SIRT2 overexpression in hepatocellular carcinoma mediates epithelial to mesenchymal transition by protein kinase B/glycogen synthase kinase-3β/β-catenin signaling


  • Juan Chen,

    1. The Key Laboratory of Molecular Biology on Infectious Diseases, Ministry of Education, Chongqing Medical University, Chongqing, China
    Search for more papers by this author
  • Anthony W.H. Chan,

    1. Departments of Anatomical and Cellular Pathology
    Search for more papers by this author
  • Ka-Fai To,

    1. Departments of Anatomical and Cellular Pathology
    2. The State Key Laboratory in Oncology in South China, The Chinese University of Hong Kong, Hong Kong, China
    Search for more papers by this author
  • Weixian Chen,

    1. Department of Clinical Laboratory, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, China
    Search for more papers by this author
  • Zhenzhen Zhang,

    1. Department of Infectious Diseases, The Children's Hospital of Chongqing Medical University, Chongqing, China
    Search for more papers by this author
  • Jihua Ren,

    1. The Key Laboratory of Molecular Biology on Infectious Diseases, Ministry of Education, Chongqing Medical University, Chongqing, China
    Search for more papers by this author
  • Chunli Song,

    1. The Key Laboratory of Molecular Biology on Infectious Diseases, Ministry of Education, Chongqing Medical University, Chongqing, China
    Search for more papers by this author
  • Yue-Sun Cheung,

    1. Departments of Surgery, The Chinese University of Hong Kong, Hong Kong, China
    Search for more papers by this author
  • Paul B.S. Lai,

    1. Departments of Surgery, The Chinese University of Hong Kong, Hong Kong, China
    Search for more papers by this author
  • Suk-Hang Cheng,

    1. Departments of Anatomical and Cellular Pathology
    Search for more papers by this author
  • Margaret H.L. Ng,

    1. Departments of Anatomical and Cellular Pathology
    Search for more papers by this author
  • Ailong Huang,

    Corresponding author
    1. The Key Laboratory of Molecular Biology on Infectious Diseases, Ministry of Education, Chongqing Medical University, Chongqing, China
    • The Key Laboratory of Molecular Biology on Infectious Diseases, Ministry of Education, Chongqing Medical University, Chongqing, China
    Search for more papers by this author
    • Fax: (86)-02368486780

  • Ben C.B. Ko

    Corresponding author
    1. The Key Laboratory of Molecular Biology on Infectious Diseases, Ministry of Education, Chongqing Medical University, Chongqing, China
    2. Departments of Anatomical and Cellular Pathology
    3. The State Key Laboratory in Oncology in South China, The Chinese University of Hong Kong, Hong Kong, China
    • Department of Anatomical and Cellular Pathology, The Chinese University of Hong Kong, Room 38019, Clinical Sciences Building, The Prince of Wales Hospital, Shatin, Hong Kong, China
    Search for more papers by this author
    • fax: 852-2637-6274

  • Potential conflict of interest: Nothing to report.

  • This work is supported by the Research Grant Council (grant nos.: CUHK 466109 and 467210) and the National Natural Science Foundation of China (grant no.; 81272764) to B.C.B.K; and National Science and Technology Major Project of China (grant no.:2013ZX10002002, 2012ZX10002005) to A.H.


Sirtuin 1 (SIRT1) has been implicated in telomere maintenance and the growth of hepatocellular carcinoma (HCC). Nevertheless, the role of other sirtuins in the pathogenesis of HCC remains elusive. We found that sirtuin 2 (SIRT2), another member of the sirtuin family, also contributes to cell motility and invasiveness of HCC. SIRT2 is up-regulated in HCC cell lines and in a subset of human HCC tissues (23/45). Up-regulations of SIRT2 in primary HCC tumors were significantly correlated with the presence of microscopic vascular invasion (P = 0.001), a more advanced tumor stage (P = 0.004), and shorter overall survival (P = 0.0499). Functional studies by short hairpin RNA–mediated suppression of SIRT2 expression in HCC cell lines revealed significant inhibition of motility and invasiveness. Depletion of SIRT2 also led to the regression of epithelial-mesenchymal transition (EMT) phenotypes, whereas the ectopic expression of SIRT2 in the immortalized hepatocyte cell line L02 promoted cell motility and invasiveness. Mechanistic studies revealed that SIRT2 regulates the deacetylation and activation of protein kinase B, which subsequently impinges on the glycogen synthase kinase-3β/β-catenin signaling pathway to regulate EMT. Conclusions: Our findings have uncovered a novel role for SIRT2 in HCC metastasis, and provide a rationale to explore the use of sirtuin inhibitors in HCC therapy. (HEPATOLOGY 2013;)

Hepatocellular carcinoma (HCC) is the fifth-most common malignancy worldwide and the second-leading cause of cancer death in Asia, generally, and in China, in particular.1 Although the cancer can be eradicated by curative surgery, most HCC patients are diagnosed with intra- and extrahepatic metastases in inoperable, advanced stages, which renders surgical resection not applicable. The prognosis is very poor for patients who have unresectable tumors, with a median survival of approximately 6 months.2 At present, even the most effective forms of systemic therapy, such as doxorubicin1, 3 or sorafenib,2, 4 only minimally extend the lifespan of these patients. Therefore, a thorough understanding of the underlying mechanisms regarding tumor growth and metastasis is vital for the development of efficacious therapeutics.

Sirtuins are mammalian homologs of the yeast silent information regulator 2 (SIR2), which are histone deacetylases that utilizes nicotinamide adenine dinucleotide as a cofactor for their functions.5 The yeast, SIR2, regulates aging by maintaining transcriptional silencing of the mating-type loci, the ribosomal DNA locus, and the telomeres.6 In mammals, there are seven homologs of SIR2 (SIRT1-7), of which SIRT1 is considered to be the human ortholog of SIR2.7 SIRT1 is mainly localized to the nucleus and plays a key role in energy metabolism, telomeric maintenance, and genomic stability by targeting a variety of nonhistone proteins.8-11 The role of SIRT1 in cancer is controversial because it may act as a tumor promoter or suppressor, depending on the tumor type.12

SIRT2 acts on certain substrates of SIRT1, such as H4K16, p53, FOXO3, and p65.13-16 Nevertheless, the predominant cytoplasmic localization of SIRT2 and its role in the regulation of tubulin dynamics and neuronal motility suggested that it might have functional roles distinctive from SIRT1.17, 18 Indeed, recent findings suggested that SIRT2 is associated with mitotic apparatus during the cell cycle19, 20 and is essential for maintaining genomic stability by deacetylating CDH1 and CDC20 of the anaphase-promoting complex/cyclosome.21 Emerging evidence has also suggested that SIRT2 is involved in tumorigenesis.21 SIRT2 deficiency results in aneuploidy and mitotic cell death, and SIRT2-deficient mice have a higher propensity for developing tumors.21 Moreover, SIRT2 expression is down-regulated in some cancers,21, 22 suggestive of a tumor-suppressor function.

SIRT1 expression is up-regulated in HCC, and SIRT1 may play a role in HCC tumorigenesis through telomere maintenance.23 In this study, we showed that SIRT2 is also up-regulated in HCC. Overexpression of SIRT2 in primary HCC tumors is positively correlated with microscopic vascular invasion and adverse patient prognosis. Using HCC cell models, we uncovered a key role of SIRT2 as a tumor promoter in HCC by promoting epithelial-mesenchymal transition (EMT) and motility of HCC cells by targeting the protein kinase B/glycogen synthase kinase (Akt/GSK)3-β/β-catenin-signaling pathway. Our findings provide a rationale for the clinical exploration of the use of sirtuin inhibitors in HCC therapy.


Abs, antibodies; α-SMA, alpha smooth muscle actin; Akt, protein kinase B; APC, adenomatous polyposis coli; BrdU, bromodeoxyuridine; cDNA, complementary DNA; EMT, epithelial-mesenchymal transition; FACS, fluorescence-activated cell sorting; FBS, fetal bovine serum; GFP, green fluorescent protein; GSK, glycogen synthase kinase; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; IF, immunofluorescence; mRNA, messenger RNA; PI, propidium iodide; qPCR, qantitative polymerase chain reaction; shRNA, short hairpin RNA; SIR2, silent information regulator 2; SIRT1/2, sirtuin 1 and 2; WT, wild type.

Materials and Methods

Plasmids, Antibodies, and siRNA.

SIRT2 short hairpin RNA (shRNA) (shSIRT2-1 and shSIRT2-2) or nontargeting shRNA (shCont) was cloned into a modified pLentilox-3.7 lentivirus plasmid vector containing a blasticidin-resistant gene (provided by Dr. D.Y. Jin from The University of Hong Kong). Sequence of shSIRT2-1 and SIRT2-2 targeting shRNA is 5′-GCCAACCATCTGTCACTACTT-3′ and 5′-GCTAAGCTGGATGAAAGAGAA-3′, respectively. Sequence of shCont is 5′-GCAACAAGATGAAGAGCACCAA-3′. SIRT1 shRNA (shSIRT1-1) expressing lentivirus was generated as previously described.23 The pcDNA3.1-β-catenin and pcDNA3.1-SIRT2 expression vector was from Addgene (Cambridge, MA). SIRT2 (Sc-20966) and N-cadherin (sc-59987) antibodies (Abs) were from Santa Cruz Biotechnology (Santa Cruz, CA); β-catenin (#8480), vimentin (#3932), α-catenin (#3236), E-cadherin (#3195), AKT (#2966), and acetylated-lysine (#9441) Abs were from Cell Signaling Technology, Inc. (Danvers, MA); active β-catenin (clone 8E7, 05-665) Ab was from Millipore (Billerica, MA); and β-actin (A5316) and alpha smooth muscle actin (α-SMA; A5228) Abs were from Sigma-Aldrich (St. Louis, MO). Smartpool siRNAs against β-catenin was obtained from Thermo Fisher Scientific Inc. (Waltham, MA).

HCC Specimens.

Tumorous liver tissues and the corresponding adjacent nontumoral liver tissues were obtained from 45 patients who underwent curative surgery for HCC the Prince of Wales Hospital in Hong Kong. Patients were not subjected to any neoadjuvant therapy before surgery. Informed consent was obtained from each patient that was recruited. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the clinical research ethics committee of the Chinese University of Hong Kong. Clinical and pathology records were retrieved and the following information was obtained: age at initial diagnosis, gender, size of the tumor, American Joint Committee on Cancer (7th edition) tumor-node-metastasis stage, follow-up duration; and disease-free and overall survival. Total RNAs and proteins were extracted from these specimens.

Cell Culture.

HepG2, SK-Hep-1, and PLC5 cells were obtained from American Type Culture Collection (Manassas, VA). The Huh-7 cell line was acquired from the Health Science Research Resource Bank (Osaka, Japan). The L02 cell line was obtained from Prof. Nathalie Wong (The Chinese University of Hong Kong). HepG2 was cultured in Eagle's minimum essential medium containing 10% fetal bovine serum (FBS; Gibco BRL, Grand Island, NY). SK-Hep-1, Huh-7, PLC5, Hep3B, and L02 cells were cultured in Dulbecco's modified Eagle's medium containing 10% FBS (Gibco BRL). All cells were maintained in a humidified incubator at 37°C with 5% CO2.

Lentivirus Production.

Lentivirus expressing shSIRT2-1, shSIRT2-2, or shCont was produced in HEK-293FT cells using the corresponding shRNA-expressing pLentilox-3.7 vector with the aid of packaging plasmids pLP1, pLP2, and pLP/VSVG from the BLOCK-iT Lentiviral RNAi Expression System (Invitrogen, Carlsbad, CA). Viruses were concentrated by using PEG-it virus precipitation solution (System Biosciences, Mountain View, CA) and stored at −80°C.

Quantitative Real-Time Polymerase Chain Reaction.

Total RNA was extracted from tissue by using TRIzol reagent (Invitrogen), and complementary DNA (cDNA) was synthesized from 1 μg of total RNA using High Capacity RNA-to-cDNA Master Mix (Applied Biosystems, Foster City, CA), according to the manufacturer's instructions. Genomic DNA was digested by DNase I (New England Biolabs). Quantitative polymerase chain reaction (qPCR) experiments were performed using the SYBR Green PCR Core reagent kit (Applied Biosystems), and reactions were carried out using an ABI 7900 real-time PCR system (Applied Biosystems). The primers for quantifying SIRT2 are 5′-CCGGCCTCTATGACAACCTA-3′ and 5′-GGAGTAGCCCCTTGTCCTTC-3′. The primers for quantifying β-actin are 5′-CTCTTCCAGCCTTCCTTCCT-3′ and 5′-AGCACTGTGTTGGCGTACAG-3′.

Immunoprecipitation and Western Blotting Analysis.

Immunoprecipitation (IP) was carried out using protein G-agarose (Millipore). For western blotting analysis, protein lysates were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and immunoblotted with Abs, as indicated. Blottings were developed with ECL western blotting reagents (Pierce Biotechnology, Rockford, IL). Signal intensity was quantified by ImageJ (National Institutes of Health, Bethesda, MD).

Luciferase Reporter Assays.

The effect of SIRT2 RNA interference on β-catenin activity was determined by TOP/FOP flash luciferase reporter, as previously described.24 Eight hours after transfection of plasmids, cells were transduced with lentivirus expressing the indicated shRNA. Two days after infection, cells were harvested and assayed by the Dual Luciferase Report Assay System (Promega, Madison, WI), according to the manufacturer's instructions. pRL-RK was cotransfected with reporter plasmid to normalize transfection efficiency. Luciferase activity was determined by a GloMax microplate luminometer (Promega).

Analysis of Cell-Cycle Distribution.

Cell-cycle distribution was determined by fluorescence-activated cell sorting (FACS) analysis, as previously described.25 Cells were stained with propridium iodide (PI; Sigma-Aldrich). Flow cytometry was carried out by a FACS Calibur flow cytometer (BD Biosciences, San Jose, CA). Data acquisition and analysis were done with CellQuest (BD Biosciences).

Cell Proliferation and Bromodeoxyuridine Assay.

Cell proliferation in response to SIRT2 silencing was determined by trypan blue exclusion assay. DNA synthesis was determined by the bromodeoxyuridine (BrdU) assay, according to the manufacturer's instructions (Roche Diagnostics, Basel, Switzerland). The result was expressed as a percentage maximum absorbance at 450 nm, based on three independent experiments.

Colony Formation Assay.

Cells were grown in blasticidin-containing medium (3.5 μg/mL) for 14 days, and colony formation assay was carried out as previously described.26 Crystal violet-stained colonies were scored, and results from duplicate assays were expressed as the mean from four independent experiments.

Migration and Invasion Assays

Cell motility and invasive abilities were assessed by way of transwell (Corning Life Sciences, Bedford, MA) and Matrigel invasion (BD Biosciences), respectively. For transwell migration assay, 2 × 104 cells were seeded, whereas 5 × 104 cells were seeded for the invasion assay. Cells migrated to the underside of the membrane were fixed and stained with 0.1% crystal violet and were enumerated for 10 microscope fields. Mean values of migrating or invading cells were expressed as percentages relative to mock or vector control. Each experiment was performed in replicate inserts, and mean value was calculated from three independent experiments.

Wound-Healing Assay.

Lentivirus-transduced SK-Hep-1 cells were seeded onto six-well plates. Cells were grown to confluency and were gently scratched with a pipette tip to create a mechanical wound. Images were taken after 24 hours, and the subsequent recolonization of the stripped surface was quantified by measuring the distance between the wound edges. Experiments were carried out in triplicate wells from three independent experiments.

Immunofluorescence Microscopy.

Cells grown on glass coverslips were fixed in 4% paraformaldehyde and permeabilized using 0.5% Triton X-100. β-catenin was stained with rabbit polyclonal anti-β-catenin Ab overnight at 4°C. Cells were then rinsed with phosphate-buffered saline and incubated with goat antirabbit fluorescein isothiocyanate Abs. Filamentous actins were stained with TRITC-labeled Phalloidin (Sigma-Aldrich). Cells were counterstained with 4′,6-diamidino-2-phenylindole and examined by fluorescence microscopy (Zeiss Axiovert 200 M; Carl Zeiss, Oberkochen, Germany).

Statistical Analysis.

SIRT2 expression in HCC and nontumoral liver tissues was compared using the paired Student t test. Correlation between SIRT2 and individual clinicopathologic parameters was evaluated with the nonparametric chi-square test, Spearman's σ rank test, and the Student t test. Kaplan-Meier's method was used to estimate the survival rates for SIRT2 expression. Equivalences of the survival curves were tested by log-rank statistics. All statistical analyses were carried out by the statistical program, SPSS version 16.0 (SPSS, Inc., Chicago, IL). A two-tailed P value <0.05 was regarded as statistically significant.


SIRT2 Expression in HCC.

We first determined the expression of SIRT2 using a panel of HCC cell lines. Consistent with earlier findings that SIRT2 utilizes alternate ATGs for translation,18, 27 two SIRT2 isoforms of variable abundance were detected in most HCC cell lines examined (SK-Hep-1, PLC5, HepG2, Hep3B, and Huh-7) (Fig. 1A). The level of SIRT2 was low in L02 cells, an immortalized human liver cell line (Fig. 1A). On the other hand, only one SIRT2 isoform expressed at low level was detected in normal liver specimens, whereas it was highly expressed in HCC tissues (Fig. 1A). These data suggest that the translation of SIRT2 is subjected to differential regulation, and SIRT2 is expressed at a higher level in HCC cells and tissue. Different from SIRT1, which is localized exclusively in the nucleus in HCC cells,23 both SIRT2 isoforms were predominantly localized in the cytoplasm in HCC cell lines (Fig. 1B), suggesting that SIRT1 and SIRT2 may elicit distinctive functions in HCC.

Figure 1.

Expression and subcellular localization of SIRT2 in HCC. (A) Expression of SIRT2 protein was determined by using SIRT2 Abs. β-actin was used as a loading control. Total protein extracts were used for the analysis. Normal livers were obtained from patients that were free of HCC and hepatitis. T, liver specimen from HCC patient. (B) SIRT2 was predominately localized in the cytoplasmic compartment in HCC cells. The purity of cytoplasmic and nuclear extracts was confirmed by Abs against Fyn and Nup88, respectively. N, nuclear; C, cytoplasmic. (C) Expression of SIRT2 in paired HCC was determined by using SIRT2 Abs. Relative SIRT2 protein expression was determined by normalizing SIRT2 signal to the β-actin signal. NT, adjacent nontumoral liver; T, tumor. (D) Significant up-regulation of SIRT2 protein expression was suggested in HCC tumors, relative to the adjacent nontumoral liver (n = 45) (*P <0.001). (E) Kaplan-Meier's analyses were performed according to SIRT2 protein expression on HCC patients. High tumoral SIRT2 expression was regarded at >1.5-fold up-regulation, relative to the adjacent nontumoral liver. Survival (P = 0.0499) of patients who had high SIRT2 expression was shorter.

To gain a better understanding of the role of SIRT2 in HCC, we tried to determine the expression level of SIRT2 using clinical specimens of HCC. We evaluated a panel of commercially available SIRT2 Abs for immunohistochemistry, but the lack of specificity of these Abs precluded the analysis of a SIRT2 expression pattern using HCC tissue microarrays. Alternatively, we determined protein level and significance of SIRT2 expression in 45 pairs of primary HCC and adjacent nontumoral liver. Clinicopathological parameters of these patients are summarized in Table 1. Western blotting analysis revealed that SIRT2 expression can be detected in the majority of nontumoral liver specimens, but a significant portion of patients showed elevated SIRT2 level in tumor tissues (23 of 45 cases) (Fig. 1C). Moreover, the average level of SIRT2 was found to be significantly higher (P < 0.001) in the tumor group (median, 1.68; quartiles, 1.05-2.24), relative to the nontumoral liver group (median, 1.00; quartiles, 0.66-1.68) (Fig. 1D). However, analysis of SIRT2 messenger RNA (mRNA) levels from these patients by real-time qPCR reveals that average SIRT2 mRNA levels in tumor and nontumoral liver did not differ significantly (data not shown), suggesting that SIRT2 expression in HCC is regulated by transcription-independent mechanisms. Correlative analysis of SIRT2 protein levels with clinicopathologic features suggested significant association between increased SIRT2 expression and the histologic presence of microscopic vascular invasion (P = 0.001) and more-advanced tumor stages (P = 0.004) (Table 1). Up-regulation of SIRT2 in HCC was also found to predict shorter overall survival (P = 0.0499) of patients (Fig. 1E).

Table 1. Correlative Analysis of SIRT2 Protein Levels With Clinicopathological Features
  SIRT2 Expression (Tumor/Nontumoral) 
Clinicopatholigic ParametersNo. of SpecimensLowHighP Value
  1. High tumoral SIRT2 expression was regarded at >1.5-fold up-regulation, relative to the adjacent nontumoral liver.

  2. Abbreviation: NS, no significant difference.

Age4557.4 ± 11.454.7 ± 11.1NS
 2 or 320515 
Multiple tumor    
Vascular invasion (macro)    
Vascular invasion (micro)    

Effect of SIRT2 Silencing on HCC Cell Proliferation and Apoptosis.

An earlier study suggested that there was a role for SIRT2 in the motility of mouse embryonic fibroblasts.18 The association between SIRT2 expression and microscopic vascular invasion in HCC also suggested a role of SIRT2 in the cell motility of HCC cells. Therefore, we carried out lentivirus-mediated shRNA knockdown to elucidate the cellular functions of SIRT2. Two independent shRNAs (shSIRT2-1 and shSIRT2-2) showed efficient SIRT2 knockdown in p53 wild-type (WT) (SK-Hep-1 and HepG2) and mutated (PLC5 and Huh-7) HCC cells, compared with scrambled shRNA (shCont)-transduced cells (Fig. 2A).

Figure 2.

Effect of SIRT2 gene silencing on growth of HCC cells. (A) SIRT2 protein expression in HCC cells was significantly reduced by lentiviruses expressing SIRT2-targeting (shSIRT2-1 and shSIRT2-2) versus control (shCont) shRNA. Cells were harvested 4 days after viral transduction. β-actin was used as a loading control. (B) Proliferation of HCC cells was suppressed by the depletion of SIRT2. SK-Hep-1 cells were transduced with the respective lentivirus, and cell number was determined by trypan blue exclusion assay at the indicated number of days thereafter. Plots are cumulative cell numbers versus days in culture (*P < 0.001). (C) Depletion of SIRT2 reduced colony formation. SK-Hep-1 cells were transduced with lentiviruses expressing the indicated shRNA. Cells were grown for 14 days in the presence of blasticidin (3.5 μg/mL) and stained with crystal violet. (D) Depletion of SIRT2 reduced DNA incorporation. Cells expressing the indicated shRNA were pulsed with BrdU (10 μmol/L) for 4 hours, followed by staining with anti-BrdU Ab. Signal incorporation was measured at 450 nm. The O.D. values were expressed as percentage relative to the untreated group (*P < 0.001). (E) Cell-cycle analysis of SIRT2 knockdown cells by using PI and flow cytometric analysis. *P < 0.01, when compared to shCont.

Down-regulation of SIRT2 inhibited the growth of the above HCC cells over a course of 6 days, and this was independent of their p53 status (Fig. 2B and Supporting Fig. 1). Knockdown of SIRT2 also reduced the number and size of blasticidin-resistant SK-Hep-1 cells colonies, as determined by colony formation assay (Fig. 2C). Concordantly, SIRT2 depletion reduced DNA synthesis by 50% (P < 0.001), as measured by BrdU incorporation assay (Fig. 2D). Cell-cycle distribution, as determined by FACS analysis, revealed that SIRT2 silencing significantly induced G1 and G2 arrest in p53 WT (SK-Hep-1 and HepG2) and in p53-mutated (Huh-7 and PLC5) cells, respectively (Fig. 2E). Nevertheless, unlike the knockdown of SIRT1,23 knockdown of SIRT2 neither resulted in cellular senescence nor apoptosis of HCC cells (data not shown). Taken together, these data suggested that SIRT2 depletion might inhibit cell growth by inducing cell-cycle delay.

Functional Role of SIRT2 in Cell Migration and Invasion.

Next, we evaluated whether SIRT2 plays a role in the motility of HCC cells. Depletion of SIRT2 (shSIRT2-1 and shSIRT2-2), but not SIRT1 (shSIRT1-1), markedly reduced cell migration through transwell (P < 0.01) (Fig. 3A). Concordantly, knockdown of SIRT2 also diminished wound-healing capacity (P < 0.01) (Fig. 3B) and impaired cell invasion through Matrigel (P < 0.01) (Fig. 3C). In contrast, ectopic expression of WT SIRT2 promoted migration and invasion capacity in L02 cells (Fig. 3D). Together, these data suggested a role of SIRT2 in the motility and invasiveness of HCC cells. EMT, the sequence of events that converts adherent epithelial cells into migratory cells, which invade the extracellular matrix,28 is associated with tumor metastasis.29 Therefore, we determined whether EMT is responsible for SIRT2-mediated change in cell motility. Knockdown of SIRT2 in HCC cells induced the expression of epithelial markers E-cadherin and α-catenin that was accompanied by a concomitant reduction of mesenchymal marker N-cadherin and α-SMA. Nevertheless, the expression of vimentin, another mesenchymal marker, was not altered (Fig. 3E). F-actin distribution was also rearranged in SIRT2-depleted cells from a stress-fiber to a cortical pattern, suggestive of a conversion to the epithelial phenotype (Fig. 3F). Therefore, our data suggest a loss of mesenchymal-like features and reacquisition of epithelial characteristics in SIRT2-depleted HCC cells. The role of SIRT2 in EMT was further supported by the reduced expression of E-cadherin and alpha-catenin, as well as the enhanced expression of N-cadherin and α-SMA in L02 cells, which SIRT2 was ectopically expressed (Fig. 3E).

Figure 3.

Functional effects of SIRT2 on HCC migration, invasion, and EMT. (A) Significant reduction in cell migration was found in cells expressing shSIRT2-1 and shSIRT2-2, compared with cells expressing shCont and shSIRT1-1, respectively (*P < 0.001). Representative images of migrated SK-Hep-1 and Huh7 cells in each group are shown. (B) Significant impairment of wound-healing ability was found in SK-Hep-1 cells expressing the shSIRT2-1 and shSIRT2-2, compared to cells expressing shCont. (C) Significant reduction in cell invasion was found in SK-Hep-1 and Huh-7 cells expressing shSIRT2-1 and shSIRT2-2, compared to cells expressing shCont (*P < 0.001). (D) Ectopic expression of SIRT2 promoted migration and invasion in L02 cells (*p < 0.001). (E) SIRT2 promotes EMT in HCC. Western blottings showed a significant increase in the expression of the epithelial markers, E-cadherin and alpha-catenin, and a corresponding reduction of the mesenchymal markers, N-cadherin, and α-SMA in SIRT2-depleted cells, relative to control. Ectopic expression of SIRT2 in L02 cells suppressed the expression of E-cadherin and α-catenin, but increased the expression of N-cadherin and α-SMA. (F) Phalloidin staining (red) in SK-Hep-1 cells expressing shSIRT2-1 and shSIRT2-2 showed a reduction of actin stress-fiber filaments.

SIRT2 Regulates β-Catenin Signaling in HCC Cells.

Reduced expression of E-cadherin and the activation of WNT signaling lead to the accumulation and nuclear import of β-catenin, where it interacts with TCF/LEF to induce the expression of genes responsible for the EMT process.30, 31 Therefore, we have elucidated whether SIRT2 plays a role in β-catenin signaling or not. Expression of SIRT2 shRNAs in SK-Hep1 and Huh7 cells markedly reduced the level of the total, as well as the active (dephosphorylated on Ser37 and Ser41), β-catenin (Fig. 4A). The effect of SIRT2 depletion on β-catenin was blocked in the presence of proteasome inhibitor MG132, suggesting that the degradation of β-catenin is mediated by the proteasome (Fig. 4B). Knockdown of SIRT2 also caused a redistribution of cytoplasmic and nuclear β-catenin to the membranous localization (Fig. 4C). Concordantly, TOPflash and FOPflash luciferase reporter analysis revealed that the transactivation of TCF reporter was inhibited by the depletion of SIRT2 (Fig. 4D).

Figure 4.

SIRT2 regulates β-catenin signaling. (A) Depletion of SIRT2 reduced the level of active and total β-catenin. (B) Proteasome inhibitor MG132 (10 μM; 8 hours) prevented the degradation of β-catenin induced by SIRT2 depletion in SK-Hep-1 cells. (C) IF staining of β-catenin (green) in SK-Hep-1 cells expressing SIRT2 shRNA showed a redistribution of β-catenin to membrane. Arrows indicate nuclear localization of β-catenin in control cells. (D) TOP/FOP luciferase activity assay revealed that β-catenin/Tcf4 transcription activity was suppressed by the depletion of SIRT2 (*P < 0.01).

To further determine whether SIRT2 exerts its function by β-catenin signaling, we ectopically expressed β-catenin or green fluorescent protein (GFP) in SIRT2-depleted SK-Hep-1 cells. Importantly, ectopic expression of β-catenin, but not GFP, significantly restored cell proliferation (Fig. 5A), as well as enhanced cell migration (Fig. 5B) and invasion (Fig. 5C). In contrast, ectopic expression of SIRT2 in nontumorigenic L02 cells promoted their migration and invasion that was inhibited by depletion of β-catenin (Fig. 5D). Collectively, these data suggested that SIRT2 regulates HCC cell growth and motility through regulating β-catenin signaling.

Figure 5.

SIRT2 regulates HCC cell growth and motility through β-catenin. (A) Ectopic expression of β-catenin, but not GFP, significantly rescued the proliferation of SIRT2-depleted SK-Hep1 cells over a time course of 4 days (*P < 0.01) (left). Western blotting analysis confirmed the up-regulation of β-catenin in cells transfected with β-catenin cDNA (right). (B and C) Ectopic expression of β-catenin, but not GFP, significantly rescued cell migration (B) and invasion (C) in SIRT2-depleted cells (*P < 0.01). (D) Ectopic expression of SIRT2 in L02 cells significantly promoted cell migration and invasion that was inhibited by the depletion of β-catenin (*P < 0.01) (left). Western blotting analysis shows the expression of SIRT2 and β-catenin in control cells (lane 1), SIRT2-overexpressing cells (lane2), and SIRT2-overexpressing cells with β-catenin knockdown (lane 3) (right).

To elucidate the underlying mechanism of SIRT2-dependent β-catenin inactivation, we determined the status of GSK-3β, which forms a destruction complex with Axin and adenomatous polyposis coli (APC) for the phosphorylation and degradation of β-catenin.31 Depletion of SIRT2 increased the abundance of unphosphorylated (activated) and total GSK-3β, whereas it reduced the level of phosphorylated (activated) Akt (Fig. 6A). Because Akt phosphorylates and inactivates GSK-3β,32 our results suggested that SIRT2 may affect EMT by regulating the Akt/GSK-3β/β-catenin-signaling axis. An earlier study suggested that phosphorylation and activity of Akt is regulated by SIRT1-dependent deacetylation33; therefore, we determined whether SIRT2 plays a role in the acetylation of Akt, GSK-3β, and β-catenin proteins. These proteins were first immunoprecipitated by the corresponding Abs, respectively, and their acetylation status was determined by anti-acetylated-lysine Abs. Our data showed that β-catenin was neither acetylated when SIRT2 was expressed nor depleted, whereas GSK-3β was constitutively acetylated under both conditions (Fig. 6B). On the other hand, although Akt was also constitutively acetylated, its acetylation level was markedly up-regulated by the depletion of SIRT2, whereas depletion of SIRT1 did not alter Akt acetylation (Fig. 6B). More important, SIRT2, but not SIRT1, was coimmunoprecipitated with AKT (Fig. 6C). Taken together, these data revealed a novel role of SIRT2 in the β-catenin signaling pathway by regulating Akt acetylation in HCC cells.

Figure 6.

Effect of SIRT2 on Akt/GSK3β/β-catenin signaling. (A) Depletion of SIRT2 reduced the level of phosphorylated GSK-3β and Akt, whereas it promoted the level of total GSK-3β. (B) Depletion of SIRT2 enhanced Akt acetylation. β-catenin, GSK-3β, and Akt was immunoprecipitated with the respective Abs and probed with anti-acetylated-lysine Abs. SK-Hep-1 cells were depleted for SIRT1 and SIRT2 using lentivirus expressing shSIRT1-1 and shSIRT2-1 respectively. (C) Coimmunoprecipitation of AKT and SIRT2. SK-Hep1 cells were immunoprecipitated using AKT Abs and immunoblotted with SIRT2, SIRT1, and AKT Abs, respectively.


Sirtuins are involved in various aspects of biological processes, such as the regulation of gene expression, cellular stress response, DNA repair and metabolism, and so on. Despite there being a growing interest in elucidating the functions of sirtuins, how this group of deacetylases is involved in tumorigenesis is still poorly understood.34 Among these, SIRT1 has been the one studied most extensively, yet its role in tumor development remains controversial. Overexpression of SIRT1 apparently protects against tumor formation in the mouse model of colon tumor,35 and SIRT1 expression was substantially reduced in human breast cancer.36 Concordantly, an increase genomic instability and susceptibility to the development of spontaneous tumors was observed in SIRT1+/−;p53+/− mice.36 However, SIRT1 is also found to be overexpressed in a variety of cancers, including acute myeloid leukemia, HCC, skin tumor, and prostate tumor.23, 37-39 These data suggested that SIRT1 might act as a tumor promoter or suppressor in a cell-type–specific manner. On the other hand, a recent study by Kim et al. showed that SIRT2-deficient mice are more susceptible to the development of tumors.21 Moreover, they showed that SIRT2 mRNA expression is reduced in breast cancer and HCC, respectively.21 Contrary to their findings, we found that SIRT2 mRNA is expressed at a similar level between tumorous and adjacent nontumorous tissue in HCC, and SIRT2 protein is evidently expressed at a higher level in tumors. Unfortunately, Kim et al.'s study did not determine SIRT2 protein level in HCC, which precluded a direct comparison between the two studies. Nevertheless, together, these data suggest that SIRT2 may play a dual function in carcinogenesis similar to SIRT1. The deficiency of SIRT2 may induce genomic instability and drive tumorigenesis.21 However, in tumors such as HCC, where SIRT2 expression remains intact, up-regulation may play a protumorigenic role.

The correlation between SIRT2 expression and microscopic vascular invasion of HCC prompted our investigation into the role of SIRT2 in cell motility and invasive phenotypes. Concordantly, our data suggested that SIRT2 plays a role in EMT. This is supported by the fact that gene knockdown of SIRT2 in HCC cells reverses, whereas ectopic expression of SIRT2 in nontumorgenic liver cells promotes, EMT features, including a change in the expression of mesenchymal and epithelial markers. During tumorigenesis, EMT is associated with aberrant activation of canonical Wnt or the phosphoinositide 3-kinase/Akt pathway, which inactivates GSK-3β and stabilizes β-catenin and Snail, respectively.40, 41 SIRT1 has been implicated in a role in Wnt/Akt pathway by regulating the expression and activity of Dishevelled proteins42 and in glucose-induced Akt deacetylation and activation.33 However, our data suggested that in HCC cells, SIRT2, but not SIRT1, regulates Akt deacetylation and activity. As a result, it impinges on the GSK-3β/β-catenin-signaling cascade to regulate EMT and cell migration. Aberrant activation of the Wnt/β-catenin-signaling pathway is frequently observed in HCC. The latest HCC genomic analysis suggested that although β-catenin/Axin1/APC mutations are found in 50% of all HCC cases (HBV, hepatitis C virus, nonalcoholic steatohepatitis, alcohol-related, and hemochromatosis), mutation of these genes only accounts for 28% of all cases in the HBV subgroup.43 Because most of the HCC cases in our locality are associated with HBV infection, our data suggest that increased SIRT2 activity might activate β-catenin, and hence metastasis of HCC, in these patients of which the mutation rate of β-catenin/Axin1/APC is relatively lower. Earlier, we demonstrated that SIRT1 may promote HCC cell growth through its role in telomere maintenance23; our current data further demonstrate a role for SIRT2 in cell migration that may contribute to HCC metastasis. Together, these studies provide a rationale to explore whether pharmacological inhibition of SIRT1 and SIRT2 by a dual SIRT1/SIRT2 inhibitor, such as sirtinol,44 splitomicin,45 and cambinol,46 or their analogs, will be a novel strategy for targeted therapy of HCC overexpressing these sirtuins.


We would also like to thank CUHK's Academic Editor, Dr. David Wilmshurst, for commenting on a draft of this manuscript.