Purpose: We aimed to develop prognostic biomarkers for gastrointestinal stromal tumors (GIST) using a proteomic approach.

Experimental Design: We examined the proteomic profile of GISTs using two-dimensional difference gel electrophoresis. The prognostic performance of biomarker candidates was examined using a large-scale sample set and specific antibodies.

Results: We identified 43 protein spots whose intensity was statistically different between GISTs with good and poor prognosis. Mass spectrometric protein identification showed that the 43 spots corresponded to 25 distinct gene products. Eight of the 43 spots derived from pfetin, a potassium channel protein, and four of the eight pfetin spots had a high discriminative power between the two groups. Western blotting and real-time PCR showed that pfetin expression and tumor metastasis were inversely related. The prognostic performance of pfetin was also examined by immunohistochemistry on 210 GIST cases. The 5-year metastasis-free survival rate was 93.9% and 36.2% for patients with pfetin-positive and pfetin-negative tumors, respectively (P < 0.0001). Univariate and multivariate analyses revealed that pfetin expression was a powerful prognostic factor among the clinicopathologic variables examined, including risk classification and c-kit– or platelet-derived growth factor receptor A mutation status.

Conclusions: These results establish pfetin as a powerful prognostic marker for GISTs and may provide novel therapeutic strategies to prevent metastasis of GIST.

Gastrointestinal stromal tumors (GIST) are the most common primary mesenchymal tumors of the digestive tract, with a prevalence of 15 to 20 per 1,000,000 (1, 2). The clinical course of GISTs spans a wide spectrum from a curable disorder to a highly malignant disease that leads to metastasis and death. Thus, the molecular background of GISTs has been extensively studied to predict the behavior of individual tumors. GISTs are characterized by the presence of mutations and overexpression of c-kit and, clinically, by their response to treatment with imatinib (37). Tumor location and certain molecular aberrations, including c-kit, platelet-derived growth factor (PDGFR), and p16 alterations, have been found to correlate with patient prognosis and response to treatment with imatinib (812). However, none of these variables has been proven to be clinically useful in improving patient outcomes yet. GISTs arise from the intestinal cells of Cajal, which are the mesenchymal pacemaker cells of the guts (13), their biological characteristics, however, remain largely obscure.

Recent comprehensive studies offered a global view of molecular aberrations associated with the malignant spectrum of GISTs. Genomic studies using fluorescence in situ hybridization and array-based comparative genomic hybridization identified chromosomal regions frequently amplified and target genes within these regions, the copy number status of which correlated with tumor behavior (14, 15). Global mRNA expression studies using DNA microarrays identified the genes that are involved in the signaling pathways specific to kit or PDGFR and aberrantly regulated in GISTs (16), the genes associated with histologic features denoting malignancy (17), and the genes differentially expressed based on the KIT genotype and GIST anatomic site (18). These comprehensive studies will further increase our understanding of the biology of GIST and lead to the development of practical tumor markers to support individualized therapy (8). Emerging technologies that examine the overall features of the expressed proteins, namely the proteome, have identified many candidate proteins associated with early diagnosis (19), differential diagnosis (20), prognosis (21), and response to chemotherapy (22) in various diseases. Many lines of evidence have indicated that DNA copy number and mRNA expression levels do not necessarily correspond to the protein contents, and that posttranslational modifications cannot be predicted by DNA sequences (23, 24), suggesting that proteomic studies offer unique data that cannot be obtained by other approaches. The proteomic profile of GISTs has not been established yet, and a proteomic study using a large-scale clinical sample set would complement the genome and transcriptome studies.

In this report, we did a comprehensive quantitative expression study on the intact proteins of GIST clinical samples using two-dimensional difference gel electrophoresis and mass spectrometry. Proteomic studies on peptides have been used to develop tumor markers, but intact proteins have not been considered for this purpose, with a few exceptions. Two-dimensional difference gel electrophoresis, as the most advanced form of two-dimensional gel electrophoresis, has the great advantage of being able to be used to study intact proteins. We found that the expression levels of 43 proteins, including eight variants of pfetin (predominantly fetal-expressed tetramerization domain; potassium channel tetramerisation domain containing protein 12), which was originally reported as a protein highly expressed in fetal cochlea and brain (25), correlated with prognosis. We verified the prognostic value of pfetin expression on 210 GIST cases using immunohistochemistry. Our findings indicate that the use of pfetin expression as a prognostic indicator may facilitate tailored medical care for GIST patients.

Patients and clinical information. We examined the tumor tissues of 212 GIST patients who underwent surgery at the National Cancer Center Hospital consecutively from October 1977 to December 2005. All patients underwent resection with curative intent and were not treated with adjuvant chemotherapy, including treatment with imatinib, until distant metastasis was diagnosed. Histologic features of the tissues were reviewed by three board-certified pathologists (K.S., T.S., and T.H.). Diagnosis was based on the WHO classification system for soft-tissue tumors (26), including the examination of tumor size, presence of necrosis, differentiation, mitotic rate, MIB-1 index, presence of epithelioid cells, and CD34 and CD117 expression. Using this large, well-characterized single hospital–based sample set, we were able to identify proteomic features that differ significantly when examined in relation to certain clinicopathologic variables. This project was approved by the institutional review board of the National Cancer Center.

Previous reports indicated that GIST patients that were histologically classified as of being at low or intermediate risk did not develop metastases within 2 y postsurgery, whereas GIST patients histologically classified as of being at high risk developed metastases within 1 y postsurgery (27). For proteomic analysis, we grouped the GIST samples into two groups. GISTs that had metastases at diagnosis or developed metastases within 1 y postsurgery and were categorized in the high-risk group based on their histologic features were defined as poor-prognosis GISTs (P-GIST; Table 1, samples 1-8). GISTs that did not have metastases within 2 y postsurgery and were grouped in the low- or intermediate-risk group based on the histologic features were defined as good-prognosis GISTs (G-GIST; Table 1, samples 9-17). The samples listed in Supplementary Table S1 were excluded from this classification; samples 18, 19, 24, and 25 were excluded because RNA data were not available for the validation study and the other samples because they did not meet the criteria for classification either as P-GISTs or G-GISTs.

Table 1.

Clinicopathologic features of the training set samples (17 cases)

Sample NoAgeGenderSiteHistologic typesSize (cm)Risk classification*Type of KIT mutationMetastatic site (first development)Metastasis time after diagnosis (mo)Follow-up time after diagnosis (mo)Follow-up status
68 Stomach Spindle 19 High Wild-type Peritoneal metastasis 16 DOD 
56 Stomach Spindle 38 High EX11 deletion Peritoneal metastasis DOD 
58 Stomach Spindle 13 High EX11 deletion Peritoneal metastasis 11 DOD 
50 Rectum Spindle High EX11 deletion Peritoneal metastasis 11 60 AWD 
51 Stomach Mixed (spindle main) 12 High EX11 deletion Peritoneal metastasis 69 AWD 
34 Small intestine Spindle 18 High EX11 deletion Liver metastasis At diagnosis 31 AWD 
68 Small intestine Mixed (spindle main) High EX9 insertion Peritoneal metastasis At diagnosis AWD 
72 Stomach Spindle 25 High EX11 559 V-D Peritoneal metastasis AWD 
64 Stomach Spindle 3.5 Low Wild-type — — 68 NED 
10 64 Stomach Spindle Low Wild-type — — 81 NED 
11 54 Stomach Spindle 10 Intermediate EX11 560 V-G — — 77 NED 
12 68 Small intestine Spindle 3.7 Low EX9 insertion — — 50 NED 
13 77 Small intestine Spindle Low Wild-type — — 69 NED 
14 40 Stomach Spindle 10 Intermediate EX11 559 V-D — — 88 NED 
15 52 Stomach Spindle Intermediate EX11 576 L-P — — 62 NED 
16 76 Small intestine Spindle Intermediate EX11 deletion — — 48 NED 
17 81 Stomach Spindle 5.5 Intermediate EX11 559 V-D — — 43 NED 
Sample NoAgeGenderSiteHistologic typesSize (cm)Risk classification*Type of KIT mutationMetastatic site (first development)Metastasis time after diagnosis (mo)Follow-up time after diagnosis (mo)Follow-up status
68 Stomach Spindle 19 High Wild-type Peritoneal metastasis 16 DOD 
56 Stomach Spindle 38 High EX11 deletion Peritoneal metastasis DOD 
58 Stomach Spindle 13 High EX11 deletion Peritoneal metastasis 11 DOD 
50 Rectum Spindle High EX11 deletion Peritoneal metastasis 11 60 AWD 
51 Stomach Mixed (spindle main) 12 High EX11 deletion Peritoneal metastasis 69 AWD 
34 Small intestine Spindle 18 High EX11 deletion Liver metastasis At diagnosis 31 AWD 
68 Small intestine Mixed (spindle main) High EX9 insertion Peritoneal metastasis At diagnosis AWD 
72 Stomach Spindle 25 High EX11 559 V-D Peritoneal metastasis AWD 
64 Stomach Spindle 3.5 Low Wild-type — — 68 NED 
10 64 Stomach Spindle Low Wild-type — — 81 NED 
11 54 Stomach Spindle 10 Intermediate EX11 560 V-G — — 77 NED 
12 68 Small intestine Spindle 3.7 Low EX9 insertion — — 50 NED 
13 77 Small intestine Spindle Low Wild-type — — 69 NED 
14 40 Stomach Spindle 10 Intermediate EX11 559 V-D — — 88 NED 
15 52 Stomach Spindle Intermediate EX11 576 L-P — — 62 NED 
16 76 Small intestine Spindle Intermediate EX11 deletion — — 48 NED 
17 81 Stomach Spindle 5.5 Intermediate EX11 559 V-D — — 43 NED 

NOTE.PDGFR mutations: All samples lacked of PDGFR mutations. Detail data: Supplementary Table S1.

Abbreviations: NED, no evidence of disease; AWD: alive with disease; DOD, dead of disease.

*

Prognostic classification based on tumor size and MIB-1 grade (Hasegawa, T. et al. Hum Pathol. 2002, 33:669-676).

For the immunohistochemical study, we selected 210 patients who did not have distant metastases at the time of surgery.

Protein expression profiling. Frozen samples were crushed to powder with a CryoPress (Microtech Nichion) under cooling with liquid nitrogen. The frozen powder was then treated with urea lysis buffer (6 mol/L urea, 2 mol/L thiourea, 3% CHAPS, and 1% Triton X-100). After centrifugation at 15,000 rpm for 30 min, the supernatant was used as the source of cellular proteins for protein expression studies.

Two-dimensional difference gel electrophoresis was done as described previously (20, 28, 29). In brief, the internal control sample was prepared by mixing a portion of all individual samples. Five micrograms of the internal control sample and of each individual sample were labeled with Cy3 and Cy5, respectively (CyDye DIGE Fluor saturation dye; GE Healthcare Biosciences) according to the manufacturer's instructions. The differently labeled protein samples were mixed and separated by two-dimensional gel electrophoresis, which was achieved using IPG DryStrip gels for the first dimension separation (length, 24 cm; isoelectric point range, between 4 and 7; GE Healthcare Biosciences) and SDS-PAGE for the second dimension separation (EttanDalt II; GE Healthcare Biosciences). The gels were scanned using laser scanners (Tyhoon Trio; GE Healthcare Biosciences) at appropriate wavelengths. For all spots, the intensity of the Cy5 image was normalized by that of the Cy3 image in the identical gel so that gel-to-gel differences were compensated, using the DeCyder image software (GE Healthcare Biosciences). System reproducibility was verified by comparing the protein profiles obtained from three independent separations of the identical sample (sample 22; Supplementary Table S1). Scatter plot analysis revealed that the standardized intensity of >96% of the spots ranged within a 2.0-fold difference (Supplementary Fig. S1). Representative two-dimensional images with the numbers of the identified spots are shown in Fig. 1A and Supplementary Fig. S3.

Fig. 1.

Identification of proteins differentially expressed in GISTs. A, representative two-dimensional image of proteins detected in GIST tissues. The 43 spots identified in this study are circled and numbered. The spot numbers correspond to those in B, Table 2, and Supplementary Table S2. The image is shown enlarged in Supplementary Fig. S3. B, hierarchical clustering of the 17 GIST cases based on the intensity of the 43 protein spots. Yellow, P-GIST; light blue, G-GIST. Right, spot numbers and the protein names. C, principal component analysis of the 17 GIST samples based on the intensity of the 43 protein spots clearly discriminated between the P-GIST (yellow) and G-GIST (light blue) group. D, the similarity between the samples was measured by calculating the correlation coefficiency of the intensity of the 43 spots; the results are summarized in the correlation matrix shown.

Fig. 1.

Identification of proteins differentially expressed in GISTs. A, representative two-dimensional image of proteins detected in GIST tissues. The 43 spots identified in this study are circled and numbered. The spot numbers correspond to those in B, Table 2, and Supplementary Table S2. The image is shown enlarged in Supplementary Fig. S3. B, hierarchical clustering of the 17 GIST cases based on the intensity of the 43 protein spots. Yellow, P-GIST; light blue, G-GIST. Right, spot numbers and the protein names. C, principal component analysis of the 17 GIST samples based on the intensity of the 43 protein spots clearly discriminated between the P-GIST (yellow) and G-GIST (light blue) group. D, the similarity between the samples was measured by calculating the correlation coefficiency of the intensity of the 43 spots; the results are summarized in the correlation matrix shown.

Close modal

Data analysis. We identified protein spots whose intensity was statistically (Wilcoxon test, P < 0.001) different between the groups examined. Hierarchical clustering, principal component analysis, correlation matrix study, and spot ranking were done using the Expressionist software (Genedata).

Protein identification by mass spectrometry. Proteins corresponding to the spots of interest were identified by mass spectrometry according to our previous report (20, 30). Cy5-labeled proteins separated by 2D-PAGE were recovered in gel plugs and digested with modified trypsin (Promega). The trypsin digests were subjected to liquid chromatography coupled with tandem mass spectrometry equipped with a nanoelectrospray ion source (Paradigm MS4 dual solvent delivery system; Michrom BioResources, Inc.) for microflow high performance liquid chromatography, an HTS PAL auto sampler (CTC Analytics), and a Finnigan LTQ linear ion trap mass spectrometer (Thermo Electron Co.) equipped with a nanoelectrospray ion source (AMR, Inc.). The Mascot software (version 2.1; Matrix Science) was used to search for the mass of the peptide ion peaks against the SWISS-PROT database (Homo sapiens, 12867 sequence in Sprot_47.8 fasta file). Proteins with a Mascot score of 35 or more were used for protein identification. When multiple proteins were identified in a single spot, the proteins with the highest number of peptides were considered as those corresponding to the spot.

Western blotting and immunohistochemistry. Protein samples were separated by SDS-PAGE and subsequently blotted on a nitrocellulose membrane. The membrane was incubated with rabbit polyclonal antibody against pfetin (1:1,000 dilution) kindly provided by Dr. Morton (25), and then horseradish peroxidase–conjugated secondary antibody (1:1,000 dilution; GE Healthcare Biosciences). Pfetin was detected using an emhanced chemiluminescence system (GE Healthcare Biosciences) and LA 1000 (Fuji film).

Pfetin expression was examined immunohistochemically using paraffin-embedded tissues. In brief, 4-μm-thick tissue sections were autoclaved in 10 mmol/L citrate buffer (pH 6.0) at 121°C for 30 min and incubated with the antibody against pfetin (1:500 dilution). Immunostaining was done according to the streptavidin-biotin peroxidase method using the Strept ABC Complex/horseradish peroxidase kit (DAKO). One pathologist (K. S.) and one medical doctor (Y. S.) reviewed the sections stained with antipfetin antibody in a blinded fashion regarding clinical data (age, sex, anatomic site, and outcome). Positively stained cells were defined as those that had higher staining intensity than that of vascular endothelial cells, which served as positive controls. Cases with >20% of tumor cells stained positively with the antipfetin antibody were considered as pfetin positive, whereas cases with <20% pfetin-positive tumor cells were considered as pfetin negative. In most cases, the difference was so obvious that two reviewers had consistent results. Inconsistencies, if any, were resolved by discussion, as a usual process of pathologic diagnosis in the hospital.

Mutation study for c-kit and PDGFRA. We examined the c-kit and PDGFRA genes for the presence of mutations as previously described (31) in the 39 cases where DNA samples were available. In brief, DNA was extracted from the frozen tissues, and the exons including the frequent mutation sites for c-kit and PDGFRA were amplified by PCR. The PCR products were purified with 2% agarose gel electrophoresis, extracted with an QIAquick PCR Purification kit (Qiagen), and sequenced using an ABI Prism 3100 Genetic Analyzer (Applied Biosystems). The primer sets for c-kit were as follows: 5′-TCTAGTGCATTCAAGCACAATGG-3′ and 5′-CATGACTGATATGGTAGACAGAG-3′ for exon 9, and 5′-CCAGAGTGCTCTAATGACTGAGAC-3′ and 5′-AAAGGTGACATGGAAAGCCCCTG-3′ for exon 11.The primer set for exon 12 of the PDGFRA gene was 5′-CCTGGTCATTTATAGAAACCGAG-3′ and 5′-CTCCCATCTTGAGTCATAAGGCA-3′. PCR cycling variables were as follows: cycle at 96°C for 1 min; 50 cycles at 94°C for 30 s, 56°C for 30 s, and 68°C for 2.5 min; and finally 1 cycle at 72°C for 5 min.

Quantitative reverse transcription-PCR. We extracted mRNA and generated cDNA using the Super Script III kit (Invitrogen) in the 39 cases where mRNA samples were available. The quantitative amplification was monitored with Taq Man Gene Expression Assays using premade primers for pfetin and Human glyceraldehyde-3-phosphate dehydrogenase, and Taq Man Universal PCR Master Mix with 7500 Real-time PCR system (Applied Biosystems) according to the manufacturer's instructions.

Statistical analysis. The tumor-specific and metastasis-free survival times were calculated from the first resection of the primary tumor to death from tumor-specific causes or to first evidence of metastasis, respectively. All time-to-event end points were computed by the Kaplan-Meier method (32). Patients who died of causes unrelated to GISTs were censored at the time of death. Potential prognostic factors were identified by univariate analysis using the log-rank test. Independent prognostic factors were evaluated using a Cox's proportional hazards regression model and a stepwise selection procedure. To arrive at a parsimonious multivariate model, covariates were selected into the model only if they contributed significantly to the fit of the model based on the χ2 test value. P value differences of <0.05 were considered to be significant. Statistical analyses were done using the SPSS statistical package (SPSS).

We compared the protein expression profiles between nine G-GISTs and eight P-GISTs using two-dimensional difference gel electrophoresis. We selected 1,513 protein spots that appeared in at least 75% of the images of the Cy3-labeled internal control sample to decrease irrelevant expression data. The G- and P-GIST samples were not classified into their respective groups based on the overall protein expression features of the samples (Supplementary Fig. S2). However, we found that 43 protein spots had significantly different intensity between the two groups (P < 0.01). The localization of the 43 spots on the two-dimensional image is shown in Fig. 1A (enlarged image in Supplementary Fig. S3). Hierarchical clustering and principal component analysis accurately classified the 17 GIST samples into either the G- or the P-GIST group based on the intensity of the 43 selected spots (Fig. 1B and C). The profiles of the 43 spots were similar between samples of the same group and different between the two groups (Fig. 1D). Mass spectrometric protein identification revealed that the 43 protein spots corresponded to 25 distinct gene products (Fig. 1B; Table 2; Supplementary Table S2).

Table 2.

A list of identified protein

Spots no*Accession noIdentified proteinWilcoxon test P valueFold difference (ration of means)Overall RankpI (obs)pI (cal)§MW (obs; kDa)MW (cal; kD)§Protein scorePeptide matchesSequence coverage (%)
1378 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 8.23E-05 4.862 6.4 5.5 35.1 35.7 406 20.0 
1314 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 8.23E-05 5.805 5.4 5.5 44.0 35.7 338 19.7 
1,371 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 8.23E-05 3.781 6.2 5.5 35.7 35.7 146 8.3 
1413 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 1.65E-04 3.198 26 6.5 5.5 34.5 35.7 306 23.4 
575 P17987 T-complex protein 1, α subunit 3.29E-04 1.405 28 5.1 5.8 73.3 60.3 267 12.4 
893 O00148 ATP-dependent helicase DDX39 6.99E-04 2.321 41 5.5 5.5 51.8 49.1 109 3.7 
1367 P11177 Pyruvate dehydrogenase estrone component β subunit 8.94E-04 3.669 24 6.4 6.2 35.1 39.2 174 9.7 
1303 Q96CX2 Pfetin 9.87E-04 2.917 5.2 5.5 44.7 35.7 236 14.5 
1919 P00441 Superoxide dismutase (Cu-Zn) 1.33E-03 2.547 12 6.6 5.7 15.3 15.8 60 7.8 
611 P17987 T-complex protein 1, α subunit 1.56E-03 1.982 4.9 5.8 67.9 60.3 667 11 22.5 
1184 P60709 Actin, cytoplasmic 1 1.56E-03 1.902 20 5.9 5.3 46.3 41.7 276 11.2 
1376 Q96CX2 Pfetin 1.56E-03 2.850 38 6.5 5.5 35.1 35.7 396 11 23.7 
834 P14625 Endoplasmin precursor 1.75E-03 3.487 4.7 4.8 57.2 92.5 977 18 20.5 
483 P26038 Moesin 1.86E-03 1.558 15 5.2 6.1 84.0 67.7 208 6.9 
369 P08238 Heat shock protein HSP 90-β 1.86E-03 1.977 16 4.4 5.0 89.4 83.1 263 7.3 
922 P05091 Aldehyde dehydrogenase, mitochondrial precursor 2.04E-03 1.359 31 5.6 6.6 50.0 56.4 323 10.1 
273 Q9Y4L1 150 kDa oxygen-regulated protein precursor 3.32E-03 1.384 36 4.7 5.2 113.7 111.3 102 2.4 
1436 P07355 Annexin A2 3.32E-03 1.590 39 6.9 7.6 34.5 38.5 262 13.3 
1,748 P62993 Growth factor receptor-bound protein 2 3.32E-03 1.545 40 6.0 5.9 25.5 25.2 79 7.4 
205 P12814 α-actinin 1 3.70E-03 1.993 11 4.4 5.3 114.1 103.1 593 12 12.9 
1710 P04792 Heat-shock protein β-1 3.70E-03 2.078 13 6.1 6.0 27.5 22.8 491 15 54.1 
574 P20700 Lamin B1 3.70E-03 2.575 14 4.2 5.1 75.1 66.3 1616 33 36.4 
1555 P08758 Annexin A5 3.70E-03 1.438 19 6.1 4.9 31.5 35.8 96 5.3 
454 P26038 Moesin 3.70E-03 1.441 37 5.4 6.1 82.3 67.7 366 9.2 
1315 P14625 Endoplasmin precursor 3.73E-03 2.895 23 5.5 4.8 44.3 92.5 216 4.7 
1276 Q9Y3F4 Serine-threonine kinase receptor-associated protein 3.85E-03 1.650 29 5.1 5.0 45.0 38.4 250 15.4 
304 Q01432 AMP deaminase 3 4.00E-03 1.691 22 4.8 6.5 113.4 88.8 185 4.6 
1372 Q96CX2 Pfetin 4.04E-03 2.972 30 6.3 5.5 35.1 35.7 118 8.3 
1478 P40925 Malate dehydrogenase, cytoplasmic 4.08E-03 1.534 27 6.2 6.9 33.6 36.3 116 5.7 
1916 P09382 Galectin-1 4.21E-03 1.917 6.1 5.3 16.9 14.6 244 35.8 
1209 P30740 Leukocyte elastase inhibitor 4.33E-03 2.380 18 5.2 5.9 46.0 42.7 551 23.0 
1884 P00441 Superoxide dismutase (Cu-Zn) 5.25E-03 3.121 6.8 5.7 18.4 15.8 139 17.0 
77 P42704 130 kDa leucine-rich protein 5.51E-03 1.456 21 4.5 5.5 195.2 145.2 120 1.7 
1166 P60709 Actin, cytoplasmic 1 5.51E-03 3.971 42 5.9 5.3 46.3 41.7 246 15.5 
1921 P09382 Galectin-1 6.22E-03 1.912 6.1 5.3 15.8 14.6 90 11.9 
1421 Q96CX2 Pfetin 6.22E-03 1.665 34 6.4 5.5 34.5 35.7 336 13 23.4 
366 P08238 Heat shock protein HSP 90-β 7.02E-03 1.705 25 4.4 5.0 89.4 83.1 420 13.8 
1566 P82979 Nuclear protein Hcc-1 7.02E-03 1.526 43 6.1 6.1 30.6 23.5 196 13.9 
473 P26038 Moesin 7.59E-03 1.323 32 5.6 6.1 80.5 67.7 775 15 18.8 
1205 Q9NYL9 Tropomodulin-3 7.90E-03 1.935 10 5.2 5.1 46.0 39.6 168 9.1 
283 P08238 Heat shock protein HSP 90-β 7.90E-03 1.300 33 4.1 5.0 113.1 83.1 907 17 22.1 
463 P26038 Moesin 7.90E-03 1.524 35 5.4 6.1 82.3 67.7 356 9.7 
841 P14625 Endoplasmin precursor 9.45E-03 2.478 17 4.7 4.8 57.2 92.5 159 4.0 
Spots no*Accession noIdentified proteinWilcoxon test P valueFold difference (ration of means)Overall RankpI (obs)pI (cal)§MW (obs; kDa)MW (cal; kD)§Protein scorePeptide matchesSequence coverage (%)
1378 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 8.23E-05 4.862 6.4 5.5 35.1 35.7 406 20.0 
1314 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 8.23E-05 5.805 5.4 5.5 44.0 35.7 338 19.7 
1,371 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 8.23E-05 3.781 6.2 5.5 35.7 35.7 146 8.3 
1413 Q96CX2 Potassium channel tetramerisation domain containing protein 12 (Pfetin) 1.65E-04 3.198 26 6.5 5.5 34.5 35.7 306 23.4 
575 P17987 T-complex protein 1, α subunit 3.29E-04 1.405 28 5.1 5.8 73.3 60.3 267 12.4 
893 O00148 ATP-dependent helicase DDX39 6.99E-04 2.321 41 5.5 5.5 51.8 49.1 109 3.7 
1367 P11177 Pyruvate dehydrogenase estrone component β subunit 8.94E-04 3.669 24 6.4 6.2 35.1 39.2 174 9.7 
1303 Q96CX2 Pfetin 9.87E-04 2.917 5.2 5.5 44.7 35.7 236 14.5 
1919 P00441 Superoxide dismutase (Cu-Zn) 1.33E-03 2.547 12 6.6 5.7 15.3 15.8 60 7.8 
611 P17987 T-complex protein 1, α subunit 1.56E-03 1.982 4.9 5.8 67.9 60.3 667 11 22.5 
1184 P60709 Actin, cytoplasmic 1 1.56E-03 1.902 20 5.9 5.3 46.3 41.7 276 11.2 
1376 Q96CX2 Pfetin 1.56E-03 2.850 38 6.5 5.5 35.1 35.7 396 11 23.7 
834 P14625 Endoplasmin precursor 1.75E-03 3.487 4.7 4.8 57.2 92.5 977 18 20.5 
483 P26038 Moesin 1.86E-03 1.558 15 5.2 6.1 84.0 67.7 208 6.9 
369 P08238 Heat shock protein HSP 90-β 1.86E-03 1.977 16 4.4 5.0 89.4 83.1 263 7.3 
922 P05091 Aldehyde dehydrogenase, mitochondrial precursor 2.04E-03 1.359 31 5.6 6.6 50.0 56.4 323 10.1 
273 Q9Y4L1 150 kDa oxygen-regulated protein precursor 3.32E-03 1.384 36 4.7 5.2 113.7 111.3 102 2.4 
1436 P07355 Annexin A2 3.32E-03 1.590 39 6.9 7.6 34.5 38.5 262 13.3 
1,748 P62993 Growth factor receptor-bound protein 2 3.32E-03 1.545 40 6.0 5.9 25.5 25.2 79 7.4 
205 P12814 α-actinin 1 3.70E-03 1.993 11 4.4 5.3 114.1 103.1 593 12 12.9 
1710 P04792 Heat-shock protein β-1 3.70E-03 2.078 13 6.1 6.0 27.5 22.8 491 15 54.1 
574 P20700 Lamin B1 3.70E-03 2.575 14 4.2 5.1 75.1 66.3 1616 33 36.4 
1555 P08758 Annexin A5 3.70E-03 1.438 19 6.1 4.9 31.5 35.8 96 5.3 
454 P26038 Moesin 3.70E-03 1.441 37 5.4 6.1 82.3 67.7 366 9.2 
1315 P14625 Endoplasmin precursor 3.73E-03 2.895 23 5.5 4.8 44.3 92.5 216 4.7 
1276 Q9Y3F4 Serine-threonine kinase receptor-associated protein 3.85E-03 1.650 29 5.1 5.0 45.0 38.4 250 15.4 
304 Q01432 AMP deaminase 3 4.00E-03 1.691 22 4.8 6.5 113.4 88.8 185 4.6 
1372 Q96CX2 Pfetin 4.04E-03 2.972 30 6.3 5.5 35.1 35.7 118 8.3 
1478 P40925 Malate dehydrogenase, cytoplasmic 4.08E-03 1.534 27 6.2 6.9 33.6 36.3 116 5.7 
1916 P09382 Galectin-1 4.21E-03 1.917 6.1 5.3 16.9 14.6 244 35.8 
1209 P30740 Leukocyte elastase inhibitor 4.33E-03 2.380 18 5.2 5.9 46.0 42.7 551 23.0 
1884 P00441 Superoxide dismutase (Cu-Zn) 5.25E-03 3.121 6.8 5.7 18.4 15.8 139 17.0 
77 P42704 130 kDa leucine-rich protein 5.51E-03 1.456 21 4.5 5.5 195.2 145.2 120 1.7 
1166 P60709 Actin, cytoplasmic 1 5.51E-03 3.971 42 5.9 5.3 46.3 41.7 246 15.5 
1921 P09382 Galectin-1 6.22E-03 1.912 6.1 5.3 15.8 14.6 90 11.9 
1421 Q96CX2 Pfetin 6.22E-03 1.665 34 6.4 5.5 34.5 35.7 336 13 23.4 
366 P08238 Heat shock protein HSP 90-β 7.02E-03 1.705 25 4.4 5.0 89.4 83.1 420 13.8 
1566 P82979 Nuclear protein Hcc-1 7.02E-03 1.526 43 6.1 6.1 30.6 23.5 196 13.9 
473 P26038 Moesin 7.59E-03 1.323 32 5.6 6.1 80.5 67.7 775 15 18.8 
1205 Q9NYL9 Tropomodulin-3 7.90E-03 1.935 10 5.2 5.1 46.0 39.6 168 9.1 
283 P08238 Heat shock protein HSP 90-β 7.90E-03 1.300 33 4.1 5.0 113.1 83.1 907 17 22.1 
463 P26038 Moesin 7.90E-03 1.524 35 5.4 6.1 82.3 67.7 356 9.7 
841 P14625 Endoplasmin precursor 9.45E-03 2.478 17 4.7 4.8 57.2 92.5 159 4.0 
*

Spot numbers refer to those in Fig. 1A and Supplementary Fig. S3.

Accession numbers of protein were derived from SWISS-PROT and National Center for Biotechnology Information nonredundant databases.

Observed isoelectric point and molecular weight calculated according to location on the two-dimensional gel.

§

Theoretical isoelectric point and molecular weight obtained from Swiss-Prot and the ExPASy database. (http://au.expasy.org).

Mascot score for the identified proteins based on the peptide ions score (P < 0.05) (http://www.matrixscience.com).

We aimed to prioritize the protein spots according to their discriminative power for the two groups. We created a classifier based on a support vector machine algorithm that used the intensity of the 43 spots and ranked the 43 spots according to their contribution to the classification using support vector machine weight algorithm (Table 2). We found that 4 of the 8 identified pfetin spots were ranked within the top 10 protein spots whose intensity was different between the groups (Table 2), and that pfetin spots appeared 8 times in the list of the 43 protein spots (Fig. 1B; Table 2).

Pfetin is highly expressed in fetal cochlea and brain (25), consistent with the fact that the origin of GIST is Cajal cells, neuronal cells in the gut. Thus, we further validated the relationship of the expression of pfetin with the malignant potential of GISTs. SDS-PAGE/Western blotting showed that the expression of pfetin was lower in the P-GIST compared with the G-GIST group (Fig. 2A). Two bands on SDS-PAGE/Western blotting corresponded to the location of protein spots for pfetin variants on the two-dimensional image (Fig. 1A; Supplementary Fig. S3). These results were further validated in an additional four GIST samples that were not included in the initial proteome study (Supplementary Fig. S4A). Pfetin expression was not observed in the 6 liver metastases examined or in primary high-risk GISTs that developed metastases between 13 and 30 months postsurgery (Supplementary Fig. S4B; Supplementary Table S1).

Fig. 2.

Validation of the differential expression of pfetin. G-GISTs expressed pfetin at significantly higher levels than P-GISTs. A, Western blotting. Case numbers correspond to those in Fig. 1. B, immunohistochemistry; pfetin is overexpressed in G-GISTs (top), whereas it is not expressed in P-GISTs (bottom). C, the pfetin mRNA expression levels detected in the 17 GIST samples examined. P-GIST, 1:8; G-GIST, 9:17.

Fig. 2.

Validation of the differential expression of pfetin. G-GISTs expressed pfetin at significantly higher levels than P-GISTs. A, Western blotting. Case numbers correspond to those in Fig. 1. B, immunohistochemistry; pfetin is overexpressed in G-GISTs (top), whereas it is not expressed in P-GISTs (bottom). C, the pfetin mRNA expression levels detected in the 17 GIST samples examined. P-GIST, 1:8; G-GIST, 9:17.

Close modal

We used immunohistochemistry to evaluate pfetin expression in situ. Positive cells were diffusely stained with antipfetin antibody in membrane and cytoplasm. All nine G-GISTs expressed pfetin, whereas none of the eight P-GISTs did (Fig. 2B). Pfetin expression was not observed in neighboring host cells or in the intestinal cells of Cajal (Supplementary Fig. S5).

Real-time RT-PCR revealed pfetin mRNA levels were higher in G-GISTs than in P-GISTs (Fig. 2C). However, the difference between the P- and G- GIST group was less obvious at the mRNA than at the protein level, suggesting that pfetin expression is partially regulated at the mRNA level, and that posttranscriptional regulation may also play an important role in pfetin expression.

As pfetin expression has been reported to correlate with c-kit mutation status (33), we examined 39 primary GISTs for the presence of c-kit and PDGFRA gene mutations and monitored their pfetin expression levels by Western blotting (Supplementary Tables S1 and S3). Overexpression of pfetin was observed in 12 of 29 c-kit mutation positive cases and in 6 of 10 negative cases (P = 0.389; Supplementary Table S3). PDGFRA mutations were not detected in the series. We observed no significant correlation between pfetin expression and c-kit or PDGFR mutation status.

The immunohistochemical study of 210 GISTs revealed a strong correlation between pfetin expression and a number of clinicopathologic variables including the tumor size, mitotic index, MIB-1 index, degree of differentiation, and risk classification (P < 0.0001; Table 3). Moreover, distant metastasis was observed in a significantly higher proportion of patients with pfetin-negative tumors compared with those with pfetin-positive tumors (24 of 39 versus 12 of 171 cases; P < 0.0001), with a median follow-up period of 73 months. The 5-year metastasis-free survival rate was significantly higher in the pfetin-positive than in the negative group overall (93.9% versus 36.2%; P < 0.0001; Fig. 3A; Table 3) as well as within each risk group (Fig. 3B-D). Multivariate analysis revealed that pfetin expression was a powerful predictor of disease-specific survival (Table 3). Note that high-risk cases were divided into two groups, the pfetin-positive and the pfetin-negative group, the latter having a worse prognosis. Furthermore, tumor-specific survival was statistically significantly longer in the pfetin-positive compared with the pfetin-negative group (P < 0.0001; Table 3; Supplementary Fig. S6). These data clearly indicate that prognosis relying solely on the established risk classification system is not sufficiently accurate to determine the post-operative therapeutic strategy for GIST patients, and the use of pfetin expression may further refine the prognostic criteria so as to identify patients who may benefit from additional therapy.

Table 3.

Univariate and multivariate analysis of prognostic factor and the relationship between clinicopathologic variables and pfetin expression

VariableNumber of casesMetastasis-free survival
Tumor-specific survival
Multivariate analysis of metastasis free survival by Cox regression
Pfetin positive (no. cases)Pfetin negative (no. cases)Correlation (pfetin) χ2 (P)
5 y (%)Log-rank (P)5-y (%)Log-rank (P)PRelative risk95% CI
Age   0.3290  0.8350      0.5220 
    <60 112 85.8 ± 3.7  91.8 ± 3.0     93 19  
    60< 98 80.7 ± 4.2  95.6 ± 2.2     78 20  
Sex   0.4420  0.0393      0.8909 
    F 99 86.5 ± 3.7  96.1 ± 2.2     81 18  
    M 111 80.4 ± 4.2  91.6 ± 2.8     90 21  
Site   0.0001  0.1655      0.4776 
    Stomach 170 87.8 ± 2.7  93.9 ± 2.0     140 30  
    Nonstomach 40 60.7 ± 9.2  92.2 ± 5.3  0.0270 2.21 1.09-4.49 31  
Histology   0.1003  0.2068      0.5153 
    Spindle 189 85.4 ± 2.8  93.6 ± 2.0     155 34  
    Epithelioid 21 67.7 ± 10.9  95.2 ± 4.6     16  
Size   <0.0001  <0.0001      <0.0001 
    <5 cm 128 92.4 ± 2.6  98.1 ± 1.3     112 16  
    5-10 cm 63 76.6 ± 0.4  93.7 ± 3.5     51 12  
    15 cm< 19 45.0 ± 11.9  60.5 ± 13.0  0.0070 2.05 1.22-3.44 11  
Necrosis   <0.0001  0.0034      0.0070 
    + 19 43.0 ± 12.7  76.9 ± 11.7     10 30  
    − 191 86.9 ± 2.7  95.1 ± 1.7     161  
Miosis   <0.0001  <0.0001      <0.0001 
    <5/50HPF* 148 95.5 ± 2.0  98.2 ± 1.3     136 12  
    5-10/50HPF 33 80.4 ± 7.2  96.8 ± 3.2     26  
    5/50HPF< 29 29.9 ± 9.0  68.2 ± 9.4     20  
MIB-1   <0.0001  <0.0001      <0.0001 
    <9% 164 96.0 ± 1.8  98.4 ± 1.1     152 12  
    10-29% 19 51.4 ± 12.6  85.6 ± 9.7     11  
    <30% 27 32.5 ± 9.6  70.8 ± 9.4     19  
Differentiation   <0.0001  <0.0001      <0.0001 
    Score 1 161 95.9 ± 1.8  98.4 ± 1.1     149 12  
    Score 2 49 43.5 ± 7.7  78.1 ± 6.6  <0.0001 10.40 3.68-29.45 22 27  
Risk classification   <0.0001  <0.0001      <0.0001 
    Low 110 97.7 ± 1.6  98.9 ± 1.1     100 10  
    Intermediate 46 90.6 ± 5.2  97.0 ± 3.0     44  
    High 54 48.5 ± 7.4  80.6 ± 6.0     27 27  
Pfetin   <0.0001  <0.0001       
    Positive 171 93.9 ± 2.0  97.2 ± 1.4        
    Negative 39 36.2 ± 8.7  76.5 ± 7.9  0.0020 3.75 1.60-8.81    
VariableNumber of casesMetastasis-free survival
Tumor-specific survival
Multivariate analysis of metastasis free survival by Cox regression
Pfetin positive (no. cases)Pfetin negative (no. cases)Correlation (pfetin) χ2 (P)
5 y (%)Log-rank (P)5-y (%)Log-rank (P)PRelative risk95% CI
Age   0.3290  0.8350      0.5220 
    <60 112 85.8 ± 3.7  91.8 ± 3.0     93 19  
    60< 98 80.7 ± 4.2  95.6 ± 2.2     78 20  
Sex   0.4420  0.0393      0.8909 
    F 99 86.5 ± 3.7  96.1 ± 2.2     81 18  
    M 111 80.4 ± 4.2  91.6 ± 2.8     90 21  
Site   0.0001  0.1655      0.4776 
    Stomach 170 87.8 ± 2.7  93.9 ± 2.0     140 30  
    Nonstomach 40 60.7 ± 9.2  92.2 ± 5.3  0.0270 2.21 1.09-4.49 31  
Histology   0.1003  0.2068      0.5153 
    Spindle 189 85.4 ± 2.8  93.6 ± 2.0     155 34  
    Epithelioid 21 67.7 ± 10.9  95.2 ± 4.6     16  
Size   <0.0001  <0.0001      <0.0001 
    <5 cm 128 92.4 ± 2.6  98.1 ± 1.3     112 16  
    5-10 cm 63 76.6 ± 0.4  93.7 ± 3.5     51 12  
    15 cm< 19 45.0 ± 11.9  60.5 ± 13.0  0.0070 2.05 1.22-3.44 11  
Necrosis   <0.0001  0.0034      0.0070 
    + 19 43.0 ± 12.7  76.9 ± 11.7     10 30  
    − 191 86.9 ± 2.7  95.1 ± 1.7     161  
Miosis   <0.0001  <0.0001      <0.0001 
    <5/50HPF* 148 95.5 ± 2.0  98.2 ± 1.3     136 12  
    5-10/50HPF 33 80.4 ± 7.2  96.8 ± 3.2     26  
    5/50HPF< 29 29.9 ± 9.0  68.2 ± 9.4     20  
MIB-1   <0.0001  <0.0001      <0.0001 
    <9% 164 96.0 ± 1.8  98.4 ± 1.1     152 12  
    10-29% 19 51.4 ± 12.6  85.6 ± 9.7     11  
    <30% 27 32.5 ± 9.6  70.8 ± 9.4     19  
Differentiation   <0.0001  <0.0001      <0.0001 
    Score 1 161 95.9 ± 1.8  98.4 ± 1.1     149 12  
    Score 2 49 43.5 ± 7.7  78.1 ± 6.6  <0.0001 10.40 3.68-29.45 22 27  
Risk classification   <0.0001  <0.0001      <0.0001 
    Low 110 97.7 ± 1.6  98.9 ± 1.1     100 10  
    Intermediate 46 90.6 ± 5.2  97.0 ± 3.0     44  
    High 54 48.5 ± 7.4  80.6 ± 6.0     27 27  
Pfetin   <0.0001  <0.0001       
    Positive 171 93.9 ± 2.0  97.2 ± 1.4        
    Negative 39 36.2 ± 8.7  76.5 ± 7.9  0.0020 3.75 1.60-8.81    

Abbreviation: 95% CI, 95% confidence interval.

*

HPF, high-power field.

Fig. 3.

Pfetin expression in the primary tumor samples was predictive of the metastasis-free survival period. Statistically significant differences in the metastasis-free survival period were observed between the pfetin-positive and pfetin-negative groups (P < 0.0001) both for the M0 GIST patients overall (A) and within each risk group of patients (B-D).

Fig. 3.

Pfetin expression in the primary tumor samples was predictive of the metastasis-free survival period. Statistically significant differences in the metastasis-free survival period were observed between the pfetin-positive and pfetin-negative groups (P < 0.0001) both for the M0 GIST patients overall (A) and within each risk group of patients (B-D).

Close modal

Employing proteomics tools, we identified 43 protein variants corresponding to 25 distinct gene products that distinguished GIST patients according to their clinical outcome. The discriminating power of this set of proteins may be developed for use in a clinical setting. However, albeit useful to describe complex clinical variables, cutting-edge proteomic technologies cannot be transferred easily to a hospital setting, considering the high installation costs and labor intensity, the low throughput, and the required operational skills. A smaller number of proteins, measurable by simpler techniques, may be preferable for use in practice. With this notion, we showed that pfetin expression can be examined by SDS-PAGE/Western blotting, immunohistochemistry, and quantitative RT-PCR. Moreover, using a large-scale sample set, we showed that the expression levels of pfetin, as evaluated by immunohistochemistry, are predictive of patient outcome. Therefore, evaluation of pfetin expression can be applied in a clinical setting using the existing examination protocol and may allow the identification of a high-risk patient group that may benefit from adjuvant therapy, such as treatment with imatinib, while it may also help spare low-risk patients unnecessary treatment. As mass spectrometric global surveys for phosphorylated proteins identified pfetin as a phosphorylated protein (34, 35), the multiple protein spots of pfetin may correspond to the different phosphorylation variants.

Recently, Kang et al. (33) did a proteomic study on 12 GIST samples using two-dimensional gel electrophoresis and reported that pfetin overexpression (C13orf2 in their report) correlated with histologic grading and the presence of c-kit mutations. In contrast, our results indicated that pfetin expression is inversely correlated with histologic grading (Figs. 1 and 2), and that pfetin expression levels are not associated with c-kit mutation status (Supplementary Tables S1 and S3). Moreover, the proteins reported to correlate with histologic malignancy by Kang et al. (33), including Annexin V, HMGB1, glutamate dehydrogenase 1, and fibrinogen β chain RoXaN (33), were not identified as such in our study, whereas 24 gene products identified in this study were not listed in their report (33). As these discrepancies may be due to differences in patient populations and proteomic modalities used, an international project to integrate all reported proteomic data in a common proteomic platform is needed to elucidate the molecular background of GISTs.

Although pfetin is known to contain a voltage-gated potassium (K+) channel tetramerization domain (25), its function in the process of cancer development and progression is unknown. Although GISTs originate from Cajal cells, immunohistochemistry revealed that pfetin was absent in Gajal cells (Supplementary Fig. S5). Proteomic analysis of pfetin-associated proteins may provide clues to understanding the role of pfetin in GIST development and progression.

Our study has the limitation of not detecting proteins expressed in low levels. We did not observe overexpression of the kit (3641) or PDGFRA gene products, or loss of CD44 (42) or p16 (11, 12). In addition, we did not detect CD34 (18) or connexin 43 (43) expression, reported to be commonly up-regulated in stomach and small intestinal GISTs, respectively. Aberrant regulation of these gene products was initially detected at the mRNA level and was later confirmed at the protein level using specific antibodies. Presently, any global approach to protein expression cannot uncover the whole proteome in a quantitative and reproducible way. The continuing efforts to improve the sensitivity of proteomic modalities have enabled the uncovering of several thousands of proteins with posttranslational modifications (24, 44, 45). We believe that such efforts will overcome some of the inherent limitations of proteomics and lead to a more detailed understanding of the disease mechanisms and to novel therapeutic strategies in the near future.

In conclusion, we identified a possible correlation of 43 protein variants corresponding to 25 distinct gene products with variables of clinical interest in GIST and validated pfetin expression using a specific antibody. From this study, pfetin expression is predictive of metastasis and survival of patients with GISTs and, as such, may be used in clinical practice to improve existing therapeutic strategies. Assessment of the prognostic power of the combined use of pfetin and the other 24 proteins as well as more extensive validation of pfetin using additional samples are now under way in our laboratory.

Grant support: Ministry of Health, Labor, and Welfare and by the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation of Japan (patent no. 2006-286087).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

Current address for T. Hasegawa: Department of Clinical Pathology, Sapporo Medical University School of Medicine.

We thank Kiyoaki Nomoto, Chizu Kina, and Sachiko Miura for their excellent technical support in the immunohistochemical study, and Kano Nishiyama and Yukiko Fujie in electrophoresis.

1
Joensuu H, Kindblom LG. Gastrointestinal stromal tumors-a review.
Acta Orthop Scand Suppl
2004
;
75
:
62
–71.
2
Miettinen M, El-Rifai W, HL Sobin L, Lasota J. Evaluation of malignancy and prognosis of gastrointestinal stromal tumors: a review.
Hum Pathol
2002
;
33
:
478
–83.
3
Corless CL, Fletcher JA, Heinrich MC. Biology of gastrointestinal stromal tumors.
J Clin Oncol
2004
;
22
:
3813
–25.
4
Heinrich MC, Rubin BP, Longley BJ, Fletcher JA. Biology and genetic aspects of gastrointestinal stromal tumors: KIT activation and cytogenetic alterations.
Hum Pathol
2002
;
33
:
484
–95.
5
Miettinen M, Majidi M, Lasota J. Pathology and diagnostic criteria of gastrointestinal stromal tumors (GISTs): a review.
Eur J Cancer
2002
;
38
Suppl 5:
S39
–51.
6
Demetri GD, von Mehren M, Blanke CD, et al. Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors.
N Engl J Med
2002
;
347
:
472
–80.
7
van Oosterom AT, Judson I, Verweij J, et al. Safety and efficacy of imatinib (STI571) in metastatic gastrointestinal stromal tumours: a phase I study.
Lancet
2001
;
358
:
1421
–3.
8
Heinrich MC, Corless CL, Blanke CD, et al. Molecular correlates of imatinib resistance in gastrointestinal stromal tumors.
J Clin Oncol
2006
;
24
:
4764
–74.
9
Debiec-Rychter M, Sciot R, Le Cesne A, et al. KIT mutations and dose selection for imatinib in patients with advanced gastrointestinal stromal tumours.
Eur J Cancer
2006
;
42
:
1093
–103.
10
Debiec-Rychter M, Dumez H, Judson I, et al. Use of c-KIT/PDGFRA mutational analysis to predict the clinical response to imatinib in patients with advanced gastrointestinal stromal tumours entered on phase I and II studies of the EORTC Soft Tissue and Bone Sarcoma Group.
Eur J Cancer
2004
;
40
:
689
–95.
11
Schneider-Stock R, Boltze C, Lasota J, et al. Loss of p16 protein defines high-risk patients with gastrointestinal stromal tumors: a tissue microarray study.
Clin Cancer Res
2005
;
11
:
638
–45.
12
Schneider-Stock R, Boltze C, Lasota J, et al. High prognostic value of p16INK4 alterations in gastrointestinal stromal tumors.
J Clin Oncol
2003
;
21
:
1688
–97.
13
Rumessen JJ, Peters S, Thuneberg L. Light- and electron microscopical studies of interstitial cells of Cajal and muscle cells at the submucosal border of human colon.
Lab Invest
1993
;
68
:
481
–95.
14
Tornillo L, Duchini G, Carafa V, et al. Patterns of gene amplification in gastrointestinal stromal tumors (GIST).
Lab Invest
2005
;
85
:
921
–31.
15
El-Rifai W, Sarlomo-Rikala M, Andersson LC, Knuutila S, Miettinen M. DNA sequence copy number changes in gastrointestinal stromal tumors: tumor progression and prognostic significance.
Cancer Res
2000
;
60
:
3899
–903.
16
Kang HJ, Nam SW, Kim H, et al. Correlation of KIT and platelet-derived growth factor receptor α mutations with gene activation and expression profiles in gastrointestinal stromal tumors.
Oncogene
2005
;
24
:
1066
–74.
17
Koon N, Schneider-Stock R, Sarlomo-Rikala M, et al. Molecular targets for tumour progression in gastrointestinal stromal tumours.
Gut
2004
;
53
:
235
–40.
18
Antonescu CR, Viale A, Sarran L, et al. Gene expression in gastrointestinal stromal tumors is distinguished by KIT genotype and anatomic site.
Clin Cancer Res
2004
;
10
:
3282
–90.
19
Petricoin EF, Ardekani AM, Hitt BA, et al. Use of proteomic patterns in serum to identify ovarian cancer.
Lancet
2002
;
359
:
572
–7.
20
Suehara Y, Kondo T, Fujii K, et al. Proteomic signatures corresponding to histological classification and grading of soft-tissue sarcomas.
Proteomics
2006
;
6
:
4402
–9.
21
Chen G, Gharib TG, Wang H, et al. Protein profiles associated with survival in lung adenocarcinoma.
Proc Natl Acad Sci U S A
2003
;
100
:
13537
–42.
22
Okano T, Kondo T, Fujii K, et al. Proteomic signature corresponding to the response to gefitinib (Iressa, ZD1839), an epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor, and mutation in EGFR in lung adenocarcinoma.
Clin Cancer Res
2007
;
13
:
799
–805.
23
Orntoft TF, Thykjaer T, Waldman FM, Wolf H, Celis JE. Genome-wide study of gene copy numbers, transcripts, and protein levels in pairs of non-invasive and invasive human transitional cell carcinomas.
Mol Cell Proteomics
2002
;
1
:
37
–45.
24
Varambally S, Yu J, Laxman B, et al. Integrative genomic and proteomic analysis of prostate cancer reveals signatures of metastatic progression.
Cancer Cell
2005
;
8
:
393
–406.
25
Resendes BL, Kuo SF, Robertson NG, et al. Isolation from cochlea of a novel human intronless gene with predominant fetal expression.
J Assoc Res Otolaryngol
2004
;
5
:
185
–202.
26
Stankey RH LA. Pathology and Genetics of Tumours of the Digestive System. Lyon: IARC Press; 2000.
27
Hasegawa T, Matsuno Y, Shimoda T, Hirohashi S. Gastrointestinal stromal tumor: consistent CD117 immunostaining for diagnosis, and prognostic classification based on tumor size and MIB-1 grade.
Hum Pathol
2002
;
33
:
669
–76.
28
Unlu M, Morgan ME, Minden JS. Difference gel electrophoresis: a single gel method for detecting changes in protein extracts.
Electrophoresis
1997
;
18
:
2071
–7.
29
Yokoo H, Kondo T, Fujii K, Yamada T, Todo S, Hirohashi S. Proteomic signature corresponding to α fetoprotein expression in liver cancer cells.
Hepatology
2004
;
40
:
609
–17.
30
Okano T, Kondo T, Kakisaka T, et al. Plasma proteomics of lung cancer by a linkage of multi-dimensional liquid chromatography and two-dimensional difference gel electrophoresis (2D-DIGE).
Proteomics
2006
;
6
:
3938
–48.
31
Yoshida Y, Shibata T, Kokubu A, et al. Mutations of the epidermal growth factor receptor gene in atypical adenomatous hyperplasia and bronchioloalveolar carcinoma of the lung.
Lung Cancer
2005
;
50
:
1
–8.
32
Kaplan EL, Meier P. Nonparametric estimation from incomplete observations.
J Am Stat Assoc
1958
;
53
:
457
–81.
33
Kang HJ, Koh KH, Yang E, et al. Differentially expressed proteins in gastrointestinal stromal tumors with KIT and PDGFRA mutations.
Proteomics
2006
;
6
:
1151
–7.
34
Molina H, Horn DM, Tang N, Mathivanan S, Pandey A. Global proteomic profiling of phosphopeptides using electron transfer dissociation tandem mass spectrometry.
Proc Natl Acad Sci U S A
2007
;
104
:
2199
–204.
35
Matsuoka S, Ballif BA, Smogorzewska A, et al. ATM and ATR substrate analysis reveals extensive protein networks responsive to DNA damage.
Science
2007
;
316
:
1160
–6.
36
Lasota J, Jasinski M, Sarlomo-Rikala M, Miettinen M. Mutations in exon 11 of c-Kit occur preferentially in malignant versus benign gastrointestinal stromal tumors and do not occur in leiomyomas or leiomyosarcomas.
Am J Pathol
1999
;
154
:
53
–60.
37
Ernst SI, Hubbs AE, Przygodzki RM, Emory TS, Sobin LH, O'Leary TJ. KIT mutation portends poor prognosis in gastrointestinal stromal/smooth muscle tumors.
Lab Invest
1998
;
78
:
1633
–6.
38
Hirota S, Isozaki K, Nishida T, Kitamura Y. Effects of loss-of-function and gain-of-function mutations of c-kit on the gastrointestinal tract.
J Gastroenterol
2000
;
35
Suppl 12:
75
–9.
39
Wardelmann E, Losen I, Hans V, et al. Deletion of Trp-557 and Lys-558 in the juxtamembrane domain of the c-kit protooncogene is associated with metastatic behavior of gastrointestinal stromal tumors.
Int J Cancer
2003
;
106
:
887
–95.
40
Lasota J, Dansonka-Mieszkowska A, Stachura T, et al. Gastrointestinal stromal tumors with internal tandem duplications in 3′ end of KIT juxtamembrane domain occur predominantly in stomach and generally seem to have a favorable course.
Mod Pathol
2003
;
16
:
1257
–64.
41
Antonescu CR, Sommer G, Sarran L, et al. Association of KIT exon 9 mutations with nongastric primary site and aggressive behavior: KIT mutation analysis and clinical correlates of 120 gastrointestinal stromal tumors.
Clin Cancer Res
2003
;
9
:
3329
–37.
42
Montgomery E, Abraham SC, Fisher C, et al. CD44 loss in gastric stromal tumors as a prognostic marker.
Am J Surg Pathol
2004
;
28
:
168
–77.
43
Nishitani A, Hirota S, Nishida T, et al. Differential expression of connexin 43 in gastrointestinal stromal tumours of gastric and small intestinal origin.
J Pathol
2005
;
206
:
377
–82.
44
Klose J, Nock C, Herrmann M, et al. Genetic analysis of the mouse brain proteome.
Nat Genet
2002
;
30
:
385
–93.
45
Kondo T, Hirohashi S. Application of highly sensitive fluorescent dyes (CyDye DIGE Fluor saturation dyes) to laser microdissection and two-dimensional difference gel electrophoresis (2D-DIGE) for cancer proteomics.
Nat Protoc
2006
;
1
:
2940
–56.