Subjects
Blood samples were collected from breast cancer patients and normal healthy volunteers at the Seoul National University Hospital (Seoul, Korea). The use of human samples for research purpose was authorized by the Institutional Review Board of Seoul National University Hospital, and all the patients and volunteers agreed to take part in the experiment with the name signed on the informed consent document. The plasma sample was depleted of top six abundant serum proteins using a multiple-affinity MARS column (Agilent Technologies, Palo Alto, CA, USA) [12], and precipitated with trichloroacetic acid. The pellet was dissolved in ICAT denaturation buffer (6 M urea, 0.05% SDS, 5 mM EDTA, 50 mM Tris-HCl, pH 8.3).
ICAT labeling and sample preparation
A pooled plasma sample from 6 breast cancer patients was labeled with a 'heavy (H)' ICAT reagent (Applied Biosystems, Framingham, MA, USA), whilst another pooled sample from 6 normal healthy women was labeled with a 'light (L)' reagent. We pooled equal amount of proteins from individual samples. Proteins (100 μg) in the denaturantion buffer were first reduced with 250 mM tris(2-carboxyethyl)phosphine for 30 min. ICAT-labeling was then performed using 350 nmol ICAT reagent with gentle shaking for 2 hr at 37°C, and terminated with 1.75 μmol DTT for additional 5 min. The H- and L-ICAT-labeled samples were mixed, diluted 10 fold with 50 mM Tris (pH 8.0), and digested with 5 μg of trypsin (Promega, Madison, WI, USA) for 16 hr at 37°C. The reaction was quenched at 0.5% phosphoric acid. The tryptic digest was applied on a polysulfoethyl A column (Western Analytical, Murrieta, CA, USA) equilibrated with 10 mM KH2PO4 in 25% ACN (pH 3.0) using an ÄKTA Explorer system (GE Healthcare Biosciences, Uppsala, Sweden), eluted with a 40-min gradient from zero to 0.4 M KCl, and collected on 40 fractions. The SCX fractions were neutralized by the addition of 10 volumes of 2× PBS, loaded on an ICATR avidin-catridge (Applied Biosystems), and then washed with PBS followed by 50 mM ammonium bicarbonate in 20% methanol, pH 8.3. ICAT-labeled peptides were eluted with a solution of 0.4% TFA in 30% acetonitrile, dried in vacuo, redissolved in 90 μl of 95% TFA, incubated at 37°C for 2 hr to cleave off the biotin moiety from the ICAT label, and finally dried again.
Liquid chromatography and tandem mass spectrometry
An Agilent nanoflow-1200 series HPLC system was connected to a linear ion trap mass spectrometer (LTQ, Thermo Electron, San Jose, CA, USA). The dried ICAT-labeled peptide sample was reconstituted with 20 μL of 0.4% acetic acid, and an aliquot (1 μL) was injected to a reverse-phase Magic C18aq column (13 cm × 75 μm) equilibrated with 95% buffer A (0.1% formic acid in H2O) + 5% buffer B (0.1% formic acid in acetonitrile). The peptides were eluted in a linear gradient of 10 to 40% acetonitrile over 75 min. The MS survey was scanned from 300 to 2000 m/z, and followed by three data-dependent MS/MS scans with the following options: isolation width, 1.5 m/z; normalized collision energy, 25%; dynamic exclusion duration, 180 sec.
Database searches
Peak lists were generated using Extract-msn program in Bioworks package v3.1 (Thermo Electron) with the following parameters: minimum ion count threshold, 15; minimum intensity, 100. The peak lists were compared against the human International Protein Index database including known contaminants (IPI, versions 3.24, European Bioinformatics Institute, http://www.ebi.ac.uk/IPI) using the SEQUEST (TurboSequest version 27, revision 12) allowing two missed cleavages (trypsin) and ±0.5 and ±3 Da mass tolerance for MS/MS and MS respectively. ICAT option (+227.26 Da fixed modification plus +9 Da variable modification) on cysteine residue was used and a variable modification of methionine oxidation (+16 Da) was allowed. Peptide assignment and quantification were performed with the Trans-Proteomic Pipeline provided by Institute for Systems Biology (TPP, version 4.0, http://www.proteomecenter.org). The SEQUEST search output was used as an input for Peptide-Prophet module and peptides with probabilities greater than 0.05 were included in the following Protein-Prophet. Proteins with probabilities greater than 0.5 were put into manual inspection to evaluate MS/MS spectral quality [14]. False discovery rate was 10% at the cut-off value of 0.5 before manual inspection. From a list of 238 proteins, 30 proteins were removed due to lack of quantification information and 53 proteins were removed due to their unreliable mass spectra during manual inspection. The number of removed proteins (53 ea) by manual inspection exceeded the number of estimated false positives (~24 ea). As a result, the false discovery rate for the final data set containing 155 proteins would be almost zero.
Western blot analysis
Plasma samples were resolved on 10% SDS-PAGE gel, and electro-transferred to nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA, USA). In order to handle a large number of samples that exceeded loading sites of a gel, two or more gels were placed on a transfer membrane to minimize experimental bias of western blots. Immunobloting analyses were performed as described previously [12], using antibodies against neural cell adhesion molecule L1 (CHL1; Atlas, Stockholm, Sweden), α1-acid glycoprotein 2 (ORM2; Proteintech Group Inc., Chicago, IL, USA), monocyte differentiation antigen CD14 (CD14; Abcam, Cambridge, MA, USA), BTD (GeneTex Inc., San Antonio, TX, USA), and glutathione peroxidase 3 (GPX3; Abcam).
Analysis of gene expression microarray data
Large cohort tissue microarray datasets of breast cancer patients analyzed by Human Genome U133A platform (GPL96) were downloaded from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/projects/geo/). The samples included 1,715 cases of biopsied breast cancer tissues (GSE1456, GSE2034, GSE2990, GSE3494, GSE4922, GSE5364 and GSE11121) and 95 cases of laser-capture microdissected (LCM) breast cancer tissues (GSE5847). The latter 95 samples were considered to be positive controls for breast cancer, since the LCM would be effective to get rid of contamination of normal tissues or blood cells in the breast cancer biopsies. As references, 39 microarray data of normal human breast tissues analyzed by the same or upgrade version (GPL570) were obtained from the GEO database. The obtained microarray data were analyzed by the R-package 2.7.2 using an Affy package and a gcRMA package to make normalization and adjustments of the background and average intensities. The average values of gene expression were calculated from replicate probes. From the log2-transformed values of each tumor sample, the log2 medians of included normal samples were subtracted to calculate a tumor-to-normal ratio (fold change) conveniently.
Statistical analysis
Band intensities of Western blot images were quantified using ImageQuant version 5.2. (GE Healthcare Biosciences), and compared by Wilcoxon rank-sum test using SPSS 12.0 (SPSS, Chicago, IL, USA) due to the small sample size. For the statistical analysis of gene expression data, Wilcoxon rank-sum tests were performed with a 95% confidence interval, as the sample size of breast cancer tissues was greater than the size of normal controls and also the two data sets had been collected independently from large cohort breast cancer studies and standard human tissue microarray studies.
Comments
View archived comments (1)