HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) All Versions of this Article: 19/9/1125 04-3000fjev1    most recent Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mauri, P. Articles by Sorio, C. Search for Related Content PubMed PubMed Citation Articles by Mauri, P. Articles by Sorio, C.

Identification of proteins released by pancreatic cancer cells by multidimensional protein identification technology: a strategy for identification of novel cancer markers

Pierluigi Mauri^*,1, Aldo Scarpa^,1, Anna Chiara Nascimbeni, Louise Benazzi^*, Emanuela Parmagnani, Andrea Mafficini, Marco Della Peruta, Claudio Bassi, Kaoru Miyazaki and Claudio Sorio^,2

^* Istituto Tecnologie Biomediche (ITB-CNR), Segrate, Milano, Italy;
Dipartimento di Patologia, and
Dipartimento di Scienze Chirurgiche e Gastroenterologiche, Università di Verona,Verona, Italy;
Kihara Institute for Biological Research, Yokohama City University, Yokohama, Japan

² Correspondence: Università di Verona, Dipartimento di Patologia, Sezione di Patologia Generale, Strada Le Grazie, Verona 37134, Italy. E-mail: claudio.sorio{at}univr.it

SPECIFIC AIMS

The aims of this study were to 1) find an efficient and rapid approach to identify proteins released by cancer cells using small sample volumes and automated technology; and 2) identify novel proteins released by pancreatic cancer cells using culture supernatant and validate the findings on a panel of representative cell lines, primary cancers, and xenografted cancer cells.

PRINCIPAL FINDINGS

1. Optimal conditions and time point for collection of supernatants
To determine optimal treatment and supernatant collection times, multidimensional protein identification technology (MudPIT) was applied to different preparations. The protocol showing the lower number of cytosolic proteins, suggestive of cell damage, was from Kratchmarova et al. with modifications, consisting of 6 gentle washes with serum-free medium, followed by 18 h incubation in serum-free medium. We directly subjected 50 µL of the 18 h conditioned medium to protease digestion and 10 µL were analyzed by MudPIT.

2. MudPIT identified 30 proteins released by resting pancreatic cancer Suit-2 cells
Analysis of serum-free supernatant (10 µL) from Suit-2 cells consistently identified from 4 independent experiments 30 proteins, some of which linked to relevant cancer cell features such as modification of the extracellular environment and angiogenesis. Fifteen of these are classified as secreted proteins in public databases. Indicating the validity of this approach, 8 have already been associated with pancreatic cancer and include MMP-1, MMP-7, TIMP1, SERPINE2, TGFBI, Mac-2 binding protein, clusterin, and glycerol-3-phosphate dehydrogenase. Many of these proteins are thought to play a role in ECM degradation/remodeling and thus promote invasion and metastasis. The remaining 7 proteins had not been previously associated with pancreatic cancer cells; these included CSPG2/versican, Mac25/angiomodulin, IGFBP-1, HSPG2/perlecan, FAM3C, APLP2, and cyclophilin B. We validated our MudPIT findings for two of these, CSPG2/versican and Mac25/angiomodulin, as they were among the proteins most represented in our samples according to the number of peptides detected in supernatants.

3. CSPG2/versican is released by pancreatic cancer cell lines and primary adenocarcinoma cells
CSPG2/versican mRNA was detected by RT-PCR in 5 of 6 pancreatic cancer cell lines probed (Fig. 1 C). In primary pancreatic cancers, a high amount of CSPG2/versican was detected in the desmoplastic stroma while cancer cells were immunonegative (Fig. 1A ). To demonstrate that CSPG2/versican is released by cancer cells in vivo, we set up an experimental model consisting in Suit-2 cells resuspended in anamorphous matrix (Matrigel®), xenografted in the flank of nu/nu mice, and allowed to proliferate for 1 wk. The formalin-fixed paraffin-embedded implant was then immunostained with an anti-human versican antibody. The analysis clearly demonstrated that secretion does occur, as the proteoglycan accumulates at the interface between cells and the matrix before diffusing in the matrix itself (Fig. 1B ). The cells themselves did not stain with the antibody, indicating a low intracellular accumulation of the proteoglycan compatible with rapid release.

View larger version (68K): [in this window] [in a new window]   Figure 1. CSPG2/versican and Mac25/angiomodulin expression in pancreatic adenocarcinoma. CSPG2/versican: A) immunohistochemistry of primary pancreatic adenocarcinoma shows strongly positive peritumoral stroma while cancer cells are immunonegative (arrow); B) the same pattern of reactivity is detected in Suit-2 cells implanted in nu/nu mice in a Matrigel® matrix, showing strong immunostain with anti-versican antibody at the cell-matrix interface, while cancer cells expression is undetectable, demonstrating that cancer cells produce and immediately release the protein; and C) RT-PCR analysis shows the presence of versican transcripts in 5 of 6 cell lines, T3M4 being negative. Actin expression is shown to demonstrate equal amount of starting RNA. Mac25/angiomodulin: D) cancer cells in primary pancreatic adenocarcinoma show strong cytoplasmic immunostaining; E) immunohistochemistry of Suit-2 cells implanted in nude mice in Matrigel® matrix, showing immunostain with anti-mac25 antibody in the cytoplasm of cells and matrix; and F) Western blot analysis of supernatants showing that angiomodulin is released by 5 of 6 pancreatic cancer cell lines (anti-cdc42 was the negative control antibody, and Ponceau staining indicates the relative amount of proteins loaded on each lane).

4. Mac25/angiomodulin is a major secreted protein in pancreatic cancer cell lines and is overexpressed in primary pancreatic cancer
Western blot analysis showed that 5 of 6 pancreatic cancer cell lines released Mac25 protein in the supernatant (Fig. 1F ); the presence of the protein within the cells was confirmed by immunoprecipitation and Western blot analysis of cell lysates (not shown). Immunohistochemistry showed that Mac25 was clearly expressed in primary pancreatic cancer cells (Fig. 1D ) and xenografted Suit-2 (Fig. 1E ). In cancer tissues, Mac25/angiomodulin positivity was also associated with small vessels (Fig. 2 ). Normal pancreas showed Mac25 expression in the insulae of Langerhans, while a faint signal was present in small ducts (not shown).

View larger version (51K): [in this window] [in a new window]   Figure 2. Schematic diagram of the strategy used to discover proteins released by cancer cells. Serum-free supernatant from a cancer cell line is subjected to MudPIT analysis. The approach involves the generation of peptides from enzymatic digestion of a complex protein mixture, their separation by means of two micro-HPLC columns (cation exchange and reversed phase, 2DC) and direct analysis of eluted peaks by tandem mass spectrometry. The identification of the corresponding proteins is then obtained through an automated database search with appropriate software, such as the SEQUEST algorithm for data handling of mass spectra. The presence and localization of the proteins identified is then searched in primary cancers and body fluids.

5. Additional proteins were identified upon phorbol ester activation of cancer cells
To maximize detection of secreted polypeptides, we treated Suit-2 cells with phorbol ester/ionophore, which induces changes in the release/synthesis of proteins. The MudPIT analysis of supernatant of activated Suit-2 reproducibly identified, from 4 independent experiments, 16 proteins in addition to those detected in resting conditions. Among these, 6 proteins are classified as secreted in public databases. Of these, uPA, TSP-1, and syndecan-1 have already been associated with pancreatic cancer whereas syndecan 4, ß2 microglobulin, and ICA69 have not.

6. Score values assigned by SEQUEST software to proteins identified by MudPIT correlates with the amount of proteins in the sample
SEQUEST algorithm assigns to each identified protein a score value that corresponds to a confidence index in the matching between experimental MS spectra and theoretical spectra obtained from the database. However, the variation in score values for several proteins in the supernatants of activated Suit-2 with respect to those of resting cells suggested that SEQUEST score value might be correlated with the amount of protein in the sample.

To verify this, we focused on three proteins—Mac25/angiomodulin, CSPG2/versican, and MMP-1—whose score values were decreased, unchanged, and increased in activated Suit-2 cells, respectively. We extracted the ion chromatograms corresponding to each peptide identified by MudPIT for these proteins and calculated the related peak area (i.e., a quantitative parameter in activated and resting samples). Score values and their ratio (activated/resting cells, A/R) for the proteins were then compared with the peak areas and their A/R ratio for each corresponding peptide. There was a good correlation between the ratios from score values and those from peak areas of individual peptides for the 3 proteins considered. In activated cells, peptides associated to Mac25 protein displayed a decrease in A/R score value ratio (0.56), in good agreement with peak area ratios of T_98-114 (0.57) and T_227-240 (0.61) peptides. The A/R score ratio of CSPG2/versican (0.95) was in good agreement with the calculated peak area ratio of peptides T_80-92 (1.15) and T_101-119 (1.02). Finally, MMP-1 showed an increase in A/R score ratio (22.5) and peak area ratios of tryptic peptides T_137-151 (20.0), and T_348-362 (23.0). The relationship between score values and quantity of protein in the sample was further confirmed by specific assays for the 3 proteins considered. We can therefore conclude that the SEQUEST score value is predictive of the amount of peptides in the sample.

CONCLUSIONS AND SIGNIFICANCE

We provide an efficient and rapid protocol that allows the identification of major proteins released by cancer cells in few days using a small quantity (10 µL) of culture supernatants that, after minimal handling, is analyzed by automated MudPIT technology.

This approach efficiently identified novel as well as known proteins secreted by pancreatic cancer cells: a total of 46 proteins were found to be released by resting and activated cells. The fact that 21 of these proteins are classified as secreted in public databases demonstrates the validity of our approach. Among these proteins, 11 have already been shown to be expressed by pancreatic cancer, and our present findings suggest that they are released by cancer cells. Ten of the secreted proteins we identified have not been described in pancreatic cancer so far, and therefore represent novel findings.

A relevant finding is that cancer cells produce and release a series of proteoglycans, including CSPG2/versican, HSPG2/Perlecan, Syndecan-1, and Syndecan-4. The demonstration that CSPG2/versican isreleased in vivo, including primary cancers and xenografted cancer cells, challenge the common view that fibroblasts of tumor stroma are the sole source of proteoglycans usually found overexpressed in cancer tissues. CSPG2/versican is not only a structural protein of the stroma, but seems to have an active role in the modulation of cellular functions.

We provide evidence that the score value assigned by SEQUEST software to proteins identified by MudPIT correlates with the protein amount in the sample. The fact that MudPIT delivers quantitative information is highly relevant, as this allows a direct and high throughput comparison of samples in a fully automated setting.

FOOTNOTES

¹ These authors contributed equally to this work.

To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.04-3000fje; doi: 10.1096/fj.04-3000fje

This article has been cited by other articles:

				Y. Arvidsson, E. Andersson, A. Bergstrom, M. K Andersson, G. Altiparmak, A.-C. Illerskog, H. Ahlman, D. Lamazhapova, and O. Nilsson Amyloid precursor-like protein 1 is differentially upregulated in neuroendocrine tumours of the gastrointestinal tract Endocr. Relat. Cancer, June 1, 2008; 15(2): 569 - 581. [Abstract] [Full Text] [PDF]

				V. Castronovo, D. Waltregny, P. Kischel, C. Roesli, G. Elia, J.-N. Rybak, and D. Neri A Chemical Proteomics Approach for the Identification of Accessible Antigens Expressed in Human Kidney Cancer Mol. Cell. Proteomics, November 1, 2006; 5(11): 2083 - 2091. [Abstract] [Full Text] [PDF]

				G. W. Yip, M. Smollich, and M. Gotte Therapeutic value of glycosaminoglycans in cancer. Mol. Cancer Ther., September 1, 2006; 5(9): 2139 - 2148. [Abstract] [Full Text] [PDF]

				M. Maiorino, A. Roveri, L. Benazzi, V. Bosello, P. Mauri, S. Toppo, S. C. E. Tosatto, and F. Ursini Functional Interaction of Phospholipid Hydroperoxide Glutathione Peroxidase with Sperm Mitochondrion-associated Cysteine-rich Protein Discloses the Adjacent Cysteine Motif as a New Substrate of the Selenoperoxidase J. Biol. Chem., November 18, 2005; 280(46): 38395 - 38402. [Abstract] [Full Text] [PDF]

This Article Full Text (PDF) All Versions of this Article: 19/9/1125 04-3000fjev1    most recent Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mauri, P. Articles by Sorio, C. Search for Related Content PubMed PubMed Citation Articles by Mauri, P. Articles by Sorio, C.