Research Article | Open Access
Wei-Chao Chang, Chuan-Kai Chou, Chih-Chiang Tsou, Sheng-Hsiang Li, Chein-Hung Chen, Yu-Xing Zhuo, Wen-Lian Hsu, Chung-Hsuan Chen, "Comparative Proteomic Analysis of Proteins Involved in the Tumorigenic Process of Seminal Vesicle Carcinoma in Transgenic Mice", International Journal of Proteomics, vol. 2010, Article ID 726968, 14 pages, 2010. https://doi.org/10.1155/2010/726968
Comparative Proteomic Analysis of Proteins Involved in the Tumorigenic Process of Seminal Vesicle Carcinoma in Transgenic Mice
We studied the seminal vesicle secretion (SVS) of transgenic mice by using one-dimensional gel electrophoresis combined with LTQ-FT ICR MS analysis to explore protein expression profiles. Using unique peptide numbers as a cut-off criterion, 79 proteins were identified with high confidence in the SVS proteome. Label-free quantitative analysis was performed by using the IDEAL_Q software program. Furthermore, western blot assays were performed to validate the expression of seminal vesicle proteins. Sulfhydryl oxidase 1, glia-derived nexin, SVS1, SVS3, and SVS6 showed overexpression in SVS during cancer development. With high sequence similarity to human semenogelin, SVS2 is the most abundance protein in SVS and is dramatically decreased during the tumorigenic process. Our results indicate that these protein candidates could serve as potential targets for monitoring seminal vesicle carcinoma. Moreover, this information can provide clues for investigating seminal vesicle secretion-containing seminal plasma for related human diseases.
Primary seminal vesicle carcinoma is an extremely rare neoplasm; only a few cases have been reported [1, 2]. Seminal vesicle carcinoma is usually associated with diffuse carcinomas of the bladder, prostate, or upper tracts. Very few studies have been published on seminal vesicle invasion , and the tumor biology of seminal vesicle carcinoma is not well understood. The diagnoses of seminal vesicle carcinoma are generally based on a combination of morphologic, immunohistochemical, and radiological examinations. However, it is difficult to make a definitive diagnosis using limited biopsy material.
The seminal vesicle makes up part of the male accessory sexual glands. After puberty, the glands produce a fluid called seminal vesicle secretion (SVS), which accumulates in the lumen of the seminal vesicles. SVS contains both protein and nonprotein components and composes the majority of seminal plasma. The extirpation of seminal vesicles from adult rodents greatly reduces fertility, indicating that SVS plays an important role in sperm activity and modification . Some SVS proteins have been analyzed to characterize their functions and physiological activities [5–9], but the constituents of the SVS proteome have not been well studied.
Proteomic studies are broadly applied to the diagnostic, prognostic, and therapeutic fields. In terms of disease diagnosis and prognosis, body fluid analysis proves to be more attractive than tissue analysis because it provides several advantages including low invasiveness, minimum cost, and easy sample collection and processing . The linear ion trap-Fourier transform ion cyclotron resonance mass spectrometer (LTQ-FT ICR MS) has been exploited as a tool for identifying and characterizing isolated proteins . Combining LTQ-FT ICR MS with liquid chromatography, large-scale and high-quality proteomic analyses of several body fluids are achievable, including tears, urine, and seminal plasma [12–14]. By virtue of its capacity for exhaustive investigation, proteomic research has aroused high expectations for the discovery of biomarkers of various diseases [15, 16].
In this work, we used the transgenic mouse TAg as an animal model to study seminal vesicle carcinoma. Transgenic mouse TAg (C57BL/6-TgN (TRAMP) 8247Ng) expresses an SV40 large T antigen that is driven by the probasin promoter . In addition to being models for prostate cancer, TAg mice also develop other tumors, including neuroendocrine tumors in the prostate and neoplasms in the seminal vesicles. We analyzed the SVS proteome by using one-dimensional PAGE and LC LTQ-FT ICR MS. Accurate masses of tryptic peptides were measured by FT ICR MS, and tandem mass (MS/MS) experiments were performed in LTQ to acquire adequate data for proteomic identification. The resulting mzXML format data and Mascot search results were applied to perform label-free quantitative analysis. Moreover, we used western blot assays to validate the quantitative results.
2. Materials and Methods
2.1. Animals and Histopathology
TAg mice were purchased from the Jackson Laboratory (Maine, USA), and wild type C57BL/6JNarl normal control mice were obtained from the National Laboratory Animal Center (Taipei, Taiwan). They were housed in a specifically pathogen-free facility and handled in accordance with the guidelines of Guide for the Care and Use of Laboratory Animals. All of the mice were given free access to a standard murine chow diet and reverse osmosis water and were maintained on a 14:10 hours light-dark cycle at 21– In this study, TAg mice ( for each subgroup) of various ages (32–40 weeks) and normal control mice = were euthanized and the seminal vesicles and prostate glands were processed for histopathologic examination.
2.2. Sample Collection and Preparation
The seminal vesicles were collected by careful dissection to free them from the adjacent coagulating glands, and SVS was squeezed directly into 5 mL of 8 M urea solution. Small amounts of coagulated pellets were removed through centrifugation at 15,000 g for 10 minutes. The protein concentration of the supernatant was determined using the Bio-Rad Protein Assay (Bio-Rad Laboratories, Calif, USA) by the measurement of Each 20 g of SVS proteins was applied to 13% PAGE, and the gel was subsequently visualized through coomassie blue-staining. Gel lanes were divided into 10 sections and all slices were cut into small gel pieces (1 mm3), followed by in-gel digestion. Additionally, five of the observed major bands were cut for independent assays. Briefly, the procedure of in-gel digestion included the following steps in order: (a) destaining with 50% acetonitrile (ACN) and 25 mM ammonium bicarbonate (ABC); (b) reduction using freshly prepared 10 mM dithiothreitol for 45 minutes at (c) alkylation using freshly prepared 55 mM iodoacetamide for 45 min at room temperature in the dark; (d) enzyme digestion with 4 ng/L sequencing grade trypsin in 25 mM ABC solution at for 16–18 hours; (e) extraction of tryptic peptides using a 60% ACN/1% trifluoroacetic acid solution. After drying to remove the solvent, the re-dissolved tryptic peptides were subjected to LC LTQ-FT ICR MS analysis.
2.3. LC LTQ-FT ICR MS Analysis
MS/MS experiments were performed with an LTQ-FT ICR MS (Thermo Electron, Calif, USA) equipped with a nanoelectrospray ion source (New Objective, Mass, USA), an Agilent 1100 Series binary HPLC pump (Agilent Technologies, Calif, USA) and a Famos autosampler (LC Packings, Calif, USA). Tryptic peptide mixtures were injected at a 10 L/min flow rate into a self packed precolumn in line with a reverse phase C18 nanocolumn (75 m I.D. 200 mm) that used Magic C18AQ resin (particle size, 5 m; pore size, 200 Å; Michrom Bioresources, Calif, USA). The analytic program was set at a linear gradient from 10% to 50% ACN with a 60 minutes running cycle and a split flow rate of 300 nL/min. The full-scan survey MS experiment ( 320–2,000) was executed in FT ICR MS with a mass resolution of 100,000 at 400. The top ten most abundant multiply charged ions, if they were above a minimum threshold of 1,000 counts, were sequentially isolated for MS/MS by LTQ. Singly charged ions were rejected for MS/MS sequencing.
2.4. Mascot Search
The raw files of spectra were converted to mgf files with Mascot Daemon (data import filter: mass range, 600–5400; grouping tolerance, 1.4) and merged into a single file for searching by the MASCOT (version 2.1, Matrix Science Ltd., London, UK) software platform based on the IPI mouse database (v 3.36). The following MASCOT parameter settings were used; the peptide tolerance was 15 ppm with and peptide charges and the MS/MS tolerance was 0.6 Da. Two missed cleavages by trypsin were allowed, carbamidomethyl (C) was used as a fixed modification and oxidation (M) and deamidated (NQ) were used as variable modifications. The significance threshold for the identification was set to
2.5. Label-Free Quantitative Analysis
A software program IDEAL-Q (ID-based Elution time Alignment by Linear regression Quantification) was developed in-house to analyze LC-MS/MS data for label-free quantitative analysis . The program was used to process the LC-MS data and the search results obtained from the Mascot search engine to extract the quantification information. The whole quantitative analysis consisted of the following tasks. (I) Data preparation and construction of the protein list. The raw data files generated from the mass instrument were converted into the mzXML data format by the ReAdW program (http://tools.proteomecenter.org/wiki/index.php?title=Software:ReAdW). The data of each fractional LC-MS/MS runs coming from the same sample were merged and then searched by the Mascot search engine to establish a protein list, which contained identified proteins and their related peptide information. The mzXML files coupled with the peptide and protein identification results were input to the IDEAL-Q program. (II) Extracting quantitative information from each LC-MS run. For quantitative analysis of a peptide in an LC-MS run, we extracted the LC-MS data within the range of ±1.5 minutes of its elution time and ±3.5 Da of the precursor value. The peak clusters located within the selected elution time and the precursor value from the extracted data underwent a peptide validation process. For peptide validation, the following three criteria were applied: signal-to-noise ratio (S/N), charge state (CS), and isotope pattern (IP). The S/N criterion checks whether the precursor peak has a valid S/N ratio 2 The CS criterion eliminated the peak clusters with an incorrect charge state by examining whether the distance between adjacent peaks is equal to 1/z (tolerance ±1/10 1/z). Finally, the IP criterion examined the correlation between isotopic distribution of the observed peak intensities and the theoretical isotopic distribution of the peptide. The correlation was then evaluated by a Chi-square goodness of fit test 0.218 The purpose of peptide validation was to filter out false peptide signals; only the peptide passing the validation criteria were processed to subsequent quantification. We used the extracted ion chromatogram (XIC) to determine peptide abundance in an LC-MS run. (III) Peptide abundance and peptide ratio processing. First we determined the abundances of valid peptides in each LC-MS run. Then we calculated the peptide abundance in a fraction by averaging the peptide abundances of all repeated runs. We summed the peptide abundances in all fractions to represent the peptide quantity in the sample. Following, the peptide ratio between samples could be calculated. (IV) Protein abundance processing. We selected nondegenerate unique peptides and performed Dixon’s test to eliminate outliers of peptide ratios for each protein. We then used the top three highly abundant peptides of one protein to represent the quantity of this protein by a weighted average.
2.6. Western Blot
Polyclonal antibodies against SVS1, SVS2, SVS3 , SVS5, SVS7 , sulfhydryl oxidase 1, glia-derived nexin, carcinoembryonic antigen-related cell adhesion molecule 10 (CEACAM 10) , Lysozyme C-type M, secreted seminal-vesicle Ly-6 protein 1 (SSLP-1) , serine protease inhibitor kazal-like protein (SPINKL) , and serine protease inhibitor Kazal-type 3 (SPINK3)  were raised in New Zealand White rabbits. Antibody against human albumin, which also crossreacted with mouse albumin, was purchased from Calbiochem (Darmstadt, Germany). After PAGE separation, proteins were transferred to a PVDF membrane using a semidry blotter (ATTA, Tokyo, Japan). The electrophoresis program was set with a constant current (1.5 mA/cm2) for 1 hour. The membrane was blocked with 5% (w/v) skim milk in phosphate-buffered saline (PBS) at room temperature for O/N reaction, and then incubated with primary antibody (1 : 5000) in PBS with 2% (w/v) skim milk for 1 hour. After gentle agitation in four changes of PBST (PBS with 0.05% Tween 20) for 15 minutes each, the membrane was immuno-reacted with secondary antiserum (horseradish peroxidase-conjugated goat antirabbit IgG, Amersham Pharmacia) diluted to 1 : 10000 in PBS with 2% (w/v) skim milk for 1 hour. Immuno-reactive bands were revealed using an enhanced ECL substrate according to the manufacturer’s instructions (Pierce, Illinois, USA).
3.1. Histopathology and Tumor Categories
According to the histopathologic results, the neoplasm development of the accessory sex gland was categorized into three stages: (I) hyperplasia of the seminal vesicle (Hp), (II) adenoma of the seminal vesicle (Ad), and (III) adenocarcinoma of the seminal vesicle with prostate cancer (Ac) (Figure 1). The Hp stage showed a typical hyperplasia area. These areas, which were characterized by small size and uniformity of cells and nuclei, appeared to have increased glandular enfolding and lack of mitoses. The Ad stage showed the replacement of most of the normal glandular parenchyma with tumor masses encompassing two main components, epithelial cells and stromal cells. The cuboidal to columnar epithelial cells were arranged in multiple papillary fronds and glandular structures, which were separated and supported by abundant, immature, and fibrovascular stroma. However, mitotic figures were infrequent. At the Ac stage, the seminal vesicular tumor showed an appearance similar to the Ad stage, except that sometimes hemorrhage occurred in the seminal vesicles. Prostatic adenocarcinoma of the Ac stage appeared as intraluminar cribriform to papillary epithelial proliferations that completely or almost completely filled the lumen of several adjacent alveoli. The epithelium that lined the tumors was composed of bland-appearing, cuboidal to tall columnar epithelial cells with maintenance of nuclear polarity and sometimes with basophilic cytoplasm.
3.2. Protein Pattern and Proteomic Analysis
Three SVS samples of each group were pooled for the proteomic measurements. SVS proteins were separated by PAGE and their profile was presented by coomassie blue staining (Figure 2). The protein profile was dominated by a small number of highly expressed proteins. The protein distribution of normal SVS was different from those coming from the tumorous SVS, which displayed more complex protein contents. Based on protein pattern comparison, five major bands were chosen as reference indicators. Bands A and E showed upregulated expression, while band B showed downregulated expression in the tumorous SVS. MS/MS was used to analyze the major proteins in bands A to E, which were identified as albumin, seminal vesicle secretory protein 2 (SVS2), SVS4, SVS5, and hemoglobin beta, respectively. The serum proteins, albumin, and hemoglobin beta, are obviously increased in SVS in the early stages of tumorigenesis.
The proteomic analyses of four groups of SVS, which were normal, Hp, Ad, and Ac, were performed by repeated experimental runs. As a consequence of these analyses, excluding single peptide or single unique peptide matched proteins and keratins that were frequent contaminants in proteomic analysis, 179 proteins met the search criterion and were identified in the proteome by combining four kinds of samples. In order to further improve confidence in the identified proteins, the reproducibility of the matched unique peptides found in every SVS sample was considered as an index to sift through the protein list. The determination of unique characteristics for a peptide was based on the proteotypic peptide sequence that could only match to one protein through the IPI mouse database. The number of unique peptides for one protein was summed with the results of proteomic analyses of the four groups of SVS; therefore, not all peptide signals could be found in the four groups of SVS. Generally, a smaller number of unique peptides coupled with a lower detection rate. We defined the undetectable rate as the average percentage that could not find the matched unique peptides at the designated unique peptide number in the four groups of SVS. With the unique peptide numbers 2, 3, 4, and 5, the undetectable rate was 27.7%, 28.0%, 27.6%, and 12.5%, respectively. When the unique peptide number was more than 5, the undetectable rate decreased to zero (Figure 3). Thus, we chose the unique peptide number 5 as a cut-off criterion to finalize the SVS proteome. Based on this condition, 79 proteins were included in the SVS proteome (Table 1). Among this proteome, 47 proteins were secreted or serum proteins that were the most abundant in this proteome and 16 proteins belonged to seminal vesicle proteins that have been identified or characterized at the previous studies. In addition, a small number of intracellular proteins, including cytoplasm protein, lysosome protein, endoplasmic reticulum protein, and membrane protein were also found in this proteome.