ARTICLE pubs.acs.org/jpr
Search for the Tumor-Associated Proteins of Oral Squamous Cell Carcinoma Collected in Taiwan using Proteomics Strategy Kuo-An Liao,† Yeou-Guang Tsay,‡ Li-Chien Huang,§ Hsuan-Ying Huang,|| Chien-Feng Li,*,^,# and Ting-Feng Wu*,§ †
Department of Oral and Maxillofacial Surgery, Chi-Mei Medical Center, Tainan, 710, Taiwan Institute of Biochemistry and Molecular Biology, National Yang-Ming University, Taipei, 112, Taiwan Department of Pathology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, 833, Taiwan ^ Department of Pathology, Chi-Mei Medical Center, Tainan, 710, Taiwan # Institute of Biomedical Science, National Sun Yat-Sen University, Kaohsiung, 804, Taiwan § Department of Biotechnology, Southern Taiwan University, Tainan,710, Taiwan
)
‡
bS Supporting Information ABSTRACT: Squamous cell carcinoma (SCC) accounts for more than 90% of malignant tumors of the oral cavity. In Taiwan, oral squamous cell carcinoma (OSCC) is among the most frequent malignancies, largely due to betal quid chewing. Despite the recent improvement in treatment results, the longterm outcome of OSCC generally remains poor, especially for those with advanced diseases. It is therefore desirable to identify potential biomarkers that may aid in risk stratification and perhaps the development of therapeutic targets. In this study, we exploited two-dimensional gel electrophoresis (2-DE) coupled with mass spectrometry to compare the proteome maps of 10 OSCC specimens with their adjacent nontumorous epithelia to identify differentially expressed proteins. Comparative proteomics indicated that 17 proteins were differentially expressed in OSCC with 11 up-regulated and 6 down-regulated proteins. These deregulated proteins participated in cytoskeletal functions, cell signaling, antiapoptosis, angiogenesis, lipid metabolism, drug metabolism, and protein translation/turnover. They were all associated with tumor development in various cancers. Among the dys-regulated proteins, the immunoexpression of three proteins including nicotinamide N-methyltransferase, apolipoprotein AI, and 14-3-3 zeta were evaluated in 38 OSCCs of testing cohort to confirm the proteomics data. Subsequently, the expression of 14-3-3 zeta, as the most relevant to OSCC progression determined by testing cohort, was further assessed in 80 OSCCs of independent validation cohort to identify the clinical relevance of its expression. By this comprehensive study, we identified 14-3-3 zeta as the only prognosticator of local recurrence-free survival (LRFS) and also an independently predicted factor of disease-specific survival (DSS). KEYWORDS: proteomics, 2-DE, 14-3-3 zeta, oral squamous cell carcinoma, NNMT, oral cancer, LCMS/MS
’ INTRODUCTION Squamous cell carcinoma (SCC) accounts for more than 90% of malignant tumors of the oral cavity.1 In Taiwan, oral squamous cell carcinoma (OSCC) is among the most frequent malignancies, largely due to betal quid chewing.2,3 Despite the recent improvement in treatment results, the long-term outcome of OSCC generally remains poor, especially for those with advanced diseases.4 The current effective therapy to cure OSCC is complete surgical resection at the early stage,5 while early diagnosis is seldom achieved, given that many patients initially present with locally advanced or metastatic diseases.6 Moreover, even after extensive surgery, a considerable number of patients develop local recurrences or distal metastases that are resistant to conventional r 2011 American Chemical Society
chemotherapy. It is therefore desirable to identify potential biomarkers that may aid in risk stratification and perhaps the development of therapeutic targets. Several proteins related to tumor progression, metastasis and differentiation have been evaluated for the prognosis of OSCC.710 Recent investigations using gene expression profiling, such as DNA microarray and proteomics have provided more extensive gene expression profile of OSCC. Extensive gene profiling studies demonstrated that various genes related to cell adhesion, motility, invasion, cell cycle, metabolism and angiogenesis are dys-regulated in Received: November 16, 2010 Published: February 16, 2011 2347
dx.doi.org/10.1021/pr101146w | J. Proteome Res. 2011, 10, 2347–2358
Journal of Proteome Research OSCC.1116 However, few if any of these studies comprehensively validate the deregulated candidate genes at the protein level. Proteomics has been recently recognized as a powerful tool to search for the potential diagnostic and prognosis biomarkers for various cancers and to study the carcinogenesis, progression and metastasis.17,18 Recent gel-based comparative proteomics investigations involved with OSCC have implicated that RACK1,19 related to OSCC invasiveness as well as metastasis and Lin-7C20 whose down-regulation is associated with OSCC metastasis may be two potential biomarkers for early detection, prognosis and monitoring the therapy of OSCC. In this study, we exploited two-dimensional gel electrophoresis (2-DE) coupled with mass spectrometry to compare the proteome maps of 10 OSCC specimens with their adjacent nontumorous epithelium to identify differentially expressed proteins. Comparative proteomics indicated that 17 proteins were differentially expressed in OSCC. Of these, the immunoexpression of three proteins including nicotinamide N-methyltransferase, apolipoprotein AI, and 14-3-3 zeta were evaluated in 38 OSCCs of testing cohort to confirm the proteomics data. Subsequently, the expression of 14-3-3 zeta, as the most relevant to OSCC progression determined by testing cohort, was further assessed in 80 OSCCs of independent validation cohort to identify the clinical relevance of its expression. By this comprehensive study, we identified 14-3-3 zeta as the only prognosticator of local recurrence-free survival (LRFS) and also an independently predicted factor of disease-specific survival (DSS).
’ MATERIALS AND METHODS Patients, Tumor Samples, and Cell Lines
The institutional review board had approved this study (CLH0063). We first utilized proteomics to profile aberrant protein expression of 10 fresh stage IV OSCC specimens (screening set, Supplementary Table S1, Supporting Information), including 4 well differentiated, 4 moderately differentiated, and 2 poorly differentiated tumors. To identify the protein expression most relevant to tumor progression in OSCC, protein expression of NNMT1, APOA1 and 14-3-3 zeta was assessed by using immunohistochemistry (IHC) in the testing cohort (Supplementary Table S2, Supporting Information), comprising 38 cases of surgically resected OSCC. To independently validate the prognostic impact of 14-3-3 zeta, IHC was further performed on an independent validation cohort comprising 80 OSCCs received wide tumor excision with curative intent (Table 3). To precisely obtain adjacent nontumorous epithelia for proteomic analyses, surgically removed cut margins were submitted to the pathologists to confirm there is neither tumor nor remarkable dysplasia. In the Immunohistochemical study, safe margins without morphological abnormality were enrolled for comparison. Five OSCC cell lines including HSC3, UMSCC1, SCC25, Tca83, and KB were cultured as previously reported2123 and their total protein lysates were collected for the evaluation of in vitro expression of 14-3-3 zeta. Preparation of OSCC Protein Lysates
Fresh OSCC specimens were ground under liquid nitrogen in mortar and then the powder was dissolved in lysis buffer (700 μL lysis buffer per 40 mg OSCC) [7 M urea, 2 M thiourea, 100 mM dithiothreitol (DTT), 4% (v/v) CHPAS, 40 mM tris-base (pH 10), 1 mM PMSF, and 1 Complete Mini protease inhibitor cocktail tablet (Roche, Diagnostics, Indianapolis, USA) per liter] with shaking at room temperature for 1 h. Then the lysate was
ARTICLE
centrifuged at 349 000 g for 2 h at 15 C in Type 90 Ti rotor (Beckman Coulter, Fullerton, CA) and the supernatant was precipitated with 2-D cleanup kit (Amersham Bioscience Corp. Piscataway, NJ) according to the manufacture’s suggestion. The resulting protein lysate was measured by Bio-Rad DC protein assay. Isoelectric Focusing (IEF)
The pH 47, 18-cm immobibline dry strips (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) were rehydrated for 16 h at 20 C with 300 μL rehydration buffer [7 M urea, 2 M thiourea, 4% (v/v) CHAPS, 2% (w/v) DTT, 0.5% (v/v) IPG buffer and trace of bromophenol blue]. After rehydration, 100 μg protein lysates prepared from each of OSCC samples and their adjacent nontumorous tissues were cup-loaded onto the rehydrated gel strips with Ettan IPGphor Cup Loading Manifold (AmershamPharmacia Biotech Inc., Piscataway, NJ). The proteins were then focused at 20 C at 50, 100, 200, 500, 1000, 5000 and 8000 V, respectively, with a total of 81 434 voltage-hours. SDS-Polyacrylamide Gel Electrophoresis (SDS-PAGE)
After isoelectric focusing, the gel strips were equilibrated in equilibration buffer [6 M urea, 30% (v/v) glycerol, 2% (w/v) SDS] containing 2% (w/v) DTT for 15 min and then in equilibration buffer containing 5% (w/v) iodoacetamide for a further 15 min. The equilibrated gel was loaded onto the top of a 12.5% (w/v) polyacrylamide gel and sealed with 0.5% (w/v) agarose and the proteins were separated at 420 V using BioRad Protean IIxi until bromophenol blue reaching the bottom of the gel. Silver Staining
The silver staining was performed according to the protocol described in the 2D plus one silver staining kit (AmershamPharmacia Biotech Inc., Piscataway, NJ) with some modifications as described elsewhere.24 Briefly, the gel was fixed in fixation solution (ethanol/water/acetic acid, 4/5/1, v/v/v) after electrophoresis and treated with sensitizing solutions [0.5 M sodium acetate, 0.5% (w/v) sodium thiosulphate] for 30 min. After sensitization, the gels were washed and incubated in 0.25% (w/v) silver nitrate solution for 20 min and then developed by incubating with the developing solution [2.5% (w/v) sodium carbonate and 0.015% (v/v) formaldehyde] until the protein spots appeared. Glutardialdehyde was not used in the silver staining due to the protein identification by mass spectrometry. Image Analysis
The images of 2-DE gel subjected to silver staining were captured using BioRad GS800 Densitometer. To search for proteins showing disparity in OSCC, the proteomes of OSCC and their corresponding adjacent nontumoruos tissues were analyzed by PDQuest 8.0.1 (BioRad) software. Two replica gel pairs were collected from each clinical sample except that only one gel pair was acquired from each of OSCC#3 and OSCC#8 because the protein amount was not enough for the replica analysis. A total of 18 pairs of well-focused gels from 10 OSCC and 10 nontumorous epithelium tissues were compared. Differentially expressed spots identified by computer analysis were further confirmed by visualization. The intensity of the spot was computed and normalized as a percentage of the total intensity of all spots in a gel and analyzed with Student’s t-test (QI Macros SPC software 2010, KnowWare International Inc., Denver, CO). For each deregulated protein spot volumes of individual protein spots across replica gels of nontumorous or OSCC tissues were analyzed by the normal distribution test and then Student’s t-test was performed when the normal distribution was obtained. 2348
dx.doi.org/10.1021/pr101146w |J. Proteome Res. 2011, 10, 2347–2358
Journal of Proteome Research
ARTICLE
However, when the normal distribution was not observed, log transformation was carried out followed by the normal distribution test and Student’s t-test. In all cases, statistical variance of the OSCC:normal spot intensity ratio within 95% (Student’s t-Test; p < 0.05) was considered to be significantly different. Furthermore the differentially expressed proteins present at least in 9 out of 18 gel pairs were regarded as OSCC-associated proteins. In-Gel Digestion
The in-gel digestion and mass spectrometric analysis were performed as described previously with some modifications.24 The spots excised from the gel were incubated in a solution containing 15 mM of potassium ferricynate and 50 mM sodium thiosulfate until the brownish color disappeared. The destained gel piece was washed with 25 mM ammonium bicarbonate for 10 min and then with 25 mM ammonium bicarbonate/50% acetonitrile for 10 min. After drying in SpeedVac (ThermoSavant, Milford, MA), the gel was incubated with 50 μL of 2% (v/v) 2-mercaptoethanol in darkness for 20 min and an equal volume of 10% (v/v) vinylpyridine in 25 mM ammonium bicarbonate/ 50% acetonitrile was added and incubated further for 20 min. Then the gel was washed three times with 25 mM ammonium bicarbonate and dehydrated in 25 mM ammonium bicarbonate/ 50% acetonitrile. The gel was dried and treated with 50 ng of modified trypsin (Amersham-Pharmacia Biotech Inc., Piscataway, NJ) in 100 μL of 25 mM ammonium bicarbonate at 37 C overnight. The supernatant was collected after digestion and the gel was extracted with 200 μL of 0.1% formic acid. The extracts were combined and dried in SpeedVac and resuspended in 0.1% (v/v) formic acid immediately for mass spectrometric analysis or stored at 20 C until use. Protein identification analysis via liquid chromatographytandem mass spectrometry (LCMS/MS)
The protein digest was analyzed in LTQ-Orbitrap hybrid tandem mass spectrometer (ThermoFisher, San Jose, CA) inline coupled with Agilent 1200 nanoflow HPLC system equipped with LC Packing C18 PepMap 100 (length: 5 mm; internal diameter: 300 μm; bead size: 5 μm) as the trap column and Agilent ZORBAX XDB-C18 (length: 50 mm; internal diameter: 75 μm; bead size: 3.5 μm) as the separating column. File Converter in Xcalibur 2.0SR package (ThermoFisher, San Jose, CA) and an in-house program were used to extract the MS/MS information as well as to compute the charge and mass for each analyzed peptide. TurboSequest program (ver. 27, rev. 11) was then used to search the best matched peptides from a nonredundant protein database whose FASTA sequences were downloaded from National Center for Biotechnology Information (ftp://ftp.ncifcrf.gov/pub/nonredun/) on 2008/05/05 with 440 000 entries. While only the tryptic peptides with e2 missed cleavages were considered, the mass ranges during the database search were 1 and 3.5 m/z for fragment and precursor ions respectively. The protein identities were verified only when there were at least two peptides matched and both search results had high Xcore (i.e., g 2.0 for doubly charged peptides and g3.0 for triply charged ones) and with minimal differences between observed and hypothetical masses (i.e., ΔM
