A Proteomic Analysis of the Plasma Glycoproteins of a MCF-7 Mouse

To develop a better animal model of human breast cancer, a group of ovarectimized 6-week-old female Ncr nude mice were implanted with a MCF-7 tumor...
0 downloads 0 Views 2MB Size
Download PDF
A Proteomic Analysis of the Plasma Glycoproteins of a MCF-7 Mouse Xenograft: A Model System for the Detection of Tumor Markers Christina I. Orazine,† Marina Hincapie,† William S. Hancock,*,† Maureen Hattersley,‡ and Jeff H. Hanke‡ Barnett Institute, Northeastern University, 341 Mugar Building, Boston, Massachusetts 02115, and AstraZeneca R&D Boston, PLC, 35 Gatehouse Drive, Waltham, Massachusetts 02451 Received December 14, 2007

In this study, we report a plasma proteomic analysis of a mouse MCF7 xenograft, using a novel platform named M-LAC (multilectin affinity chromatography), in an attempt to identify putative serum biomarkers of tumor presence and response to therapy. The use of the M-LAC platform enabled us to focus on secreted proteins as well as remove interference from serum albumin and other nonglycosylated proteins. The study focused on the MCF7 human xenograft tumor model which enabled us to distinguish tumor proteins (human peptide sequences) from host-derived murine proteins, potentially discriminating tumor- versus supporting tissue-derived markers. A large set of murine proteins was identified in this study, including several signaling molecules such as EGFR, interleukin-6 receptor, protein-kinase C, and phosphatidylinositol kinase which changed in plasma levels relative to tumor-free animals. We also detected in the samples with maximal tumor growth a number of human tumor-derived proteins linked to cell signaling, immune response, and transcriptional regulation. This is the first report where tumor-derived peptides could be detected in the serum of a xenograft model. We conclude that the M-LAC approach may be used to detect plasma proteins of potential biological significance in tumorbearing animals and warrants further study in terms of increasing the sensitivity of the method for the characterization of low level tumor markers and to explore the applicability of these markers for human studies. Keywords: Lectin • MCF-7 • glycoprotein • proteomics • breast cancer • biomarker

Introduction

study of genetically modified animal models16 as well as xenografts.17

The plasma proteome has been a major focus of the Human Proteome Initiative because of the pivotal connection of this proteome to the disease-based initiatives of the organization.1 In addition, clinical chemistry has made this study a priority in terms of better understanding disease and the search for biomarkers for the early detection of cancer.2 Quantitative proteomics can also be used to characterize the human secretome, and in particular the altered secretion of proteins in disease with the goal of discovering cancer biomarkers3 and to characterize the tumor secretome.4 We5 and others6–9 have chosen as a focus the study of the plasma glycoproteins because of the intimate association of this post-translational modification with the secretion process. Altered glycosylation has often been associated with development, progression, and metastasis of cancer,10–12 and thus, the study of the glycoproteins is of undoubted significance to the cancer field.13–15 Proteomics technology has also been applied to plasma analysis in the

For cancer, one anticipates the release of proteins associated with the development of the tumor into the blood stream and that such proteins can give a signature for early diagnosis. The release of tumor-specific proteins can be related to processes such as the increased synthesis and secretion of glycoproteins, the cleavage of matrix or membrane-associated proteins, and the release of intracellular proteins from cells that have undergone apoptosis.2 The proteomic analysis of the interstitial fluid from breast tumors has suggested that cell death and the release of cytoplasmic or nuclear-associated proteins makes a contribution, albeit small, to the blood proteome of cancer patients.18 Recently, we reported on significant changes in the plasma glycoproteins in breast cancer13 and autoimmune disease.19 However, plasma consists of an extremely complex mixture of proteins that are released from a variety of tissues including leukocytes and thrombocytes, and thus. it is a challenge to understand the relevance and source a given protein. To begin to tease apart tumor-specific proteins from other plasma proteins, we focused on studying the glycoproteins in an MCF-7 nude-mouse xenograft tumor model to see if we could identify proteins related to tumor biology and

* To whom correspondence should be addressed. E-mail: wi.hancock@ neu.edu. † Northeastern University. ‡ AstraZeneca R&D Boston.

1542 Journal of Proteome Research 2008, 7, 1542–1554 Published on Web 03/13/2008

10.1021/pr7008516 CCC: $40.75

 2008 American Chemical Society

research articles

Proteomics of Plasma Glycoproteins of a MCF-7 Mouse Xenograft progression of the disease. The use of closely controlled animal models greatly reduces the complexity and variation of samples relative to studies in humans.20,21 Implanted human tumors also have several advantages over spontaneous tumors as they have distinct origins and genetic backgrounds. For example, the MCF-7 cell line used in this study has an extensive history22–31 and was derived from a pleural effusion of a patient with breast cancer.20 The cell line is estrogen receptor-positive, although such estrogen-dependent lines rarely metastasize in nude mice.31 It is considered to be uninvasive in both the Matrigel outgrowth study and in the nude mouse.22 Several studies have attempted to identify proteins likely to be associated with the development of proliferating MCF-7 cell lines.,24–30 such as with mitogenic concentrations of 17-beta-estradiol (E2),27 or MCF-7 cell lines selected for resistance to anticancer drugs (mitoxantrone),22 and following treatment with doxorubicin29 or adriamycin.26 Recently, LC/MS proteomic studies were performed on nuclear fractions isolated from cultured MCF-7 cells and identified 3715 putative proteins, while a related study of MCF-7 plasma membrane proteins characterized 540 proteins28,30 and thus set the stage for xenograft studies. The study, however, that is closest to our report was reported by Juan et al.17 in which the authors studied a xenotransplantation model in the nude mouse of 5 different human cancer cell lines. While this report did not include the MCF-7 cell line, and only identified a few mouse acute phase proteins and no human proteins, the authors did speculate on the advantage of using a well-controlled inbred animal model to identify tumor-derived serum proteins that were of human origin. We used a more sensitive proteomics approach based on LC/MS and sample prefractionation focused on the glycoproteins to see if we could indeed identify tumor-specific proteins based on the murine host response to tumor growth as well as unique human sequences. In our study, ovarectomized mice received an injection with MCF-7 cells, which were implanted with an estrogen pellet. The orthotopic placement of the MCF-7 tumor cells in the mouse mammary pad is considered to be a model of human breast cancer.20,21,31 The animals were separated into groups that were treated with estrogen, tamoxifen, or both. We then performed an analysis of the glycoproteins using multilectin affinity chromatography (M-LAC). In contrast to the narrow specificity of antibodies, lectins have a general affinity to glycosylation motifs which are similar across species. Another advantage of our approach is that serum albumin, which makes up approximately 50% of plasma proteins, is not glycosylated and is depleted. Thus, the M-LAC platform has been shown5 to increase the dynamic range of plasma measurement and provide differential quantitative information of a significant number of plasma proteins. We also used bovine fetuin as an internal standard which enabled us to normalize measurements across the sample set using label free quantitation in the mass spectrometer.32 Nano LC-LTQ-FT analysis was then used to identify tumor-specific proteins based on species-specific peptide identification. While the sequence homology between human and mouse is high, it is possible to detect speciesspecific peptides using the discriminating power of the high mass accuracy measurement. We have shown in this study that we could indeed detect tumor-specific proteins derived from the murine host as well as human-specific proteins from the tumor. In addition, this study supports the continued development of the M-LAC approach in animal models as well as humans to identify

markers of tumor presence and response to therapy. On the basis of the results of this report, we have identified future studies that will enable insights into the release of tumor markers into the blood stream and thus facilitate the search for biomarkers for the early detection of cancer.

Experimental Procedures Materials. Concanavalin A (Con A), Jacalin, and Wheat germ agglutinin (WGA) agarose bound lectins were purchased from Vector laboratories (Burlingame, CA). Disposable polypropylene columns were purchased from Pierce (Rockford,IL). Bovine fetuin and all other chemicals were purchased from SigmaAldrich (St. Louis, MO). The BCA protein assay reagent kit was from Pierce (Rockford, IL). Sequencing grade, modified trypsin was purchased from Promega (Madison, WI). The 5 kDa Amicon molecular weight cutoff filters were purchased from Millipore (Billerica, MA). Discovery BIO Wide Pore C18 cartridges (C18, 2 cm × 4.00 mm, 3 µm particles) were from Supelco (Bellefonte, PA). Xenograft Study. The experimental design for the xenograft study is shown in Figure 1. Estrogen receptor-positive MCF-7 cells were obtained from the ATCC and maintained in culture in RPMI media supplemented with 10% fetal bovine serum and 1% L-glutamine at 37 °C, 5% carbon dioxide. MCF-7 xenografts were established by surgically implanting 8 × 106 cells into the third mammary fat pad of 6-week-old female, NCr Nude mice (Taconic Farms, Germantown, NY). Growth of tumors was supplemented with 0.72 mg of 60 day-release estrogen pellets (Innovative Research of America, Sarasota, FL) which were implanted subcutaneously on the back of the animals 24 h prior to cell implantation. When tumors reached an average volume of 120 mm3 (day 19 after implantation), the mice were randomized by tumor volume (N ) 10 per group) into four groups: no treatment (estrogen deprivation), estrogen only, tamoxifen citrate only, or estrogen and tamoxifen citrate. In all mice, the original estrogen pellet was removed on the day of randomization. The estrogen only mice were implanted with a new estrogen pellet. The tamoxifen citrate only mice were implanted with a 5 mg, 60 day-release tamoxifen citrate pellet (Innovative Research, Sarasota, FL). The combined estrogen and tamoxifen treated mice received both estrogen and tamoxifen pellets. Tumor growth was assessed using digital calipers, and tumor volume and body weight were measured twice a week. After 3 weeks of treatment, five mice from each group were sacrificed, and tumor tissue and blood were collected from each mouse. After 6 weeks of treatment, the remaining mice from each group were sacrificed, and tumor tissue and blood were harvested. Blood was collected into EDTA microtainer tubes (Beckton Dickinson) then spun at 10 000 rpm for 3 min to separate plasma. The plasma from each group of mice was pooled. All tumor and plasma samples were stored at –80 °C. Mice were maintained and sacrificed according to Institutional Animal Care and Use guidelines. Affinity Capture of Glycoproteins with M-LAC. Pooled samples were randomized prior to analysis to avoid bias. Samples were stored at -75 °C until analysis. One hundred microliters of sample was diluted to 500 µL with binding buffer (25 mM Tris, 0.15 M sodium chloride, 1 mM calcium chloride, 1 mM magnesium chloride, and 0.055 mM sodium azide, at pH 7.4). Bovine fetuin (25 µg/mL) was added to each plasma sample prior to lectin fractionation to serve as an internal standard. Multi lectin columns were gravity-packed as previously deJournal of Proteome Research • Vol. 7, No. 4, 2008 1543

research articles

Orazine et al.

Figure 1. Experimental design for the study of the Ncr nude mouse xenograph as a model of human breast cancer. Initially, 56 ovarectimized mice were split into two groups: nontumor and tumor. A MCF-7 tumor was implanted subcutaneously in the mammary fat pad of those mice belonging to the tumor group, while the control group was not implanted. Each of the two groups was further split into 4 treatment groups: no treatment, estrogen, tamoxifen, or both estrogen and tamoxifen. Estrogen and/or tamoxifen pellets were surgically implanted in mice of the estrogen groups and the estrogen plus tamoxifen groups. For the individual time points, plasma from 5 individual mice was pooled for each analysis. Plasma was collected at the beginning of the study, 3 weeks into the study, and at 6 weeks.

scribed5 with a 1:1:1 mixture of the three lectins. A single M-LAC column was used for each individual sample to avoid possible contamination. The sample was allowed to penetrate into the column for an incubation period of 15 min, and then proteins not bound were washed from the column by the addition of two subsequent 5 mL portions of M-LAC binding buffer, as described above. Washes were collected and subjected to the analysis of protein concentration using a Bradford assay. The measurement of protein concentration in the unbound fraction provided an important quality control point. Elution of the glycoprotein fraction was accomplished by the addition of 4 mL of Elution Buffer (25 mM Tris, 0.5 M sodium chloride, 0.2 M methyl-R-mannopyrannoside, 0.2 M methylR-glucopyrannoside, 0.8 M galactose, 0.5 M N-acetyl-glucosamine, and 0.05% sodium azide at pH 7.4). A portion of this fraction was also retained for protein concentration analysis. Digestion of Glycoprotein Fraction. The glycoproteinenriched, bound fraction was concentrated down to 50 µL using 5 kDa Amicon molecular weight cutoff filters. Fifty microliters of sample was denatured with 7.2 M guanidine chloride in 0.1 M ammonium bicarbonate (pH 8.0), added to make a concentration of 5.8 M guanidine chloride. The reduction of protein disulfide bonds in the sample was achieved by the addition of 5 mM DTT and incubation at 60 °C for 30 min. Protein sulfhydryl groups were then alkylated with 15 mM iodoacetamide in darkness for 30 min. The alkylation reaction was quenched by adding a second aliquot of 5 mM of DTT. Samples were diluted to decrease the concentration of guanidine chloride to 1.2 mM with 50 mM ammonium bicarbonate buffer (pH 8.0). Trypsin was added to samples at a 1:40 (w/w) ratio. Samples were incubated for 18 h at 37 °C. A second aliquot of trypsin was added at a 1:25 ratio, and the samples were incubated at 37 °C for an additional 4 h. Digestion was stopped by the addition of formic acid to a final concentration of 1%. Reversed-Phase Desalting Using HPLC. Peptides were separated from salts and any undigested material by chromatography on a Discovery BIO C18 column installed on a Shimadzu HPLC (Shimadzu Scientific Instruments, Columbia, MD). Mobile phase A was composed of 0.1% TFA in HPLC grade water. Mobile phase B was composed of 0.1% TFA in 1544

Journal of Proteome Research • Vol. 7, No. 4, 2008

HPLC grade acetonitrile. The step gradient method employed for the separation consisted of 3 min steps of: 0% mobile phase B to wash away salts remaining from the digestion, 30% B to elute peptides to be used for LC/MS analysis, and 90% B to wash the column of any undigested proteins or large peptides. The flow rate was set at 1.5 mL/min. Eluting peptides and proteins were monitored at 214 and 280 nm. The 30% fraction was collected and concentrated on a speed vacuum to remove AcCN and prepare it for analysis. The organic solvent was removed from each fraction under vacuum, and samples were not taken to complete dryness in order to minimize losses. Each fraction was reconstituted to the same volume (20 µL buffer) and then stored at -80 °C. For LC/MS analysis, a 2.5 µL aliquot was removed. Proteomic Analysis by NanoLC-MS/MS. All nanoLC-MS/ MS experiments were performed on an Ettan MDLC system (GE Healthcare, Piscataway, NJ) coupled with a Thermo Finnigan linear ion trap mass spectrometer (Thermo_electron, San Jose, CA). A 15 cm long, 75 µm i.d. capillary column (purchased from New Objective, Woburn, MA) packed in house with 5 µm, 200 Å pore size Magic C18 stationary phase was used for all LC-MS/MS experiments. Mobile phase A and mobile phase B were 0.1% formic acid in HPLC grade water and 0.1% formic acid in HPLC grade acetonitrile, respectively. Prior to injection, each sample was concentrated to 20 µL. Duplicate injections of 2.5 µL were loaded onto a Peptide Captrap column (Michrom Bioresources, Auburn, CA) using the MDLC autosampler. The trap column was washed with mobile phase A for 10 min at 10 µL/min, and then the flow rate was reduced to 280 nL/min, the trap column was placed in-line with the capillary column, and the gradient method was started. A linear gradient method beginning at 2% mobile phase B, after 30 min of equilibration, which preceded to 40% B over 160 min, to 90% B after an additional 20 min, and remained constant at 90% for 20 min, comprised the LC separation method. The GE software package Unicorn (GE Healthcare, Piscataway, NJ) allowed the operation of the Ettan MDLC. The ion transfer tube of the LTQ during the analysis was 245 °C, and the electrospray ionization voltage was set to 2.0 kV. Normalized collision energy was 35% for MS/MS analysis. MS/

Proteomics of Plasma Glycoproteins of a MCF-7 Mouse Xenograft MS was triggered automatically by operating in data-dependent mode. The 7 most intense peaks were selected from the full MS scan of 400-2000 m/z for MS/MS. Precursor ion exclusion time was 1 min. NanoLC-LTQ-FT Analysis. An UltimMate NanoHPLC system (LC Packings-Dionex, Marlton, NJ) and LTQ-FT mass spectrometer (Thermo Electron, San Jose, CA) were used for additional nanoLC-MS/MS analyses of samples. The capillary LC column and mobile phases were the same as described above. Electrospray voltage was 1.8 kV. The normalized collision energy was 28% for MS/MS. The ion transfer tube temperature was 245 °C. A medium resolution preview MS scan was generated after the ions were injected into the ICR cell. The Excalibur software selected the 8 most abundant ions for MS/ MS analysis. While the LTQ fragmented these ions, the FT performed a full high resolution MS scan. Precursor ions were excluded from subsequent fragmentation for 1 min. Data Processing and Analysis. Peptides derived from the internal standard bovine fetuin were identified by searching MS2 spectra against a bovine fetuin database downloaded from Swiss-Prot September 2005. Murine and human proteins were identified by searching MS2 spectra against a murine data downloaded from Swiss-Prot July 2005 or a human database which was downloaded from Swiss-Prot September 2005. For all database searches, trypsin was selected as the enzyme, and two missed cleavages were allowed. Carbamidomethylation of cysteine was included in the search parameters. Tolerances were set at (1.4 Da for precursor ion mass and (1.0 Da for product ion mass. Peptide Prophet Software was used to filter the results to a minimum probability of 95%. Only proteins identified with 2 or more unique peptides were considered. Relative Quantitation. Preliminary investigation of the complete data set was performed by analyzing the average of the total sequencing events (i.e., spectral count) for each protein.33 Biologically interesting proteins were selected for manual integration of peak areas. For each protein, two high confidence peptides were chosen for peak area quantitation of the extracted ion chromatogram. For the reported peptides, the correlation of variation between the replicate runs was 15% or less. Normalization of the data was achieved by performing the same process of peak area quantitation on 2 peptides having the best signal-to-noise ratio which had been uniquely identified for bovine fetuin. Of the two peptides, the one with the lowest variability was then averaged across all data points (RSD for this peak area was 9.7%). This average value was utilized for normalization of the data. The overall relative standard deviation was 15% or less between duplicate injections (5% for higher intensity peaks).

Results M-LAC Fractionation. We used the multi lectin affinity chromatographic (M-LAC) approach to fractionate mouse plasma into a bound, glycosylated protein rich portion and a flow through fraction which contained mainly nonglycosylated proteins, such as serum albumin. While each of the lectins have affinities to a broad range of carbohydrate structures, we will describe the major binding determinants in the following sentence for purposes of simplicity. The fractionation was achieved with a mixture of three agarose-bound lectins, Concanavalin A (specific for R-mannose type structures), Jacalin (O-linked N-acetylglucosamine), and Wheat germ agglutinin (sialic acid). The combination of multiple lectins has been

research articles

shown to give more complete capture of the glycoprotein fraction than the use of single lectins.5 Since the mouse model is immunoglobulin-deficient, we did not need to deplete the immunoglobulin fraction. In this manner, we avoided the use of a depletion step which improves throughput of the study and avoids possible losses of material. In this study, each treatment group contained 5 animals, plasma samples from each group were pooled, and the resulting 20 pools, originating from control groups, tamoxifen, and estrogen treatment at different time points, were randomized prior to the M-LAC step (see legend to Figure 1 for the experimental design). Bovine fetuin (25 µg) was added as an internal standard to each pool and then loaded onto the M-LAC column. The overall recovery (unbound and bound fraction) of the M-LAC step for all mouse plasma samples was determined by a Bradford protein analysis and averaged 96% with a CV of 10%. The split between the glycosylated and nonglycosylated fractions was 9% (CV 28%) versus 87% (CV of 14%). The Bradford assay gives only an estimate of protein concentration but is useful in monitoring the consistency of the M-LAC fractionation step. The variation in amount of the glycosylated fraction is not unexpected in a study which contains both tumor-implanted and control animals and different treatment groups, all of which may have effects on the degree of glycosylation of plasma proteins; for example, differences in glycosylation patterns have been noted in certain cancers.10–12 Preparation of Tryptic Peptide Digests for LC/MS Analysis. At all steps in a proteomic analysis, variability can be introduced especially when one considers the complexity and dynamic range of the plasma samples. The trypsin digestion step is no exception; for example, disease-related changes in plasma may affect the rate of cleavage of certain proteins. As described in Experimental Procedures, we used two aliquots of trypsin (ratio of 4% (w/w) at 0 and 24 h) at 37 °C to digest the reduced and alkylated glycoprotein fraction (50 µL). In addition, the sample was denatured with guanidine-HCl to minimize incomplete trypsin digestion of associated or aggregated proteins. Despite these precautions, our experience is that one cannot expect 100% digestion of all proteins in such a complex sample, and thus, we have instituted a reversedphase HPLC (RPLC) cleanup step. Such a step has the advantage of desalting the sample as well as removing partially digested proteins. The latter is particularly important in improving the lifetime of the capillary reversed-phase column and the consistency of a series of LC/MS analyses. We also used the UV peak area at 214 nm of the peptide fraction, collected in an isocratic step at 40% organic solvent, to analyze the reproducibility of the digestion and perform any necessary adjustment to the amount of sample analyzed in the LC/MS step. For the entire sample set (20 pools), the coefficient of variation of peak area was 27% and consistent with the expected biological viability of the study. In addition, the RPLC step is relatively fast, and the column was reusable, unlike many filtration devices. LC/MS Analysis of the Sample Set. Since we use the label free approach to give relative quantitation across a sample set, it was important to control bias in the analytical protocol. A key aspect of our approach was to improve column lifetime with our RPLC cleanup step, which also minimizes carry over from sample to sample. It was, however, necessary to perform a blank gradient between each analysis. We then monitored variability of LC/MS response over the entire sample set by measuring the integrated base peak between 40 and 160 min Journal of Proteome Research • Vol. 7, No. 4, 2008 1545

research articles

Orazine et al.

Table 1. Abundant Proteins Found in both the Nude Mouse Control Sample and a Human Nondisease Samplea spectral countb

descriptionc

479 89 56 56 49 39 51 36 26 25 25

Alpha-2-macroglobulin Hemopexind Complement C3 Complement factor Hd Kininogend Fibronectind Fibrinogen Serum albumind Apolipoprotein A-I Beta-2-glycoprotein I Phosphatidylinositol-glycan-specific phospholipase D 1 Alpha-2-HS-glycoproteind Fetuin-Bd Vitronectind Kallikreind Interalpha-trypsin inhibitor heavy chain H Prothrombind Alpha-1-antitrypsind Alpha-1-microglobulin Clusterind Haptoglobin Afamind Complement C4d Complement component C9 Complement factor Bd Serotransferrin Ceruloplasmind Apolipoprotein E Plasminogen Serum amyloid P-componentd Complement component C8 Myosin

24 24 24 20 17 16 15 14 14 13 12 11 3 10 10 9 8 7 7 6 6

a As reported in Yang et al.13 b Measured for control mouse plasma sample as determined by LC/MS analysis of a tryptic digest. c The following 5 proteins were not found in the human glycoproteome: Apolipoprotein C-III, Leukemia inhibitory factor receptor, Liver carboxylesterase N, Murinoglobulin, and Serine protease inhibitor A3K. d Proteins also found in another mouse glycoproteome study.34

(elution window for the majority of peptides), and the amount of variation was again consistent with the degree of biological variation (CV approximately 30%). In an effort to minimize false-positive identifications, we used criteria suggested by HUPO for the identification of peptides by MS/MS sequencing, as well as a requirement for a probability greater than 0.9 (Protein Prophet). The final criteria was to list proteins identified with 2 or more unique peptides; this gave a total of 1877 proteins by combining the total data set. In addition, the final time point for the group of 6weekold mice with implanted tumors receiving either estrogen or tamoxifen were also analyzed on the hybrid mass spectrometer, linear ion trap FT MS. The additional analysis by the FT MS system allowed the use of accurate mass and retention time information to confirm identifications, such as macrophage colony-stimulating factor 1 receptor, insulin receptor, DNAdependent protein kinase, and vitronectin (data not shown). Table 1 compares the glycoproteins of the control group of the nude mouse with of the corresponding human glycoproteins. It should be noted that the control group consists of animals which have been ovarectomized, implanted with the MCF-7 cell line, but then deprived of the estrogen implant. In this situation, the growth of the xenograft is not sustained, and 1546

Journal of Proteome Research • Vol. 7, No. 4, 2008

thus, this group acts as a control. The data shows a high degree of commonality at the level of abundant glycoproteins between the mouse and the human samples. In addition, apolipoprotein CIII and carboxylesterase N could be presumed to be present in the human sample at concentrations below the detection limit for this study. One protein, murinoglobin, is speciesspecific, while two proteins, leukemia inhibitory factor receptor (LIFR) and serine protease inhibitor A3K, appear to be elevated in this animal model. Many of the proteins listed in the table were also reported in the study of the C57BL mouse plasma glycoproteins34 which used an affinity capture of the glycopeptides (common identifications are labeled in Table 1). One difference with the athymic mouse model from human and other mouse studies is the very low level of immunoglobulins in the plasma samples of the nude mouse. The strong degree of correlation with previous studies does give confidence about the relevance of this animal model for the study of human disease. These results also confirm that the M-LAC approach can be used as a method for the enrichment of glycoproteins from a variety of mammalian species, which is consistent with the observations of the conservation of general glycosylation motifs across mammals.35 Differential Quantitation. The MS/MS data was first searched against the murine database (February 2006). We used spectral counts33 as an initial measure of either up- or downregulation of proteins in the large data set resulting from this study. Table 2 shows a comparison of proteins that had a substantial difference (ratio of >2 or 3 (or the raw signal intensity was >1 × 104). The relative peak areas were either reported as the mean of measured peak areas or as the area recorded for the peptide with the best S/N value (cross-checked with the

other peptides for consistency). If the areas disagreed by >30%, then the values were not reported. The results of these measurements with relation to the biology of this cancer model will be reported in the Discussion. As an example, EGFR was shown to be decreased approximately 3-fold in the 6-weekold mice which were treated with estrogen (maximum tumor volume) relative to the control group (Table 2) which was consistent with the spectral count measurement. Identification of Tumor-Specific Peptides. Previous studies on xenografts of human cell lines in athymic mice have been unable to characterize human proteins in the presence of abundant mouse plasma proteins.17 For the characterization of human-specific proteins, we therefore selected the sample which represented the mice having the greatest tumor mass and with the highest level of secreted tumor proteins. This group, as can be seen in Figure 2, is the mice which received estrogen at the 6 week time point with a tumor volume of 736 versus 1800 identifications) for proteins associated with signaling pathways and identified 69 kinases (data not shown). This list was curated for disease association, and then proteins were selected for label-free differential quantitation (an important criteria was that the candidate was present at a level sufficiently high for robust measurement19). As described in Experimental Procedures, the quantitation process included either spectral count values (Table 2) or peak area measurements (Table 3). The changes in protein concentration listed in Table 3 are given for the different time points (0, 3, and 6 weeks) for the no treatment, estrogen, tamoxifen, and estrogen plus tamoxifen treatment group (all groups contain the MCF-7 implant). Tables 2 and 3 show that some proteins increased during the time course of the estrogen treatment (insulin receptor substrate 2 (IRS2), leukemia inhibitory factor 1 receptor (LIF), some showed an increase at the intermediate time point and then a decrease (tyrosine-protein kinase JAK2, serine protein kinase ATM, and Interleukin-1 receptor-associated kinase 4 (IRAK)) and the rest showed decreases (PI3-K, DNA-dependent protein kinase 4 (DNA-PK), interleukin-1 receptor associated kinase 4 (IL-1RK), JAK1, EGFR, vitronectin). In our quantitative measurements, the variability between analytical replicates was less than 10%, but to account for biological variability, we generally consider a 2-fold change as significant. The following proteins met this criteria across the time points for the animals with estrogen stimulation (all decreases): JAK1 (14-fold), PI 3-kinase (below our detection limit after the initial time point), and EGFR (3.7-fold). These proteins showed no significant changes in the control animal which suggested that decreased levels of these proteins in the estrogen group could be associated with tumor growth. The lipid kinase phosphoinositol-4-phosphate 3-kinase (PI 3-kinase or mouse PK3CA, class II, containing the p110delta and C2 subunits) and related members represent one of the most important regulatory proteins that control key cellular functions and the resulting phosphoinositide products activate a host of signaling proteins.51 Several pivotal studies have shown that PI 3K has an integral role in tumorigenesis via 1550

Journal of Proteome Research • Vol. 7, No. 4, 2008

Orazine et al. association with oncoproteins, by genetic analysis and by mouse transgenic and knockout studies.58 Importantly, other studies have shown recruitment of PI 3-kinase with signaling molecules identified in this study: epidermal growth factor receptor (EGFR),52 interleukin-1 receptor associated kinase (IRAK),54 ephrin receptor EphA8, A3 and B2,54 and members of the PI 3-kinase like kinase family (PIKK members DNA-PK and ATM).55 Also, it has been shown that ATM heterozygotes have an increased risk of developing breast cancer56 and is known to be associated with EGFR (see 3a). A related protein found in this study is GPI-PLD which cleaves the GPI anchor sequence attached to many cell surface proteins including receptors.57 GPI-PLD was downregulated in the estrogen and combined treatment groups, which is consistent with reports that this protein was decreased during inflammation58 and which can be related to tumor growth. The observations of decreases in JAK1 and EGFR are supported by other reports of reduced JAK1 expression in breast cancer tissues versus matched noncancer tissues,59 and serum levels of the extracellular domain of EGFR were observed to be decreased in breast cancer patients.60 Furthermore, studies of MCF-7 cell lines have found the presence of estrogen can suppress expression of EGFR,61 which is consistent with our observation that human EGFR was not observable in the estrogen-treated group (see next section). An advantage of the xenograft model is that we could identify EGFR with mousespecific peptides (see Table 3) and thus demonstrate a nontumor source. While the tissue(s) of origin for the EFGR observed in the plasma samples is not known, a possible candidate is the adjacent stromal tissue which is known to express growth factors and related receptor kinases.46,62,63 Future studies with this model system will be required to understand the potential role of stromal proteins in the development of the cancer-related plasma glycoproteins, but interestingly, such signaling can occur with decreased receptor expression as was observed in this study.63–66 To further explore this point, Figure 3 shows that a comparison of the two treatment groups (animals with no tumor and with tumor) can allow one to estimate the effects of estrogen on the mouse versus the more complex interaction with animal and tumor. Interestingly, the figure shows that there are opposite changes in levels of the kinases for the two groups, which mimics the heterotypic expression reported above and is consistent with the presence of stromal tissue in the tumor bearing animals. In the estrogen plus tamoxifen treatment group, in which we observed intermediate tumor growth, a number of proteins, serine-protein kinase ATM, Ephrin receptor A8, interleukin-1 receptor associated kinase (IRAK), and EGFR were decreased, while phosphatidylinositol-4-phosphate 3-kinase C2 domain (PI 3-kinase) showed a significant increase (3-fold). These results were consistent with previous studies which have shown that tamoxifen treatment of MCF-7 xenografts reduced tumor growth rates, increased apoptosis significantly,67 downregulated EGFR, and that this receptor is further decreased by estrogen.38 The final treatment group (tamoxifen) is the least relevant to this study as no tumor growth was observed in this animal group. Table 3 shows that substantial increases and decreases were observed with all proteins, except for phosphatidylinositol-glycan-specific phospholipase D1 (PHLD), including a 15fold increase in PI 3-kinase, which could be attributed to additional, nonestrogen-receptor mechanisms.68,69 The change over time observed for this key enzyme in the combined treatment group (increase with estrogen plus tamoxifen treat-

Proteomics of Plasma Glycoproteins of a MCF-7 Mouse Xenograft

research articles

Figure 3. Effects of estrogen on mouse (M) and mouse with tumor (Tu). In the figure, 0, 3, and 6 stand for the time of sample collection. All of the peak areas measurements are normalized by dividing by 10 × 5. In this analysis, the corresponding control (mouse, no tumor, no implant or mouse, tumor, no treatment) was subtracted from the reported values for the effects of estrogen on the mouse (M) or mouse with growing tumor (Tu), respectively, and the relative errors have been summed. Other proteins which show similar changes such as EPHB3, JAK2, and PRKDC are not shown.

ment vs decrease with estrogen treatment) could be attributed to such nonreceptor effects of tamoxifen. Characterization of Tumor (Human)-Specific Proteins. As described in Results, the mouse group with the expected highest level of tumor proteins (6 weeks with estrogen treatment) was selected for the attempted characterization of human-specific peptide sequences. While such proteins have not been previously characterized for mouse xenografts, other studies have successfully studied mixed proteomes via the measurement of unique species-specific peptides.70 Table 4 gives a conservative list of human-specific peptides that were measured in the FTMS (see Results). To meet our conservative criteria, each protein had to be identified with at least two peptides with at least one human specific-peptide (shown in bold). The table also groups the proteins into the following classifications: growth factor and cytoskeletal signaling, immune response, and transcriptional regulation. It was of interest to observe that the proteins discovered had in many cases extensive association with breast cancer (literature associations are listed as footnotes to Table 4). As shown in Table 4, many of the lower level proteins were only detected by specific human peptides without identification of corresponding mouse sequences. In the case of 4 abundant proteins (spectral count greater than 20), both mouse and human proteins were observed which suggests both mouse and tumor origins. The remaining proteins, which were present at lower levels (spectral count g2) are probably glycosylated and in many cases potentially secreted and are thus candidates for tumor-specific markers, such as LDL receptor-related protein 1B and ATP binding protein A2. Conversely, known tumorassociated proteins such as EGFR, JAK, and PI3-K were found in this study only with mouse-specific peptides and thus must have a nontumor origin. While the biosynthesis of such proteins may indeed be tumor-related, such as a stromal origin (see Figure 3), these proteins cannot be directly associated with the tumor unless expression profiling studies are performed to demonstrate an unique site of synthesis. The value of the

mouse model in studying human tumor secretion is emphasized by the difficulty in performing such profiling studies in man, and furthermore, even if a protein is expressed, it may not be secreted. In the future, we believe that higher levels of stringency could be achieved by performing an integrated protein and glycan analysis, as cancer has been associated with the generation of abnormal or rare glycan motifs. The corresponding protein analysis has the potential of determining the tissue of origin for the glycan. Our approach to such an investigation of tumorspecific glycoproteins will be to use affinity isolations with biospecific ligands such as antibodies or lectins, followed by both protein and glycan characterization using methodology previously described.71,72 We are hopeful that such an approach will identify tumor-specific biomarkers in mouse plasma samples that can be translated to human studies.

Conclusions In this study, we report an in-depth plasma proteomic analysis of a mouse xenograft of the human breast cancer cell line MCF-7. We focused on the glycoproteins by using a novel platform named M-LAC (multilectin affinity chromatography) which both increased the dynamic range of the study and primarily selected for secreted proteins. A large set of proteins was identified including the following which exhibited more than a 2-fold change under the condition of tumor growth: phosphatidylinositol-4-phosphate 3-kinase, tyrosine-protein kinase JAK1, and epidermal growth factor receptor. The use of a xenograft mouse model has the potential of distinguishing tumor proteins (human peptide sequences) from host response (murine). In this study, we demonstrated for the first time that high accuracy mass measurements could detect tumor-specific proteins derived from the murine host as well as humanspecific proteins from the tumor. In addition, this study supports the continued development of the M-LAC approach in animal models as well as humans to identify markers of Journal of Proteome Research • Vol. 7, No. 4, 2008 1551

research articles tumor presence and response to therapy. On the basis of the results presented in this report, we have identified future studies directed at generating insights into the release of tumor markers into blood and thus facilitate the search for biomarkers for the early detection of cancer.

Orazine et al.

(14)

(15)

Acknowledgment. We thank Lihua Yu, Dr. Rejtar, and Dr. Wu for technical expertise. We thank Astra Zeneca, Inc. for the financial support of this research. We also thank Thermo Electron and GE Healthcare for support with instrumentation and software.

Supporting Information Available: Supplementary Table 1 is a list of all of the protein identifications made in this study along with the total number of peptides identified from each protein in each individual LC-MS analysis, the number of unique peptides identified from each protein in each analysis, and the percentage of sequence coverage. This material is available free of charge via the Internet at http:// pubs.acs.org. References (1) Omenn, G. S.; States, D. J.; Adamski, M.; Blackwell, T. W.; Menon, R.; Hermjakob, H.; Apweiler, R.; Haab, B. B.; Simpson, R. J.; Eddes, J. S.; Kapp, E. A.; Moritz, R. L.; Chan, D. W.; Rai, A. J.; Admon, A.; Aebersold, R.; Eng, J.; Hancock, W. S.; Hefta, S. A.; Meyer, H.; Paik, Y. K.; Yoo, J. S.; Ping, P.; Pounds, J.; Adkins, J.; Qian, X.; Wang, R.; Wasinger, V.; Wu, C. Y.; Zhao, X.; Zeng, R.; Archakov, A.; Tsugita, A.; Beer, I.; Pandey, A.; Pisano, M.; Andrews, P.; Tammen, H.; Speicher, D. W.; Hanash, S. M. Overview of the HUPO Plasma Proteome Project: results from the pilot phase with 35 collaborating laboratories and multiple analytical groups, generating a core dataset of 3020 proteins and a publicly-available database. Proteomics 2005, 5 (13), 3226–45. (2) Davis, M. A.; Hanash, S. High-throughput genomic technology in research and clinical management of breast cancer. Plasma-based proteomics in early detection and therapy. Breast Cancer Res. 2006, 8 (6), 217. (3) Gronborg, M.; Kristiansen, T. Z.; Iwahori, A.; Chang, R.; Reddy, R.; Sato, N.; Molina, H.; Jensen, O. N.; Hruban, R. H.; Goggins, M. G.; Maitra, A.; Pandey, A. Biomarker discovery from pancreatic cancer secretome using a differential proteomic approach. Mol. Cell. Proteomics 2006, 5 (1), 157–71. (4) Khwaja, F. W.; Svoboda, P.; Reed, M.; Pohl, J.; Pyrzynska, B.; Van Meir, E. G. Proteomic identification of the wt-p53-regulated tumor cell secretome. Oncogene 2006, 25 (58), 7650–61. (5) Yang, Z.; Hancock, W. S. Monitoring glycosylation pattern changes of glycoproteins using multi-lectin affinity chromatography. J. Chromatogr., A 2005, 1070 (1–2), 57–64. (6) Liu, T.; Qian, W. J.; Gritsenko, M. A.; Camp, D. G., II; Monroe, M. E.; Moore, R. J.; Smith, R. D. Human plasma N-glycoproteome analysis by immunoaffinity subtraction, hydrazide chemistry, and mass spectrometry. J. Proteome Res. 2005, 4 (6), 2070–80. (7) Zhang, H. The plasma proteome: high abundance versus low abundance. January 10–11, 2006, San Diego, CA, USA. Expert Rev. Proteomics 2006, 3 (2), 175–8. (8) Durham, M.; Regnier, F. E. Targeted glycoproteomics: serial lectin affinity chromatography in the selection of O-glycosylation sites on proteins from the human blood proteome. J. Chromatogr., A 2006, 1132 (1–2), 165–73. (9) Madera, M.; Mechref, Y.; Klouckova, I.; Novotny, M. V. Semiautomated high-sensitivity profiling of human blood serum glycoproteins through lectin preconcentration and multidimensional chromatography/tandem mass spectrometry. J. Proteome Res. 2006, 5 (9), 2348–63. (10) Miyamoto, S. Clinical applications of glycomic approaches for the detection of cancer and other diseases. Curr. Opin. Mol. Ther. 2006, 8 (6), 507–13. (11) Brockhausen, I. Mucin-type O-glycans in human colon and breast cancer: glycodynamics and functions. EMBO Rep. 2006, 7 (6), 599– 604. (12) Kobata, A.; Amano, J. Altered glycosylation of proteins produced by malignant cells, and application for the diagnosis and immunotherapy of tumours. Immunol. Cell Biol. 2005, 83 (4), 429–39. (13) Yang, Z.; Harris, L. E.; Palmer-Toy, D. E.; Hancock, W. S. Multilectin affinity chromatography for characterization of multiple glyco-

1552

Journal of Proteome Research • Vol. 7, No. 4, 2008

(16)

(17)

(18)

(19)

(20) (21) (22) (23) (24) (25)

(26)

(27)

(28) (29) (30) (31)

(32)

(33) (34) (35)

protein biomarker candidates in serum from breast cancer patients. Clin. Chem. 2006, 52 (10), 1897–905. Zhou, H. J.; Liu, Y. K.; Chui, J. F.; Sun, Q. L.; Lu, W. J.; Guo, K.; Jin, H.; Wei, L. M.; Yang, P. Y. A glycoproteome database of normal human liver tissue. J. Cancer Res. Clin. Oncol. 2007, 133 (6), 679– 87. Miyoshi, E.; Noda, K.; Ko, J. H.; Ekuni, A.; Kitada, T.; Uozumi, N.; Ikeda, Y.; Matsuura, N.; Sasaki, Y.; Hayashi, N.; Hori, M.; Taniguchi, N. Overexpression of alpha1–6 fucosyltransferase in hepatoma cells suppresses intrahepatic metastasis after splenic injection in athymic mice. Cancer Res. 1999, 59 (9), 2237–43. Hung, K. E.; Kho, A. T.; Sarracino, D.; Richard, L. G.; Krastins, B.; Forrester, S.; Haab, B. B.; Kohane, I. S.; Kucherlapati, R. Mass spectrometry-based study of the plasma proteome in a mouse intestinal tumor model. J. Proteome Res. 2006, 5 (8), 1866–78. Juan, H. F.; Chen, J. H.; Hsu, W. T.; Huang, S. C.; Chen, S. T.; YiChung, L. J.; Chang, Y. W.; Chiang, C. Y.; Wen, L. L.; Chan, D. C.; Liu, Y. C.; Chen, Y. J. Identification of tumor-associated plasma biomarkers using proteomic techniques: from mouse to human. Proteomics 2004, 4 (9), 2766–75. Celis, J. E.; Gromov, P.; Cabezon, T.; Moreira, J. M.; Ambartsumian, N.; Sandelin, K.; Rank, F.; Gromova, I. Proteomic characterization of the interstitial fluid perfusing the breast tumor microenvironment: a novel resource for biomarker and therapeutic target discovery. Mol. Cell. Proteomics 2004, 3 (4), 327–44. Plavina, T.; Wakshull, E.; Hancock, W. S.; Hincapie, M. Combination of abundant protein depletion and multi-lectin affinity chromatography (M-LAC) for plasma protein biomarker discovery. J. Proteome Res. 2007, 6 (2), 662–71. Cespedes, M. V.; Casanova, I.; Parreno, M.; Mangues, R. Mouse models in oncogenesis and cancer therapy. Clin. Transl. Oncol. 2006, 8 (5), 318–29. Kelland, L. R. Of mice and men: values and liabilities of the athymic nude mouse model in anticancer drug development. Eur. J. Cancer 2004, 40 (6), 827–36. Simstein, R.; Burow, M.; Parker, A.; Weldon, C.; Beckman, B. Apoptosis, chemoresistance, and breast cancer: insights from the MCF-7 cell model system. Biol. Med. 2003, 228 (9), 995–1003. Rose, D. P.; Connolly, J. M. Dietary fat and breast cancer metastasis by human tumor xenografts. Breast Cancer Res. Treat. 1997, 46 (2–3), 225–37. Sandhu, C.; Connor, M.; Kislinger, T.; Slingerland, J.; Emili, A. Global protein shotgun expression profiling of proliferating mcf-7 breast cancer cells. J. Proteome Res. 2005, 4 (3), 674–89. Yavelow, J.; Tuccillo, A.; Kadner, S. S.; Katz, J.; Finlay, T. H. Alpha 1-antitrypsin blocks the release of transforming growth factoralpha from MCF-7 human breast cancer cells. Clin. Endocrinol. Metab. 1997, 82 (3), 745–52. Schiemann, S.; Schwirzke, M.; Brunner, N.; Weidle, U. H. Molecular analysis of two mammary carcinoma cell lines at the transcriptional level as a model system for progression of breast cancer. Clin. Exp. Metastasis 1998, 16 (2), 129–39. Malorni, L.; Cacace, G.; Cuccurullo, M.; Pocsfalvi, G.; Chambery, A.; Farina. . Proteomic analysis of MCF-7 breast cancer cell line exposed to mitogenic concentration of 17beta-estradiol. Proteomics 2006, 6 (22), 5973–82. Rahbar, A. M.; Fenselau, C. Unbiased examination of changes in plasma membrane proteins in drug resistant cancer cells. J. Proteome Res. 2005, 4 (6), 2148–53. Chen, S. T.; Pan, T. L.; Tsai, Y. C.; Huang, C. M. Proteomics reveals protein profile changes in doxorubicin--treated MCF-7 human breast cancer cells. Cancer Lett. 2002, 181 (1), 95–107. Hathout, Y.; Gehrmann, M. L.; Chertov, A.; Fenselau, C. Proteomic phenotyping: metastatic and invasive breast cancer. Cancer Lett. 2004, 210 (2), 245–53. Kerbel, R. S. Human tumor xenografts as predictive preclinical models for anticancer drug activity in humans: better than commonly perceived-but they can be improved. Cancer Biol. Ther. 2003, 2 (4), S134–9. Shen, Y.; Kim, J.; Strittmatter, E. F.; Jacobs, J. M.; Camp, D. G., 2nd; Fang, R.; Tolie, N.; Moore, R. J.; Smith, R. D. Characterization of the human blood plasma proteome. Proteomics 2005, 5 (15), 4034–45. Sadygov, R. G.; Cociorva, D.; Yates, J. R., III. Large-scale database searching using tandem mass spectra: looking up the answer in the back of the book. Nat. Methods 2004, 1 (3), 195–202. Bernhard, O. K.; Kapp, E. A.; Simpson, R. J. Enhanced analysis of the mouse plasma proteome using cysteine-containing tryptic glycopeptides. J. Proteome Res. 2007, 6 (3), 987–95. Williams, A. F.; Parekh, R. B.; Wing, D. R.; Willis, A. C.; Barclay, A. N.; Dalchau, R.; Fabre, J. W.; Dwek, R. A.; Rademacher, T. W.

Proteomics of Plasma Glycoproteins of a MCF-7 Mouse Xenograft

(36)

(37)

(38) (39)

(40)

(41) (42) (43)

(44) (45) (46)

(47) (48)

(49)

(50) (51) (52)

(53)

(54)

(55)

Comparative analysis of the N-glycans of rat, mouse and human Thy-1. Site-specific oligosaccharide patterns of neural Thy-1, a member of the immunoglobulin superfamily. Glycobiology 1993, 3 (4), 339–48. Hoffman, S.; Chuong, C. M.; Edelman, G. M. Evolutionary conservation of key structures and binding functions of neural cell adhesion molecules. Proc. Natl. Acad. Sci. U.S.A. 1984, 81 (21), 6881–5. Koibuchi, Y.; Iino, Y.; Uchida, T.; Andoh, T.; Horii, Y.; Nagasawa, M.; Horiguchi, J.; Maemura, M.; Takei, H.; Yokoe, T.; Morishita, Y. Regulation of estrogen receptor and epidermal growth factor receptor by tamoxifen under high and low estrogen environments in MCF-7 cells grown in athymic mice. Oncol. Rep. 2000, 7 (1), 135–40. Wolf, D. M.; Jordan, V. C. Characterization of tamoxifen stimulated MCF-7 tumor variants grown in athymic mice. Breast. Cancer Res. Treat. 1994, 31 (1), 117–27. Goswami, S.; Sahai, E.; Wyckoff, J. B.; Cammer, M.; Cox, D.; Pixley, F. J.; Stanley, E. R.; Segall, J. E.; Condeelis, J. S. Macrophages promote the invasion of breast carcinoma cells via a colonystimulating factor-1/epidermal growth factor paracrine loop. Cancer Res. 2005, 65 (12), 5278–83. Lee, A. V.; Jackson, J. G.; Gooch, J. L.; Hilsenbeck, S. G.; CoronadoHeinsohn, E.; Osborne, C. K.; Yee, D. Enhancement of insulinlike growth factor signaling in human breast cancer: estrogen regulation of insulin receptor substrate-1 expression in vitro and in vivo. Mol. Endocrinol. 1999, 13 (5), 787–96. Jordan, V. C.; Fritz, N. F.; Gottardis, M. M. Strategies for breast cancer therapy with antiestrogens. J. Steroid Biochem. 1987, 27 (1–3), 493–8. Molloy, C. A.; May, F. E.; Westley, B. R. Insulin receptor substrate-1 expression is regulated by estrogen in the MCF-7 human breast cancer cell line. J. Biol. Chem. 2000, 275 (17), 12565–71. Stuelten, C. H.; DaCosta Byfield, S.; Arany, P. R.; Karpova, T. S.; Stetler-Stevenson, W. G.; Roberts, A. B. Breast cancer cells induce stromal fibroblasts to express MMP-9 via secretion of TNF-alpha and TGF-beta. J. Cell Sci. 2005, 118 (10), 2143–53. Harkonen, P. L.; Vaananen, H. K. Monocyte-macrophage system as a target for estrogen and selective estrogen receptor modulators. Ann. N.Y. Acad. Sci. 2006, 1089, 218–27. Hussein, M. R. Tumour-associated macrophages and melanoma tumourigenesis: integrating the complexity. Int. J. Exp. Pathol. 2006, 87 (3), 163–76. Creighton, C. J.; Bromberg-White, J. L.; Misek, D. E.; Monsma, D. J.; Brichory, F.; Kuick, R.; Giordano, T. J.; Gao, W.; Omenn, G. S.; Webb, C. P.; Hanash, S. M. Analysis of tumor-host interactions by gene expression profiling of lung adenocarcinoma xenografts identifies genes involved in tumor formation. Mol. Cancer Res. 2005, 3 (3), 119–29. Yoshimura, A. Signal transduction of inflammatory cytokines and tumor development. Cancer Sci. 2006, 97 (6), 439–47. Dirkx, A. E.; Oude Egbrink, M. G.; Kuijpers, M. J.; van der Niet, S. T.; Heijnen, V. V.; Bouma-ter Steege, J. C.; Wagstaff, J.; Griffioen, A. W. Tumor angiogenesis modulates leukocyte-vessel wall interactions in vivo by reducing endothelial adhesion molecule expression. Cancer Res. 2003, 63 (9), 2322–9. Kairouz, R.; Daly, R. J. Tyrosine kinase signaling in breast cancer: modulation of tyrosine kinase signaling in human breast cancer through altered expression of signaling intermediates. Breast Cancer Res. 2000, 2 (3), 197–202. Olayioye, M. A. Update on HER-2 as a target for cancer therapy: intracellular signaling pathways of ErbB2/HER-2 and family members. Breast Cancer Res. 2001, 3 (6), 385–9. Krasilnikov, M. A. Phosphatidylinositol-3 kinase dependent pathways: the role in control of cell growth, survival, and malignant transformation. Biochemistry (Moscow) 2000, 65 (1), 59–67. Le Page, C.; Koumakpayi, I. H.; Lessard, L.; Saad, F.; Mes-Masson, A. M. Independent role of phosphoinositol-3-kinase (PI3K) and casein kinase II (CK-2) in EGFR and Her-2-mediated constitutive NF-kappaB activation in prostate cancer cells. Prostate 2005, 65 (4), 306–15. Reddy, S. A.; Lin, Y. F.; Huang, H. J.; Samanta, A. K.; Liao, W. S. The IL-1 receptor accessory protein is essential for PI 3-kinase recruitment and activation. Biochem. Biophys. Res. Commun. 2004, 316 (4), 1022–8. Gu, C.; Park, S. The EphA8 receptor regulates integrin activity through p110gamma phosphatidylinositol-3 kinase in a tyrosine kinase activity-independent manner. Mol. Cell. Biol. 2001, 21 (14), 4579–97. Friedmann, B. J.; Caplin, M.; Savic, B.; Shah, T.; Lord, C. J.; Ashworth, A.; Hartley, J. A.; Hochhauser, D. Interaction of the

(56) (57)

(58)

(59)

(60)

(61)

(62)

(63)

(64) (65)

(66)

(67)

(68) (69)

(70)

(71)

(72) (73)

(74)

research articles

epidermal growth factor receptor and the DNA-dependent protein kinase pathway following gefitinib treatment. Mol. Cancer Ther. 2006, 5 (2), 209–18. Prokopcova, J.; Kleibl, Z.; Banwell, C. M.; Pohlreich, P. The role of ATM in breast cancer development. Breast Cancer Res. Treat. 2007, 104 (2), 121–8. Mann, K. J.; Hepworth, M. R.; Raikwar, N. S.; Deeg, M. A.; Sevlever, D. Effect of glycosylphosphatidylinositol (GPI)-phospholipase D overexpression on GPI metabolism. Biochem. J. 2004, 378 Pt 2, 641–8. Arreaza, G.; Brown, D. A. Sorting and intracellular trafficking of a glycosylphosphatidylinositol-anchored protein and two hybrid transmembrane proteins with the same ectodomain in MadinDarby canine kidney epithelial cells. J. Biol. Chem. 1995, 270 (40), 23641–7. Yeh, Y. T.; Ou-Yang, F.; Chen, I. F.; Yang, S. F.; Su, J. H.; Hou, M. F.; Yuan, S. S. Altered p-JAK1 expression is associated with estrogen receptor status in breast infiltrating ductal carcinoma. Oncol. Rep. 2007, 17 (1), 35–9. Muller, V.; Witzel, I.; Pantel, K.; Krenkel, S.; Luck, H. J.; Neumann, R.; Keller, T.; Dittmer, J.; Janicke, F.; Thomssen, C. Prognostic and predictive impact of soluble epidermal growth factor receptor (sEGFR) protein in the serum of patients treated with chemotherapy for metastatic breast cancer. Anticancer Res. 2006, 26 (2B), 1479–87. Yarden, R. I.; Wilson, M. A.; Chrysogelos, S. A. Estrogen suppression of EGFR expression in breast cancer cells: A possible mechanism to modulate growth. J. Cell Biochem. 2001, 81 (S36), 232–46. Ellis, M. J.; Singer, C.; Hornby, A.; Rasmussen, A.; Cullen, K. J. Insulin-like growth factor mediated stromal-epithelial interactions in human breast cancer. Breast Cancer Res. Treat. 1994, 31 (2–3), 249–61. Guddo, F.; Fontanini, G.; Reina, C.; Vignola, A. M.; Angeletti, A.; Bonsignore, G. The expression of basic fibroblast growth factor (bFGF) in tumor-associated stromal cells and vessels is inversely correlated with non-small cell lung cancer progression. Hum. Pathol. 1999, 30 (7), 788–94. Fromigue, O.; Kheddoumi, N.; Body, J. J. Bisphosphonates antagonise bone growth factors’ effects on human breast cancer cells survival. Br. J. Cancer 2003, 89 (1), 178–84. Olapade-Olaopa, E. O.; MacKay, E. H.; Taub, N. A.; Sandhu, D. P.; Terry, T. R.; Habib, F. K. Malignant transformation of human prostatic epithelium is associated with the loss of androgen receptor immunoreactivity in the surrounding stroma. Clin. Cancer Res. 1999, 5 (3), 569–76. Fiuraskova, M.; Brychtova, S.; Sedlakova, E.; Benes, P.; Zalesak, B.; Hlobilkova, A.; Tichy, M.; Kolar, Z. Molecular changes in PDEGF and bFGF in malignant melanomas in relation to the stromal microenvironment. Anticancer Res. 2005, 25 (6B), 4299– 303. Hawkin, R. A.; Arends, M. J.; Ritchie, A. A.; Langdon, S.; Miller, W. R. Tamoxifen increases apoptosis but does not influence markers of proliferation in an MCF-7 xenograft model of breast cancer. Breast 2000, 9 (2), 96–106. Mandlekar, S.; Kong, A. N. Mechanisms of tamoxifen-induced apoptosis. Apoptosis 2001, 6 (6), 469–77. Wilmes, P.; Bond, P. L. The application of two-dimensional polyacrylamide gel electrophoresis and downstream analyses to a mixed community of prokaryotic microorganisms. Environ. Microbiol. 2004, 6 (9), 911–20. Sam-Yellowe, T. Y.; Florens, L.; Wang, T.; Raine, J. D.; Carucci, D. J.; Sinden, R.; Yates, J. R., III. Proteome analysis of rhoptry-enriched fractions isolated from Plasmodium merozoites. J. Proteome Res. 2004, 3 (5), 995–1001. Wang, Y.; Wu, S. L.; Hancock, W. S. Approaches to the study of N-linked glycoproteins in human plasma using lectin affinity chromatography and nano-HPLC coupled to electrospray linear ion trap--Fourier transform mass spectrometry. Glycobiology 2006, 16 (6), 514–23. Ren, J. M.; Rejtar, T.; Li, L.; Karger, B. L. N-Glycan structure annotation of glycopeptides using a linearized glycan structure database (GlyDB). J. Proteome Res. 2007, 6 (8), 3162–73. den Hollander, P.; Kumar, R. Dynein light chain 1 contributes to cell cycle progression by increasing cyclin-dependent kinase 2 activity in estrogen-stimulated cells. Cancer Res. 2006, 66 (11), 5941–9. Niculescu, F.; Rus, H. G.; Retegan, M.; Vlaicu, R. Persistent complement activation on tumor cells in breast cancer. Am. J. Pathol. 1992, 140 (5), 1039–43.

Journal of Proteome Research • Vol. 7, No. 4, 2008 1553

research articles (75) Carriero, M. V.; Del Vecchio, S.; Franco, P.; Potena, M. I.; Chiaradonna, F.; Botti, G.; Stoppelli, M. P.; Salvatore, M. Vitronectin binding to urokinase receptor in human breast cancer. Clin. Cancer Res. 1997, 3 (8), 1299–308. (76) Li, Y.; Knisely, J. M.; Lu, W.; McCormick, L. M.; Wang, J.; Henkin, J.; Schwartz, A. L.; Bu, G. Low density lipoprotein (LDL) receptorrelated protein 1B impairs urokinase receptor regeneration on the cell surface and inhibits cell migration. J. Biol. Chem. 2002, 277 (44), 42366–71. (77) Bourguignon, L. Y.; Zhu, H.; Zhou, B.; Diedrich, F.; Singleton, P. A.; Hung, M. C. Hyaluronan promotes CD44v3-Vav2 interaction with Grb2-p185(HER2) and induces Rac1 and Ras signaling during ovarian tumor cell migration and growth. J. Biol. Chem. 2001, 276 (52), 48679–92. (78) Johnson, J.; Albarani, V.; Nguyen, M.; Goldman, M.; Willems, F.; Aksoy, E. Protein kinase CR is involved in IRF-3 activation and type I IFN-β synthesis. J. Biol. Chem. 2007, 282 (20), 15022–32. (79) Grossoni, V. C.; Falbo, K. B.; Kazanietz, M. G.; de Kier Joffe, E. D.; Urtreger, A. J. Protein kinase C δ enhances proliferation and survival of murine mammary cells. Mol. Carcinog. 2007, 46, 381– 90. (80) Carey, I.; Williams, C. L.; Ways, D. K.; Noti, J. D. Overexpression of protein kinase C-alpha in MCF-7 breast cancer cells results in differential regulation and expression of alphavbeta3 and alphavbeta5. Int. J. Oncol. 1999, 15 (1), 127–36. (81) Mazurkiewicz, M.; Opolski, A.; Wietrzyk, J.; Radzikowski, C.; Kleinrok, Z. GABA level and GAD activity in human and mouse

1554

Journal of Proteome Research • Vol. 7, No. 4, 2008

Orazine et al.

(82) (83)

(84)

(85)

(86)

(87)

normal and neoplastic mammary gland. J. Exp. Clin. Cancer Res. 1999 Jun, 18 (2), 247–53. Thompson, D. K.; Haddow, J. E.; Smith, D. E.; Ritchie, R. F. Elevated serum acute phase protein levels as predictors of disseminated breast cancer. Cancer 1983, 51 (11), 2100–4. Zhai, Y.; Wu, R.; Schwartz, D. R.; Darrah, D.; Reed, H.; Kolligs, F. T.; Nieman, M. T.; Fearon, E. R.; Cho, K. R. Role of beta-catenin/Tcell factor-regulated genes in ovarian endometrioid adenocarcinomas. Am. J. Pathol. 2002, 160 (4), 1229–38. Berquin, I. M.; Pang, B.; Dziubinski, M. L.; Scott, L. M.; Chen, Y. Q.; Nolan, G. P.; Ethier, S. P. Y-box-binding protein 1 confers EGF independence to human mammary epithelial cells. Oncogene 2005, 24 (19), 3177–86. van Kooij, M.; de Groot, K.; van Vugt, H.; Aten, J.; Snoek, M. Genotype versus phenotype: conflicting results in mapping a lung tumor susceptibility locus to the G7c recombination interval in the mouse MHC class III region. Immunogenetics 2001, 53 (8), 656– 61. Doyle, L. A.; Yang, W.; Abruzzo, L. V.; Krogmann, T.; Gao, Y.; Rishi, A. K.; Ross, D. D. A multidrug resistance transporter from human MCF-7 breast cancer cells. Proc. Natl. Acad. Sci. U.S.A. 1998, 95 (26), 15665–70. Wilson, A. C.; Boutros, M.; Johnson, K. M.; Herr, W. HCF-1 aminoand carboxy-terminal subunit association through two separate sets of interaction modules: involvement of fibronectin type 3 repeats. Mol. Cell. Biol. 2000, 20 (18), 6721–30.

PR7008516