SIRT2 Interacts with B-Catenin to Inhibit Wnt Signaling Output in Response to Radiation-Induced Stress

Published OnlineFirst May 27, 2014; DOI: 10.1158/1541-7786.MCR-14-0223-T

Molecular Cancer DNA Damage and Repair Research

SIRT2 Interacts with b-Catenin to Inhibit Wnt Signaling Output in Response to Radiation-Induced Stress

Phuongmai Nguyen1, Sunmin Lee2, Dominique Lorang-Leins1, Jane Trepel2, and DeeDee K. Smart1

Abstract Wnt signaling is critical to maintaining cellular homeostasis via regulation of cell division, mitigation of cell stress, and degradation. Aberrations in Wnt signaling contribute to carcinogenesis and metastasis, whereas sirtuins have purported roles in carcinogenesis, aging, and neurodegeneration. Therefore, the hypothesis that sirtuin 2 (SIRT2) directly interacts with b-catenin and whether this interaction alters the expression of Wnt target genes to produce an altered cellular phenotype was tested. Coimmunoprecipitation studies, using mouse embryonic ﬁbroblasts (MEF) from Sirt2 wild-type and genomic knockout mice, demonstrate that b-catenin directly binds SIRT2. Moreover, this interaction increases in response to oxidative stress induced by ionizing radiation. In addition, this association inhibits the expression of important Wnt target genes such as survivin (BIRC5), cyclin D1 (CCND1), and c-myc (MYC).InSirt2 null MEFs, an upregulation of matrix metalloproteinase 9 (MMP9) and decreased E-cadherin (CDH1) expression is observed that produces increased cellular migration and invasion. Together, these data demonstrate that SIRT2, a tumor suppressor lost in multiple cancers, inhibits the Wnt signaling pathway in nonmalignant cells by binding to b-catenin and that SIRT2 plays a critical role in the response to oxidative stress from radiation.

Implications: Disruption of the SIRT2–b-catenin interaction represents an endogenous therapeutic target to prevent transformation and preserve the integrity of aging cells against exogenous stressors such as reactive oxygen species. Mol Cancer Res; 12(9); 1244–53. 2014 AACR.

Introduction includingcolon,breast,prostate,ovary,andmelanoma The Wnt signaling pathway is crucial in regulating cell cancers (10). Sirtuin 2 (SIRT2) belongs to the family of sirtuins that are proliferation, apoptosis, migration, and differentiation (1). þ b-Catenin is a critical coactivator of Wnt-mediated gene NAD -dependent class III protein deacetylases, and has expression. When Wnt ligands are present, b-catenin been shown to play a major role in metabolism and longev- accumulates in the cytoplasm and is transported to the ity. SIRT2 expresses mainly in the brain, particularly the hippocampus and oligodendrocytes. SIRT2 is reported to nucleus whereupon it binds to the lymphocyte enhancer a factor (LEF)/T-cell factor (TCF) complex to recruit chro- deacetylate -tubulin, thereby regulating microtubule dynamics. Consequently, SIRT2 is proposed to play a role matin remodeling complexes and activates gene expression – (2–4). Disruptions of Wnt signaling are associated with in mitotic checkpoint to maintain genomic integrity (11 accelerated aging (5, 6) and abnormal neural development 13). It has also been observed that SIRT2 may act as a tumor (7) as well as several major diseases, including schizophre- suppressor because SIRT2 is downregulated in gliomas (14), nia, autism, Alzheimer disease, and cardiovascular disease while high levels of SIRT2 in tumors confer resistance to (3, 8, 9). In addition, aberrant regulation of the b-catenin chemotherapy (15). In the central nervous system, SIRT2 is pathway has been shown to promote tumorigenesis, believed to participate in oligodendrocyte differentiation (16), and SIRT2 has been found to accumulate in the aging rodent brain (17). In this study, we show that SIRT2 directly interacts b 1Radiation Oncology Branch, NCI, NIH, Bethesda, Maryland. 2Medical with -catenin, and the interaction increases following Oncology Branch, NCI, NIH, Bethesda, Maryland. ionizing radiation exposure. Genetic deletion of Sirt2 leads c-myc cyclin D1 Note: Supplementary data for this article are available at Molecular Cancer to activation of Wnt target genes such as , , Research Online (http://mcr.aacrjournals.org/). and survivin. Furthermore, we observe that SIRT2 is critical Corresponding Author: DeeDee K. Smart, Radiation Oncology Branch, for regulating cell migration through induction of matrix NCI, NIH, 9000 Rockville Pike, Bethesda, MD 20892. Phone: 301-496- metalloproteinase 9 (MMP-9) and downregulation of E- 5457; Fax: 301-480-5439; E-mail: [email protected] cadherin. These ﬁndings provide a mechanistic insight into doi: 10.1158/1541-7786.MCR-14-0223-T SIRT2 as a potential tumor suppressor and in cellular 2014 American Association for Cancer Research. response to ionizing radiation-induced oxidative stress in

1244 Mol Cancer Res; 12(9) September 2014

SIRT2 Inhibits Wnt Following Radiation-Induced Stress

nontransformed cells. Furthermore, MMP9 and E-cadherin plasmic b-catenin and nuclear b-catenin into band inten- may potentially act as molecular markers for cells lacking sities of manganese SOD2 and lamin B, respectively. Sirt2 in regulating cell migration/invasion. Immunofluorescence microscopy Materials and Methods Immunofluorescent staining was performed as previously Cell culture, radiation exposure, and clonogenic survival described (19). Photographs were taken with a 63 oil assay immersion objective. Primary mouse embryonic fibroblasts (MEF) derived from wild-type (WT) and Sirt2 knockout mice [KO; back- Coimmunoprecipitation and immunoblot analysis ground C57B (60%)/C129 (40%)], a kind gift of Cells were lysed in TNN buffer (50 mmol/L Tris-HCl, Drs. Hyun-Seok Kim and Chu-Xia Deng (NIH, Bethesda, pH 8.0, 120 mmol/L NaCl, 1% NP-40) for immuno- MD) in 2011, were immortalized and described previously precipitation experiments or in RIPA buffer (Thermo fi (18). As the cells were previously characterized, we did not Scienti c) for immunoblot analysis as previously described perform further testing or authentication. MEFs were cul- (20). tured in DMEM supplemented with 1% penicillin-strep- tomycin, 15% FBS, and 1% nonessential amino acid (Invi- Measurement of intracellular reactive oxygen species fi trogen). Cells that were passage greater than 37 were used in Intracellular reactive oxygen species (ROS) quanti cation all experiments. PC-3 and U87 cells were obtained from was carried out as described by Slane and colleagues (21). fl Sirt2 ATCC in 2011 and authenticated by STR analysis upon Brie y, WT and KO MEFs were labeled with either carboxy-DCFDA or H2DCFDA (Invitrogen). After incu- purchase before preservation. Cells were passaged within 4 months of resuscitation according to the ATCC protocol bation for 30 minutes at 37 C, the dye was removed, and and routinely tested for mycoplasma infection using MycoA- cells received either sham or 10 Gy radiation. Thirty minutes fl lert Mycoplasma Detection Kit (Lonza). after exposure, the uorescent intensity was measured. Mouse Sirt2 shRNA (Origene) was electroporated into H2DCFDA:DCFDA ratio was calculated, normalized to Sirt2 Sirt2 WT MEFs using Amaxa Nucleofector Kit (Lonza). either WT or KO control, respectively, and presented fl Colonies were selected in 3 mg/mL puromycin, and as normalized uorescent intensity. To inhibit ROS gener- colonies with decreased SIRT2 expression were screened ated by radiation, cells were pretreated with 200 U/mL PEG- by immunoblot analysis. Human Flag-SIRT2 (Addgene) catalase (Sigma) for 2 hours before labeling the cells. PEG- was electroporated into U87 glioma cells. Stable transfec- catalase treatment continued throughout the labeling and tants were selected in G-418, and positive clones were radiation exposure. screened by immunoblot analysis with anti-Flag antibody In vitro b (Origene). deacetylation assay of -catenin In vitro Cells were exposed to ionizing radiation in an X-Rad 320 deacetylation was performed as described by fl Sirt2 biologic irradiator (Precision X-ray, Inc.), at a dose rate of Zhang and colleagues (22). Brie y, KO MEFs were 2.1 Gy/minute. lysed in TNN buffer and 1 mg of cell lysates were used to b To determine radiosensitivity, cells were plated at clonal immunoprecipitate -catenin. Immunoprecipitates were fi density and irradiated 6 hours after plating. Experiments resuspended in the presence or absence of puri ed recom- > binant human SIRT2 (BPS Bioscience) with or without were terminated when visible colonies ( 50 cells/colony) þ were present. Cells were stained with 0.5% crystal violet, NAD . The reaction was stopped and samples analyzed by the number of colonies counted, and the surviving fractions immunoblot analysis with antibodies against acetylated b were calculated and normalized to unirradiated controls, lysine, -catenin, or SIRT2. respectively. RNA extraction and RT-PCR Antibodies RNA was isolated using an RNeasy Mini Kit (Qiagen), The following antibodies were purchased: SIRT2 (Sigma), and reverse transcription and RT-PCR were carried out as b-catenin, cyclin D1, c-jun and survivin (Santa Cruz Bio- previously described (23). Wnt pathway PCR Array and technology), phospho-b-catenin Y142 (Abcam), GSK3b reagents were from Qiagen. Primers used for RT-PCR are and pGSK3b (BD Biosciences), acetylated lysine, acetylated shown in Supplementary Table S1. b-catenin, MMP-9, E-cadherin, Akt, pAkt (Ser473), and c-myc (Cell Signaling Technology). pTOPFLASH reporter assays Sirt2 WT and KO MEFs were cotransfected with 1.5 mg Subcellular fractionation of pTOPFLASH luciferase reporter, which contains four Cells were subjected to subcellular fractionation using TCF-responsive elements in tandem, or pGL2-Basic empty NE-PER Nuclear and Cytoplasmic Extraction Reagents vector and 150 ng of b-galactosidase expression plasmid Kit (Thermo Scientific). Band intensities from subcellular (pCMV-b-gal) using FuGENE 6 (Roche Diagnostics). fractionation experiments were quantified using Image- Total amount of luciferase and b-galactosidase activities Quant software (GE Healthcare). Relative fold-change were measured and relative fold induction of luciferase was determined by dividing band intensities of cyto- activity was normalized to b-galactosidase activity. The

www.aacrjournals.org Mol Cancer Res; 12(9) September 2014 1245

Nguyen et al.

results are represented as ratio of pTOPFLASH-luciferase Results over empty-vector pGL2. Nuclear b-catenin level is elevated in Sirt2 KO MEFs Because transcriptionally active b-catenin accumulates Cell invasion assay fi Cell culture inserts with an 8-mm pore-size PET mem- in the nucleus, we rst determined the subcellular localization of b-catenin in WT Sirt2 MEFs compared with brane (Fisher Scientific) were coated with collagen IV KO Sirt2 MEFs generated through genomic deletion. (Becton-Dickinson) and placed into 24-well plates contain- Using lamin B as a nuclear marker and manganese super- ing 0.5 mL of DMEM medium supplemented with 15% 3 oxide dismutase as a cytoplasmic marker, we demonstrate FBS. A total of 6.25 10 cells in serum-free DMEM were by immunoblot analysis of nuclear- and cytoplasmic- plated into the inserts and incubated for 18 hours at 37 C. enriched fractions that there is a significant increase in Cells from the upper surface of the membrane were then nuclear b-catenin protein present in the Sirt2 KO MEFs removed and the remaining cells adhered to the bottom of relative to WT (Fig. 1A). This increase is greater than the membrane were Wright stained (Fisher Scientific). Six 2-fold as analyzed by densitometry (Fig. 1B, columns 1 random fields per membrane were photographed at 40 and counted. and 2). Conversely, there is a corresponding increase of cytoplasmic b-catenin present in the Sirt2 WT MEFs (Fig. 1A) compared with KO; this increase is also greater than Chromatin immunoprecipitation 2-fold as determined by densitometry (Fig. 1B, columns 3 Chromatin immunoprecipitation (ChIP) assays were and 4). Total b-catenin protein level remains constant carried out as described previously (20). Purified DNA between the two cell types (Fig. 1D). sampleswereanalyzedbyqRT-PCRusingSYBRGreen Immunofluorescence analysis of Sirt2 WT and KO cells PCR Master Mix (Applied Biosystems) with primers to using b-catenin antibody shows that roughly 70% of the the promoters of Myc, cyclin D1,andsurvivin (Supple- Sirt2 KO cells observed have strong immunoreactivity in the mentary Table S1). nucleus with weak but detectable immunoreactivity in the cytoplasm (Fig. 1C). In contrast, less than 5% of the Sirt2 Statistical analyses WT fibroblasts showed strong nuclear b-catenin staining, Unless indicated otherwise, all observations were con- fi whereas the majority exhibited low nuclear reactivity for rmed by at least three independent experiments. The b fl -catenin. Immuno uorescent staining provides additional results are presented as the mean SEM. We used two- b t data to demonstrate a stronger nuclear -catenin presence in tailed Student tests as determined by GraphPad Prism Sirt2 P < the KO cells. This is probably due to the fact that not all software for statistical analysis. 0.05 was considered as the Sirt2 KO cells have increased b-catenin nuclear statistically significant.

ABWT KO Nuc Cyto kD: Cyto Nuc Cyto Nuc 2.5 100 β-Catenin 2.0 Figure 1. Nuclear b-catenin levels 80 1.5 are elevated in Sirt2 KO MEFs. A, 80 immunoblot analysis of enriched Lamin B 1.0 nuclear and cytoplasmic fractions

60 change Fold for b-catenin, lamin B (nuclear 0.5 marker), or manganese superoxide 30 dismutase (cytoplasmic marker) MnSOD 0.0 WT KO WT KO showing increased nuclear 20 b-catenin protein level in Sirt2 KO MEFs. B, densitometric analysis of the subcellular fractionation. CD , P < 0.001. C, b-catenin kD: WT KO immunoﬂuorescent staining 100 showing markedly increased level β-Catenin of nuclear b-catenin in Sirt2 KO 80 MEFs versus WT. D, immunoblot WT 100 analysis showing increased pY142 b-catenin in Sirt2 KO MEFs. β-Catenin pY142 80

1246 Mol Cancer Res; 12(9) September 2014 Molecular Cancer Research

SIRT2 Inhibits Wnt Following Radiation-Induced Stress

localization and thus may have diluted out during the of SIRT2 with b-catenin is also present in PC-3 cells, a subcellular fractionation procedure. human prostate adenocarcinoma cell line that endogenously Previous studies have shown that phosphorylation of expresses SIRT2 (Fig. 2B). In addition, we overexpressed b-catenin at tyrosine residue 142 correlates with an Flag-SIRT2 or empty-vector pcDNA3 in the U87 glioblas- increase in nuclear b-catenin–dependent gene expression toma cell line, which expresses low levels of SIRT2 (Fig. 2C). (24). In the absence of SIRT2, we found that phospho- Anti-Flag pull-down and b-catenin immunoblot analysis b-catenin Y142 is significantly elevated (Fig. 1D). Taken demonstrates a positive interaction between Flag-SIRT2 and together, these results indicate that there is more nuclear, b-catenin, which is not detected in the empty-vector transcriptionally active b-catenin present in Sirt2 KO pcDNA3 cells. MEFs. To test whether the association between SIRT2 and b-catenin can be modulated, we exposed Sirt2 WT MEFs SIRT2 interacts with b-catenin to ionizing radiation at 2, 5, or 10 Gy delivered as a single To determine whether SIRT2 interacts with b-catenin, we dose before performing coimmunoprecipitation experi- performed coimmunoprecipitation experiments using cell ments (Fig. 2A). We observe that binding of SIRT2 to lysates from Sirt2 WT MEFs (Fig. 2A). SIRT2 antibody b-catenin significantly increases in response to radiation readily immunoprecipitates b-catenin (lane 2) as detected by (lanes 3–5) compared with unirradiated samples (lane 2), immunoblot analysis with b-catenin antibody. In contrast, indicating that the interaction is not static and can be rabbit IgG does not immunoprecipitate b-catenin (lane 1), regulated. The level of b-catenin remains constant between confirming the specificity of the pull-down. The interaction unirradiated (Fig. 2A, bottom, lane 1) and irradiated samples

0 Gy A SIRT2 IP D 10 Gy 10 Gy + PEG-cat kD: IgG Ctrl 2 Gy 5 Gy 10 Gy 3 100 P = 0.054 β-Catenin 80 2 β-cat IP Ctrl 2 Gy 5 Gy 10 Gy 1 β-Catenin Normalized mean fluorescence intensity 0 Sirt2 WT Sirt2 KO B IgG SIRT2 Input E 100 β-Catenin 1.2 80 1.0 Sirt2 WT IB: β-Catenin Sirt2 KO 0.8 C 0.6 Vector Flag-Sirt2 0.4 IgGkD: Flag β-cat IgG Flag β-cat 100 0.2 0.0

80 Normalized surviving fraction IB: β-Catenin 0 Gy 6 Gy

0 Gy + PEG-cat 6 Gy + PEG-cat

Figure 2. SIRT2 interacts directly with b-catenin. A, coimmunoprecipitation showing association of SIRT2 and b-catenin increases with ionizing radiation. Sirt2 WT MEFs were exposed to single fractions of 2, 5, and 10 Gy, and compared with sham controls. Cells were harvested 24 hours after radiation and immunoprecipitated either with normal rabbit IgG control, SIRT2, or b-catenin antibodies. Bound proteins were analyzed by immunoblot analysis with b-catenin antibody. B, coimmunoprecipitation analysis showing SIRT2 associates with b-catenin in PC-3 cells. C, coimmunoprecipitation analysis showing overexpressed Flag-SIRT2 binds to b-catenin. Protein lysates from U87 glioma cells stably expressing either pcDNA3 or Flag-SIRT2 were subjected to immunoprecipitation using rabbit IgG, DDK, or b-catenin antibodies. Bound proteins were analyzed by immunoblot analyzed with b-catenin antibody. D, ROS measurement. Sirt2 WT and KO MEFs were labeled with H2DCFDA or carboxy DCFDA for 30 minutes before exposure to a single fraction of 10 Gy compared with sham controls. Cells were harvested after 30 minutes and analyzed. Treatment with PEG-catalase was used to scavenge ROS generated by radiation. E, clonogenic survival assay. Sirt2 WT and KO cells were plated for clonogenic survival assay and treated with 200 U/mL PEG-cat for 2 hours, followed by radiation. PEG-cat was left for the duration of the assay. Values, mean SEM. , P < 0.05; , P < 0.005.

www.aacrjournals.org Mol Cancer Res; 12(9) September 2014 1247

Nguyen et al.

(Fig. 2A, bottom, lanes 2–4), confirming that the enhanced b-Catenin is acetylated and the Wnt signaling pathway is interaction between SIRT2 and b-catenin is not artifact activated in Sirt2 KO MEFs secondary to a change in total b-catenin levels. Acetylated b-catenin has been shown to play a role in ROS are a well-known by-product of IR exposure, and oncogenesis due to increasing b-catenin/TCF transcrip- ROS levels correlate to DNA damage that ultimately leads to tional activity (25, 26). Firestein and colleagues showed cell mortality. We examined the differences in ROS levels in that deacetylation of b-catenin by SIRT1 can suppress the presence and absence of SIRT2 at baseline and following colon tumorigenesis (27). To determine whether there is a a single exposure to 10 Gy (Fig. 2D). The level of ROS difference in the level of acetylated b-catenin, Sirt2 WT increases following ionizing radiation exposure in the pres- and KO MEFs were subjected to immunofluorescence ence and absence of SIRT2 to a similar extent (P ¼ 0.054). staining with acetylated b-catenin antibody (Fig. 3A). Interestingly, the addition of pegylated catalase (PEG-cat) Indeed, there is markedly more nuclear acetylated b-cate- was able to quench the generation of ROS in the Sirt2 KO nin present in the Sirt2 KO MEFs compared with WT, MEFs, but not in Sirt2 WT. suggesting that SIRT2 in the WT MEFs deacetylates To determine whether PEG-cat is capable of protecting b-catenin, giving rise to a lower level of acetylated b-cate- the cells from ionizing radiation-induced cell death by nin immunoreactivity. Concomitantly, coimmunopreci- Sirt2 scavenging H2O2, we performed clonogenic survival assays pitation experiments were also carried out using WT using Sirt2 WT and KO MEFs. Cells were pretreated with and KO MEFs. Whole-cell lysates were immunoprecipi- 200 U/mL PEG-cat before exposure with 6 Gy radiation. As tated with b-catenin antibody and immunoblotted with shown in Fig. 2E, without PEG-cat pretreatment, Sirt2 KO either acetylated lysine or acetylated b-catenin antibody cells were significantly more resistant to ionizing radiation- (Fig. 3B). Acetylated b-catenin is present in the Sirt2 KO induced cytotoxicity than Sirt2 WT cells. The addition of MEFs (lane 4, Ac-K and Ac-bcat panels) but not in WT. PEG-cat did not affect the survival of Sirt2 WT cells, but Furthermore, Flag-SIRT2 expression in U87 glioma slightly increased cell survival of Sirt2 KO cells upon radi- abolishes the acetylation of b-catenin (Fig. 3B, lane 8, ation treatment. Therefore, these data suggest that although Ac-K and Ac-bcat panels) further confirming that SIRT2 both Sirt2 WT and KO cells have increased ROS upon is responsible for deacetylating b-catenin. ionizing radiation exposure, Sirt2 KO cells are more capable To verify that SIRT2 directly deacetylates b-catenin, we of handling the ROS level than Sirt2 WT cells resulting in performed in vitro deacetylation assays with b-catenin in increased cellular survival following radiation exposure. the presence or absence of purified recombinant SIRT2

Figure 3. Acetylation of b-catenin ACAc-β-catenin DAPI IgG β-cat IP and activation of the Wnt signaling pathway in Sirt2 KO MEFs. A, Sirt2 ––++ acetylated b-catenin NAD –––+ immunoﬂuorescent staining 120 showing markedly increased level of acetylated b-catenin present in 80 Sirt2 KO MEFs compared with WT. WT IB: Ac-K B, increased acetylated b-catenin is present in Sirt2 KO MEFs. 120 Whole-cell lysates from Sirt2 WT and KO cells or from pcDNA3 and 80 Flag-SIRT2 transfected U87 cells IB: β-cat were immunoprecipitated with rabbit IgG control or b-catenin KO 50 antibody. Pulled-down complexes were immunoblotted with 40 acetylated lysine or acetylated IB: Sirt2 b-catenin antibody. C, SIRT2 deacetylates b-catenin in vitro. BDWT KO pcDNA3 Flag SIRT2 Immunoprecipitated b-catenin from Sirt2 KO MEFs was incubated β β β β WT KO kD: IgG -cat IgG-cat IgG-cat IgG -cat kD: with or without puriﬁed 50 120 recombinant SIRT2 in the presence þ pGSK3-β or absence of NAD . The reaction 80 40 mixtures were separated by SDS- PAGE and immunoblotted with IB: Ac-K 50 antibodies against acetylated GSK3-β 120 lysine, b-catenin, or SIRT2. D, 40 immunoblot analysis showing upregulated level of activated, 80 IB: Ac-β-cat phosphorylated GSK-3-b in Sirt2 KO MEFs compared with WT.

1248 Mol Cancer Res; 12(9) September 2014 Molecular Cancer Research

SIRT2 Inhibits Wnt Following Radiation-Induced Stress

þ and with or without the addition of NAD . b-Catenin from expression (2). Because there seems to be increased "active," Sirt2 KO MEFs was used because it is highly acetylated as acetylated b-catenin in Sirt2 KO MEFs relative to WT compared with b-catenin from WT MEFs (Fig. 3B). Con- MEFs, we decided to evaluate the protein level of GSK3b sistent with the result in Fig. 3B, without the addition of and phosphorylated GSK3b (pGSK3b) in the two cell types SIRT2, b-catenin is strongly acetylated (Fig. 3C, lane 2). (Fig. 3D). Although the total amount of GSK3b is similar However, in the presence of purified recombinant SIRT2, with no significant difference between Sirt2 WT and KO acetylation of b-catenin was greatly diminished (Fig. 3C, MEFs (Fig. 3D, bottom), there is clear upregulation of lanes 3 and 4). Interestingly, even without the addition of pGSK3b in Sirt2 KO cells (Fig. 3D, top). Therefore, it þ NAD to the reaction, SIRT2 was able to deacetylate seems that a loss of SIRT2 activates the Wnt pathway in b-catenin fully (Fig. 3C, lane 3). We hypothesized that nontransformed cells. because we used whole-cell lysates as the source of acetylated b-catenin, the pull-down may also contain endogenous Downstream Wnt target gene expression is increased in þ NAD that can serve as cofactor for the purified recombinant Sirt2 KO MEFs þ SIRT2. Exogenous NAD added to the reaction mixture Given that loss of SIRT2 was demonstrated to activate the may overwhelm the system and slow down the enzymatic Wnt pathway, we investigated whether genomic loss and activity of recombinant SIRT2, thus decreasing b-catenin knockdown of endogenously expressed Sirt2 produced an deacetylation (Fig. 3C, lane 4). effect on the expression of Wnt target genes. To address this GSK3b is a negative upstream regulator of b-catenin in question, we used a commercially available Wnt Pathway the Wnt pathway. When the Wnt signaling pathway is in the PCR Array (Qiagen) to identify Wnt pathway genes upre- off state, phosphorylation of b-catenin by GSK3b leads to gulated or downregulated in the Sirt2 WT versus KO MEFs ubiquitination and proteasomal degradation. Activation of (Supplementary Table S2). Interestingly, we observe coor- the Wnt pathway inhibits GSK3b, allowing b-catenin to dinated regulation of Wnt pathway genes. For example, accumulate and translocate into the nucleus to turn on gene genes that have prosurvival functions such as cyclin D1, Fosl1,

ABWT KO Sirt2 WT 6 60 c-Myc Sirt2 KO

40 c-Jun 4 20 Survivin

Figure 4. SIRT2 inhibits Wnt target 2 Cyc D1 gene expression. A, RT-PCR mRNA level Rel. 30 showing upregulated mRNA levels 50 of cyclin D1, c-myc, and survivin in Actin 40 Sirt2 KO MEFs compared with WT. 0 , P < 0.005; , P < 0.0005. B, immunoblot analysis showing b -cat c-myc increased protein levels of c-myc, Cyc D1 survivin cyclin D1, survivin, and c-Jun in Sirt2 KO compared with WT. C, RT- CD PCR showing that knockdown of KD Sirt2 upregulates cyclin D1, c-myc, 10 survivin Sc. shRNASirt2 and mRNA levels. 50 SIRT2 Sc. shRNA , P < 0.001. D, immunoblot p43 analysis showing knockdown of 8 Sirt2 KD 40 p39 p37 Sirt2 upregulates c-myc, cyclin D1, and survivin protein expression. 6 40 Cyc D1 30 4 60 c-Myc

Rel. mRNA level Rel. 2 20 Survivin 0

50 Tubulin c-myc Cyc D1 survivin

www.aacrjournals.org Mol Cancer Res; 12(9) September 2014 1249

Nguyen et al.

c-Jun, and c-Myc along with transcriptional coactivators of Lef-1 Tcf-7 the Wnt pathway such as and are upregulated in A Sirt2 KO MEFs relative to WT. In contrast, genes that are Luciferase negative regulators of the Wnt pathway such as Pitx2, Nkd1, TRE Sfrp1, Sfrp4, and Tle1 are down-regulated in the Sirt2 KO 100 MEFs compared with WT. For further studies, we decided to focus on the prosurvival genes. Additional experiments 80 ﬁ using RT-PCR con rmed increased mRNA expression of 60 c-myc, cyclin D1, and survivin in Sirt2 KO cells (Fig. 4A). Similarly, immunoblot analysis revealed elevated Wnt target 40 gene expression for c-myc, c-jun, survivin, and cyclin D1 in 20 Sirt2 Sirt2 KO MEFs (Fig. 4B). Knockdown of endogenous pTOPFlash-luci/pGL2 Sirt2 0 in WT MEFs using shRNA also produced increases in WT KO cyclin D1, c-myc, and survivin by RT-PCR (Fig. 4C) and immunoblot analysis (Fig. 4D). Taken together, loss of B 0.6 SIRT2 by either genomic deletion or shRNA inhibition IgG c-Myc results in elevated expression of Wnt target genes. β-Cat SIRT2 Represses b-catenin transcriptional activity 0.4 To verify that increased expression of Wnt target genes in

the absence of SIRT2 occurs at the transcriptional level, we % Input examined the effect of SIRT2 on b-catenin–mediated tran- 0.2 scription using TOPFLASH luciferase reporter analysis (Fig. 5A). We observed that deletion of Sirt2 results in more than 2-fold increase in luciferase activity, suggesting tran- 0.0 scriptional activation of the promoter containing TCF- WT KO binding sites. Furthermore, we examined the effect of SIRT2 on the 0.8 recruitment of b-catenin to TCF-binding sites of Wnt target IgG Cyclin D1 genes by ChIP (Fig. 5B). Indeed, ChIP showed an increase in 0.6 β-Cat b-catenin DNA-binding activity to the cyclin D1, c-Myc, and survivin promoters in Sirt2 KO MEFs compared with WT. Increased DNA binding to the Wnt target promoters in the 0.4 absence of SIRT2 matches the increased expression of Wnt % Input target genes. Taken together, the data suggest that SIRT2 0.2 acts as a repressor of transcription of Wnt target prosurvival genes. 0.0 WT KO SIRT2 inhibits cell migration/invasion Microtubules have also been shown to be required for cell 0.8 migration and tail retraction (28). SIRT2 has been shown to IgG play a role in microtubule deacetylation (16) thereby raising Survivin β the possibility that SIRT2 might regulate cell motility. We 0.6 -Cat determined whether the absence or presence of SIRT2 changes the chemotactic movement and invasive property 0.4 Sirt2 of MEFs. To investigate whether WT and KO cells % Input differ in motility, we subjected both cell lines to an assay that 0.2 measured the ability of cells to migrate through pores in a PET membrane. As shown in Fig. 6, Sirt2 KO ﬁbroblasts migrate 2-fold faster than WT in response to serum (Fig. 6A), 0.0 suggesting that SIRT2 is required to regulate cell motility. To WT KO examine the effect of SIRT2 on invasion, the cells are required to cross through a collagen layer as well as migrate Figure 5. SIRT2 inhibits Wnt signaling. A, pTOPFLASH luciferase through the pores of the chamber (Fig. 6B). Consistent with promoter analysis showing upregulated TCF/LEF transcriptional activity the migration result, the presence of SIRT2 markedly inhib- in Sirt2 KO MEFs compared with WT. , P < 0.005. B, ChIP analysis ited the invasion of MEFs through the collagen coating. showing increased binding of b-catenin to the promoters of c-myc, cyclin D1 survivin P < P < To further verify the effect of SIRT2 on invasive capacity, , and . , 0.5; , 0.01. we performed invasion assays using Sirt2 WT MEFs

1250 Mol Cancer Res; 12(9) September 2014 Molecular Cancer Research

SIRT2 Inhibits Wnt Following Radiation-Induced Stress

A 0.8 0.6 Figure 6. SIRT2 inhibits cell motility/ invasion and MMP9 expression 0.4 and upregulates E-cadherin OD 560 nm 0.2 expression. A, cell motility analysis WT KO showing Sirt2 KO MEFs migrated 0.0 faster through uncoated Boyden WT KO chambers compared with WT in B response to serum. B, cell invasion 800 analysis showing Sirt2 KO ﬁbroblasts migrated faster through 600 collagen-coated Boyden 400 chambers in response to serum. P < , 0.0001. C, cell invasion 200

analysis showing that knockdown # Cells migrated of Sirt2 slowed the migration of WT KO 0 Sirt2 WT MEFs through collagen- WT KO coated Boyden chambers. , P,< 0.005. D, RT-PCR showing CD upregulated Mmp9 mRNA in Sirt2 1,500 100 KO MEFs compared with WT. , P < 0.005. E, immunoblot analysis 80 Sirt2 1,000 showing increased MMP9 in 60 KO cells compared with WT. F, immunoblot analysis showing 500 40 markedly decreased E-cadherin in Rel. mRNA level Rel. Sirt2 KO cells compared with WT. # Cells migrated 20 ﬁ Data from A to C in this gure are 0 0 representative of one of three WT KO KD independent experiments. Data ﬁ Sirt2 from D to F in this gure show Sc. shRNA mean SEM from three EF independent experiments. WT KO kD: WT KO kD: 120 100 MMP9 E-cadherin 100

transfected with either scrambled shRNA or Sirt2 shRNA. E-cadherin is decreased in Sirt2 KO MEFs Knocking down Sirt2 consistently increased the invasive Because we observed a difference in adherence property capability of the cells in migrating through a collagen layer between Sirt2 WT and KO MEFs in culture, we investigated (Fig. 6C), thus further supporting results obtained with Sirt2 the effect of SIRT2 on E-cadherin expression. Immunoblot KO MEFs. analysis for E-cadherin demonstrates reduced expression in Sirt2 KO MEFs (Fig. 6E). Therefore, in addition to down- Expression of MMP9 is dramatically elevated in Sirt2 stream targets of the Wnt pathway, the regulation of the KO MEFs adherens junction protein E-cadherin is also dependent on GiventhatmigrationandinvasionofMEFswere the presence or absence of SIRT2. In summary, we have determined to be a SIRT2-dependent phenomenon, we demonstrated that the loss of SIRT2 gives rise to the examined whether the capacity of the cell to invade upregulation of prosurvival genes through the activation of extracellular matrix was also dependent on SIRT2. Pre- b-catenin–dependent Wnt signaling. Concurrently, inva- vious studies have shown that MMP-9 (29, 30) is an sion and migratory capabilities become more prominent. important effector of wound healing, tumor invasion, and MMP9 and E-cadherin may potentially act as molecular angiogenesis, and has also been characterized as a target of markers for cells lacking Sirt2 in regulating cell migration/ b-catenin–dependent Wnt signaling. As shown in Fig. invasion. All of these events are the hallmarks of a bona ﬁde 6D, Mmp-9 mRNA is markedly elevated in the Sirt2 KO malignancy and therefore support the hypothesis that SIRT2 cells relative to WT (Fig. 6D). Immunoblot analysis functions as a tumor suppressor protein in response to further conﬁrmed this increase in MMP-9 expression ionizing radiation in normal tissues. (Fig. 6E). Thus, the loss of SIRT2 results in increased Mmp-9 mRNA and protein levels, and this observation is Discussion consistent with observed differences in the migratory and Recent developments implicate SIRT2 as a major player in invasive behavior of the cells. several diseases, including cancer and neurodegenerative

www.aacrjournals.org Mol Cancer Res; 12(9) September 2014 1251

Nguyen et al.

disorders. In this study, we describe a tumor-suppressive role suggest that SIRT2 plays a significant role in mediating the for SIRT2 using WT and KO Sirt2 MEFs as a model. These cell's response to environmental stressors such as oxidative findings support other published data that link the processes damage from ionizing radiation. of tumor suppression and aging to SIRT2 and its interaction Here, we connect endogenous SIRT2 to the interaction with the acetylome (31). We present evidence that SIRT2 between the processes of aging, response to extracellular interacts with b-catenin and represses b-catenin–mediated stress by ionizing radiation exposure, de-differentiation transcriptional gene activation. Upon Sirt2 inactivation by and carcinogenesis. In contrast with Chen and colleagues genetic deletion, this repression is lifted. We demonstrated (32), who demonstrated overexpression of SIRT2 interacts that the level of transcriptionally active b-catenin increases with b-catenin in hepatocellular carcinoma, this is the first significantly. This is established by two approaches. First, we description of endogenous SIRT2 acting directly on showed that there is more b-catenin accumulation in the b-catenin in untransformed cells of advanced chronologic nucleus of Sirt2 KO MEFs by immunofluorescent staining age, demonstrating the potential for greater oxidative and by subcellular fractionation, whereas total whole-cell stress by ionizing radiation and ROS mimics in aging b-catenin remains constant between WT and Sirt2 KO tissues, specifically aging central nervous system, where MEFs. Second, we observed that there is an increase in SIRT2 is the predominant sirtuin isoform. Recently, phospho-b-catenin Y142 as well as an increase in acetylated Simic and colleagues demonstrated that SIRT1 may be b-catenin level in the Sirt2 KO MEFs. Previous studies have a therapeutic target to foster mesenchymal stem cell shown that the phosphorylated Y142 form and acetylated differentiation to preserve the integrity of mesenchymal form of b-catenin have increased transcriptional activity. stem cell–derived tissues in aging animals (33). In con- In accordance with increased active b-catenin in the trast, this manuscript demonstrates for the first time that absence of SIRT2, we showed that TCF/LEF transcription SIRT2 may be potentiating the damage induced by factor activity increased. Furthermore, ChIP analysis dem- ionizing radiation in untransformed aging cells and that onstrated that the presence of SIRT2 significantly attenuates disruption of the SIRT2–b-catenin interaction may be an b-catenin's binding to upstream promoter regions of Myc, endogenous therapeutic target to preserve the integrity of cyclin D1, and survivin. These findings correlate with obser- aging cells against exogenous stressors such as ionizing vations that both message and protein levels of these targets radiation. More importantly, this demonstrates differen- are significantly elevated. Concurrently, we also showed that tial regulation of the same pathway by differing sirtuin SIRT2 has a significant inhibitory effect on cell migration isoforms. and invasion abilities. The apparent increase in invasive and migratory potential when SIRT2 is absent may be due to Disclosure of Potential Conflicts of Interest increased expression of MMP-9 and decreased expression of No potential conflicts of interest were disclosed. E-cadherin. Our findings place SIRT2 as the second sirtuin to interact b Authors' Contributions with -catenin and regulate its transcriptional activity. Fire- Conception and design: P. Nguyen, J. Trepel, D.K. Smart stein and colleagues have shown that SIRT1 plays a major Development of methodology: P. Nguyen, J. Trepel, D.K. Smart role in inhibiting tumorigenesis and colon cancer growth by Acquisition of data (provided animals, acquired and managed patients, provided b via facilities, etc.): P. Nguyen, S. Lee, D. Lorang-Leins, J. Trepel inhibiting -catenin transcriptional activity deacetyla- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, compu- tion (27). Similar to Firestein and colleagues, we observe a tational analysis): P. Nguyen, S. Lee, J. Trepel, D.K. Smart b Sirt2 Writing, review, and/or revision of the manuscript: P. Nguyen, J. Trepel, difference in -catenin acetylation between our WT D.K. Smart and KO MEFs. We presented evidence that SIRT2 interacts Administrative, technical, or material support (i.e., reporting or organizing data, with b-catenin, deacetylates and suppresses b-catenin from constructing databases): D.K. Smart becoming active. When SIRT2 is absent, b-catenin no Study supervision: D.K. Smart longer binds and thus becomes activated. Taken together, these results strongly suggest that SIRT2 can act as a growth Grant Support This work was supported by the Intramural Research Program of the NIH, NCI, suppressor and mediate the response to ionizing radiation- and Center for Cancer Research. induced oxidative stress in normal, nontranformed tissues. The costs of publication of this article were defrayed in part by the payment of page Indeed, we demonstrated that in the presence of SIRT2, charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. radiation-induced H2O2 production was responsible for cell fi death. These ndings suggest that SIRT2 is critical to Received May 13, 2014; accepted May 14, 2014; published OnlineFirst May 24, maintaining the homeostasis of nontransformed cells, and 2014.

References 1. Logan CY, Nusse R. The Wnt signaling pathway in development and 3. Reya T, Clevers H. Wnt signalling in stem cells and cancer. Nature disease. Annu Rev Cell Dev Biol 2004;20:781–810. 2005;434:843–50. 2. Gordon MD, Nusse R. Wnt signaling: multiple pathways, multiple 4. Daniels DL, Weis WI. Beta-catenin directly displaces Groucho/TLE receptors, and multiple transcription factors. J Biol Chem 2006;281: repressors from Tcf/Lef in Wnt-mediated transcription activation. Nat 22429–33. Struct Mol Biol 2005;12:364–71.

1252 Mol Cancer Res; 12(9) September 2014 Molecular Cancer Research

SIRT2 Inhibits Wnt Following Radiation-Induced Stress

5. Brack AS, Conboy MJ, Roy S, Lee M, Kuo CJ, Keller C, et al. Increased histone dimethylation and gene expression. Mol Cell Biol 2008; Wnt signaling during aging alters muscle stem cell fate and increases 28:6720–9. ﬁbrosis. Science 2007;317:807–10. 21. Slane BG, Aykin-Burns N, Smith BJ, Kalen AL, Goswami PC, Domann 6. Liu H, Fergusson MM, Castilho RM, Liu J, Cao L, Chen J, et al. FE, et al. Mutation of succinate dehydrogenase subunit C results in Augmented Wnt signaling in a mammalian model of accelerated aging. increased O2.-, oxidative stress, and genomic instability. Cancer Res Science 2007;317:803–6. 2006;66:7615–20. 7. Ye F, Chen Y, Hoang T, Montgomery RL, Zhao XH, Bu H, et al. HDAC1 22. Zhang Y, Zhang M, Dong H, Yong S, Li X, Olashaw N, et al. Deacetyla- and HDAC2 regulate oligodendrocyte differentiation by disrupting the tion of cortactin by SIRT1 promotes cell migration. Oncogene beta-catenin-TCF interaction. Nat Neurosci 2009;12:829–38. 2009;28:445–60. 8. Toledo EM, Colombres M, Inestrosa NC. Wnt signaling in neuropro- 23. Isaacs JS, Jung YJ, Mole DR, Lee S, Torres-Cabala C, Chung YL, et al. tection and stem cell differentiation. Prog Neurobiol 2008;86:281–96. HIF overexpression correlates with biallelic loss of fumarate hydratase 9. Moon RT, Kohn AD, De Ferrari GV, Kaykas A. WNT and beta-catenin in renal cancer: novel role of fumarate in regulation of HIF stability. signalling: diseases and therapies. Nat Rev Genet 2004;5:691–701. Cancer Cell 2005;8:143–53. 10. Yao H, Ashihara E, Maekawa T. Targeting the Wnt/beta-catenin 24. Brembeck FH, Rosario M, Birchmeier W. Balancing cell adhesion and signaling pathway in human cancers. Expert Opin Ther Targets Wnt signaling, the key role of beta-catenin. Curr Opin Genet Dev 2011;15:873–87. 2006;16:51–9. 11. Harting K, Knoll B. SIRT2-mediated protein deacetylation: an emerging 25. Levy L, Wei Y, Labalette C, Wu Y, Renard CA, Buendia MA, et al. key regulator in brain physiology and pathology. Eur J Cell Biol Acetylation of beta-catenin by p300 regulates beta-catenin-Tcf4 inter- 2010;89:262–9. action. Mol Cell Biol 2004;24:3404–14. 12. Taylor DM, Maxwell MM, Luthi-Carter R, Kazantsev AG. Biological and 26. Xu W, Kimelman D. Mechanistic insights from structural studies of potential therapeutic roles of sirtuin deacetylases. Cell Mol Life Sci beta-catenin and its binding partners. J Cell Sci 2007;120(Pt 19): 2008;65:4000–18. 3337–44. 13. Haigis MC, Sinclair DA. Mammalian sirtuins: biological insights and 27. Firestein R, Blander G, Michan S, Oberdoerffer P, Ogino S, Campbell J, disease relevance. Annu Rev Pathol 2010;5:253–95. et al. The SIRT1 deacetylase suppresses intestinal tumorigenesis and 14. Hiratsuka M, Inoue T, Toda T, Kimura N, Shirayoshi Y, Kamitani H, et al. colon cancer growth. PLoS ONE 2008;3:e2020. Proteomics-based identiﬁcation of differentially expressed genes in 28. Ballestrem C, Wehrle-Haller B, Hinz B, Imhof BA. Actin- human gliomas: down-regulation of SIRT2 gene. Biochem Biophys dependent lamellipodia formation and microtubule-dependent tail Res Commun 2003;309:558–66. retraction control-directed cell migration. Mol Biol Cell 2000;11: 15. Matsushita N, Takami Y, Kimura M, Tachiiri S, Ishiai M, Nakayama T, 2999–3012. et al. Role of NAD-dependent deacetylases SIRT1 and SIRT2 in 29. Crawford HC, Fingleton BM, Rudolph-Owen LA, Goss KJ, Rubinfeld B, radiation and cisplatin-induced cell death in vertebrate cells. Genes Polakis P, et al. The metalloproteinase matrilysin is a target of beta- Cells 2005;10:321–32. catenin transactivation in intestinal tumors. Oncogene 1999;18:2883– 16. Tang BL, Chua CE. SIRT2, tubulin deacetylation, and oligodendroglia 91. differentiation. Cell Motil Cytoskeleton 2008;65:179–82. 30. Wu B, Crampton SP, Hughes CC. Wnt signaling induces matrix 17. Maxwell MM, Tomkinson EM, Nobles J, Wizeman JW, Amore AM, metalloproteinase expression and regulates T cell transmigration. Quinti L, et al. The Sirtuin 2 microtubule deacetylase is an abundant Immunity 2007;26:227–39. neuronal protein that accumulates in the aging CNS. Hum Mol Genet 31. Park SH, Zhu Y, Ozden O, Kim HS, Jiang H, Deng CX, et al. SIRT2 is a 2011;20:3986–96. tumor suppressor that connects aging, acetylome, cell cycle signaling, 18. Kim HS, Vassilopoulos A, Wang RH, Lahusen T, Xiao Z, Xu X, et al. and carcinogenesis. Transl Cancer Res 2012;1:15–21. SIRT2 maintains genome integrity and suppresses tumorigenesis 32. Chen J, Chan AW, To KF, Chen W, Zhang Z, Ren J, et al. SIRT2 through regulating APC/C activity. Cancer Cell 2011;20:487–99. overexpression in hepatocellular carcinoma mediates epithelial to 19. Mushinski JF, Nguyen P, Stevens LM, Khanna C, Lee S, Chung EJ, mesenchymal transition via akt/GSK-3beta/beta-catenin signaling et al. Inhibition of tumor cell motility by the interferon-inducible GTPase (revised version). Hepatology 2013;57:2287–98. MxA. J Biol Chem 2009;284:15206–14. 33. Simic P, Zainabadi K, Bell E, Sykes DB, Saez B, Lotinun S, et al. SIRT1 20. Nguyen P, Bar-Sela G, Sun L, Bisht KS, Cui H, Kohn E, et al. BAT3 regulates differentiation of mesenchymal stem cells by deacetylating and SET1A form a complex with CTCFL/BORIS to modulate H3K4 beta-catenin. EMBO Mol Med 2013;5:430–40.

www.aacrjournals.org Mol Cancer Res; 12(9) September 2014 1253

SIRT2 Interacts with β-Catenin to Inhibit Wnt Signaling Output in Response to Radiation-Induced Stress

Phuongmai Nguyen, Sunmin Lee, Dominique Lorang-Leins, et al.

Mol Cancer Res 2014;12:1244-1253. Published OnlineFirst May 27, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-14-0223-T

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2021/03/13/1541-7786.MCR-14-0223-T.DC1

Cited articles This article cites 33 articles, 10 of which you can access for free at: http://mcr.aacrjournals.org/content/12/9/1244.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/12/9/1244.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/12/9/1244. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.