An integrated approach for comparative proteomic analysis of human bile reveals overexpressed cancer-associated proteins in malignant biliary stenosis

Biochimica et Biophysica Acta 1844 (2014) 1026–1033

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap

An integrated approach for comparative proteomic analysis of human bile reveals overexpressed cancer-associated proteins in malignant biliary stenosis☆ Natalija Lukic a,1, Rémy Visentin a,1, Myriam Delhaye b, Jean-Louis Frossard c, Pierre Lescuyer a,d, Jean-Marc Dumonceau c, Annarita Farina a,⁎ a

Biomedical Proteomics Research Group, Department of Human Protein Sciences, Geneva University, Geneva CH-1211, Switzerland Department of Gastroenterology, Erasme Hospital, Free University of Brussels, Brussels BE-1070, Belgium c Division of Gastroenterology and Hepatology, Geneva University Hospitals, Geneva CH-1211, Switzerland d Clinical Proteomics Laboratory, Department of Genetic and Laboratory Medicine, Geneva University Hospitals, Geneva CH-1211, Switzerland b

a r t i c l e

i n f o

Article history: Received 10 April 2013 Received in revised form 21 June 2013 Accepted 28 June 2013 Available online 18 July 2013 Keywords: Bile iTRAQ Biomarker Olfactomedin-4 Syntenin-2 Ras-related C3 botulinum toxin substrate 1

a b s t r a c t Proteomics is a key tool in the identification of new bile biomarkers for differentiating malignant and nonmalignant biliary stenoses. Unfortunately, the complexity of bile and the presence of molecules interfering with protein analysis represent an obstacle for quantitative proteomic studies in bile samples. The simultaneous need to introduce purification steps and minimize the use of pre-fractionation methods inevitably leads to protein loss and limited quantifications. This dramatically reduces the chance of identifying new potential biomarkers. In the present study, we included differential centrifugation as a preliminary step in a quantitative proteomic workflow involving iTRAQ labeling, peptide fractionation by OFFGEL electrophoresis and LC-MS/MS, to compare protein expression in bile samples collected from patients with malignant or nonmalignant biliary stenoses. A total of 1267 proteins were identified, including a set of 322 newly described bile proteins, mainly belonging to high-density cellular fractions. The subsequent comparative analysis led to a 5-fold increase in the number of quantified proteins over previously published studies and highlighted 104 proteins overexpressed in malignant samples. Finally, immunoblot verifications performed on a cohort of 8 malignant (pancreatic adenocarcinoma, n = 4; cholangiocarcinoma, n = 4) and 5 nonmalignant samples (chronic pancreatitis, n = 3; biliary stones, n = 2) confirmed the results of proteomic analysis for three proteins: olfactomedin-4, syntenin-2 and Ras-related C3 botulinum toxin substrate 1. This article is part of a Special Issue entitled: Biomarkers: A Proteomic Challenge. © 2013 Elsevier B.V. All rights reserved.

1. Introduction In the presence of biliary tract malignancies, human bile was shown to collect cancer-associated proteins [1–4]. This has been recognized as a promising turning point for the discovery of new cancer biomarkers, especially for those malignancies whose diagnosis is difficult to confirm [5,6]. Differentiating malignant from nonmalignant biliary stenoses represents a major challenge for gastroenterologists. All currently available noninvasive diagnostic tools, including imaging techniques and serum biomarkers, are inadequate to provide a rapid and unambiguous identification of malignant stenoses [7–10]. The growing number of studies aimed at analyzing the bile proteome has recently raised the possibility

☆ This article is part of a Special Issue entitled: Biomarkers: A Proteomic Challenge. ⁎ Corresponding author at: Biomedical Proteomics Research Group, Department of Human Protein Sciences, Faculty of Medicine, Geneva University, Rue Michel Servet, 1, Geneva CH-1211, Switzerland. Tel.: +41 22 3795451; fax: +41 22 3795502. E-mail address: [email protected] (A. Farina). 1 These authors contributed equally to this work. 1570-9639/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbapap.2013.06.023

that proteomic approaches may play a key role in the identification of new biomarkers for discriminating malignant from nonmalignant biliary tract lesions [5,6]. The chemical complexity of bile, however, as well as the presence of molecules interfering with protein analysis, represents an obstacle to the full accomplishment of this purpose. A large discrepancy is evident between the results of purely qualitative proteomic analyses and quantitative studies that aimed to compare the expression of bile proteins in malignant and nonmalignant samples. Barbhuya et al., in 2011, were able to identify 2552 proteins in human bile by applying multiple orthogonal fractionation methods (SDSPAGE, SCX and OFFGEL electrophoresis) to pretreated (e.g., delipided, desalted and immunodepleted) bile samples [11]. Conversely, Zabron et al., during the same year, carried out a quantitative label-free analysis of bile samples from malignant and nonmalignant conditions after SDSPAGE purification and identified only about 200 proteins [2,12]. Similar results were obtained by our group while performing a comparative analysis using iTRAQ labeling of delipided and desalted bile samples collected from patients with malignant and nonmalignant biliary stenoses [13]. This huge difference in the total number of identifications goes far

N. Lukic et al. / Biochimica et Biophysica Acta 1844 (2014) 1026–1033

beyond the mere performances of mass spectrometers. The abovementioned discrepancies, indeed, also affected works exploiting highresolution hybrid instruments (e.g., LTQ-Orbitrap) [11,13]. We hypothesize that the reason could more likely be sought in the simultaneous need to introduce clean-up steps and minimize the use of pre-fractionation methods when performing quantitative studies. Purification, in fact, is mandatory to remove bile contaminants interfering with proteomic analysis, but inevitably leads to a major protein loss. On the other hand, pre-fractionation methods, which has proved to represent a good way to enhance the number of identifications [11,14,15], increase sample variability due to extra sample handling and are not particularly suitable for quantitative analysis [16,17]. These issues dramatically reduce the possibility of identifying new potential biomarkers. In a previous work, we demonstrated that differential centrifugation is an effective method of bile fractionation improving bile protein identifications [18]. Since this fractionation method is purely physical and takes place without using devices which could interact with proteins (e.g., filters, resins), the protein loss and quantification bias could be minimized. On the basis of these considerations, we decided to include differential centrifugation as a preliminary pre-fractionation step in a quantitative proteomic workflow involving iTRAQ labeling, peptide fractionation by OFFGEL electrophoresis and LC-MS/MS. In the present study, comparative analysis of bile centrifugal fractions from malignant and nonmalignant biliary stenoses is presented, demonstrating the compatibility of differential centrifugation with proteomic quantitative analysis and the ability of this integrated approach to identify and quantify a large number of proteins as well as to detect new potential candidate biomarkers overexpressed in malignant samples.

2. Material and methods 2.1. Sample collection Bile samples were collected during endoscopic retrograde cholangiopancreatography (ERCP) from patients presenting with biliary stenosis of either malignant (pancreatic adenocarcinoma, PAC, n = 4; cholangiocarcinoma, CC, n = 4) or nonmalignant (chronic pancreatitis, CP, n = 3; biliary stones, BS, n = 2) etiologies (Table 1). A volume of 10–30 mL of bile was collected upstream from the bile duct stenosis before contrast medium injection. Immediately after collection, bile samples were transported on ice, aliquoted and stored at −80 °C until analysis. A total of 4 samples, from 2 malignant (PAC, n = 1; CC, n = 1) and 2 nonmalignant (CP, n = 1; BS, n = 1) conditions, were used for comparative proteomic analysis. All samples were subjected to immunoblot verification. Patient's informed consent was obtained and the protocol of this study was approved by Ethics Committees of the Geneva University Hospitals and the Erasme Hospital.

Table 1 Patient's demographic data. No.

Gender

Age (years)

Diagnosis

Proteomic analysis

1 2 3 4 5 6 7 8 9 10 11 12 13

F M F M M M M F F M M F F

78 69 75 61 50 49 54 66 54 73 77 80 78

Pancreatic adenocarcinoma Pancreatic adenocarcinoma Pancreatic adenocarcinoma Pancreatic adenocarcinoma Chronic pancreatitis Chronic pancreatitis Chronic pancreatitis Biliary stones Biliary stones Cholangiocarcinoma Cholangiocarcinoma Cholangiocarcinoma Cholangiocarcinoma

x

1027

2.2. Sample preparation for proteomic analysis Bile samples were centrifuged at 16,000 g for 20 min at 4 °C. The obtained supernatants were diluted (1:5) with Dulbecco's Phosphate Buffered Saline (D-PBS) to reduce viscosity and subjected to the following consecutive steps of differential centrifugation: i) 70,000 g for 60 min; ii) 100,000 g for 60 min; and iii) 200,000 g for 120 min at 4 °C (Fig. 1). The pellet fractions obtained after each centrifugation step were washed with an equal volume of D-PBS and centrifuged again, under the same conditions, to remove residual supernatant. Equal aliquots of each supernatant fraction, corresponding to the same initial volume (45 μL) of crude bile, were delipided with Cleanascite (Biotech Support Group, North Brunswick, NJ, USA) and ultrafiltrated using a 3 kDa filter cut-off (YM-3 centricon, Millipore, Bedford, MA, USA) according to the manufacturer's instructions. For proteomic analysis, pellets and supernatants were treated as previously described [18,19] with modifications, according to the following details. 2.3. Gel-electrophoresis and in-gel enzymatic digestion Pellets from centrifugal fractions (70,000 g; 100,000 g and 200,000 g) were dissolved with 15 μL of Laemmli buffer in order to solubilize precipitated components and extract vesicular embedded proteins [20], heated up to 90 °C and centrifuged at 16,000 g for 2 min. Protein samples were then fractionated using SDS-PAGE on a 12.5%T, 2.6%C polyacrylamide gel (8 × 8 cm) until the migration front had moved on approximately 1.5 cm. The gel was stained with Coomassie Blue (0.1% Coomassie R-250, 50% MeOH, 7.5% AcH). Gel lanes were sliced into 4 pieces and destained by incubation in 400 μL of 50 mM triethylammonium hydrogen carbonate buffer (TEAB) pH 8.0/30% acetonitrile (ACN) for 15 min at room temperature. Destaining solution was removed and fragments were then incubated for 35 min at 56 °C in 400 μL of 10 mM dithioerythritol (DTE) in 50 mM TEAB pH 8.0. The DTE solution was then replaced by 400 μL of 55 mM iodoacetamide (IAM) in 50 mM TEAB pH 8.0 and gel fragments were incubated for 45 min at room temperature in the dark. Gel pieces were washed twice for 10 min with 400 μL of 50 mM TEAB pH 8.0 and for 10 min with 400 μL of 50 mM TEAB pH 8.0/30% ACN. Gel pieces were then dried for 30 min in a Concentrator 5301 vacuum centrifuge (Eppendorf AG, Hamburg, Germany). Dried gel pieces were rehydrated for 30 min at 37 °C in 100 μL of a 0.025 U/μL PNGase F solution. After further addition of 100 μL HPLC grade water, gel pieces were incubated 30 min at 37 °C. The tubes were briefly centrifuged and digested glycans were extracted four times by 2 min sonication in 800 μL of HPLC grade water (VialTweeter UIS250V, Hielscher Ultrasonics GmbH, Teltow, Germany). Gel pieces were then dried for 45 min in a vacuum centrifuge and rehydrated for 45 min on melting ice in 200 μL of a 50 mM TEAB pH 8.0 solution containing trypsin at 6.25 ng/μL. After overnight incubation at 37 °C, the tubes were briefly centrifuged. Extraction of tryptic peptides was performed by 2 min sonication in 200 μL of 1% formic acid (FA) at room temperature. The acidic supernatant containing peptides was transferred to a polypropylene tube. A second extraction was performed by 2 min sonication in 200 μL of 0.1% FA/50% ACN at room temperature. The second acidic supernatant was pooled with the first one. The volume of pooled extracts was reduced to 5–10 μL under vacuum, diluted twice with 35 μL of 0.1% FA, and then completely evaporated. The resulting peptides were resuspended in 1 μL of 1% SDS and 99 μL of 0.1 M TEAB pH 8.0. 2.4. In-solution enzymatic digestion

x x

x

Delipided and desalted supernatants from centrifugal fractions (16,000 g; 70,000 g; 100,000 g and 200,000 g) were mixed with 0.1 M TEAB pH 8.0 to a final volume of 100 μL. Proteins were reduced by adding 1 μL of 1% SDS and 2 μL of 50 mM tris (2-carboxyethyl) phosphine (TCEP) and heating at 60 °C for 60 min. Free thiol groups of cysteine residues were alkylated by adding 1 μL of 400 mM IAM and

1028

N. Lukic et al. / Biochimica et Biophysica Acta 1844 (2014) 1026–1033

Fig. 1. Bile fractionation workflow. Crude bile was centrifuged at 16,000 g for 20 min at 4 °C. The obtained supernatants were diluted (1:5) with Dulbecco's Phosphate Buffered Saline (D-PBS) and subjected to the following steps of differential centrifugation at 4 °C: i) 70,000 g for 60 min; ii) 100,000 g for 60 min; iii) 200,000 g for 120 min. The pellet obtained from each step was washed with an equal volume of D-PBS and centrifuged again under the same conditions. Supernatants were delipidated and desalted, while pellets (excluded the one obtained at 16,000 g) were purified by SDS-PAGE. Proteins were subjected to enzymatic digestion with PNGase F and trypsin. Deglycosylated tryptic peptides were finally labeled with iTRAQ reagents, fractionated by OFFGEL electrophoresis and analyzed by LC-MS/MS.

incubating for 30 min at room temperature in the dark with gentle agitation. Digestion of N-linked oligosaccharides from glycoproteins was achieved by adding 10 μL of PNGase F from a 0.5 U/μL stock solution. The reaction mixture was incubated at 37 °C for 3 h. Proteins were then digested overnight at 37 °C with 0.2 μg/μL trypsin in 0.1 M TEAB pH 8.0 (protein/trypsin ratio 50:1 w/w). 2.5. iTRAQ labeling and peptide fractionation by OFFGEL electrophoresis Peptides were tagged with iTRAQ Reagents Multiplex Kit (AB Sciex, Foster City, CA, USA). Each sample from the same centrifugal series (supernatant or pellet) was labeled with one of the four isobaric tags, reconstituted in 70 μL of ethanol. The labeling reaction was performed by incubating for 60 min at room temperature and then stopped by adding 8 μL of 5% hydroxylamine and incubating 15 min at room temperature. The four iTRAQ labeled peptide mixtures from each series were then pooled and dried under vacuum. The lyophilized peptides

were dissolved in 400 μL of 5% ACN, 0.1% FA, and loaded onto C18 Macro SpinColumns (Harvard Apparatus, Holliston, MA, USA). Elution was performed with 2 × 200 μL of 50% ACN, 0.1% FA. The samples were then dried under vacuum and dissolved in 360 μL of deionized water. A solution containing 6% glycerol and 0.3% IPG buffer pH 3–10 (Agilent, Santa Clara, CA, USA) was added to a final volume of 1.8 mL. Peptides were fractionated according to their pI on an Agilent 3100 OFFGEL fractionator using commercial 12 cm IPG pH 3–10 linear strips (GE Healthcare, Waukesha, WI, USA). The strips were rehydrated with 20 μL of rehydration solution (4.8% glycerol, 0.24% IPG buffer pH 3–10) per well. After a 30 min incubation, 150 μL of the sample solution was loaded per well. The isoelectric focalization was carried out at 20 °C until a total voltage of 20 kV/h with a maximum current of 50 μA and a maximum power of 200 mW. After the focalization, peptidic fractions were recovered in separate tubes and pH values were measured to check for the accuracy of the pH gradient. Fractions were then dried under vacuum, dissolved in 300 μL of 5% ACN, 0.1% FA, and loaded onto

N. Lukic et al. / Biochimica et Biophysica Acta 1844 (2014) 1026–1033

C18 Micro SpinColumns (Harvard Apparatus). Elution was performed with 2 × 150 μL of 50% ACN, 0.1% trifluoroacetic acid (TFA). Eluates were dried under vacuum and stored at −20 °C until MS analysis. 2.6. Liquid chromatography-tandem mass spectrometry Lyophilized peptides obtained from OFFGEL fractionation were dissolved in 8 μL of 5% ACN/0.1% FA; 5 μL of the resulting sample was injected for LC-MS/MS analysis. MS analysis was performed on a LTQ Orbitrap velos from Thermo Electron (San Jose, CA) equipped with a NanoAcquity UPLC system from Waters (Milford, MA). Peptides were trapped on a home-made 5 μm 200 Å Magic C18 AQ (Michrom, Auburn, CA) 0.1 × 2 mm pre-column and separated on a home-made 5 μm 100 Å Magic C18 AQ (Michrom) 0.75 × 15 mm column. The analytical separation was run for 65 min using a gradient of H2O/FA 99.9%/ 0.1% (solvent A) and ACN/FA 99.9%/0.1% (solvent B). The gradient was run as follows: 0–1 min 95% A and 5% B, then to 65% A and 35% B at 55 min, and 20% A and 80% B at 65 min at a flow rate of 220 nL/min. For MS survey scans, the Orbitrap (OT) resolution was set to 60,000 and the ion population was set to 5 × 105 with an m/z window from 400 to 2000. A maximum of 3 precursors was selected for both the collision-induced dissociation (CID) in the linear trap quadrupole (LTQ) and the high-energy C-trap dissociation (HCD) with analysis in the OT. For MS/MS in the LTQ, the ion population was set to 7 × 103 (isolation width of 2 m/z) while for MS/MS detection in the OT, it was set to 2 × 105 (isolation width of 2.5 m/z), with a resolution of 7500, first mass at m/z = 100, and a maximum injection time of 750 ms. The normalized collision energies were set to 35% for CID and 60% for HCD. 2.7. Data extraction, database interrogation and relative protein quantification Peak lists were generated from raw data using the embedded software from the instrument vendor (extract_MSN.exe v5.0). After peaklist generation, the CID and HCD spectra were merged for simultaneous identification and quantification by using EasyProtConv [21]. The mgf files, combined from the 12 analyzed OFFGEL fractions, were used for protein identification and quantification with EasyProt software platform v2.2 [21]. For protein identification, parameters were specified as follows: database = uniprot_sprot (2011_2 of 08-Feb-2011); taxonomy = Homo Sapiens; precursor error tolerance = 10 ppm; variable modification = oxidized methionine; fixed modifications = carbamidomethylated cysteine, iTRAQlabeled amino terminus and lysine, deamidated asparagine; enzyme = trypsin; potential missed cleavage = 1; cleavage mode = normal; search round = 1, scoring model = CID_LTQ_scan_LTQ; and instrument type = ESI-LTQ-Orbitrap. Protein and peptide scores were set up to maintain the false positive peptide ratio below 5%. For protein quantification, the isotopic correction was applied to reporter intensities according to the iTRAQ reagent certificate of analysis. For each protein, the mean, the standard deviation, and the coefficient of variation (CV) of relative peptide intensities were obtained for the four experimental conditions by using the EasyProt Libra quantification module [21]. The following inter-experimental ratios were finally calculated: PAC versus CP + BS, CC versus CP + BS and PAC + CC versus CP + BS.

1029

proteins were estimated but not taken into account except for comparison with previously published bile proteins including single peptidebased identifications. Given the heterogeneity of previously published lists, the comparison was made in terms of both genes and gene products in order to minimize the possibility of multiple counting. The “ID mapping” module of the Universal Protein Resource (UniProt) (http:// www.uniprot.org/, UniProt release 2013_03) was used to standardize proteins symbols and to associate them with corresponding gene names, while the data analysis platform Biocompendium (http:// biocompendium.embl.de/, accessed on march 2013) was used to convert gene symbols into protein IDs. The characterization of newly identified proteins was performed by using the “GO Slim Classification” and the “Enrichment Analysis” modules of the WEB-based GEne SeT AnaLysis Toolkit (WebGestalt) (http://bioinfo.vanderbilt.edu/webgestalt/, v1/ 30/2013) [22]. For protein quantification, protein lists were extracted from the EasyProt “Libra Details” module and manually merged. The selection of overexpressed protein was made by applying the following criteria: i) only proteins for which all three ratios (PAC/CP + BS, CC/ CP + BS, PAC + CC/CP + BS) were established, have been considered; ii) in the fraction in which the protein was quantified with the highest PAC + CC/CP + BS ratio, it was required to be N 2, with both t test and Mann–Whitney tests significant at a level b 0.05, while the two other ratios (PAC/CP + BS and CC/CP + BS) were required to be N1.2; iii) in all other quantified fractions, if any, PAC + CC/CP + BS ratios were required to be N 1.2 in at least one additional fraction from the same dataset (supernatants or pellets). The characterization of overexpressed proteins was performed by using the STRING database of known and predicted protein interactions (http://string-db.org/, v9.05) [23], and the MetaCore™ software (release 6.13) licensed from GeneGo (Thomson Reuters, NY, USA). 2.9. Immunoblot analysis For each sample, 2 μL of crude bile, or pellets obtained by centrifuging 10 μL of crude bile for 120 min at 200,000 g, were loaded on NuPAGE Novex 4–12% Bis-Tris pre-cast polyacrylamide gels, 1.5-mm thick, 15-wells (Life Technologies, Carlsbad, CA, USA). According to the manufacturer's instructions, proteins were fractionated in a XCell SureLock™ Mini-Cell Electrophoresis System (Life Technologies) and transferred onto a PVDF membrane (iBlot Transfer Stacks PVDF Regular, Life Technologies) using an iBlot western blotting system (Life Technologies). The membranes were blocked for 60 min with 5% nonfat dried milk in PBS supplemented with 0.05% Tween-20 (PBS-T) and incubated overnight at room temperature with primary antibody diluted in 1% nonfat dried milk/PBS-T. Primary antibodies against olfactomedin-4 (OLFM4), syntenin-2 (SDCB2) and Ras-related C3 botulinum toxin substrate 1 (RAC1) (all three from Abcam, Cambridge, UK) were used at a final dilution of 1 μg/μL. After washing with 1% nonfat dried milk/ PBS-T, the blot was incubated for 60 min at room temperature with horseradish peroxidase (HRP)-conjugated secondary antibody diluted in 1% nonfat dried milk/PBS-T. Finally, the blot was washed with PBS-T and immunogenic proteins were visualized using the ECL Western Blotting Detection Reagents (GE Healthcare). 3. Results 3.1. Protein identification in human bile samples

2.8. Protein list building, comparison and characterization Protein identification lists were extracted from the EasyProt “Protein Details” module, manually merged and compared. In the whole list, proteins identified in more than one fraction were reported mentioning the features (e.g., Score, Coverage, #Peptides, etc.) of the fraction in which each protein was identified with the highest number of unique peptides. For any further analysis, multiple-hit proteins and single-hit proteins identified in more than one fraction were considered. Single-hit

Proteomic analysis of bile supernatant fractions allowed the identification of a total of 1318 proteins, 796 of which were identified on the basis of at least two peptides (n = 423) or one peptide in at least two fractions (n = 373) (Supplementary Table I). The number of proteins identified in individual fractions ranged between 473 and 789, with multiple-hits ranging between 196 and 337. The analysis of pellets provided a total of 1341 protein identifications, 920 of which were identified with at least two peptides (n = 673) or with a single peptide in

1030

N. Lukic et al. / Biochimica et Biophysica Acta 1844 (2014) 1026–1033

multiple fractions (n = 247) (Supplementary Table II). In individual pellets, the numbers of protein identifications ranged between 707 and 838, with a number of multiple-hits ranging between 372 and 458. Finally, the combination of all protein lists from supernatant and pellet fractions allowed to reach a total of 2210 unique proteins, with 1267 of them occurring as multiple hits (n = 884) or single hits from multiple fractions (n = 383) (Supplementary Table III). 3.2. Protein lists comparison Comparison between protein lists obtained from bile supernatants and pellets showed that 449 were common to both fractions, while 869 and 892 were identified only in supernatants or pellets, respectively. The number of proteins characteristic of each fraction corresponded to 347 and 471 if only multiple hits and single hits from multiple fractions were considered (Fig. 2). The complete catalog of proteins obtained was compared with bile proteins identified in previously published proteomic studies [1,3,4,11,18,24–29]. The cross checking for both genes and gene products allowed highlighting 322 new bile proteins, of which 137 and 185 were identified on the basis of at least two peptides or a single peptide from multiple fractions, respectively (Supplementary Table IV). 3.3. Characterization of newly identified bile proteins According to the Gene Ontology (GO) classification, newly identified bile proteins mainly belong to high-density cellular components. The most represented classes of proteins were those associated with membranes and/or with membrane-bounded organelles. In particular, the following categories accounted for more than 63% of all GO terms: membrane, nucleus, membrane-enclosed lumen, endomembrane system, endoplasmic reticulum, mitochondrion, envelope, Golgi apparatus, cell projection, vesicle, endosome and microbody. Macromolecular complexes of proteins, nucleic acids, carbohydrates or lipids, were also highly represented (13%). Among other categories, cytosol was the fourth represented accounting for 9% of all GO terms, followed by unclassified proteins (4%), cytoskeleton (3%), ribosome (3%) and extracellular space (1%) (Fig. 3). 3.4. Detection of overexpressed proteins in malignant bile samples The number of proteins for which all three ratios were established (PAC + CC/CP + BS, PAC/CP + BS and CC/CP + BS) varied between 223 and 364 in supernatant fractions and between 359 and 444 in pellet fractions. Overall, a total of 942 proteins were quantified: 488 in

Fig. 2. Comparison of proteins identified in bile supernatant and pellet fractions. Venn diagram showing the overlap between the total number of biliary proteins identified in supernatants (left) and in pellets (right). Darker circles represent proteins identified with more than one peptide or with a single peptide in more than one fraction.

Fig. 3. Gene Ontology classification of newly identified bile proteins. Bar chart showing the cellular component categories associated to newly described bile proteins according to the “GO Slim Classification” module of the WEB-based GEne SeT AnaLysis Toolkit (WebGestalt) [22].

supernatants and 658 in pellets. The selection of overexpressed proteins highlighted 147 proteins with a PAC + CC/CP + BS ratio N2 and both of the two other ratios (PAC/CP + BS and CC/CP + BS) N1.2 in at least one fraction (data not shown). Finally, by applying more stringent criteria to this list (i.e., PAC + CC/CP + BS ratio N 1.2 in at least one additional fraction from the same dataset), the number of overexpressed proteins was 104, including 16 that were newly identified in bile (Supplementary Table V). In order to avoid missing potential biomarkers, the overexpression in all fractions was not required. It is known from the literature, that proteins released by tumor tissues can be transported in body fluids through microvesicles (e.g., oncosomes) [30,31]. These structures, having specific densities, could be mainly found in a few, or a single, centrifugal fraction. 3.5. Characterization of overexpressed proteins The STRING analysis, performed to highlight physical and functional interactions, showed several interrelated nodes connecting 69 out of 104 proteins overexpressed in malignant samples. Among these nodes, the one with the more functional interactions converged on RAC1 (Supplementary Fig. I). The further MetaCore enrichment analysis, performed on the basis of the “Disease Biomarker Networks” functional ontology, highlighted the “Breast neoplasm Regulation of Progression Through Cell Cycle” and the “Breast Neoplasm Cell Motility” as the most significantly enriched networks, with p value = 6.325−3 and 7.889−3, respectively. The protein RAC1 proved to be involved in both networks, with several other partners also found overexpressed in malignant conditions (e.g., cell division control protein 42 homolog [protein ID: CDC42, gene name: CDC42], moesin [protein ID: MOES, gene name: MSN], cofilin [protein ID:COF1, gene name: CFL1]) (Supplementary Figs. II and III). Finally, the comparison with the compendium published by Harsha et al. in 2009 [32] revealed that 46 out of 104 overexpressed proteins were previously described as potential biomarkers of pancreatic cancer (Supplementary Table V). 3.6. Immunoblot verification Immunoblot verification of three cancer-associated proteins, OLFM4, SDCB2 and RAC1, was performed on the 8 malignant (PAC,

N. Lukic et al. / Biochimica et Biophysica Acta 1844 (2014) 1026–1033

n = 4; CC, n = 4) and 5 nonmalignant samples (CP, n = 3; BS, n = 2). OLFM4 was found overexpressed in all pellet fractions by quantitative proteomic analysis with similar ratios in PAC and CC samples (Fig. 4a). Immunoblots performed on crude bile samples confirmed OLFM4 overexpression, showing a strong signal in 7/8 malignant samples (4/ 4 PAC and 3/4 CC). A faint signal was observed in one CC sample and in one nonmalignant (BS) sample. SDCB2 was found overexpressed in two out of three pellet fractions (70,000 g and 100,000 g), with PAC/ CP + BS ratios about 3–5 times higher than CC/CP + BS ratios (Fig. 4a). The immunoblot analysis of crude bile showed a clear overexpression in 4 malignant samples (2/4 PAC and 2/4 CC), but the remaining signals appeared blurred by the background. The immunodetection of SDCB2 in pellets obtained from crude bile centrifugation at 200,000 g was more informative, showing strong signals in 6/8 malignant samples (4/4 PAC and 2/4 CC). Two very faint signals could also be glimpsed in the remaining CC samples, while no nonmalignant sample showed positivity. Finally, RAC1 was found overexpressed in the 70,000 g pellet fraction, showing a PAC/CP + BS ratio 5 times higher than the CC/CP + BS ratio (Fig. 4a). Immunoblot performed on bile pellets revealed a pair of contiguous bands in 3/4 PAC samples. In contrast, a very faint signal was detectable in the remaining PAC and in only one CC samples, at an intensity similar to that of the weakest PAC sample. The remaining CCs and the nonmalignant samples were negatives to RAC1 immunodetection. For all three proteins, the results of relative MS quantification strongly correlated with immunoblot data. 4. Discussion Differential centrifugation has been established as a technique for the separation of cell organelles. The principle was first applied in 1932 to the isolation of nuclei from calf heart, muscle, and thymus [33] and subsequently improved to separate major cell components [34,35]. As a function of their size and density, the subcellular constituents sediment in different centrifugal fractions depending on the force of gravity applied. In a previous work, we showed that differential centrifugation can also be used to decrease the complexity of human bile prior to proteomic analysis [18]. This technique has the advantage of fractionating the whole proteome while minimizing protein losses, which leads to a significant increase in protein identifications. Given the limited number of bile proteins quantified by proteomic studies to

1031

date, the purpose of the present work was to introduce differential centrifugation as a sample pre-treatment for bile sub-proteome comparisons. This strategy was applied on bile samples collected from patients with malignant and nonmalignant biliary stenosis, allowing the identification of 2210 proteins. Interestingly, most of these proteins were specific of pellet or supernatant fractions, while a limited number was common to both groups of fractions, highlighting the contribution of the centrifugal fractionation in achieving complementary identifications. Moreover, 332 proteins were identified that had never been described in bile before, confirming that our approach was also able to extend the knowledge of bile proteome. The characterization of newly identified proteins revealed their predominant association with highdensity membranes or membrane-bounded organelles. This result underlines that investigating pellet fractions, as well as introducing SDS-PAGE as an intermediary step before proteomic analysis, allows the detection of proteins that might be missed by traditional approaches in virtue of their hydrophobicity and/or low abundance. The subsequent quantitative analysis led to a 5-fold increase in the number of proteins quantified compared to previous comparative proteomic approaches [2,12,13]. Among proteins classified as overexpressed in malignant conditions, many of them were previously described as associated with cancer. RAC1 protein and some of its partners were specifically recognized to be involved in breast cancer neoplastic processes. Moreover, several bile markers previously proposed for differentiating malignant from nonmalignant stenosis were found overexpressed in the present analysis; namely, neutrophil gelatinase-associated lipocalin (NGAL), carcinoembryonic antigen-related cell adhesion molecule 6 (CEAM6) and galectin-3-binding protein (LG3BP). NGAL levels in bile were shown to have a good accuracy in distinguishing malignant from nonmalignant cases with a receiver operator characteristic area under the curve (ROC-AUC) of 0.74, a sensitivity of 92% and a specificity of 47% [2]. Bile CEAM6 was recently demonstrated by our group as another reliable biomarker of malignant biliary stenosis, showing higher diagnostic performances: 0.92 ROC-AUC, 93.1% sensitivity, 83.3% specificity, 93.1% positive predictive value, 83.3% negative predictive value and 90.2% accuracy [13]. Finally, bile LG3BP (also named MAC-2BP) showed the following performances: 0.70 ROC-AUC, 69% sensitivity and 67% specificity [36]. These data strongly support the reliability of our approach in the context of the quantitative detection of bile biomarkers.

Fig. 4. Overexpression of selected proteins. a) Protein expression ratios determined by iTRAQ-based relative quantification on pellet fractions obtained by differential centrifugation of crude bile (70,000 g; 100,000 g and 200,000 g). b) Protein expression intensities visualized by immunoblot analysis of crude bile and pellet fractions (obtained by centrifuging crude bile at 200,000 g). The asterisked lanes 1, 7, 9 and 13 correspond, respectively, to the PAC, CP, BS and CC samples from which iTRAQ ratios were obtained. OLFM4, olfactomedin-4; SDCB2, syntenin-2; RAC1, Ras-related C3 botulinum toxin substrate 1; PAC, pancreatic adenocarcinoma; CP, chronic pancreatitis; BS, biliary stones; CC, cholangiocarcinoma. A color code was used to differentiate malignant (red) and nonmalignant (green) samples.

1032

N. Lukic et al. / Biochimica et Biophysica Acta 1844 (2014) 1026–1033

Immunoblot analysis confirmed proteomic results for three cancerassociated proteins: OLFM4, SDCB2, and RAC1. OLFM4 is a glycoprotein highly overexpressed in several malignancies including some digestive tumors (e.g., pancreatic, gastric and colon cancers) [37–39]. Kobayashi et al., in 2007, demonstrated that OLFM4 promotes proliferation of pancreatic cancer cells by favoring transition from the S to G(2)/M phase. The anti-apoptotic activity of OLFM4 in gastric cancer cells and its correlation with tumors differentiation and progression were also established [40–44]. In cell lysates, OLFM4 has been detected in monomeric (64 kDa), dimeric (120 kDa) and multimeric (N188 kDa) forms, predominantly localized to membrane-bound organelles [41,45,46], while, in cell culture supernatants, secreted OLFM4 was observed only in the high molecular weight multimeric form [45,46]. In our analysis, this protein was identified in all high-density bile pellet fractions with elevated ratios in malignant versus nonmalignant samples. Immunoblot verification confirmed overexpression in malignant samples. SDCB2 gene expression was found to be increased in invasive ductal adenocarcinoma compared to normal whole pancreas from organ donors [32,47]. PDZ-containing proteins are involved in cell signaling and play important roles in the assembly of supramolecular complexes. SDCB2, particularly, binds to phosphatidylinositol 4, 5-bisphosphate (PIP2) and localizes at PIP2 rich cellular locations, i.e. plasma membrane, nucleoli and nuclear speckles [48]. According to a recent characterization of nuclear dynamics, three forms of SDCBP2 were recognized in nucleoplasm and nucleoli: i) a freely diffusing monomeric form; ii) a major slowly diffusing fraction interacting with nuclear structures (i.e. chromatin); iii) a 5–8% immobile fraction showing long term binding to nuclear structures [48]. In the present quantitative analysis, SDCBP2 was identified in three pellet fractions, but overexpression was observed only in two of them. Immunoblot analysis, performed on crude samples and/ or pellet fractions from bile centrifuged at 200,000 g, allowed the detection of SDCBP2 signals in all malignant samples. Finally, RAC1 is a member of the Ras superfamily whose upregulation has been observed in breast cancer and several other malignancies (e.g. colorectal cancer, respiratory papilloma, lung carcinoma and gastric cancer) [49–53]. RAC1 is a GTP-binding protein involved in membrane ruffling and lamellipodia formation, as well as in the activation of pinocytosis and the formation of superoxide [54]. In the context of tumorigenesis, RAC1 has been suggested to play a role in proliferation, tumor progression, cancer cell migration and secondary tumor establishment [55,56]. The protein sequence contains a C-terminal motif allowing the interaction with plasma membrane. The self-association of RAC1 to form homodimers and oligomers is also mediated by this region [57]. Interestingly, RAC1-effector complexes have been investigated by centrifugal fractionation. As a result, RAC1 was found to sediment as a 60 kDa molecular complex in both, cytosolic and membrane fractions, while additional larger macromolecular complexes were detected in membrane fractions only. Immunoblot analysis of membrane-extracted fractions, revealed two contiguous bands, both detected by anti-RAC1 antibodies [58]. In our proteomic study, RAC1 was identified in the pellet fractions obtained at 70,000 g. Immunoblot analysis highlighted two distinct bands strongly overexpressed in 3 out of 4 PAC, while faint signals were detected in the remaining PAC and in 1 out of 4 CC. We may hypothesize that contiguous bands represent different isoforms of RAC1, but further investigation will be necessary to clarify the exact nature of these two proteins. 5. Conclusions The first comparative analysis of bile centrifugal fractions from malignant and nonmalignant biliary stenoses is presented here, highlighting the compatibility of differential centrifugation with quantitative proteomic analysis. This integrated approach proved to be robust and reliable, allowing to: i) extend the knowledge of the bile proteome; ii) increase the number of quantified bile proteins with respect to quantitative proteomic approaches attempted so far; iii) provide a consistent

quantification of low-abundant proteins present in high-density bile fractions; and iv) identify new overexpressed proteins in malignant samples. Immunoblot analysis of crude bile samples confirmed the overexpression of several proteins in cancer. Our findings represent a starting point for further clinical validation of proteins overexpressed in malignant bile duct stenosis and give a new impetus to quantification studies of bile proteins expression in different pathological conditions affecting the biliary tract. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.bbapap.2013.06.023. Acknowledgements We thank the Proteomics Core Facility at the Faculty of Medicine in Geneva for mass spectrometry analysis and Drs. Patrice François, Adrian Fisher and Myriam Girard, at the Genomic Research Laboratory of the Geneva University Hospitals, for their valuable technical support. Funding This project was supported by funds from “Fondation Ernst et Lucie Schmidheiny” to A.F. Funding source had no role in design, execution and interpretation of the results incorporated in the manuscript. Author contributions N.L. and R.V. — workflow management and acquisition of data; M.D. and J.L.F. — providing samples, critical revision of the manuscript for important intellectual content; P.L. — study concept and supervision, critical revision of the manuscript for important intellectual content; J.M.D. — study concept and supervision, providing samples, critical revision of the manuscript for important intellectual content; A.F. — study concept and supervision, study design, analysis and interpretation of data, manuscript writing, obtaining funding. Competing interests The authors declare that they have no competing interests. References [1] A. Farina, J.M. Dumonceau, J.L. Frossard, A. Hadengue, D.F. Hochstrasser, P. Lescuyer, Proteomic analysis of human bile from malignant biliary stenosis induced by pancreatic cancer, J. Proteome Res. 8 (2009) 159–169. [2] A.A. Zabron, V.M. Horneffer-van der Sluis, C.A. Wadsworth, F. Laird, M. Gierula, A.V. Thillainayagam, P. Vlavianos, D. Westaby, S.D. Taylor-Robinson, R.J. Edwards, S.A. Khan, Elevated levels of neutrophil gelatinase-associated lipocalin in bile from patients with malignant pancreatobiliary disease, Am. J. Gastroenterol. 106 (2011) 1711–1717. [3] T.Z. Kristiansen, J. Bunkenborg, M. Gronborg, H. Molina, P.J. Thuluvath, P. Argani, M.G. Goggins, A. Maitra, A. Pandey, A proteomic analysis of human bile, Mol. Cell Proteomics 3 (2004) 715–728. [4] S.G. Farid, R.A. Craven, J. Peng, G.K. Bonney, D.N. Perkins, P.J. Selby, K. Rajendra Prasad, R.E. Banks, Shotgun proteomics of human bile in hilar cholangiocarcinoma, Proteomics 11 (2011) 2134–2138. [5] G.K. Bonney, R.A. Craven, R. Prasad, A.F. Melcher, P.J. Selby, R.E. Banks, Circulating markers of biliary malignancy: opportunities in proteomics? Lancet Oncol. 9 (2008) 149–158. [6] A. Farina, J.M. Dumonceau, P. Lescuyer, Proteomic analysis of human bile and potential applications for cancer diagnosis, Expert. Rev. Proteomics 6 (2009) 285–301. [7] P. Charatcharoenwitthaya, F.B. Enders, K.C. Halling, K.D. Lindor, Utility of serum tumor markers, imaging, and biliary cytology for detecting cholangiocarcinoma in primary sclerosing cholangitis, Hepatology 48 (2008) 1106–1117. [8] J.M. Dumonceau, T. Koessler, J.E. van Hooft, P. Fockens, Endoscopic ultrasonographyguided fine needle aspiration: relatively low sensitivity in the endosonographer population, World J. Gastroenterol. 18 (2012) 2357–2363. [9] J.M. Dumonceau, C. Macias Gomez, C. Casco, M. Genevay, M. Marcolongo, M. Bongiovanni, P. Morel, P. Majno, A. Hadengue, Grasp or brush for biliary sampling at endoscopic retrograde cholangiography? A blinded randomized controlled trial, Am. J. Gastroenterol. 103 (2008) 333–340. [10] J.M. Dumonceau, M. Polkowski, A. Larghi, P. Vilmann, M. Giovannini, J.L. Frossard, D. Heresbach, B. Pujol, G. Fernandez-Esparrach, E. Vazquez-Sequeiros, A. Gines,

N. Lukic et al. / Biochimica et Biophysica Acta 1844 (2014) 1026–1033

[11]

[12]

[13]

[14]

[15]

[16] [17]

[18]

[19]

[20] [21]

[22] [23]

[24]

[25]

[26]

[27]

[28] [29]

[30] [31]

[32]

Indications, results, and clinical impact of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline, Endoscopy 43 (2011) 897–912. M.A. Barbhuiya, N.A. Sahasrabuddhe, S.M. Pinto, B. Muthusamy, T.D. Singh, V. Nanjappa, S. Keerthikumar, B. Delanghe, H.C. Harsha, R. Chaerkady, V. Jalaj, S. Gupta, B.R. Shrivastav, P.K. Tiwari, A. Pandey, Comprehensive proteomic analysis of human bile, Proteomics 11 (2011) 4443–4453. A.A. Zabron, V.M. Horneffer-Van der Sluis, M. Geirula, C.A. Wadsworth, F. Laird, A.V. Thillainayagam, P. Vlavianos, D. Westaby, S.D. Taylor-Robinson, R.J. Edwards, S.A. Khan, Proteomic analysis of bile in benign and malignant pancreatobiliary disease, Gut 59 (2010) A82–A83. A. Farina, J.-M. Dumonceau, P. Antinori, I. Annessi-Ramseyer, J.-L. Frossard, D.F. Hochstrasser, M. Delhaye, P. Lescuyer, Bile carcinoembryonic cell adhesion molecule 6 (CEAM6) as a biomarker of malignant biliary stenoses, BBA — Proteins Proteomics (2013). http://dx.doi.org/10.1016/j.bbapap.2013.06.010. [Epub ahead of print]. Y. Fang, D.P. Robinson, L.J. Foster, Quantitative analysis of proteome coverage and recovery rates for upstream fractionation methods in proteomics, J. Proteome Res. 9 (2010) 1902–1912. P.G. Righetti, A. Castagna, P. Antonioli, E. Boschetti, Prefractionation techniques in proteome analysis: the mining tools of the third millennium, Electrophoresis 26 (2005) 297–319. S. Surinova, R. Schiess, R. Huttenhain, F. Cerciello, B. Wollscheid, R. Aebersold, On the development of plasma protein biomarkers, J. Proteome Res. 10 (2011) 5–16. G. Zhang, D. Fenyo, T.A. Neubert, Evaluation of the variation in sample preparation for comparative proteomics using stable isotope labeling by amino acids in cell culture, J. Proteome Res. 8 (2009) 1285–1292. A. Farina, J.M. Dumonceau, M. Delhaye, J.L. Frossard, A. Hadengue, D.F. Hochstrasser, P. Lescuyer, A step further in the analysis of human bile proteome, J. Proteome Res. 10 (2011) 2047–2063. V. Fetaud-Lapierre, C.M. Pastor, A. Farina, D.F. Hochstrasser, J.L. Frossard, P. Lescuyer, Proteomic analysis of heat shock-induced protection in acute pancreatitis, J. Proteome Res. 9 (2010) 5929–5942. U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. F. Gluck, C. Hoogland, P. Antinori, X. Robin, F. Nikitin, A. Zufferey, C. Pasquarello, V. Fetaud, L. Dayon, M. Muller, F. Lisacek, L. Geiser, D. Hochstrasser, J.C. Sanchez, A. Scherl, EasyProt — an easy-to-use graphical platform for proteomics data analysis, J. Proteomics 79C (2012) 146–160. B. Zhang, S. Kirov, J. Snoddy, WebGestalt: an integrated system for exploring gene sets in various biological contexts, Nucleic Acids Res. 33 (2005) W741–W748. B. Snel, G. Lehmann, P. Bork, M.A. Huynen, STRING: a web-server to retrieve and display the repeatedly occurring neighbourhood of a gene, Nucleic Acids Res. 28 (2000) 3442–3444. L. Guerrier, S. Claverol, L. Finzi, F. Paye, F. Fortis, E. Boschetti, C. Housset, Contribution of solid-phase hexapeptide ligand libraries to the repertoire of human bile proteins, J. Chromatogr. A 1176 (2007) 192–205. H. Zhou, B. Chen, R.X. Li, Q.H. Sheng, S.J. Li, L. Zhang, L. Li, Q.C. Xia, H.Y. Wang, R. Zeng, Large-scale identification of human biliary proteins from a cholesterol stone patient using a proteomic approach, Rapid Commun. Mass Spectrom. 19 (2005) 3569–3578. L. Reinhard, C. Rupp, H.D. Riedel, T. Ruppert, T. Giese, C. Flechtenmacher, K.H. Weiss, P. Kloeters-Plachky, W. Stremmel, P. Schirmacher, P. Sauer, D.N. Gotthardt, S100A9 is a biliary protein marker of disease activity in primary sclerosing cholangitis, PLoS One 7 (2012) e29821. J. Shen, W. Wang, J. Wu, B. Feng, W. Chen, M. Wang, J. Tang, F. Wang, F. Cheng, L. Pu, Q. Tang, X. Wang, X. Li, Comparative proteomic profiling of human bile reveals SSP411 as a novel biomarker of cholangiocarcinoma, PLoS One 7 (2012) e47476. W. Wang, K.X. Ai, Z. Yuan, X.Y. Huang, H.Z. Zhang, Different expression of S100A8 in malignant and benign gallbladder diseases, Dig. Dis. Sci. 58 (2013) 150–162. D. Zhang, J. Xiang, L. Wang, Z. Xu, L. Sun, F. Zhou, X. Zha, D. Cai, Comparative proteomic analysis of gallbladder bile proteins related to cholesterol gallstones, PLoS One 8 (2013) e54489. N.S. Barteneva, E. Fasler-Kan, M. Bernimoulin, J.N. Stern, E.D. Ponomarev, L. Duckett, I.A. Vorobjev, Circulating microparticles: square the circle, BMC Cell Biol. 14 (2013) 23. T.H. Lee, E. D'Asti, N. Magnus, K. Al-Nedawi, B. Meehan, J. Rak, Microvesicles as mediators of intercellular communication in cancer—the emerging science of cellular ‘debris’, Semin. Immunopathol. 33 (2011) 455–467. H.C. Harsha, K. Kandasamy, P. Ranganathan, S. Rani, S. Ramabadran, S. Gollapudi, L. Balakrishnan, S.B. Dwivedi, D. Telikicherla, L.D. Selvan, R. Goel, S. Mathivanan, A. Marimuthu, M. Kashyap, R.F. Vizza, R.J. Mayer, J.A. Decaprio, S. Srivastava, S.M. Hanash, R.H. Hruban, A. Pandey, A compendium of potential biomarkers of pancreatic cancer, PLoS Med. 6 (2009) e1000046.

1033

[33] M. Behrens, Untersuchungen an isolierten Zell- und Gewebsbestandteilen. I. Mitteilung: Isolierung von Zellkernen des Kalbsherzmuskels, in: F. Hoppe-Seyler (Ed.), Zeitschrift für physiologische Chemie, vol. 209, K.J. Trübner, Strassburg, 1932, pp. 59–74. [34] A. Claude, Fractionation of mammalian liver cells by differential centrifugation; experimental procedures and results, J. Exp. Med. 84 (1946) 61–89. [35] A. Claude, Fractionation of mammalian liver cells by differential centrifugation; problems, methods, and preparation of extract, J. Exp. Med. 84 (1946) 51–59. [36] J. Koopmann, P.J. Thuluvath, M.L. Zahurak, T.Z. Kristiansen, A. Pandey, R. Schulick, P. Argani, M. Hidalgo, S. Iacobelli, M. Goggins, A. Maitra, Mac-2-binding protein is a diagnostic marker for biliary tract carcinoma, Cancer 101 (2004) 1609–1615. [37] S. Koshida, D. Kobayashi, R. Moriai, N. Tsuji, N. Watanabe, Specific overexpression of OLFM4(GW112/HGC-1) mRNA in colon, breast and lung cancer tissues detected using quantitative analysis, Cancer Sci. 98 (2007) 315–320. [38] N. Oue, K. Sentani, T. Noguchi, S. Ohara, N. Sakamoto, T. Hayashi, K. Anami, J. Motoshita, M. Ito, S. Tanaka, K. Yoshida, W. Yasui, Serum olfactomedin 4 (GW112, hGC-1) in combination with Reg IV is a highly sensitive biomarker for gastric cancer patients, Int. J. Cancer 125 (2009) 2383–2392. [39] A. Cuoghi, A. Farina, K. Z'Graggen, J.M. Dumonceau, A. Tomasi, D.F. Hochstrasser, M. Genevay, P. Lescuyer, J.L. Frossard, Role of proteomics to differentiate between benign and potentially malignant pancreatic cysts, J. Proteome Res. 10 (2011) 2664–2670. [40] K.K. Kim, K.S. Park, S.B. Song, K.E. Kim, Up regulation of GW112 Gene by NF kappaB promotes an antiapoptotic property in gastric cancer cells, Mol. Carcinog. 49 (2010) 259–270. [41] X. Zhang, Q. Huang, Z. Yang, Y. Li, C.Y. Li, GW112, a novel antiapoptotic protein that promotes tumor growth, Cancer Res. 64 (2004) 2474–2481. [42] W. Liu, J. Zhu, L. Cao, G.P. Rodgers, Expression of hGC-1 is correlated with differentiation of gastric carcinoma, Histopathology 51 (2007) 157–165. [43] L. Yu, M. He, Z. Yang, G. Chen, M. Li, L. Wang, S. Chen, Olfactomedin 4 is a marker for progression of cervical neoplasia, Int. J. Gynecol. Cancer 21 (2011) 367–372. [44] L. Yu, L. Wang, S. Chen, Olfactomedin 4, a novel marker for the differentiation and progression of gastrointestinal cancers, Neoplasma 58 (2011) 9–13. [45] P.K. Grover, J.E. Hardingham, A.G. Cummins, Stem cell marker olfactomedin 4: critical appraisal of its characteristics and role in tumorigenesis, Cancer Metastasis Rev. 29 (2010) 761–775. [46] W. Liu, L. Chen, J. Zhu, G.P. Rodgers, The glycoprotein hGC-1 binds to cadherin and lectins, Exp. Cell Res. 312 (2006) 1785–1797. [47] A.W. Lowe, M. Olsen, Y. Hao, S.P. Lee, K. Taek Lee, X. Chen, M. van de Rijn, P.O. Brown, Gene expression patterns in pancreatic tumors, cells and tissues, PLoS One 2 (2007) e323. [48] A. Geeraerts, F. Hsiu-Fang, P. Zimmermann, Y. Engelborghs, The characterization of the nuclear dynamics of syntenin-2, a PIP(2) binding PDZ protein, Cytometry A (2013). http://dx.doi.org/10.1002/cyto.a.22246. [Epub ahead of print]. [49] A. Schnelzer, D. Prechtel, U. Knaus, K. Dehne, M. Gerhard, H. Graeff, N. Harbeck, M. Schmitt, E. Lengyel, Rac1 in human breast cancer: overexpression, mutation analysis, and characterization of a new isoform, Rac1b, Oncogene 19 (2000) 3013–3020. [50] S.Y. Zhao, Y. Sun, Z.S. Lai, Q.Z. Nan, K. Li, Z.S. Zhang, Inhibition of migration and invasion of colorectal cancer cells via deletion of Rac1 with RNA interference, Mol. Cell. Biochem. 322 (2009) 179–184. [51] R. Wu, S.J. Coniglio, A. Chan, M.H. Symons, B.M. Steinberg, Up-regulation of Rac1 by epidermal growth factor mediates COX-2 expression in recurrent respiratory papillomas, Mol. Med. 13 (2007) 143–150. [52] A. Gastonguay, T. Berg, A.D. Hauser, N. Schuld, E. Lorimer, C.L. Williams, The role of Rac1 in the regulation of NF-kappaB activity, cell proliferation, and cell migration in non-small cell lung carcinoma, Cancer Biol. Ther. 13 (2012) 647–656. [53] H. Zhan, H. Liang, X. Liu, J. Deng, B. Wang, X. Hao, Expression of Rac1, HIF-1alpha, and VEGF in gastric carcinoma: correlation with angiogenesis and prognosis, Onkologie 36 (2013) 102–107. [54] A. Best, S. Ahmed, R. Kozma, L. Lim, The Ras-related GTPase Rac1 binds tubulin, J. Biol. Chem. 271 (1996) 3756–3762. [55] N.A. Mack, H.J. Whalley, S. Castillo-Lluva, A. Malliri, The diverse roles of Rac signaling in tumorigenesis, Cell Cycle 10 (2011) 1571–1581. [56] E. Wertheimer, A. Gutierrez-Uzquiza, C. Rosemblit, C. Lopez-Haber, M.S. Sosa, M.G. Kazanietz, Rac signaling in breast cancer: a tale of GEFs and GAPs, Cell. Signal. 24 (2012) 353–362. [57] D. Sun, D. Xu, B. Zhang, Rac signaling in tumorigenesis and as target for anticancer drug development, Drug Resist. Updat. 9 (2006) 274–287. [58] M.D. Hansen, W.J. Nelson, Serum-activated assembly and membrane translocation of an endogenous Rac1:effector complex, Curr. Biol. 11 (2001) 356–360.