Background
Gastric adenocarcinomas comprise one of the common types of cancers in Asian countries including Japan, being second only to lung cancer as to the number of deaths it causes. In spite of the recent development of diagnostic techniques, most gastric cancer patients are diagnosed at an advanced stage and have a very low five-year survival rate (less than 10%) [
1]. This is partially due to a lack of specific and sensitive biomarkers for the diagnosis and monitoring of disease progress at an early stage, although some gastric tumor markers, including the carcinoembryonic antigen, have been used and are partly effective. As gastric carcinogenesis is a multistep process, comprehensive analysis is also required for individual cases, in which different molecular events occur in each carcinogenic process.
Recently, proteomic analysis was utilized to comprehensively examine protein expression in bodily fluids, tissues and cells [
2‐
4]. This approach, as clinical proteomics, is very useful for identifying disease-associated proteins that show changes in expression and modification corresponding to a disease condition [
5,
6]. These disease-related proteins are expected to be biomarkers for diagnosis and putative targeted proteins for treatment [
7‐
9]. On the other hand, comprehensive analyses of transcriptomes in tumor tissues from various cancer patients using DNA microarrays and gene chips have been performed in recent years [
10]. However, a lack of correlation between changes in mRNAs and carcinogenesis has been demonstrated, and quantitative and qualitative changes of post-translationally modified proteins as final gene products are considered to be more informative than those of mRNAs in tumor tissues for studying the molecular events in carcinogenesis. Proteomic studies for the identification of tumor-associated proteins in gastric cancer are increasing, and proteome databases for gastric tissues [
11] and cell lines [
12] have been constructed. Most of them concern specific proteins or antigens that reflect the chemo- and thermo-resistant properties of stomach cancer [
13‐
15], and that are associated with
Helicobactor pylori [
16,
17]. In the present study, we performed comprehensive proteome analysis of tumor and nontumor tissues in Japanese patients with gastric carcinomas, and identified several proteins of which the expression levels are commonly altered in clinical cases. In particular, the expression of gastrokine-1 (GKN-1) was suggested to be under both transcriptional and translational control.
Discussion
In this study we performed proteomic analysis of tumor and nontumor tissues derived from Japanese gastric adenocarcinoma patients. Sixty-nine different proteins in the tumor tissue from a gastric cancer case were identified by 2D-EP and LC-ESI-MS/MS, which included stress proteins, Hsp70, Hsp90 and chaperonine-containing TCP1 (CCT), for self-protection; glycolytic enzymes, triose phosphate isomerase 1, α-enolase and PGK-1, for the growing energy requirement; cytoskeletal proteins, ezrin, gelsolin isoform b and vimentin; proteins involved in cell differentiation and proliferation, galectin-3 and transferrin; and proteins exhibiting putative tumor suppressor activity, FOV. Many of these proteins have been reported to be associated with the tumorigenesis of gastric adenocarcinomas involving multiple steps and factors [
18,
19].
We also demonstrated that the expression levels of several proteins in tumor tissues were commonly altered in five Japanese patients with gastric adenocarcinomas, compared to those in nontumor tissues. A common increase in protein expression in tumor tissues occurred for MnSOD, HMG-1 and PGK-1, whereas a common decrease in tumor tissues occurred for FOV, GST and AST.
Superoxide (O
2-), a free radical, is essential for the anti-microbial action of granulocytes and monocytes. Superoxide dismutase (SOD) rapidly removes the excess amount of superoxide produced in the stress and biological reactions
in vivo through catalysis of the conversion of superoxide with H
+ to H
2O
2 and O
2. MnSOD, one of the SODs, is located in mitochondria and contributes to the protection of mitochondrial DNA from damage. It has been reported that enhanced expression of MnSOD in progressive gastric cancer should be related to the 5-year survival rate after surgery [
20] and sensitivity to chemotherapy [
21]. In this study, protein expression of MnSOD was up-regulated in 4 of 5 gastric cancer patients, suggesting a self-protecting response. On the other hand, the up-regulation of MnSOD in tumor tissues has been considered to interfere with effective chemotherapy based on radical production, which may cause a decrease in sensitivity to anti-cancer drugs and determine the severity of the cancer.
HMG-1 regulates the transcription of various genes and the structural stabilization of chromosomes as a DNA-binding protein. HMG-1 has been reported to be associated with carcinogenesis and metastasis in colorectal and breast cancer [
22].
Transcriptional up-regulation of HMG-1 in gastric cancer has also been demonstrated [
23]. Expression of HMG-1 in tumor tissues was suggested to be related to resistance to cisplatin [
24]. We demonstrated here an increase in this protein in three cases.
GST is a drug-metabolizing enzyme that catalyzes the conjugation of reduced glutathione to drugs and metabolites, and also contributes to the detoxification of carcinogens. Therefore, a decrease in the activity may be a risk factor for carcinogenesis. Infection by Helicobacter pylori has been reported to cause a decrease in GST expression [
25], suggesting that a decrease in GST activity might be a cause of gastric carcinogenesis. In this study, a decrease in GST protein expression was observed in three cases. A decrease in GST protein expression may be utilized as a biomarker for the diagnosis of gastric tumors.
AST is an aminotransferase that acts on amino acids and α-keto acid, that is distributed widely in human organs. AST is released into the blood stream due to enhanced permeability and destruction of tissues. A elevated concentration of AST protein has been reported in the blood of patients with hepatitis, malignant tumors including hepatomas, and leuchemia. Gene expression of GST has also been demonstrated to be up-regulated in colorectal tumor tissues [
26]. However, a decrease in AST protein expression was commonly observed in all tumor tissues derived from the present five patients with gastric adenocarcinomas. Further study should be performed to elucidate whether a decrease in the AST protein is specifically observed in gastric tumor tissues or not.
Elevated expression of PGK-1, which is one of the enzymes in the glycolytic pathway and which catalyzes the dephosphorylation of 1,3-bisphosphoglycerate to produce ATP, has been observed in many malignant tumor tissues that are dependent on ATP as a major energy source. Solid tumors are thought to need the overproduction of ATP to maintain the enhanced proliferation. The up-regulation of glycolytic enzymes, including PGK-1 in lung cancer [
27] and M2-type pyruvate kinase in colorectal cancer [
28], has been reported and suggested to be useful for cancer screening. However, in the present study, a significant relation between PGK-1 protein expression and the phenotypes of gastric adenocarcinomas was hardly detected. Therefore, PGK-1 was considered to be not suitable for the diagnosis of gastric cancer.
Carbonic anhydrases (CAs) are zinc-containing enzymes that are distributed widely in various organs and that comprise a large family including CA-I to CA-IX. They catalyze the hydration of CO
2 for intermediate metabolism, and maintain the pH and ion equilibrium in the body. So far a direct relationship has been demonstrated between malignant transformation and protein expression for CAs I through VII [
29]. Two earlier studies revealed that the expression of both CA-I and CA-II was significantly reduced in colorectal tumors compared to in normal colorectal epithelia or mucosae [
30]. Another report presented results showing that reduced expression of CA-I and CA-II was correlated with the biological aggressiveness of colorectal cancer and synchronous distant metastasis [
31]. As suggested previously [
32], gastric and colorectal carcinomas may share a similar mechanism of cell proliferation and mucosal malignancy, and may become a biomarker for these carcinomas, because common decreases in the protein expression of CA-I and CA-II were also observed in this study.
FOV, which is identical to a stomach-specific 18-kDa antrum mucosa protein! (AMP-18) [
33,
34], GKN1 [
35], and trefoil factor interactions(z) (TFIZ) [
36], was demonstrated to be dramatically down-regulated or absent in the gastric adenocarcinoma patients in this study. The human AMP-18 and human carcinoantigenic CA11 gene product, which encodes an amino acid sequence that differs from that of AMP-18 in only a single residue [
37,
38], function as secreted growth factors partly responsible for maintaining a normal functional gastric epithelium [
33]. Both AMP-18 and the CA11 gene product have been reported to be intensively expressed in normal stomach tissue but not in most gastric cancers [
33,
37,
38]. Immunohistochemical studies demonstrated that AMP-18 appears to be present in mucosal epithelial cells of the normal human gastric antrum and duodenum [
33]. Recently, TFIZ was demonstrated to act as a growth regulator of gastric epithelial cells through the formation of a heterodimer with trefoil factor 1 (TFF1) [
36]. Our present data further confirmed the high expression of FOV in nontumor gastric tissues and the significant suppression of the protein in gastric adnocarcinomas, indicating that FOV play roles in the maintenance of normal differentiation of epithelial cells and tumor suppression but not in tumor tissues. Furthermore, we demonstrated that transcription of FOV mRNA was also commonly down-regulated in gastric cancer patients, indicating that a marked decrease in FOV protein was caused by suppression of FOV gene expression. Furthermore, in one patient the protein was not detected in nontumor tissues, suggesting that the expression level of FOV protein in individuals may determine the gastric adenocarcinoma phenotype. Accordingly, the expression level of FOV as a biomarker may be informative for the assessment of gastric cancer.
Methods
Materials
DNase I, RNase A, 2-mercaptoethanol (2-ME), glass beads (212–300 μm), Nonidet P-40, acrylamide, N,N,N',N'-tetramethylenediamine (TEMED), sodium dodecylsulfate (SDS), iodoacetamide and dithiothreitol (DTT) were purchased from Sigma (St. Louis, MO). Agarose for isoelectronic focusing (IEF), and Pharmalyte pI 3–10, 4–6.5 and 8–10.5 were from Amersham Bioscience (Piscataway, NJ). Trypsin (sequencing grade) was from Roche (Manheim, Germany). Phenylmethylsulfonyl fluoride (PMSF), thiourea, sorbitol, sodium pyrophosphate, ammonium persulfate, D,L-aspartic acid, trichloroacetic acid, sulfosalicylic acid dihydrate, acetic acid, acetonitrile, formic acid and trifluoroacetic acid (TFA) were from Wako Pure Chemicals (Osaka, Japan). Urea was from Katayama Chemicals (Osaka, Japan). Pepstatin A and leupeptin were from the Peptide Institute (Osaka, Japan). NH4HCO3 and N,N'-methylenebisacrylamide were from Nacalai Tesque (Kyoto, Japan). CBB R-250 was from ICN Biomedicals Inc. (Aurora, OH). Molecular mass standards were from APRO Science, Inc. (Tokushima, Japan). TRIZOL reagent was from Life Technologies (Frederick, MD). Oligo(dT)12–18 primer, deoxynucleotides (dNTPs), and RNaseOUT were from Invitrogen (Carlsbad, CA). M-MLV reverse transcriptase and Taq DNA polymerase were from Promega (Madison, WI).
Tissues and sample preparation
Primary gastric adenocarcinomas and adjacent nontumor mucosae were collected on gastrectomy and provided by the Dept. of Surgery, Graduate School of Medicine, The University of Tokushima, Tokushima, Japan. The research was carried out in accordance with the Declaration of Helsinki of the World Medical Association, and was approved by the ethical committee of the University of Tokushima. Informed consent was also given by all of the patients who provided the clinical samples. Tissues were frozen in a dry ice-methanol bath as soon as possible after dissection and stored in a deep freezer (-80°C) before use. For mRNA analysis, tissues were first immersed in RNAlater (Takara, Tokyo, Japan) before freezing. Detailed clinicopathological data including the tumor stage (according to the AJCC system), site and differentiation, and histological data on the tissue samples are listed in Table
3. None of these cases were classified in the scirrhous type category, and tumor tissue was clearly distinguishted from non-tumor one in each case. For two-dimensional gel electrophoresis (2D-EP), protein extraction from tissues was carried out by the following procedure. Frozen blocks (20–30 mg wet weight) were homogenized with a plastic pestle (Toyobo, Tokyo, Japan) in the presence of glass beads in 10 vol/wet weight of dialysis buffer comprising 5 M urea, 1 M thiourea, 10 mM NaPPi, 1.67 μL/mL 2-ME, 0.005% DNase I, 0.05 mg/ml RNase A, 20 μM leupeptin, 1 mM EDTA, 2 mM PMSF and 20 μM pepstatinA, and then centrifuged at 50,000 rpm for 30 min at 4°C (Beckman -Coulter, Fullerton, CA). The resultant supernatant was used as the tissue extract. Protein concentrations were determined with a Bradford protein assay kit (Bio-Rad, Hercules, CA) using bovine gamma-globulin as a standard.
Table 3
Clinical features of the patients with gastric adenocarcinomas.
A | 63 | Male | UM | G3 | IIIA | T2N2M0H0 |
B | 68 | Female | L | G2 | II | T2N2M0 |
C | 76 | Male | ML | G2 | IV | T4N3M0 |
D | 78 | Female | L | G2 | III | T2N0M0H0 |
E | 77 | Male | ML | G2 | II | T2N1M0 |
Two-dimesional gel electrophoresis
2D-EP was carried out as described previously [
39]. The first-dimensional isoelectric focusing was performed in an 1% (w/v) agarose gel (ϕ 2.6 × 180 mm) with a pH 3–10 gradient at 700 V for 18 hr at 4°C, and the second-dimensional SDS gel electrophoresis was performed with a 5–15% (w/v) acrylamide gradient (
M r range, 6–200 kDa) in a standard slab gel (20 × 13 cm) at 15 mA for 3 h, and then at 70 mA for 2 h at room temperature. Gels were stained with CBB R-250.
Protein samples (500 μg) extracted from the tumor center and surrounding histologically normal mucosa were subjected to 2D-EP and run in pairs side by side.
Some of the stained spots were excised from the 2D-gel, in-gel digested with trypsin and then subjected to LC-ESI-MS/MS analysis as described previously [
40]. The peptide mixture was separated with a reversed phase nanoLC system (Famous, Swichos II, Ultimate, LC Packings, Sunnyvale, CA). The eluted peptides were sprayed directly into an ESI mass spectrometer (Esquire3000 Plus, Bruker-Daltonics, Fremont, CA).
A large volume of MS/MS data was acquired with DataAnalysis 3.1 software (Bruker-Daltonics), converted to text files listing the mass values of the parent ions, and intensities and masses of fragment ions, and then processed with the MASCOT algorithm (Matrix Science Ltd, London, U.K.) to assign peptides in the NCBI non-redundant sequence database using a taxonomic restriction, 'human'. The database search was performed with the parameters described by Yoshimura et al. [
40].
Image analysis and MS peptide sequencing
Image acquisition and analysis were performed with Molecular Imager FXProPlus (Bio-Rad) and ImageMaster software (Bio-Rad). Comparisons were made between gel images of tumor and matched nontumor samples pair by pair. Normalized volume differences were statistically calculated for all five cases. The content of glyceraldehyde 3-phosphodehydrogenase (GAPDH) protein was used as a reference to evaluate the fold alteration of protein expression between tumor and matched nontumor tissues as the levels of GAPDH mRNA and protein were not changed between tissues derived from patients. Consistently and significantly different spots were selected for analysis by LC-ESI-MS/MS. Protein spots!were cut out from gels in small pieces, and subjected to in-gel tryptic digestion overnight [
40]. Peptide mass spectra were recorded, and parameters for spectra acquisition were used as stated previously [
40]. Accuracy in database protein matching using Mascot Search [
41] was judged as a score over 37, which was obtained in most of the analyses.
RNA isolation and RT-PCR analysis
Tumor and matched nontumor samples (approx. 50 mg wet weight) were minced, and then homogenized manually in 1 ml of TRIZOL reagent (Invitrogen) on ice. RNA was isolated without DNase I treatment according to the manufacturer's protocol. Briefly, 0.2 ml of CHCl
3 was added to the homogenate, followed by centrifugation at 20,600 × g for 15 min. An equal volume of 2-propanol was added to the resultant supernatant to precipitate RNA. After centrifugation, the pellet was rinsed with 75% ethanol/diethylpyridylchloride (DEPC)-treated water, followed by drying. The pellet was dissolved in an appropriate volume of DEPC-treated water as the total RNA fraction. For reverse transcription (RT), 2 μg of RNA from each sample was transcribed at 37°C for 1 h in the presence of 200U of Molony leukemia virus reverse transcriptase (Promega), oligo(dT)
12–18 primer, 0.5 mM dNTPs and 50U of RNaseOUT. The PCRs for carbonic anhydrase-I and II, glutatione-S-transferase, FOV, and GAPDH were performed within a linear range of amplification using the selected primer set and conditions, and expected size of products, as summarized in Table
4. The PCR products were analyzed by 1.5% agarose gel electrophoresis and stained with ethidium bromide.
Table 4
Primer sets used for RT-PCR analysis.
MnSOD | sense | 5' - acgcggcctacgtgaacaacctgaa - 3' |
| antisense | 5' - aaccccaacctgagccttggacacc - 3' |
HMG-1 | sense | 5' - cgggaggagcataagaagaagcacc - 3' |
| antisense | 5' - caatggacaggccaggatgttctcc - 3' |
GAPDH | sense | 5' - gtcatccatgacaactttgg - 3' |
| antisense | 5' - tgctgtagccaaattcgttg - 3' |
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TY mainly carried out the molecular studies, participated in the proteomic determination and RT-PCR analysis. YK carried out the proteomic studies, partly participated in the proteomic determination. YY supported the proteomic studies by TY. YT partly participated in the RT-PCR for gastric cancer-related protein. NT parly participated in the proteomic determination. HO carried out the operation of gastric cancer patiens, participated in providing the cancer tissues. AU carried out the operation of gastric cancer patients, participated in providing clinical information. TY supported the proteomic studies and advised the methods. KI conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.