Regulation of Macrophage Inhibitory Factor (MIF) by Epidermal Growth Factor Receptor (EGFR) in the MCF10AT Model of Breast Cancer Progression Simin Lim,†,‡ Lee-Yee Choong,†,‡ Chong Poh Kuan,† Chen Yunhao,† and Yoon-Pin Lim*,†,§,| Cancer Science Institute of Singapore, National University of Singapore, Department of Biological Sciences, Faculty of Science, National University of Singapore, and Bioinformatics Institute, Agency for Science, Technology and Research, Singapore Received May 14, 2009
Genetic aberration of EGFR is one of the major molecular characteristics of breast cancer. However, the molecular changes associated with EGFR signaling during different stages of breast cancer development have not been studied. In this study, complementary two-dimensional-DIGE and iTRAQ technologies were used to profile the expression level of proteins in 4 isogenic cell lines in the MCF10AT model of breast cancer progression following a time course of EGF stimulation. A total of 80 proteins (67 from iTRAQ, 15 from DIGE, 2 common in both) were identified to be up- or down-regulated by EGF treatment. Following EGF stimulation, the expression level of MIF, a cytokine that has been implicated in many human cancers, was decreased in MCF10A1 normal breast mammary epithelial cells, increased in MCF10AT1k preneoplastic and MCF10CA1h low grade breast cancer cells, but showed no obvious difference in the MCF10CA1a high grade cancer cells. The increase in MIF expression level following EGF treatment could also be observed in A431 cervical cancer cells. EGF-induced increases of MIF expression levels in CA1h breast cancer cells were abrogated when MEK, but not PIK3CA, was knocked down. In addition, silencing of MIF diminished the proliferation of EGF-stimulated CA1h cells when compared to control cells. Taken together, our data suggested an EGFR f MEK f MIF proliferative pathway that has never been reported previously and that this pathway “evolves” during disease progression as modeled by the MCF10AT system. Revelation of the novel relationship between MIF and EGF may contribute to an integrated understanding of the roles of these oncogenic factors during breast cancer development. Keywords: EGF • MIF • breast cancer • proteomics • 2D-DIGE, iTRAQ
1. Introduction The epidermal growth factor receptor (EGFR) expression and its downstream signaling pathways are important components in regulating proliferation and behavior in many epithelial tumors, including lung and breast cancer, which is the second leading cause of cancer deaths in women in the United States.1 Despite the relatively large collection of knowledge on EGFR signaling defects in breast cancer,2 there have been no comprehensive studies investigating how the EGFR signaling network evolves during breast cancer progression. Targetdirected drugs like Gefitinib that selectively inhibit EGFR are now used in clinical therapy for non-small cell lung cancer (NSCLC) but displayed limited efficacy only in patients with EGFR activation mutations, which are rare in sporadic breast * To whom correspondence should be addressed. Yoon-Pin Lim, Cancer Science Institute of Singapore, National University of Singapore, Centre for Life Sciences, 28 Medical Drive, #02-14C, Singapore 117456. Tel: (65) 65161313. Fax: (65) 68739664. E-mail:
[email protected]. † Cancer Science Institute of Singapore, National University of Singapore. ‡ These authors contributed equally. § Faculty of Science, National University of Singapore. | Agency for Science, Technology and Research.
4062 Journal of Proteome Research 2009, 8, 4062–4076 Published on Web 06/16/2009
cancers.3-6 Clinical trials on the use of other EGFR inhibitors in breast cancer have also not been promising and the reasons are poorly understood.7 One main recurring obstacle in targetdirected therapies is the poor understanding of the molecular etiology of disease. Hence, elucidating the molecular changes associated with EGFR signaling during breast cancer progression will be of strategic significance in EGFR-directed therapeutics. Experimental cancer models are invaluable to studying the molecular events during disease development due to the intrinsic difficulties (e.g., diverse genetic and environmental backgrounds in patients) in studying disease progression in the clinical setting. For example, in vitro cell line variants have been used to study the molecular basis of osteolytic bone metastasis.8 A useful breast cancer model is the MCF10AT series of isogenic, xenograft-derived cell lines developed in Fred Miller’s laboratory.9,10 The MCF10AT model has several salient features of proliferative breast disease in human and it comprises isogenic cell lines that mimic the different stages of breast cancer progression.11,12 The cell lines are MCF10A1, MCF10AT1K.cl2, MCF10CA1h, and MCF10CA1a.cl1 that represent normal, pre10.1021/pr900430n CCC: $40.75
2009 American Chemical Society
MIF is a Novel Target of EGFR malignant epithelium, low grade, and high grade lesions, respectively.9,10 MCF10A1 cells were not tumorigenic in nude mice whereas MCF10AT1K.cl2 cells could form simple ducts that progress into benign hyperplasia and occasionally carcinoma. MCF10CA1h formed largely well-differentiated carcinoma whereas MCF10CA1a.Cl1 produced poorly differentiated carcinoma and could metastasize to the lung in tail vein injection assay. Other molecular characteristics of the MCF10AT model including its karotype and status of molecular determinants such as p53, ER, etc. have been reviewed elsewhere.12 The cell lines in the MCF10AT model have been studied in terms of cytogenetics, apoptosis, TGF-β signaling, and basal protein expressions.13-19 We have also previously conducted a phosphoproteomic study with the MCF10AT model and found aberrant expressions of several novel phosphoproteins.20,21 One of our findings was the diminution of EGFR expression during breast cancer progression in vitro (MCF10AT model) and ex vivo (clinical samples),21 an observation also made by others following immunohistochemistry of EGFR in clinical samples.22,23 Having established aberrant EGFR expression in breast cancer, we aimed to examine how the status of EGFR expression in cancer cells at different stages of breast cancer development affects their response to EGF stimulation. To this end, we profiled the EGF-induced proteomics changes between the different cell lines in the MCF10AT model using complementary two-dimensional differential gel electrophoresis (2DDIGE) and isotope tags for relative and absolute quantification (iTRAQ). Our goal was to identify potentially novel targets of EGF/EGFR signaling and select one that exhibited different responses between various MCF10AT cell lines following EGF treatment. Further studies would establish the role of the candidate protein in EGF signaling and breast cancer biology.
2. Experimental Procedures 2.1. Reagents. Horse radish peroxidase (HRP)-conjugated antimouse IgG and antirabbit IgG secondary antibodies were purchased from Sigma-Aldrich, St. Louis, MO while the antigoat IgG-HRP secondary antibody was from Santa Cruz Biotechnology. Mouse monoclonal anti-PIK3CA and anti-Ras antibodies (used in Supplementary Figures, Supporting Information) and PY20H antiphosphotyrosine antibodies were from BD Biosciences (San Jose, CA). Goat antimacrophage migration inhibitory factor (MIF) polyclonal antibody was from Novus Biologicals (Littleton, CO); Anti-c-Met (used in Supplementary Figures, Supporting Information) and anti-actin horseradish peroxidase conjugated antibodies and mouse monoclonal antiMEK1 antibodies were from Santa Cruz Biotechnology; Antiphospho-p44/42 MAPK (T202/Y204) rabbit polyclonal antibodies were from Cell Signaling Technology, Inc. (Danvers, MA). Prestained molecular weight protein markers and PolyVinylidene DiFluoride membranes were from Bio-Rad; protease inhibitors cocktail was from Roche (Mannheim, Germany). 2.2. Cell Lines and EGF Treatment. Xenograft-derived breast cancer cell lines (MCF10A1, MCF10AT1KCl.2, MCF10CA1h and MCF10CA1aCl.1) were obtained from Dr. Fred Miller at the Barbara Ann Karmanos Cancer Institute (Detroit, MI) and were maintained in DMEM/F12 (1:1) supplemented with 5% horse serum, 100 U/mL penicillin, 292 mg/mL streptomycin, 20 ng/mL epidermal growth factor (Upstate), 10 µg/mL insulin (Sigma-aldrich), 100 ng/mL cholera toxin (Calbiochem) and 0.5 µg/mL hydrocortisone (Sigma-Aldrich) as previously described elsewhere.18 Cells were incubated at 37 °C with a 5% CO2 humidified atmosphere and were starved with serum- and
research articles additives-free media overnight prior to stimulation with 50 ng/ mL of EGF (Sigma-Aldrich) for either 4 or 16 h. 2.3. Protein Extraction and Condition Media Collection. Cells were rinsed with ice-cold PBS and lysed on ice for protein extraction with two different lysis buffers for (i) 2D-DIGE (8 M urea, 4% w/v CHAPS, 30 mM pH 9.0 Tris-Cl, protease inhibitor and 1 mM sodium orthovanadate) and (ii) iTRAQ (0.2% IGEPAL, 0.2% Triton X, 0.2% w/v CHAPS, 75 mM NaCl, 1 mM EDTA, 50 mM sodium fluoride, protease inhibitor, and 1 mM sodium orthovanadate). Protein lysates were then clarified by centrifugation at 4 °C at 14 000 rpm for 10 min. The total protein was determined using the 2-D Quant Kit (GE Healthcare) for 2DDIGE samples and a bicinchoninic acid assay (BCA) kit (Pierce Biotechnology, Rockford, IL) for iTRAQ samples. For condition media analysis, cells were grown in phenol red-free media to prevent interference with protein estimation. Cell debris was first removed using the MilexGP Filter Unit 0.22 µm (Millipore). The condition media was then concentrated using the AmiconUltra centrifugal filter (Millipore) at 5000RCF, 4 °C in a refrigerated centrifuge (Thermo Electron Corporation) for 1 h and 30 min. The filtrate was removed and concentration was carried out for another 20 min. Protein quantification of the concentrated condition media (retentate) was performed with the BCA assay. Typically 2% of total proteins in lysates or conditioned media were used for Western blot analysis. 2.4. Isobaric Peptide Labeling and nanoLC-MS/MS Analysis. A total of 100 µg of protein from each sample was reduced, alkylated, digested and labeled with iTRAQ reagents according to the manufacturer’s protocol (Applied Biosystems, Framingham, MA). The dried, labeled peptides were then constituted with 14 mL Buffer A, prior to strong cation-exchange fractionation using a PolySULFOETHYL A Column (PolyLC, Columbia, MD) 5 µm of 200 mm length × 2.1 mm ID, 200 Å pore size, on a Agilent 1100 quaternary HPLC unit (Agilent Technologies, Santa Clara, CA), at a flow-rate of 300 µL/min. Buffer A consisted of 5 mM KH2PO4 and 25% acetonitrile, pH 2.7 and Buffer B consisted of 5 mM KH2PO4, 25% acetonitrile and 350 mM KCl, pH 2.7. The 120 min gradient consisted of 100% A for 30 min, 10% to 45% B for 70 min, 45% to 100% B for 5 min, 100% B for 10 min and finally 100% A for 5 min. A total of 23 fractions were collected and these fractions were dried in a vacuum concentrator and stored at -20 °C prior to mass spectrometric analysis. Mass spectrometry was performed using a QStar XL Hybrid ESI Quadrupole time-of-flight tandem mass spectrometer, ESIqQ-TOF-MS/MS (Applied Biosystems, Framingham, MA; MDSSciex, Concord, Ontario, Canada) coupled with an online nanoflow liquid chromatograph (Agilent 1100 system from Agilent, Santa Clara). The SCX fractions were resuspended in 100 µL of loading buffer (0.1% triflouroacetic acid and 2% acetonitrile) and 39 µL of the peptide mixture was loaded onto a reverse phase Peptide Captrap (Michrom Bioresources) for desalting at 10 µL/min for 13 min. Subsequently the desalted peptides were subjected to online separation on ProteCol column (SGE Analytical Science, Victoria, Australia), a C18 capillary column 3 µm of 150 µm ID × 10 cm length, 300 Å. The buffers used in the gradient were Buffer A, consisting of 0.1% formic acid and Buffer B, consisting of 0.1% formic acid in 90% acetonitrile. The nano-LC gradient was 170 min in length comprising 5 min of 5% Buffer B, a linear gradient step ramping from 5 to 90% Buffer B in 120 min, 100% Buffer B for 15 min and finally 100% Buffer A for 30 min. The mass spectrometry was set to perform data acquisition in the positive Journal of Proteome Research • Vol. 8, No. 8, 2009 4063
research articles
Lim et al.
ion mode; with a selected mass range of 370-1600 m/z. Multiply charged ions were selected for tandem mass spectrometry, with three most abundant charged ions above 50 counts. The time of summation of MS/MS events was set to 2 s. Protein identification and quantification for iTRAQ samples were carried out using ProteinPilot software (version 2.0; Applied Biosystems, MDS-Sciex). The search was performed against the human genome (16 602 human sequences), Uniprot Swiss-Prot database release 11 (with a total entries of 269,293), downloaded from Uniprot (May 2007, http://au.expasy.org/).24 Database search was performed by setting cysteine modification by MMTS as a fixed modification and also considering a set of 94 possible modifications through the Paragon Algorithm.25 Consideration of other parameters including miscleavages, MS and MS/MS tolerances are built-in features of the Paragon Algorithm. Only proteins identified with at least 95% confidence, that is, p e 0.05, were taken into account. The results were then exported into Excel for manual data interpretation. Statistical calculation for iTRAQ-based detection and relative quantification were calculated via the Paragon Algorithm. To estimate the false positive rate (FDR) in the data set obtained, we employed a database search strategy against a concatenated pseudoreverse database.26 This database was created in-house, consisting of 16 602 human sequences and their pseudo reverse sequences. Here we defined FDR as the percentage of decoy proteins identified against the total protein identification. 2.5. Two-Dimensional DIGE. There were 12 biological conditions in this study. Biological duplicates were used for 2D-DIGE analyses, and this gave rise to 24 test samples. An internal standard consisting of 100 µg of proteins in total was created by combining equal amounts of proteins (4.17 µg) from each of the 24 test samples. Each analytical gel was loaded with 2 different test samples plus the internal standard each with 100 µg of protein labeled with 400 Fmol of either Cy2 (for internal standard), Cy3 (for test sample) or Cy5 (for test sample) (GE Healthcare). A dye-swapping scheme was used to ensure that the two duplicates from any condition were not all labeled with the same dye to prevent the occurrence of any dye-specific labeling artifacts (see Figure 1C). Details of the concept of internal standard and labeling protocol are described in the Ettan DIGE User Manual (18-1173-17 Edition AA, GE Healthcare). Prior to iso-electric focusing (IEF), pH4-7, 18 cm immobilized pH gradient (IPG) strips (GE Healthcare) were passively rehydrated overnight with the labeled samples and rehydration buffer containing 8 M urea, 2% w/v CHAPS, 20 mM dithiothreitol (DTT), 0.5% IPG buffer and trace amount of bromophenol blue. IEF was performed using the Ettan IPGphor II (GE Healthcare) with the following focusing parameters: (i) 150 V, 75 Vhr; (ii) 300 V, 150Vhr; (iii) 500 V, 500 Vhr; (iv) 1000 V, 500 Vhr; (v) 1000-8000 V, 8000 Vhr; (vi) 8000 V, 48000 Vhr. After IEF, IPG strips were equilibrated in a buffer containing 50 mM Tris-HCl (pH 8.8), 6 M urea, 30% w/w glycerol, 2% SDS and 65 mM DTT for 15 min to chemically reduce the samples. DTT was then replaced with 135 mM iodoacetamide for a further 15 min to alkylate the samples. Second dimension separation was performed on 10% polyacrylamide gels and conducted using the PROTEAN II XL cell (Biorad Laboratories). Each gel was run at 10 mA overnight in the dark. Gels were scanned directly between gel plates on the Typhoon Trio (GE Healthcare) at 100 µm resolution. Fluorescent images were captured using optimal excitation/emission 4064
Journal of Proteome Research • Vol. 8, No. 8, 2009
Figure 1. (A) Antiphosphotyrosine immunoblots of the lysates from cells from the MCF10AT model untreated or stimulated with 50 ng/mL of EGF for 5 min. (B and C) Schematic diagram of the experimental design for expression profiling of proteins in the MCF10AT model. Cells were either untreated or stimulated with EGF for 4 h or 16 h. Detection and relative quantification of proteins were performed using (B) iTRAQ labeling and ESI-LCMS/MS analysis and (C) 2D-DIGE and MALDI-TOF/TOF. Experimental duplicates were labeled 1 and 2 within parentheses.
wavelength for each dye: Cy2 (488/520 nm), Cy3 (532/580 nm) and Cy 5 (633/670 nm). Scan settings were optimized to achieve similar maximum signal intensity from all 3 fluorescent dyes on a single gel. Image analyses were first performed using the DIA (differential in-gel analysis) module of the DeCyder 2-D Differential Analysis Software v6.5 (GE Healthcare) for automatic spot detection, spot volume determination, normalization, and background subtraction of proteins within every image. Protein spots were then matched across all 36 images (Cy2, Cy3, and Cy5 images from all 12 gels). Gels containing the biological duplicates from the same condition were grouped together and the degree of difference (expression ratio) in the standardized abundance between 2 protein spot groups (e.g.,
research articles
MIF is a Novel Target of EGFR 4 h EGF stimulation versus nonstimulated) was compared and Student’s t test performed with the Decyder BVA (biological variation analysis) module to identify proteins with statistically significant difference in expression levels between 2 conditions (p < 0.05). In DeCyder output, an increase in protein abundance is expressed as a positive value (e.g., “1.5” for a 1.5-fold increase) and a negative value for a decrease in protein abundance (e.g., “-1.5” for a 1.5-fold decrease). For the purpose of standardizing data presentation with the iTRAQ data, decrease in protein abundance was converted to a range of 0 to
