary Methods.

MDA-MB231 cells were grown in black 96-well plates overnight at 37°C. The next day, the cells were treated with H₂O₂ for 30 minutes at 37°C. The media were aspirated, and the cells were washed with 1× PBS and DCF-DA in phenol red–free media at a final concentration of 10 μmol/L was added to the cells and incubated for 30 minutes. The dye was washed with 1× PBS and the fluorescence was read using a BioTek fluorescence plate reader (19).

S100 cytosolic extracts were prepared from HEK293 vector and SIRT2 KD cell lines. Equal amounts of protein were subjected to immunoprecipitation with anti-acetyl lysine antibody. The proteins were eluted with glycine buffer (pH 2.7). The proteins were then subjected to in vitro labeling with Cy-3 and Cy-5 N-hydroxysuccinimidyl ester. Cy-2 was used as an internal standard. The samples were subjected to isoelectric focusing and then separated in a second dimension by SDS-PAGE. The gels were fixed, stained, and protein spots were analyzed using GE Healthcare DeCyder software. The protein spots of interest were subjected to automated in-gel tryptic digestion and MALDI/MS/MS spectra were performed with 4800 Proteomics Analyzer MALDI-TOF/TOF mass spectrometer (Applied Biosystems; refs. 24, 25).

MDA-MB-231, MCF7 and HEK293 vector, and SIRT2 O/E cells were exposed to the indicated concentration of H₂O₂ for different time points and the lysates were resolved in a 8% native PAGE electrophoresis. The formation of high molecular weight oligomers were visualized with antioxidized-Prdx-1 or anti-Prdx-1 antibody, as described previously (18, 26). In vivo overoxidation of Prdx-1 was determined, as described previously (27).

HEK293, MDA-MB-231, and MCF7 cells stably transfected with vector, or SIRT2 cDNA were treated with H₂O₂ for 30 minutes, respectively, and the cells were harvested and lysed with native lysis buffer. Protein concentrations were measured using a BCA Kit. Endogenous oligomers and multimers of Prdx-1 protein were immunoprecipitated using Dynabeads M-280 sheep anti-mouse IgG (Invitrogen) and anti-Prdx-1 (LF-MA0214; AbFrontier) according to the manufacturer's instructions. The immunoprecipitated Prdx-1 and Dynabeads were used for chaperone activity assay as described previously (18, 26). Briefly, each reaction contained 2 μmol/L malate dehydrogenase (MDH) and immunoprecipitated chaperone with Dynabeads in 200 μL of 50 mmol/L HEPES-KOH buffer (pH 7.5). The chaperone activity was monitored by measuring the absorbance (320 nm) in a BioTeck SynergyMx plate reader at 43°C for 2 hours.

Wild-type (AB) zebrafish were maintained using standard procedures (28). Zebrafish embryos and larvae were obtained by natural mating. Morpholino oligonucleotides were injected into yolk at the one-cell stage using an IM300 microinjector (Narishige; ref. 29). The SIRT2 MO (Genetools LLC) were designed against the splice-donor sites of exon 6 of SIRT2: 5′-TATGTAAAGTCAGACCTGTTTGTG-3′. SIRT2 MO was injected (0.5 nL) into the yolk of 1-cell-stage zebrafish embryos at a final quantity of 4 ng. To validate the KD, total RNA was extracted from 48 hpf embryos using TRIzol reagent. Two hundred nanograms of total RNA was reverse transcribed using a High-Capacity Reverse Transcription Kit (Applied Biosystems) following manufacturer's protocol. qRT-PCR was carried out using SYBR green. For evaluation of toxicity, 48 hpf embryos injected with SIRT2 MO or a 5-bp mismatch control were treated with the indicated concentrations of H₂O₂. At the end of 48 hours of treatment no mortality was observed but there were morphologic abnormalities, as described previously (29). The embryos were placed in tricaine solution and imaged using an epifluorescence microscope. For estimation of the ROS, 48 hpf embryos injected with the SIRT2 MO or 5-bp mismatch control were treated with H₂O₂ for 30 minutes. The embryos were then incubated with 50 μmol/L DCF-DA (Invitrogen). Individual embryos were transferred to each well of a 96-well plate and read using a plate reader. At least six embryos per experimental condition were used.

Among the known cytosolic targets of deacetylation by SIRT2 is α-tubulin (2). As shown in Fig. 1A, KD of SIRT2 with two separate shRNAs stably transduced into HEK-293 cells induced the acetylation of α-tubulin, without altering the levels of the total α-tubulin. In addition, lysates from cells transduced with the control shRNA or SIRT2 shRNA were also immunoblotted with anti-acetyl-lysine antibody, again demonstrating increased acetylation of several proteins, including α-tubulin and histone H3 (Fig. 1B). We next performed the two-dimensional differential in-gel electrophoresis (DIGE) analysis on the S100 cytosolic extracts from HEK293 shRNA control and SIRT2 KD cell lines. Figure 1C demonstrates a representative two-dimensional-gel image showing the protein spots exhibiting more than a 2-fold difference in the mobility between the vector control and SIRT2 KD–treated samples. One of the spots demonstrating a 2.15-fold change in its mobility was identified by mass spectrometry to be Prdx-1. We also identified phosphoglycerate kinase 1 (PGK1) and MDH (Fig. 1C). To confirm that SIRT2 interacts with Prdx-1, we introduced the FLAG-tagged Prdx-1 and HA-tagged SIRT2 into HEK293 cells. The cells were lysed, and the anti-HA antibody immunoprecipitates were then immunoblotted with the anti-FLAG and anti-HA antibodies. As shown in Fig. 2A, the epitope-tagged Prdx-1 coimmunoprecipitated with the epitope-tagged SIRT2. We further confirmed that the endogenous Prdx-1 also coimmunoprecipitated with the FLAG-tagged SIRT2 introduced into MDA-MB-231 (Fig. 2B) and MCF7 cells (Fig. 2C). Next, we determined whether the genetic KD or chemical inhibition of SIRT2 affects the acetylation of the endogenous Prdx-1 in MCF7 cells. Figure 2D demonstrates that chemical inhibition of the catalytic activity of SIRT2 by nicotinamide (NA) or shRNA-mediated 90% KD of SIRT2-induced lysine acetylation of Prdx-1. We next performed the reverse immunoprecipitation with anti-acetylated lysine in cells with ectopic overexpression or KD of SIRT2. Figure 2E shows that ectopic overexpression of SIRT2 (shown above in Fig. 2B) deacetylates Prdx-1, resulting in attenuated levels of the immunoprecipitated, acetylated Prdx-1. Conversely, SIRT2 KD caused an increase in the levels of the immunoprecipitated acetylated Prdx-1 (Fig. 2E).

Next, we determined whether ectopic overexpression or activation of SIRT2 or KD of SIRT2 by shRNA perturbs acetylation of the endogenous Prdx-1. For this, we ectopically overexpressed or knocked-down SIRT2 by shRNA in MCF7 and MB-231 cells (Fig. 3A). We also ectopically expressed the catalytically inactive mutant form of SIRT2 (H187A), created by site-directed mutagenesis (10), in MB-231 cells (Fig. 3A and B). First, we confirmed that the ectopic expression of SIRT2, but not of the mutant SIRT2, increased the intracellular lysine deacetylase activity in MB-231 and MCF7 cells (Supplementary Fig. S1 and data not shown). However, enforced alterations in the levels of SIRT2 neither affected the levels of Prdx-1, nor led to any change in the levels of several other HDACs, for example, SIRT3, SIRT6, and HDAC6, or of the antioxidant proteins, including superoxide dismutase and catalase (Supplementary Fig. S2A and data not shown). Ectopically expressed wild-type and mutant SIRT2 were localized to the cytoplasm of MB-231 and MCF7 cells (Fig. 3B and Supplementary Fig. S2B). As shown in Fig. 3C, compared with the MB-231 transduced with the vector alone, by deacetylating Prdx-1, ectopically overexpressed SIRT2 significantly inhibited the antioxidant activity of Prdx-1 (P = 0.01). Conversely, MB-231 cells exhibiting shRNA-mediated KD of SIRT2 showed a significant increase in the antioxidant activity of the resulting hyperacetylated Prdx-1 (Fig. 3C). Here, the antioxidant activity of the deacetylated or hyperacetylated Prdx-1 in the cellular protein extract was assayed by estimating its reducing effect on H₂O₂-mediated NADPH oxidation, which was compared with the antioxidant activity of the recombinant Prdx-1 in the same assay (Fig. 3C). As compared with the control MB-231 (transduced with vector alone), treatment of MB-231 cells with ectopic overexpression of SIRT2 with 500 μmol/L of H₂O₂ resulted in significantly increased intracellular levels of ROS (P = 0.01; Fig. 3D). Exposure to H₂O₂ caused overoxidation of the active sulfhydryl residues in Prdx-1 to sulfinic or sulfonic acid, which was detected by utilizing the antibody that recognizes the overoxidized cysteine in Prdx-1 (27), in MB-231 cells overexpressing wild-type SIRT2 but not those expressing the catalytically dead mutant form of SIRT2 (Fig. 3E). Following exposure of the MB-231 cells expressing the catalytically dead mutant form of SIRT2 to H₂O₂, no increase in the levels of the oxidized Prdx-1 was observed (Fig. 3E). Therefore, the reduced antioxidant activity of the de-acetylated Prdx-1 was associated with increased levels of ROS as well as increase in the oxidized form of Prdx-1, whereas in the catalytically dead mutant, Prdx-1 was relatively resistant to oxidation. Collectively these findings highlight that, although the antioxidant activity of Prdx-1 regulates intracellular levels of H₂O₂, high levels of H₂O₂ in turn oxidize and regulate the antioxidant activity of Prdx-1 (15, 16).

The cytosolic 2-Cys Prdx-1 has dual function as a peroxidase and molecular chaperone (18). Upon exposure to oxidative stress, Prdx-1 transitions to high molecular weight oligomers and functions as a molecular chaperone (30). Removal of the oxidative stress switches Prdx-1 back from a chaperone to a low molecular weight peroxidase function (18, 30). Next, we determined whether in SIRT2-overexpressing cells expressing deacetylated Prdx-1, the latter is overoxidized due to increased intracellular levels of ROS, and whether Prdx-1 would form multimers and exhibit increased chaperone activity toward one of its known substrate proteins, MDH (31). Figure 4A and Supplementary Fig. S3 demonstrate that following exposure to 500 μmol/L of H₂O₂ for 30 minutes to 4 hours, as compared with their respective controls, HEK293 and MB-231 cells overexpressing SIRT2 exhibited increased levels of the high molecular weight multimers of Prdx-1 and their oxidized counterparts (right and left panel), as detected by anti-Prdx-1 and anti-oxidized Prdx-1 antibody, respectively. Prdx-1 displays a robust capacity to suppress the misfolding of MDH. Thus, the increased chaperone function of the oxidized Prdx-1 multimers in SIRT2 overexpressing versus the vector control HEK293 and MB-231 cells exposed to H₂O₂ resulted in reduced absorbance at 320 nm (less light scattering of misfolded MDH), which was due to improved folding of MDH (Fig. 4B). Similar effects were also observed in SIRT2-overexpressing MCF7 versus the vector control cells exposed to lower concentrations of H₂O₂ for 30 minutes (Fig. 4C and D).

We next determined whether reduced antioxidant activity of Prdx-1 and higher intracellular ROS levels in the SIRT2-overexpressing cells would result in increased DNA damage, especially following exposure to H₂O₂. Figure 5A and B demonstrate that treatment with H₂O₂ induced higher intracellular levels of γ-H2AX, signifying an increase in the DNA damage and response, in the SIRT2-overexpressing versus the vector control MB-231 cells. This was estimated in the nuclei by confocal immunofluorescence microscopy, as well as in the cell lysates by immunoblot analysis (Fig. 5A and B). We also determined the DNA damage and repair at an individual cell level by the Comet assay (19, 20). Treatment with H₂O₂ induced more DNA damage, determined by estimating the length of the comet tails, in the SIRT2 overexpressing versus the vector control MB-231 cells (Fig. 5C). Although a higher percentage of untreated cells exhibited shorter comet tails and lower tail movement, treatment with H₂O₂ dose-dependently increased the percentage of cells with longer comet tails and higher tail movement, more so in the SIRT2 overexpressing versus the vector control MB-231 cells (Fig. 5C and D). In contrast, exposure to H₂O₂ induced less DNA damage represented by higher % of cells with shorter tail movement in MB-231 cells expressing either the catalytically dead mutant SIRT2 or KD of SIRT2 (Fig. 5C and D).

Previous studies have demonstrated that increased intracellular ROS levels induce the nuclear localization and transcriptional activity of FOXO3A, resulting in upregulation of one of its targets, the proapoptotic, BH3 domain-only BIM protein (10, 11). Consistent with this, as compared with the untreated control cells, MB-231 cells overexpressing SIRT2 demonstrated higher accumulation of FOXO3A in the nucleus (Fig. 6A and B). Treatment with H₂O₂ further increased the levels of FOXO3A in the nucleus of SIRT2 O/E cells to a greater degree than in the vector control cells. Similar results were also obtained in MCF7 cells following ectopic overexpression of SIRT2 (Supplementary Fig. S4A). This was associated with more induction of BIM levels and increased caspase-3 cleavage and activity (Supplementary Fig. S4B). In a dose- and time-dependent manner, treatment with H₂O₂ also induced markedly higher levels of cell death in MB-231 cells overexpressing SIRT2, associated with deacetylated and relatively inactive Prdx-1, as compared with a similar treatment with H₂O₂ of MB-231 cells expressing catalytically inactive mutant SIRT2 or those with KD of SIRT2 (Fig. 6C). Next, we compared the lethal effects of exposure to the intracellular ROS-inducing agents, such as AT and menadione, in MB-231 and MCF-7 cells with the ectopic overexpression of SIRT2 versus the control MB-231 and MCF7 cells with the ectopic expression of the vector alone. As shown in Fig. 6D and E, exposure to AT or menadione induced significantly more lethality in cells with overexpression of SIRT2 versus the control MB-231 and MCF7 cells.

Previous reports have documented that embryonic toxicity due to oxidative stress induced by H₂O₂ or AT treatment of developing embryos of zebrafish (Danio rerio) is characterized by in vivo developmental abnormalities, including pericardial edema, circulation failure, looping failure, and dorsal curvature (29). Next, we determined the effects of antisense morpholinos against exon 6 of the zebrafish homolog of SIRT2 or mismatch controls, injected into single-cell stage zebrafish embryos (29). The effects of the morpholinos on mRNA level of SIRT2 and the effect of treatment with 3.0 mmol/L of H₂O₂ for 30 minutes on ROS levels at 48 hours postfertilization (hpf) were determined (Supplementary Fig. S5). Figure 7A demonstrates that compared with the mismatch control morpholino, antisense SIRT2 morpholino downmodulated mRNA levels of SIRT2. Treatment with H₂O₂ induced ROS levels in control embryos exposed to 3.0 mmol/L H₂O₂, which was markedly suppressed by cotreatment with the antioxidant N-acetylcysteine (NAC; Fig. 7B). In contrast, SIRT2 morpholino treatment significantly attenuated ROS levels due to H₂O₂ treatment, which was further reduced by cotreatment with NAC (Fig. 7B). Notably, injection of the mismatch control morpholino, followed by exposure to 3.0 mmol/L H₂O₂, showed the characteristic toxicity profile, including pericardial edema, circulation failure, and dorsal curvature in zebrafish larvae at 120 hpf. However, consistent with the attenuation of ROS levels, SIRT2 morpholino-treated embryos failed to exhibit developmental abnormalities including pericardial edema and dorsal curvature (Fig. 7C). These findings demonstrate that even a partial KD of SIRT2, by inducing acetylation and increased antioxidant activity of Prdx-1, abrogated the ROS-mediated toxic effects in zebrafish embryos.

Previous reports have identified the role of SIRT2 as a tumor suppressor, primarily due to its ability to maintain genomic fidelity during mitosis (1, 5, 6). SIRT2 achieves this by deacetylating and stabilizing BUBR1 and APC/C activity, which delays anaphase until chromosomes are attached at the mitotic spindle (1, 5, 6, 32). Recently, SIRT2 was also shown to directly interact with β-catenin and inhibit the WNT transcriptional targets including the oncoproteins survivin, c-Myc, and cyclin D1 (33). SIRT2 is located at 19q13.2, which is frequently deleted in human gliomas (34). Consistent with these observations, SIRT2 knockout mice develop cancers, and SIRT2 levels are reduced in variety of cancer types, including breast, liver, renal, and prostate cancers (1, 5). Along with SIRT2, the peroxide scavenger 2-Cys Prdx-1 has also been shown to have a tumor suppressive role and is classified as a “gerontogene” (35). Mice lacking Prdx-1 develop tumors prematurely during aging, associated with increased 8-oxo-dG levels and oxidative DNA damage (36). Prdx-1 deficiency was also shown to cause increased c-Myc and AKT activity, the latter due to ROS-mediated inactivation of PTEN (37–39). In addition, Prdx-1 was also demonstrated to inhibit TNFα-mediated NF-κB transcriptional activity (16). In the current study, we have established for the first time a direct link between SIRT2 levels and Prdx-1 deacetylation and activity. Utilizing DIGE coupled with MS/MS, we demonstrate that the predominantly cytosolic SIRT2 binds and deacetylates cytosolic Prdx-1. This is consistent with the previous reports, indicating that Prdx-1 present in the intermitochondrial space is deacetylated on lysine-197 by the mitochondria-resident SIRT3, whereas the cytosolic Prdx-1 is also a substrate for deacetylation by HDAC6 (40, 41). Importantly, we demonstrate here that SIRT2-mediated deacetylation of Prdx-1 reduces its antioxidant peroxidase activity (Fig. 7D). Consequently, SIRT2-overexpressing breast cancer cells, especially when subjected to oxidant stress induced by H₂O₂, accumulate ROS and DNA damage, as estimated by the comet assay and increased γH2AX levels, as well as demonstrate loss of cell viability. In contrast, this was not seen in breast cancer cells with ectopic expression of the catalytically inactive mutant form of SIRT2. Reduced peroxidase activity of the deacetylated Prdx-1 was also associated with ROS-induced overoxidation and multimer formation by Prdx-1, which represents a switch from the peroxidase to chaperone function of Prdx-1 (15, 18, 27). Previous studies have shown that the chaperone function of multimeric Prdx-1 enhances resistance to the lethal effects of oxidative stress and heat shock (18, 27). Despite this, our findings also show that the ultimate loss of viability caused by exposure of the SIRT2-overexpressing cells to high levels of H₂O₂ was due to increased nuclear accumulation of FOXO3A accompanied with the induction of BIM levels (Fig. 7D). This is also consistent with the tumor suppressor function of SIRT2.

Compared with their normal counterparts, cancer cells exhibit increased levels of ROS, including superoxide and hydroxyl radicals and H₂O₂, which promotes cell signaling for proliferation and other biologic functions (13, 16, 42, 43). Engagement by the ligands or cytokines of receptor tyrosine kinases or G protein–coupled receptors leads to transient generation of H₂O₂, catalyzed by the cell membrane-localized NADPH oxidases (16, 42, 43). This H₂O₂ oxidizes and inactivates cysteine residue in the nearby tyrosine phosphatases, which normally attenuate receptor signaling by dephosphorylating the pathway signaling kinases, thereby promoting the signaling for growth and proliferation (16, 17, 42, 43). Recently, membrane-associated Prdx-1 was shown to be transiently phosphorylated on its tyrosine-194 residue and thereby inactivated, allowing nearby accumulation of H₂O₂, inactivation of tyrosine phosphatases, and stimulation of the kinase-mediated signaling (42). However, excessive levels of ROS can inflict oxidative damage to lipids, proteins, and DNA (13). Among the H₂O₂ neutralizing proteins is Prdx-1, possessing a conserved N-terminal cysteine residue, which is also oxidized by H₂O₂ but reduced by thioredoxin (16, 43). Findings presented here clearly demonstrate that KD of SIRT2, by inducing acetylation of Prdx-1, increases its antioxidant peroxidase activity. This was associated with a reduction in the DNA damage and apoptosis triggered by H₂O₂-induced oxidant stress.

In contrast to its tumor suppressive role during tumorigenesis, in a variety of established tumor types, Prdx-1 levels are increased and have been shown to be transcriptionally upregulated by NRF2 (36, 38, 44). Increased Prdx-1 levels have been demonstrated in many cancers, including bladder cancer and non–small cell lung cancer (NSCLC), where they have been associated with a high grade and advanced stages (35, 45, 46). In addition, Prdx-1 has been demonstrated to potentially serve as a prognostic and therapeutic target in cancer (47, 48). Overall, these reports highlight that Prdx-1 levels and activity regulate the redox homeostasis, which controls the growth and survival of cancer cells (13). Findings presented here clearly demonstrate that KD of SIRT2, by inducing acetylation of Prdx-1, increases its antioxidant peroxidase activity. This was accompanied by decreased in vitro accumulation of DNA damage detected by the comet assay, following exposure to oxidative stress induced either by H₂O₂ or by exposure to AT or menadione. Furthermore, in vivo SIRT2 KD also exerted protection against toxicity associated with oxidative stress in zebrafish embryos. There was a reduction in the H₂O₂-induced accumulation of ROS and the embryos also failed to develop characteristic features of embryonic toxicity due to oxidative stress.

In summary, our findings demonstrate that in cancer cells, selectively activating SIRT2 would lead to deacetylation and inactivation of Prdx-1, thereby sensitizing cancer cells to agents that induce oxidative stress and promote lethal DNA damage. By inducing nuclear accumulation of FOXO3A and induction of BIM, increased SIRT2 with reduced Prdx-1 activities could also induce lethal effects through a mechanism independent of the function of the tumor suppressor TP53.

A. Melnick reports receiving a commercial research grant from Janssen, Roche, and Eli Lilly and is a consultant/advisory board member for Eli Lilly, Roche, Epizyme, and Boehringer. No potential conflicts of interest were disclosed by the other authors.

Conception and design: K.N. Bhalla

Development of methodology: H. Ma

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): W. Fiskus, V. Coothankandaswamy, J. Chen, K. Ha, D.T. Saenz, S.S. Krieger, C.P. Mill, B. Sun, P. Huang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): W. Fiskus, V. Coothankandaswamy, J. Chen, H. Ma, D.T. Saenz, S.S. Krieger, B. Sun, J.S. Mumm, K.N. Bhalla

Writing, review, and/or revision of the manuscript: W. Fiskus, P. Huang, J.S. Mumm, A.M. Melnick, K.N. Bhalla

Study supervision: K.N. Bhalla

This research was partially supported by the CCSG P30 CA016672.

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.