Identification of Guanylate-Binding Protein 1 as a Potential Oral

Identification of Guanylate-Binding Protein 1 as a Potential Oral Cancer ... If you have an individual subscription, please log in using your ACS ID t...
3 downloads 0 Views 5MB Size
ARTICLE pubs.acs.org/jpr

Identification of Guanylate-Binding Protein 1 as a Potential Oral Cancer Marker Involved in Cell Invasion Using Omics-Based Analysis Chia-Jung Yu,†,‡,§,# Kai-Ping Chang,‡,||,# Yin-Ju Chang,§ Chia-Wei Hsu,§ Ying Liang,‡ Jau-Song Yu,†,‡,§ Lang-Ming Chi,‡,z Yu-Sun Chang,‡ and Chih-Ching Wu*,‡,^ †

Department of Cell and Molecular Biology, ‡Molecular Medicine Research Center, §Graduate Institute of Biomedical Sciences, and Department of Medical Biotechnology and Laboratory Science, Chang Gung University, Taoyuan, Taiwan Department of Otolaryngology-Head Neck Surgery, zDepartment of Medical Research and Development, Chang Gung Memorial Hospital, Taoyuan, Taiwan

)

^

bS Supporting Information ABSTRACT: Oral cavity squamous cell carcinoma (OSCC) is a devastating disease that accounts for 3% of all cancer cases diagnosed annually. OSCC is usually diagnosed at advanced clinical stages, resulting in poor outcomes. To identify effective biomarkers for improved OSCC diagnosis and/or management, we simultaneously analyzed the OSCC cell secretome and tissue transcriptome. Among the 19 candidates isolated, guanylate-binding protein 1 (GBP1) was selected for further validation using serum samples from OSCC patients and healthy controls. Notably, the serum level of GBP1 was higher in OSCC patients, compared to that in healthy controls. Immunohistochemical analysis further revealed GBP1 overexpression in OSCC tissues, compared with adjacent noncancerous epithelia. Importantly, the higher GBP1 level in OSCC tissue was associated with higher overall pathological stage, positive perineural invasion, and poorer prognosis. Moreover, GBP1 modulated the migration and invasion of OSCC cells in vitro. Our results collectively indicate that integrated analysis of the cancer secretome and transcriptome is a feasible strategy for the efficient identification of novel OSCC markers. KEYWORDS: secretome, transcriptome, guanylate-binding protein 1, oral cavity squamous cell carcinoma, metastasis

’ INTRODUCTION Oral cavity squamous cell carcinoma (OSCC), the most common cancer of the head and neck, is a devastating disease that comprises 3% of all cancer cases diagnosed annually.1,2 In Taiwan, the most prevalent area of OSCC occurrence worldwide, the disease accounts for the fourth highest incidence of malignancy in men and seventh highest in women.1,3 Despite significant advances in treatment over the recent decades, 4050% of OSCC patients die within 5 years of diagnosis, mostly ascribed to metastasis and/or local recurrence.4,5 Unfortunately, the majority of patients present in advanced stages of OSCC at the time of diagnosis, which results in poor prognosis.6 Currently, none of the routinely used and clinically validated OSCC markers possess sufficient diagnostic and prognostic capabilities. Identification of carcinogenetic abnormalities in OSCC based on the presence of specific markers may therefore contribute to the development of long-term treatment strategies. Lymphatic metastasis is one of major risk factors of OSCC, leading to significant deterioration of prognosis. Only 2540% of OSCC patients with lymph node metastasis at initial diagnosis achieve 5-year survival, compared to approximately 90% of patients without metastasis.1,7 To date, no biomarkers have been included in r 2011 American Chemical Society

clinical workup strategies for the detection of nodal metastasis. Thus, it is critical to identify the molecules involved in OSCC development and improve the ability of predicting lymph node metastasis to optimize tailored therapies for individual patients.8 Proteomics-based approaches are widely employed in cancer research, and have been particularly successful in the discovery of potential cancer biomarkers.912 Identification of proteins present in the conditioned media of cultured cell lines derived from specific cancer types may represent an attractive strategy for discovering tumor biomarker candidates, as they are more likely to be detected in blood samples.1315 Previously, we identified 37 proteins released from two OSCC cell lines, OECM1 and SCC4, using MALDI-TOF mass spectrometry, and confirmed elevated serum levels of one of the proteins, Mac-2 BP, in OSCC patients.16 However, Mac-2 BP is also considered a biomarker of other cancer types, including nasopharyngeal carcinoma (NPC), colon, breast, liver, and lung cancers,1720 indicating that these proteins are abundantly secreted by various cancer cell lines and not specific for OSCC. Thus, in-depth Received: May 5, 2011 Published: June 29, 2011 3778

dx.doi.org/10.1021/pr2004133 | J. Proteome Res. 2011, 10, 3778–3788

Journal of Proteome Research analysis of OSCC secretomes using advanced protein separation and identification technologies is necessary to facilitate the discovery of novel OSCC biomarkers. In the present study, the secretomes of the two OSCC cell lines were further profiled using one-dimensional gel electrophoresis in combination with the nano liquid chromatographytandem mass spectrometry (GeLCMS/MS) approach.21 Among the secreted proteins, guanylate-binding protein 1 (GBP1) was selected for further evaluation as an OSCC biomarker candidate, based on the finding that, in the array-based data sets from the public domain, expression of this gene was also significantly elevated in OSCC tissues. Subsequent proof-of-concept experiments confirmed higher serum levels of GBP1 in pretreated OSCC patients, compared to those in healthy individuals. Immunohistochemical analyses further showed that GBP1 is overexpressed in OSCC cells, compared to adjacent noncancerous epithelia. The higher GBP1 level in cancer tissue was significantly associated with poorer prognosis of OSCC patients. Moreover, GBP1 modulated the migration and invasion of OSCC cells in vitro.

’ MATERIALS AND METHODS Cell Culture and Cell Viability

The OSCC cell lines, OEC-M1 and SCC4, were grown in RPMI-1640 supplemented with 10% fetal bovine serum (FBS), 25 mM HEPES, and antibiotics in 5% CO2 at 37 C, as described previously.16 OEC-M1 is an oral epidermal carcinoma cell line derived from the gingiva of a Chinese patient,22 whereas SCC4 is a tongue squamous cell carcinoma cell line derived from a 55year-old male (ATCC No. CRL-1624). Viability of cancer cell lines cultured in complete or serum-free media were determined as described in Supplemental Materials and Methods. Harvest of Conditioned Media from Cancer Cell Lines

Conditioned media from OSCC cell lines were collected and processed as described in Supplemental Materials and Methods. One-Dimensional SDS-PAGE and In-Gel Digestion of Proteins

Proteins (50 μg) were resolved on 10% SDS-PAGE and stained with 0.5% Coomassie Brilliant Blue G-250 (AppliChem GmbH, Darmstadt, Germany). The whole gel lane was cut into 40 pieces and subjected to in-gel tryptic digestion essentially as described in Supplemental Materials and Methods.

Reverse-Phase Liquid ChromatographyTandem Mass Spectrometry

Each peptide mixture was reconstituted in HPLC buffer A (0.1% formic acid; Sigma, St. Louis, MO), loaded onto a trap column (Zorbax 300SB-C18, 0.3  5 mm, Agilent Technologies, Wilmington, DE) at a flow rate of 0.2 μL/min in HPLC buffer A, and separated on a resolving analytical C18-packed PicoFrit column, 10 cm in length (with an inner diameter of 75 and 15 μm tip, New Objective, Woburn, MA). Peptides were eluted with a linear gradient of 010% HPLC buffer B (99.9% acetonitrile containing 0.1% formic acid) for 3 min, 1030% buffer B for 35 min, 3035% buffer B for 4 min, 3550% buffer B for 1 min, 5095% buffer B for 1 min, and 95% buffer B for 8 min at a flow rate of 0.25 μL/min across the analytical column. The LC setup was coupled online to LTQ-orbitrap discovery (Thermo Fisher, San Jose, CA) operated using Xcalibur 2.0

ARTICLE

software (Thermo Fisher). Intact peptides were detected in Orbitrap at a resolution of 30 000. Internal calibration was performed using the ion signal of (Si(CH3)2O)6H+ at m/z 445.120025 as a lock mass.23 For the MS survey scan, a datadependent procedure that alternated between one MS scan followed by six MS/MS scans for the six most abundant precursor ions was applied. The m/z values selected for MS/ MS were dynamically excluded for 180 s. The electrospray voltage applied was 1.8 kV. Both MS and MS/MS spectra were acquired using one microscan, with a maximum fill time of 1000 and 100 ms for MS and MS/MS analyses, respectively. Automatic gain control was used to prevent overfilling of the ion trap, and 5  104 ions were accumulated in the ion trap for generation of MS/MS spectra. For MS scans, the m/z scan range was 3502000 Da. Database Searching and Bioinformatics

All MS/MS samples were analyzed using the Mascot search algorithm (Version 2.1, Matrix Science, London, U.K.). Mascot was set up to search the Swiss-Prot database (released Jun 15, 2010, selected for Homo sapiens, 20 367 entries), assuming trypsin as the digestion enzyme. Mascot was searched with a fragment ion mass tolerance of 0.50 Da and parent ion tolerance of 10 ppm. Search parameters included differential amino acid mass shifts for oxidized methionine (+15.99 Da) and fixed modification for carbamidomethyl cysteine (+57 Da). We measured the false-positive rate of protein identification by searching a random database in which sequence entry from the “normal” database was randomly shuffled. The number of hits from each search was categorized based on the score. After the Mascot database search, the resulting DAT files of the MS/MS samples for each cell line were then integrated through Scaffold software (Version 3.00.07, Proteome Software, Inc., Portland, OR) to generate a nonredundant list of identified proteins. The Scaffold software includes a peptide probability score program, PeptideProphet, that aids in the assignment of peptide MS spectra24 and a ProteinProphet program that assigns and groups peptides to a unique protein or protein family in cases where peptides are shared among several isoforms.25 ProteinProphet allows the filtering of largescale data sets with assessment of predictable sensitivity and false positive identification error rates. In our analysis, peptide identifications were accepted at greater than 95.0% probability, while protein identifications were accepted if they contained at least 3 known peptides and at greater than 95.0% probability. To predict the secretion pathways of identified proteins, we employed SignalP with Hidden Markov models for estimating the presence of secretory signal peptide sequences,26 SecretomeP for deducing nonsignal peptide-triggered protein secretion,27 and the transmembrane Hidden Markov model (TMHMM) for predicting transmembrane helices in proteins.28 Patient Populations and Clinical Specimens

Tumor specimens for immunohistochemical analysis were obtained from 87 OSCC patients diagnosed at the Chang Gung Memorial Hospital (Taoyuan, Taiwan) from 2004 to 2006. The demographic data for these patients and controls are shown in Table 3. Serum samples were collected from 37 healthy controls (32 men and 5 women ranging from 35 to 85 years of age, mean age: 57.6 ( 13.3 y) and 38 OSCC patients (33 men and 5 women ranging from 39 to 89 years of age, mean age: 59.2 ( 10.9 y) at the Chang Gung Memorial Hospital. All OSCC patients were biopsy-proven and underwent routine check-ups according to 3779

dx.doi.org/10.1021/pr2004133 |J. Proteome Res. 2011, 10, 3778–3788

Journal of Proteome Research

ARTICLE

the standard protocol. The control subjects were volunteers subjected to routine health examinations or individuals presenting with otolaryngology-related, non-neoplastic diseases. Patients and controls with a history of malignant disease were excluded from the study. This research followed the tenets of the Declaration of Helsinki, and all subjects signed an informed consent form approved by the Institutional Review Board of Chang Gung Memorial Hospital before participation and permitting the use of tissue or blood samples collected before treatment. Immunohistochemistry and the Scoring System

Immunohistochemical (IHC) staining was performed with a rat anti-GBP1 monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). GBP1 expression was scored using a combined method accounting for both staining intensity and percentage of stained cells as described in Supplemental Materials and Methods. Western Blot Analysis

Proteins were separated using SDS-PAGE, transferred onto polyvinylidene difluoride (PVDF) membranes, and probed with a rat monoclonal antibody against GBP1 (1:500 dilution, Santa Cruz Biotechnology) for 16 h at 4 C. Proteins of interest were detected with horseradish peroxidase-conjugated goat anti-rat IgG antibodies (Santa Cruz Biotechnology) and visualized with the ECL system (Millipore, Billerica, MA), according to a previously published protocol.29 ELISA

An ELISA developed in house was used to measure serum level of GBP1 as described in Supplemental Materials and Methods. Cloning and Expression of GBP1

GBP1 DNA was amplified from OEC-M1 cells. GBP1 expression in Escherichia coli and mammalian cell was performed as described in Supplemental Materials and Methods. Gene Knockdown of GBP1 with Small Interfering RNA

For knockdown, 25-nucleotide long RNA duplexes targeting human GBP1 were synthesized and annealed by Invitrogen (Grand Island, NY). OEC-M1 cells were transfected with control siRNA or GBP1-specific siRNA (UUCCUUUAGUGUGAGACUGCACCGU) using Lipofectamine RNAiMAX reagents (Invitrogen), according to the protocol provided by the manufacturer. At 48 h after transfection, cell lysates were prepared for Western blotting to determine gene knockdown efficacy. Cell Proliferation Analysis

Proliferation of cells transfected with siRNA or GBP1 constructs was measured using the 3-(4,5-dimethylthiazol-2-yl)2,5-diphenyltetraz-lium bromide (MTT) colorimetric growth assay (Invitrogen) as described in Supplemental Materials and Methods.30 Cell Migration and Invasion Assays

Trans-well migration and invasion assays were performed using cells transfected with siRNA or GBP1 constructs as described in Supplemental Materials and Methods. Statistical Analysis

For comparing the IHC scores between the paired tumor and pericancerous normal mucosa samples, the Wilcoxon signedrank test was employed. Specifically, serum levels, cell growth,

Figure 1. Identification of potential OSCC markers via combined analysis of the cancer cell secretome and tissue transcriptome. (A) The strategy consists of OSCC cell secretome profiling in conjunction with OSCC tissue transcriptome analysis, followed by validation in clinical specimens. (B) Conditioned media (CM) and cell extracts (CE) of OSCC cells were collected and processed as described in Materials and Methods. Proteins (50 μg) were resolved on 10% SDS gels and stained with Coomassie blue. Protein bands in the CM lane were excised for further analysis. (C) Proteins (50 μg) from CM and CE of the cell lines were subjected to Western blot analysis using antibodies against βtubulin and R-actin. (D) The viabilities of OEC-M1 and SCC4 cells grown for 24 h in complete and serum-free media were determined.

and migration rates were compared using the Wilcoxon test. Receiver operator characteristic (ROC) curves were constructed by plotting sensitivity versus (1  specificity), and the areas under the curves (AUC) analyzed with the Hanley and McNeil method.31 Survival analysis was plotted with the KaplanMeier method and analyzed using the Log-rank test. P-values