Secretome-Based Proteomic Profiling of Ras-Transformed MDCK Cells Reveals Extracellular Modulators of Epithelial-Mesenchymal Transition Rommel A. Mathias,† Bo Wang,‡ Hong Ji,† Eugene A. Kapp,† Robert L. Moritz,† Hong-Jian Zhu,‡ and Richard J. Simpson*,† Joint Proteomics Laboratory, Ludwig Institute for Cancer Research and the Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia, and Department of Surgery, Royal Melbourne Hospital, The University of Melbourne, Parkville, Victoria, Australia Received December 21, 2008
Epithelial-mesenchymal transition (EMT) is a highly conserved morphogenetic process by which epithelial cells lose their basic morphological characteristics such as cell-cell contact and gain mesenchymal properties such as increased motility and invasiveness. To gain insights into proteins released from cells that modulate the EMT process, we compared secretome protein expression profiles of MDCK cells and Ras-transformed MDCK cells (21D1) that stably express oncogenic Ras using 2D-DIGE/LC-MS/MS. Differentially expressed secretome proteins were compared with their corresponding gene expression profiles using the Affymetrix GeneChip system. Down-regulated proteins were predominantly involved with cell-cell contact and cell-matrix adhesion (e.g., desmocollin 2, clusterin, collagen XVII and transforming growth factor-β induced protein ig-h3), while up-regulated proteins were proteases and factors that promote migration (MMP-1, kallikrein 6, TIMP-1, and S100A4/metastasin). Many of the secretome proteins identified in this study have not been previously identified in the context of EMT and may shed light on the underlying mechanisms associated with this cellular process. Keywords: epithelial-mesenchymal transition • EMT • MDCK • DIGE • Ras • secretome • quantitative proteomics
Introduction Epithelial-mesenchymal transition (EMT) is a morphogenetic process involving an orchestrated series of transcriptional reprogramming and protein modification events that allow polarized, immotile epithelial cells to convert into motile mesenchymal cells.1-4 The diminution of epithelial structural integrity is central in EMT in that it allows cells to escape constraints imposed by tissue architecture. This highly conserved cellular process involves loss of cell polarity, reduced cell-cell and cell-extracellular matrix (ECM) interactions, and cytoskeletal and ECM remodelling that facilitates motility. Although well recognized to be a central process in early embryonic morphogenesis,5 EMT over the past decade has been strongly implicated in key steps of tumor progression, especially the process by which epithelial cancers metastasize.1,3,6 EMT can be broadly defined by the following alterations in cellular phenotype: (1) morphological changes from a cobblestone-like epithelial monolayer with apical-basal polarity, to dispersed spindle-shaped mesenchymal cells containing lead* To whom correspondence should be addressed. Professor Richard J. Simpson, Ludwig Institute for Cancer Research, PO Box 2008, Royal Melbourne Hospital Royal Parade, Parkville, Victoria 3050, Australia. Tel: +61 03 9341 3155. Fax: +61 03 9341 3192. Email:
[email protected]. † Ludwig Institute for Cancer Research and the Walter and Eliza Hall Institute of Medical Research. ‡ The University of Melbourne. 10.1021/pr8010974 CCC: $40.75
2009 American Chemical Society
ing edge polarity; (2) differentiation marker changes from cellcell junction proteins (E-cadherin, claudins and occludins) and cytokeratin intermediate filaments, to a network of vimentin filaments and an N-cadherin adhesion system; and (3) functional changes associated with the conversion of immotile cells to cells that can migrate and invade through ECM.7 Although not all of these phenotypic changes invariably occur during EMT, acquisition of the latter is considered to be a hallmark of the cellular process. To better understand the fundamental mechanisms underlying EMT, several in vitro models have been derived from various tissues, summarized in.;8 of these, Madin-Darby canine kidney (MDCK) cells are one of the best characterized.9 For example, disruption to cell junctions, increased local motility, and cell scattering were observed when MDCK cells were cultured with conditioned medium from human embryo fibroblasts.10 Scatter factor/Hepatocyte growth factor (HGF) was subsequently identified as the molecule responsible for this phenomenon; HGF has also demonstrated ability to promote EMT in other human carcinoma cell lines.11 Increased migration and invasion observed in cells undergoing EMT is heavily reliant on the loss of intercellular adhesion mediated by E-Cadherin.12 Down-regulation of E-cadherin has largely been attributed to increased expression of many EMT-inducing transcriptional factors, including; Snail, Slug, ZEB-1, E47 and SIP-1 that bind to the proximal E-boxes of the promoters of Journal of Proteome Research 2009, 8, 2827–2837 2827 Published on Web 03/18/2009
research articles 13
E-cadherin. Overexpression of Snail, Slug, or E47 in MDCK cells elicits a full EMT at both the morphologic and behavioural levels,12-14 while stable silencing of overexpressed Snail by siRNA effectively restores the epithelial phenotype.15 Interestingly, Snail appears to be located at the hub of multiple signaling pathways leading to EMT,16 and its expression directly induced by transforming growth factor-β (TGF-β) and oncogenic Ras.17 MDCK cells transformed with oncogenic Ras alone exhibit loss of polarity, multilayering cell growth, and Actin cytoskeletonalterationsthroughactivatedRaf/MAPKsignaling.18,19 In addition, sustained Raf activation establishes an autocrine TGF-β loop that promotes invasiveness, while blocking the apoptotic response.20 While Raf/MAPK signaling is both necessary and sufficient to initiate early stages of EMT, recent reports suggest matrix metalloproteinases (MMPs) are necessary for later stages of progression.21 The ability of MMPs to remodel the ECM, alter cell morphology, and activate secreted growth factors highlights this family of proteases and others as important regulators of EMT. Despite progress made in identifying key molecular players and signaling pathways during EMT,7,22 a detailed knowledge of extracellular proteins involved in the process is limited. To identify novel extracellular effectors of EMT in an unbiased way, we have conducted a proteomic analysis of the secretome from MDCK and Ras-transformed MDCK cells (21D1 cells). To identify proteins that are differentially expressed as MDCK cells undergo EMT, the secretome, comprising both exosomes23 and soluble-secreted proteins, was analyzed by DIGE coupled with mass spectrometry.24 Expression levels for proteins markedly dysregulated during EMT were correlated with corresponding mRNA transcript levels. In this study, we report reduced expression levels for several proteins involved in intercell adhesion and attachment to the underlying cellular matrix. Additionally, we report diminished expression of basement membrane constituents, elevated levels of secreted proteases, and increased levels of factors implicated in cell migration and invasion. Collectively, our findings highlight, for the first time, the involvement of several proteins in the EMT cellular process.
Experimental Procedures Cell Culture. MDCK cells25 and the oncogenic mutated Rastransformed derivative 21D1, were routinely cultured in Dulbecco’s modified Eagle’s medium (DMEM) (GIBCO) supplemented with 10% FCS (CSL), 60 µg/mL penicillin (CSL), 100 µg/mL streptomycin (SIGMA) at 37 °C with 10% CO2. 21D1 cells were generated by transfecting pcDNA3 (Invitrogen) containing v-Ha-Ras under the control of a CMV promoter and a SV40driven neomycin resistance plasmid gene into MDCK cells using Fugene6 (Roche), according to manufacturer’s instructions. Clones stably expressing Ras were isolated by Western blot analysis and immunofluorescence from G418 (Invitrogen) resistant clones, and the 21D1 clone was selected as the mesenchymal model for further studies. Western Immuno-Blotting. MDCK and 21D1 cell lysates were prepared with lysis buffer (1% Triton X-100, 30 mM Hepes, 150 mM NaCl, and 5 mM EDTA) on ice for 1 h, followed by centrifugation to remove insoluble cellular debris. Supernatants were collected and quantified (Bio-Rad protein assay), and 10 µg of each lysate separated by SDS-PAGE. Proteins were then electro-transferred to polyvinylidene fluoride membranes (Schleicher & Schuell) and blocked with PBS buffer containing 0.2% Tween-20 and 5% powdered milk for 1 h. Membranes were 2828
Journal of Proteome Research • Vol. 8, No. 6, 2009
Mathias et al. probed with either a mouse anti-Ras (Upstate biotechnology), mouse anti-E-cadherin (Transduction Laboratories), rabbit anti-ZO-1 (ZYMED), or mouse antivimentin (Millipore) primary antibody for 1.5 h, followed by incubation with the corresponding goat antimouse or antirabbit IgG HRP conjugated (Bio-Rad) secondary antibody for 1 h. Antigen-antibody complexes were detected by ECL (GE Healthcare). Cell Migration, Invasion, and Viability Assays. Migration assay: the migration/motility of MDCK and 21D1 cells was analyzed in vitro using the wound-healing assay as previously described.12 Briefly, confluent cell monolayers were wounded (lightly scratched) with a pipet tip. After careful washing to remove detached cells, fresh medium was added and the cells left to culture for 24 h. Phase contrast images were taken, and migration evaluated based on the percentage of wound closure. All phase contrast microscopy was performed on a Nikon TE 2000-E microscope equipped with a DP-70 camera. Invasion assay: 1 mg/mL rat tail collagen (BD Biosciences) and 10% matrigel (BD Biosciences) were mixed in DMEM at 0 °C. Onehundred microliters of this mixture was added to wells of a 24-well plate for solidification at 37 °C for 4 h. Cells were added to the remainder of the mixture and mixed rapidly. Onehundred fifty microliter droplets of cell-containing mixture were then dispersed on-top of the solidified collagen-matrigel in each well, and the plates placed in a 37 °C incubator for 4 h. Two-hundred microliter DMEM was then overlaid and changed every 24 h. Cell invasiveness was assessed by estimating structure formation using phase-contrast microscopy, of lumen-containing cysts versus branching chords at several depths below the gel surface. Viability assays: viability of MDCK and 21D1 cells during serum-free culture for 24 h was assessed both by metabolic activity using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay,26 as well as the trypan blue dye-exclusion assay.27 All assays were performed in triplicate. Secretome Purification. MDCK and 21D1 cells were grown to 70% confluence in DMEM containing 10% FCS, washed three times with serum-free DMEM, and left to culture in this medium at 37 °C with 10% CO2 for 24 h. Cell culture medium (CM) from 10 dishes (a total of 100 mL from 5 × 107 cells for each line) was harvested and centrifuged (480 × g, 5 min) to sediment floating cells, followed by centrifugation (2000 × g, 10 min) to remove cellular debris. Complete EDTA-free protease inhibitor cocktail tablets (Roche) were added to the resultant supernatant and the solution filtered through a 0.1 µm Supor membrane VacuCap 60 Filter Unit (Pall). The filtrate was concentrated to 1 mL using a 5K NMWL Amicon ultra centrifugal filter device (Millipore), total protein precipitated using a 2-D Clean-Up Kit (GE Healthcare) and proteins resolubilized in 2-DE Sample Buffer (7 M urea, 2 M thiourea, 4% CHAPS and 20 mM Tris, pH 8.5). Protein concentrations were determined using the 2-D Quant Kit (GE Healthcare). Fluorescence Difference Gel Electrophoresis (DIGE) Labeling. Four independent MDCK and 21D1 cell secretome samples were prepared for DIGE analysis. Each secretome sample was represented four times across four DIGE gels, and coordinated by the same pooled internal standard (see Supporting Information Figure S1A). Protein samples were minimally labeled with fluorescent cyanine dyes (GE Healthcare) according to the manufacturer’s instructions. Briefly, 25 µg of each sample was labeled with 200 Fmol of either Cy3 or Cy5, while the internal standard which was generated by pooling 12.5 µg from each sample (total 100 µg) was labeled with 800
Proteomic Profiling of Ras-Transformed MDCK Cells Fmol of Cy2. A dye-swapping scheme was used such that both MDCK and 21D1 secretome samples were labeled twice with Cy3 and twice with Cy5, to eliminate any dye-specific labeling bias. Labeling was performed for 30 min on ice in the dark, after which the reactions were quenched with 10 mM lysine (1 µL for each 200 Fmol dye) for 20 min on ice in the dark. Each 50 µg pair of Cy3 and Cy5-labeled protein samples were combined and mixed with one aliquot of 25 µg Cy2-labeled internal standard. An equal volume of 2 × Sample Buffer (9 M urea, 4% CHAPS, 1% IPG buffer, and 20 mM DTT) was added to each combined Cy3/Cy5/Cy2 mixture. Thus, each of the four DIGE gels run contained 25 µg Cy3 labeled sample, 25 µg Cy5 labeled sample, and 25 µg Cy2 labeled internal standard (see Supporting Information Figure S1B). 2-DE. For IEF, Immobiline 24 cm pH 3-10 Linear DryStrip Gels (GE Healthcare) were rehydrated overnight in DeStreak rehydration solution (GE Healthcare) containing 0.5% IPG buffer. Each Cy3/Cy5/Cy2 mixture consisting of a sample pair and internal standard was loaded into a sample cup at the anode in preparation for IEF. IEF was performed using an Ettan IPGphor 3 IEF system (GE healthcare) for approximately 80 kV-h (300 V for 3 h, ramped to 1000 V over 6 h, ramped to 8000 V over 3 h, and hold at 8000 V for 8 h). Prior to second dimension SDS-PAGE, each IPG strip was equilibrated in equilibration buffer (6 M urea, 30% glycerol, 2% SDS, 50 mM Tris-HCl, pH 8.8 and 0.002% bromophenol blue) supplemented with 1% DTT for 20 min at room temperature, followed by incubation in equilibration buffer containing 2.5% iodoacetamide for 20 min at room temperature. IPG strips were placed on top of hand-cast 8-18% gradient polyacrylamide gels that had been cast with low-fluorescence glass plates using a DALTtwelve Gel Caster (GE Healthcare). Second dimension SDS-PAGE was performed using an Ettan DALTtwelve Separation Unit (GE Healthcare) at 5 W/gel for 30 min, followed by a constant 90 V at 23 °C until the dye-front reached the bottom of the gel. MDCK and 21D1 preparative 2-DE gels (500 µg protein), from which spots were excised for mass spectrometric protein identification, were prepared identically and subjected to the same IEF and SDS-PAGE conditions as the analytical 2-D DIGE gels. Following electrophoresis, the preparative sequencing gels were fixed in 40% aqueous methanol and 7% aqueous acetic acid for 30 min, washed with deionized water 3 times for 10 min. Proteins were visualized by incubating the gels with Imperial Protein Stain (PIERCE) for 1 h. DIGE Imaging and DeCyder Analysis. DIGE gel images were scanned at 100 µm resolution on a Typhoon 9410 variable mode imager (GE Healthcare) using excitation/emission wavelengths specific for Cy2 (488/520 nm), Cy3 (532/580 nm), and Cy5(633/ 670 nm). Analysis of DIGE gels was performed using DeCyder 6.5 software (GE Healthcare) in batch processor mode with an estimated number of spots set to 2500, and the spot exclusion filter set to exclude any spot with a volume less than 7500. Batch processor was used to link the Differential In-gel Analysis (DIA) and Biological Variation Analysis (BVA) modules together in an automated fashion. The gel containing the highest number of spot-features was assigned the master gel, and manual spot matching was then performed to correctly match the remaining three Cy2 gel images with the Cy2 master. In DIA spot boundaries and volumes were codetected for Cy3, Cy5, and Cy2 channels on each gel, and protein spot abundance expressed as a standard:sample ratio. In BVA, protein abundance was compared across multiple samples using the internal standard to normalize between gels, and statistical
research articles analysis was performed to provide average ratio and one-way ANOVA values between samples. Gel Excision and Tryptic Digestion. Analytical DIGE gels were matched to the preparative 2-DE gels, and 65 protein spots of interest were excised from the preparative gels and subjected to automated in-gel reduction, alkylation, and tryptic digestion28 using the MassPREP Robotic Liquid Handling Station (Micromass). Briefly, gel sections were reduced with 10 mM DTT (Calbiochem) for 30 min, alkylated for 20 min with 25 mM iodoacetamide (Fluka), and digested with 150 ng trypsin (Worthington) for 4.5 h at 36 °C. Peptide extracts were concentrated to approximately 10 µL by centrifugal lyophilization using a SpeedVac AES 1010 (Savant) for subsequent LC-MS/MS analysis. LC-MS/MS. Extracted peptides were transferred to 100 µL glass autosampler vials, and 8 µL of sample was fractionated by capillary reverse-phase liquid chromatography on a 1100 series LC system (Agilent) using an in-house manufactured column (5 µm C4MS 150 × 0.15 mm i.d. column (Vydac) developed with a linear 60-min gradient with a flow rate of 0.8 µL/min at 45 °C from 0-100% solvent B, where solvent A was 0.1% aqueous formic acid and solvent B was 0.1% aqueous formic acid/60% acetonitrile. The capillary HPLC was coupled online to an LCQ-Deca (Finnigan) for automated MS/MS analysis of individually isolated peptide ions. The Deca was operated in data-dependent mode (triple-play) to automatically switch between MS, Zoom MS (automated charge state recognition), and MS/MS acquisition, selecting the most intense precursor ion from the initial MS scan for fragmentation using collision-induced dissociation. Dynamic exclusion was employed for a period of 2.5 min when four consecutive precursor ions of the same mass were observed within half a minute. Database Searching and Protein Identification. Peak lists were extracted using extract-msn as part of Bioworks 3.3.1 (Thermo Fisher Scientific). The parameters used to generate the peak lists for the Deca data were as follows: minimum mass 700; maximum mass 5000; grouping tolerance 1.4 Da; intermediate scans 1; minimum group count 1; 10 peaks minimum and total ion current of 100 000. Peak lists for each LC-MS/ MS run were merged into a single Mascot generic file (MGF format) for database searches. The MGF format provided an embedded CHARGE parameter that allowed LCQ-Deca data to be searched as 2, 3, and 4 + species simultaneously. MGF files were searched against the LudwigNR_subset database comprising 397 867 entries using the MASCOT v2.2.01 search algorithm (Matrix Science). The search parameters consisted of carboxymethylation of cysteine as a fixed modification (+58 Da), NH2-terminal acetylation (+42 Da) and oxidation of methionine (+16 Da) as variable modifications. A peptide mass tolerance of (3 Da, fragment ion mass tolerance of ( 0.8 Da, and an allowance for up to two missed tryptic cleavages was used. Protein identifications were first clustered and analyzed by an in-house developed program MSPro, as previously described.29 Briefly, peptide identifications were deemed significant if the Ions score was g the Homology score (or Identity score if there was no Homology score). False positive protein identifications were estimated by searching MS/MS spectra against the corresponding reverse-sequence (decoy) database.30 Protein annotation was obtained from the Ensembl database, and where none was available, the protein entries amino acid sequence was protein BLAST searched against the Uniprot database, and the closest homologue’s properties inferred. Journal of Proteome Research • Vol. 8, No. 6, 2009 2829
research articles
Mathias et al.
Figure 1. MDCK cells undergo EMT following transformation with oncogenic Ras. Cell morphology, protein expression, migration and invasion ability was assessed in MDCK and 21D1 cells. (A) 21D1 cells exhibit elongated and spindle-shaped cell morphology during EMT. (B) Following Ras-transformation of MDCK cells, Western blot analysis of 21D1 cells shows diminished expression of epithelial markers E-cadherin and ZO-1, and increased expression of the mesenchymal marker vimentin. (C) Wound-healing assay reveals increased migration in 21D1 cells, compared to wild-type MDCK cells after 24 h. (D) Phase contrast images of cell growth in collagen-matrigel show spherical cyst-like MDCK structures, while 21D1 cells demonstrate increased invasiveness through formation of branching tubular extensions.
RNA Isolation and Microarray Analysis. MDCK and 21D1 cells were gown using the same culture conditions described for proteome analysis. Total RNA was isolated using an RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. RNA quality was assessed using the Agilent Bioanalyser 2100 and NanoChip protocol. A total of 7 µg was labeled using the Affymetrix One Cycle cDNA synthesis kit (Millennium Sciences) and subsequent cDNA was cleaned using the Affymetrix GeneChip Sample Cleanup kit (Millennium Sciences). Incorporation of biotin into the resultant cRNA was achieved using the Affymetrix IVT labeling kit (Millennium Sciences), and the labeled cRNA cleaned with the Cleanup kit, before being quantified on the Bioanalyser. Twenty micrograms of labeled cRNA was then fragmented to the 50-200 bp size range, and quality control checked on the Bioanalyser. 0.05 µg/µL cRNA and a probe cocktail including, 1× Hybridization Buffer (100 mM MES, 1 M NaCl, 20 mM EDTA, 0.01% Tween-20), 0.1 mg/ mL Herring Sperm DNA, 0.5 mg/ml BSA, and 7% DMSO were then prepared for hybridization to the Canine Genome 2.0 GeneChip Array (Affymetrix). A total hybridization volume of 200 µL per sample was loaded onto each Canine GeneChip, for hybridization at 45 °C for 16 h in an oven with a 60 rpm rotating wheel. Chips were then washed using the appropriate fluidics script in the Affymetrix Fluidics Station 450 and scanned using the Affymetrix GeneChip Scanner 3000. The scanners GeneChip Operating Software (GCOS), was then used to convert the signal from the chip into a DAT file. Subsequent CEL and CHP files were generated for analysis, and fold change differences calculated between samples.
Results Oncogenic Ras Induces EMT in MDCK Cells. After MDCK cell transfection with oncogenic Ras, 21D1 cells were assessed 2830
Journal of Proteome Research • Vol. 8, No. 6, 2009
for loss-of-epithelial and gain-of-mesenchymal markers and characteristics associated with EMT (Figure 1). Under phasecontrast microscopy, MDCK cells appear round and resemble cobble-stone-like growth structures (Figure 1A). They form a uniform epithelial sheet where individual cells are tightly adjoined to surrounding cells. In contrast, 21D1 cells look fibroblast-like, have an elongated spindle-shaped morphology, and are scattered in culture with reduced intercell contact. Intact adherens and tight junctions mediate the apical-basolateral polarization seen in MDCK cells, via expression of epithelial proteins E-cadherin and ZO-1. Western blot analysis of 21D1 cell lysates shows diminished expression of these epithelial markers following Ras-transformation (Figure 1B). Additionally, expression of the mesenchymal marker vimentin was observed minimally in MDCK cells, but with pronounced expression in 21D1 cells during EMT (Figure 1B). MDCK cells exhibit limited migratory capacity due to tight neighboring cell-cell interactions. Figure 1C shows before- and after-images of a wound-healing assay over 24 h, where individual MDCK cell movement is restricted, as cells move as a sheet on block. 21D1 cell motility appears enhanced and considerably more dynamic, as individual cells have migrated further into the wounded area. When invasive capability was evaluated via growth structure formation in collagen-matrigel, MDCK cells grew into lumen-containing spherical cysts, whereas 21D1 cells formed tubular extensions that penetrated deeper into the gel (Figure 1D). DIGE Analysis Identifies and Quantifies Differentially Expressed Secretome Proteins During EMT. Because serumfree culture conditions were required for CM collection, MDCK and 21D1 cell viability was assessed to ensure the secretome samples were derived from viable cells that were metabolically active. Figure 2A shows that when cultured in serum-free
research articles
Proteomic Profiling of Ras-Transformed MDCK Cells
Figure 2. MDCK and 21D1 metabolic activity and cell viability. (A) Metabolic activity of cells cultured with 10% FCS or serumfree conditions was measured using the MTT assay. (B) Cell viability as determined by the Trypan Blue dye exclusion assay after culture in serum-free conditions for 24 h.
DMEM over 24 h, the metabolic activity of both cell lines reduced slightly, to 82 and 85% for MDCK and 21D1 cells, respectively. Additionally, under both conditions 21D1 cells have increased metabolic activity compared to MDCK cells. Cell viability was also determined using the Trypan Blue dye exclusion assay. MDCK cell viability decreased from 99.2 to 96% over 24 h when cultured in DMEM containing 10% FCS, compared to serum-free DMEM; while the viability of 21D1 cells remained constant at 98.5% after culture in serum-free conditions over the same time period. Because both cell lines were metabolically active with minimal cell death after 24 h in serum-free culture, secretome purification (Figure 3) and analysis was performed under this condition. Secretome proteins differentially expressed during EMT were identified by DIGE analysis. Four individual secretome preparations were purified from both MDCK and 21D1 cells, labeled with Cy dye, and separated across four 2-D gels. Gel-to-gel variation was reduced as two samples were run on the same gel and intergel spot matching was coordinated by the internal standard. Additionally, the reverse labeling strategy revealed no dye-specific bias. While minor variability (
