IPO-38 Is Identified as a Novel Serum Biomarker of Gastric Cancer

IPO-38 Is Identified as a Novel Serum Biomarker of Gastric Cancer Based on Clinical Proteomics Technology Yuan Hao,†,# Yingyan Yu,*,† Lishun Wang,‡,# Min Yan,† Jun Ji,† Ying Qu,† Jun Zhang,† Bingya Liu,† and Zhenggang Zhu*,† Department of Surgery of Ruijin Hospital, Shanghai Institute of Digestive Surgery, and Department of Pathophysiology of School of Medicine, Shanghai Jiaotong University, 197 Ruijin second Road, Shanghai, 200025, P.R. China Received October 4, 2007

Gastric cancer is one of the most common malignancies in China. So far, there are few reliable serum biomarkers for diagnosis. The available biomarkers of CEA, CA19–9 and CA72–4 are not sufficiently sensitive and specific for gastric cancer. In this study, a high density antibody microarray was used for identifying new biomarkers from serum samples of gastric cancer. Serum samples from colorectal cancer, pancreatic cancer, hepatocellular cancer, and breast cancer were also screened for comparative study. As result, some candidate biomarkers were identified. IPO-38, an up-regulated serum protein in gastric cancer was selected for subsequent validation including serum IPO-38 expression by ELISA and IPO-38 protein expression by immunohistochemistry. The immunoprecipitation by IPO-38 for gastric cancer cell line and MALDI-TOF/TOF mass spectrometer suggested that pull-down of IPO-38 belongs to H2B histone, which was supported by co-localization study of laser scanning confocal microscope. A follow-up study showed that the survival rate of IPO-38 negative group was better than that in IPO38 positive group. The study first clarified the property of IPO-38 proliferating marker, and proposed that IPO-38 protein is a promising biomarker both for diagnosis and for predicting prognosis of gastric cancer. Keywords: Gastric carcinoma • Diagnosis • Predicting Prognosis • Antibody Microarray • Clinical Proteomics • IPO-38 • H2B Histone • Biomarker

Introduction Gastric cancer is one of the most common malignancies in China. It not only accouns for a large percentage of cancerrelated deaths in China, but is also the second leading cause of cancer death worldwide.1,2 Patients with gastric cancer have a relatively poor prognosis, especially the patients of late clinical stage with a less than 35% of 5-year survival rate.3 Although the prognosis of gastric cancer has been improved somewhat, further increasing of survival rate will rely on an early diagnosis. As we know, the progression of the gastric cancer is usually silent until it reaches more advanced stage.4 Endoscope biopsy is still a golden standard for gastric cancer diagnosis. However, endoscope examination is an uncomfortable and invasive procedure. It is not a publicly acceptable screening method. Serum specimen screening is easy and inexpensive. It is considered as an ideal diagnostic specimen in general. Some serum tumor markers have been discovered. They include carcinoembryonic antigen (CEA), carbohydrate antigen 19–9 (CA19–9), and carbohydrate antigen 72–4 (CA72–4).5,6 Accord* To whom correspondence should be addressed. E-mail address: Dr. Yingyan Yu, [email protected]; Dr. Zhenggang Zhu, [email protected]. † Department of Surgery of Ruijin Hospital. ‡ Department of Pathophysiology of School of Medicine, Shanghai Jiaotong University. # These authors contributed equally to this work.

3668 Journal of Proteome Research 2008, 7, 3668–3677 Published on Web 08/05/2008

ing to our clinical experience and other reports, the early diagnostic role of CEA, CA19–9, and CA72–4 for gastric cancer remains controversial. The sensitivity of single biomarker in tumor diagnosis is low (usually less than 40%) with high “falsepositives”.7 Some authors proposed that theses biomarkers can be used as prognostic factors of gastric cancer.8 In addition, none of the existing serological tumor markers have yet been shown to be sufficiently sensitive and specific for screening gastric cancer.9 Therefore, identifying novel serological biomarkers with higher specificity and sensitivity is eagerly desired for gastric cancer. Development of proteomics technology provided a good chance for discovering new biomarkers of gastric cancer. Protein microarray is one of the proteomics technology. Among them, antibody microarray is more commonly used in clinical proteomics. Clinical proteomics refers to the total protein profile of a cell, a tissue, an organ, serum, or body fluid from patients. Clinical proteomics emphasizes the application of proteomics technologies at the bedside, to acquire serumbased proteomics pattern characteristic of the blood-proteome of a diseased state.10 This technology enables parallel detection of the multiple proteins in low sample volumes with a good reproducibility and high sensitivity.11 Therefore, it can be used to search and identify candidate biomarkers with highthroughput protein profiling.12 In this paper, we used a kind 10.1021/pr700638k CCC: $40.75

 2008 American Chemical Society

IPO-38 Is a Novel Serum Biomarker of Gastric Cancer

research articles

of high density antibody micrarray to analyze serum protein profiling of gastric cancer. Some other types of surgical solid cancer including colorectal cancer, hepatocellular cancer, pancreatic cancer as well as breast cancer were used as a comparative study. The microarray contains 722 antibodies in triplicate, including a few house-keeping proteins as internal control and multiple antibodies belonging to important biological processes such as stress response, angiogenesis, apoptosis, cell cycle progression, DNA injury repairing, signal transduction, and lymphocyte markers, which can detect target proteins representing a broad range of functional classes involving signal transduction, cell cycle regulation, gene transcription, apoptosis, oncogenesis, and immunological function of host. One up-regulated serum protein, IPO-38, a cell proliferating marker, was chosen for further validation by enzyme-linked immunosorbent assay (ELISA), Western-blot, immunoprecipitation, and immunohistochemistry as well as MALDI-TOF/TOF-MS analysis. At the same time, the serological levels of CEA, CA19–9, and CA72–4 were detected as parallel contrast to new identified biomarker.

with two times higher or lower than normal control were used for hierarchical clustering analysis. Serum IPO-38 Levels Validation by Enzyme-Linked Immunosorbent Assay (ELISA). Another 94 sera specimen from gastric cancer were collected before operation in early morning on an empty stomach from the Department of Surgery, Ruijin Hospital, School of Medicine, Shanghai Jiaotong University between January 2006 and November 2007. There are 67 males and 27 females with ages between 23 and 80 with a mean age of 58.78 years old. Of these, 59 cases underwent subtotal or total gastrectomy with D2 lymph node dissection. Nineteen cases underwent gastrectomy without further lymph node dissection because of advanced stage. Other 16 cases underwent abdominal exploration only because of metastasis. TNM staging was appraised based on combination of postoperation histological examination and intraoperative evaluation according to TNM classification of International Union Against Cancer. There are 20 early gastric cancer (EGC) cases and 74 advanced gastric cancer cases; 56 patients had been followed up since operation day. The follow-up period was 1∼21 months with the average of 11.34 months. Classical tumor markers of CEA, CA19–9, and CA72–4 were routinely measured at the clinical ward before operation. CEA and CA19-9 were detected by chemiluminescence immunoassay (CLIA, Abbott ARCHITECT i2000), CA72–4 by electrochemiluminescence immunoassay (ECLIA, Roche Ecl 2010). Forty-one serum samples from healthy donors were enrolled as controls. The healthy control cases enrolled in our study were verified by a series of examinations, such as thoracic X-ray and abdominal ultrasound. The age of healthy controls was between 28 and 79 with mean age of 47.63 years old. Fifteen out of 41 healthy control cases have examination results of CEA, CA19–9, and CA72–4. Human IPO-38 ELISA kit (Adlitteram Diagnostic Laboratories) was used to detect sera IPO-38 levels according to manufacturer’s instructions. Briefly, 100 µL of standards was dispensed into each of eight wells and 100 µL of specimens including 56 serum samples of gastric cancer patients and 15 serum samples of healthy donors was dispensed into plate wells. After 50 µL of enzyme conjugate reagent was dispensed into each well, the solutions were gently mixed for 15 s. Then, the plate was incubated at 37 °C for 60 min. After removal of the mixture from the incubator, the microtiter wells were rinsed with deionized water and emptied five times. Then, the wells were sharply striken onto absorbent paper to remove residual water droplets. Subsequently, 50 µL of color A and color B Reagent was added into each well and the solutions were incubated at 37 °C for 15 min. The reaction was stopped with the addition of 50 µL of stop solution into each wells and gently mixed for 30 s. It is important that all the blue color changes to yellow color completely in each well and that the optical density is read at 450 nm within 30 min in microtiter plate reader. IPO-38 Protein Expression Analysis on Precancerous and Cancer Tissues of Stomach. Gastric cancer tissues from 10 out of the 56 cases of gastric cancer above-mentioned were chosen for IPO expression validation (stages II-IV). In addition, 10 cases of intestinal-metaplasia as well as 10 cases of mucosa dysplasia were also selected for immunohistochemical validation. All tissues were fixed with 10% formaldehyde and paraffinembedded. The tissues were cut at 4 µm in thickness. The mouse monoclonal antibody IPO-38 Ab-1 was purchased from Laboratory Vision & NeoMarkers Corporation. A two-step immunohistochemical technique was used. Briefly, the sections

Materials and Methods Clinical Samples Collection and Antibody Microarray Detection. Peripheral blood samples of 5 mL were drawn from 11 malignant tumor patients in early morning on an empty stomach at Ruijin Hospital, School of Medicine, Shanghai Jiaotong University, between January 2006 and January 2007. The blood was centrifuged at 2000g for 10 min within 4 h after collection, and then stored at -30 °C until detection. The blood of two healthy volunteers was used as a normal control. In every case, the malignant diagnosis was confirmed by histopathology after operation. The patients enrolled in this study included 3 with advanced gastric carcinomas, 2 with advanced colorectal carcinomas, 2 with advanced pancreatic carcinomas, 2 with advanced liver carcinomas, and 2 with advanced breast carcinomas. The patients’ ages were from 21 to 70-year old. All samples were obtained with informed consent. This study was approved by Hospital Institutional Review Boards for human subject research. High density antibody microarray (Laboratory Vision Corporation, Cat.# TAA-001, CA) was used. The operating intructions were follwed according to the user manual. It is briefly summarized as follows. Centrifuge the sera at 10 000g for 10 min at 4 °C. Using BCA protein assay kit to measure protein concentration, add 0.0143 mg of Biotinylation Reagent to 0.1 mg of protein sample. The ratio of Biotinylation Reagent to protein sample should be maintained at 1:7 at room temperature (23–25 °C) for 2 h. Purify biotin labeled proteins using spin colums. Hybridize the biotin -labeled protein to the antibody microarray in a humidity chamber at room temperature for 2 h. After hybridization, detect the microarray by streptavidin solution for 45 min at room temperature. After incubation with detection antibody-Cy3 fluorescent dye for 45 min at room temperature, insert the microarray into the GenePix 4000B (Axon Instruments) for signal scanning. The ratio of fluorescent intensities for each spot is interpreted as a ratio of concentration for the corresponding protein in blood samples. The scanned image information was converted into numeric data and exported into Microsoft Excel. Hierarchical clustering and visualization were performed using the program Cluster and Treeview (http://genome-www5.stanford.edu/ MicroArray/SMD/restech.html). The numeric data of samples

Journal of Proteome Research • Vol. 7, No. 9, 2008 3669

research articles

Hao et al.

Figure 2. Serum levels of IPO-38 in gastric cancer and in normal control. Nonparametric statistics Wilcoxon (p < 0.001).

Figure 1. Image of hierarchical clustering analysis for serum antibody microarray. The top numbers represent the sample numbers. From left to right: gastric cancer, 1-3; breast cancer, 4, 5; liver cancer, 6, 7; colorectal cancer, 8, 9; pancreatic cancer, 10, 11.

were dewaxed and hydrated, and then boiled in 10 mM citrate buffer, pH 6.0 for 10 min. Blocking was performed with nonspecific binding with 5% (v/v) bovine serum albumin (BSA) for 10 min. The sections were incubated with IPO-38 Ab-1 (1: 100 working dilution) at 4 °C for 12 h in a moist chamber. After washing with 0.02 M phosphate buffer saline (PBS) pH 7.4 three times, the sections were incubated with secondary antibody conjugated with HRP-labeled polymer (DakoCytomation EnVision+) at 37 °C for 30 min. The section were then incubated with liquid DAB substrate-chromogen for 10 min at room temperature, rinsed in distilled water, and counterstained with hematoxylin. PBS was used instead of IPO-38 Ab-1 as negative control. Lymph node was used as a positive control tissue. IPO38 is expressed in nucleus of proliferating cells. To evaluate the results, obvious positive areas were selected, and the number of positive nuclei/100 cancer cells were calculated at 10 high power fields under microscope. The mean positive 3670

Journal of Proteome Research • Vol. 7, No. 9, 2008

Figure 3. ROC curve of IPO-38. The AUC is 0.811, which means that serum IPO-38 can be used as a potential diagnostic biomarker.

nuclei/100 cancer cells was determined as the labeling index of IPO-38. Two separated pathologists made a review of the sections. Immunoprecipitation and Western-blot by IPO-38 for Gastric Cancer Cell Lines. Gastric cancer cell lines of KATOIII, SNU-1, and SNU-16 (purchased from ATCC) were cultured in DMEM containing10% (v/v) heat-inactivated fetal bovine serum (FBS) and 2 mM L-glutamine. Cells were maintained at 37 °C in a humidified atmosphere of 95% air and 5% CO2. Cells (2 × 106 cells) were collected into each tube, and 200 µL of Mammalian Protein Extraction Reagent (Pierce Corporation) was added with shaking for 10 min. Cell debris was removed by centrifugation at 14 000g for 15 min, and the supernatant was transfered to a new tube for Western blot and immunoprecipitation analysis. The Seize X mammalian immunoprecipitation Kit was purchased from Pierce Corporation. The extracted sample was diluted 1:1 with PBS, 2 µL of IPO-38 antibody was added (200 µg/mL) to each tube, and the solution was incubated for 2 h at 4 °C. Them, 100 µL of immobilized protein G Plus was added

research articles

IPO-38 Is a Novel Serum Biomarker of Gastric Cancer Table 1. Sensitivities and Specificities of IPO-38 and Classical Tumor Markers

IPO-38 CEA CA72–4 CA19–9

cancer patients positive N ) 67

sensitivity %

normal controls positive N ) 15

specificity %

AUC

std. error

38 5 19 17

56.7 7.5 28.4 25.4

1 0 2 0

93.3 100 86.7 100

0.750 0.537 0.575 0.575

0.066 0.079 0.077 0.075

into each tube and incubation proceeded for 2 h at 4 °C. The tube was centrifuged at 3000g for 60 s and the flow-through was discarded. The pellets were washed with PBS three times and centrifuged. Then, 50 µL of PAGEprep elution buffer was added into each tube at 100 °C for 5 min. The protein from cells and immunoprecipitation was infused in the lanes of two SDS-PAGE gels with 100 µg protein/lane in electrophoresis buffer for 55 min at 100 V (Bio-Rad). One SDS-PAGE gel was used for Western-blot transfer. Another was stained by Coomassie G250 stain (Bio-SafeCoomassie, Bio-Rad). The PVDF membrane (Millipore) was blocked with 1% skimmed milk for 2 h and incubated with IPO-38 monoclonal antibody (working dilution 1:200) overnight at 4 °C, then incubated with horseradish peroxidase-conjugated bovine antimouse Ig G (working dilution 1:3000, Santa Cruz Bio) and stained by Coomassie G250 for 1 h. MALDI-TOF/TOF-MS for IPO-38 Immunoprecipitation Product. The 14 kDa protein band of the gel was diced into piecemeal (1 mm3) and placed into a siliconized tube. After washing with deionized water, 100 µL of 25 mmol/L ammonium bicarbonate (NH4HCO3) per 50% acetonitrile (ACN) was added (enough to cover the gel) and mixed for 15 min. This step was repeated twice until the gel pieces become shrinked. Then, the gel pieces were vacuum-dried for 5 min, making the gel completely dry and dust-like. The gel was then re-hydrated with 10 µL of trypsin solution (10 ng/µL trypsin in 20 mM NH4HCO3) and incubated at 4 °C for 30 min. Then, 25 mM NH4HCO3 was added (exclude trypsin) and incubated at 37 °C overnight. The tube was placed at -20 °C to stop the reaction. Then, 1 µL of 1% trifluoroacetic acid (TFA) was added to 10 µL of sample. C18 ZipTips (Millipore) were rinsed with 60% ACN/0.1% TFA and repeated once. The ZipTip was

Figure 4. ROC curves of IPO-38 and other classical tumor markers. IPO-38 shows more sensitivity and specificity for gastric cancer diagnosis than other classical tumor markers.

Table 2. Comparison of Clinicopathological Factors Between IPO-38 Positive Group and IPO-38 Negative Group IPO-38 (+)

IPO-38 (-)

N ) 31 n

Male Female

25 6

N ) 25 X2

P value

61.0% 40.0%

Gender 16 39.0% 9 60.0%

1.955

Not significant

1.005

Not significant

%

n

%

e55 >55

14 17

36.4% 50.0%

Age 8 63.6% 17 50.0%

Upper Middle Lower

2 21 8

25.0% 61.8% 57.1%

Location 6 75.0% 13 38.2% 6 42.9%

3.438

Not significant

e5 cm >5 cm

16 15

56.8% 78.9%

Tumor Size 12 43.2% 4 21.1%

6.816

0.013

6 25

Lauren Type 31.6% 13 68.4% 67.6% 12 32.4%

6.579

0.013

T1∼2 T3∼4

9 22

Depth of Invasion 56.3% 7 43.8% 55.0% 18 45.0% 0.007

Not significant

Negative Positive

11 20

Lymph Node Metastasis 55.0% 9 45.0% 55.6% 16 44.4% 0.002

Not significant

Negative Positive

23 8

Distant Metastasis 50.0% 23 50.0% 80.0% 2 20.0% 2.991

Not significant

I II III IV

8 4 5 14

Stage Classification 53.3% 7 46.7% 66.7% 2 33.3% 25.7% 9 64.3% 66.7% 7 33.3% 3.549

Not significant

Yes No

17 14

Curative Resection 48.6% 18 51.4% 66.7% 7 28% 1.739

Not significant

Intestinal Diffuse

equilibrated with 0.1% TFA, and the peptides were loaded onto the Zip tip via repeat pipetting up and down 10 times (care was taken not to let the tip run dry or introduce air bubbles into the tip). The tip was rinsed with 0.1% TFA and the desalted peptide eluted with 4 µL of 50%ACN/0.1%TFA until the sample was concentrated to 1 µL. The sample was dissolved in matrix solution containing 10 mg/mL R-cyano-4-hydroxycinnamic Acid (CHCA) in 50% ACN and 0.1% TFA and spotted onto the MALDI sample target plate. The peptide mass spectra were obtained with a MALDI-TOF/TOF mass spectrometer (4700 Proteomics Analyzer, Applied Biosystems). The peptide mass fingerprinting (PMF) was obtained in the mass range between 800 and 4000 Da with approximately 5000 laser shots. To obtain the spectra with the mass accuracy of b25 ppm, trypsin Journal of Proteome Research • Vol. 7, No. 9, 2008 3671

research articles

Hao et al.

Figure 5. Kaplan–Meier survival curves and Log-Rank statistics between IPO-38 negative group and IPO-38 positive group in gastric cancer. Serum IPO-38 negative patients show better survival than IPO-38 positive patients (P ) 0.0467).

autolysis peaks were used for the internal calibration. Up to 5 of the most intense peaks, excluding those from the matrix, background, trypsin autolysis, acrylamide, or keratin peaks, were selected for subsequent MS/MS data acquisition. Database search: Protein identification was processed and analyzed by searching the Swiss-Prot protein database using MASCOT search engine of Matrix Science integrated in Global Protein Server Workstation. The mass tolerance was limited to 50 ppm. The results were accepted as a good identification when the Global Protein Server score confidence was higher than 95%. Validation of Correlation between IPO-38 and H2B Proteins by Western Blot. Protein samples extracted from gastric cancer cell line KATO-III and SNU-1 were separated by Western blot. One PVDF membrane was incubated with IPO-38 antibody and another with H2B polyclonal antibody (working dilation 1:10 000, Abcam). Then, membranes were incubated with bovine anti-mouse IgG-HRP and bovine anti-rabbit IgGHRP (Santa Cruz Biotechnology, at 1:3000 dilution), respectively, for 2 h and stained by Coomassie G250 for 1 h. Co-localization Analysis of IPO-38/H2B by Laser Scanning Confocal Microscope (LSCM). Gastric cancer cell line KATO-III was cultured in the same way as mentioned above. Cells (1 × 105 cultured in DMEM containing 10% FBS) were put onto APES-coated slides and fixed in formaldehyde for 30 min, then treated with 0.25% Triton X-100 for 15 min. The slides were

blocked by 5% BSA for 30 min and incubated with mouse monoclonal antibody IPO-38 (working dilution 1:100) and rabbit anti-human histone H2B polyclonal antibody (working dilution 1:100). respectively, for 2 h at 37 °C. Following PBS washes, slides were incubated with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG (working dilution 1:50, Dako Cytomation) and tetramethylrhodamine isothiocyanate(TRITC)-conjugated goat anti-rabbit IgG (1:50, Dako Cytomation) in the dark at room temperature for 45 min and stained with 4′-6-diamidino-2-phenylindole (DAPI) in the dark for 5 min. The slides were mounted using fluorescent mounting medium (Dako Cytomation) and observed by LSCM (LSM 5 PASCAL, Carl Zeiss, Germany). Statistical Analysis. Serum ELISA detection data of IPO-38 were analyzed by Wilcoxon 2-sample nonparametric statistics. Receiver Operating Characteristics (ROC) graph was performed. Pearson Chi-square test and Fisher exact test were used to assess the relationship between IPO-38 and clinical-pathological parameters. Kaplan–Meier and two-sided log-rank test were used to calculate and plot overall survival and survival curve differences. The statistical program SPSS version 11.0 was used for analysis.

Result Serum Protein Expression Finding by Antibody Microarray. An extensive data set obtained from antibody microarray was analyzed by hierarchical clustering algorithm. The protein expression intensity of samples with values 2 times higher or lower than normal control was used for hierarchical clustering plotting (Figure 1). The right side of hierarchical clustering plotting showed the names of serum proteins. The red or dark red represents the increase compared to normal control. The green color represents the decrease compared to normal control. The top numbers represent the sample numbers. It is clearly shown that gastric cancer has a different serum protein expression profile from colorectal cancer as well as pancreatic cancer. But the serum protein expression profile of gastric cancer showed some similarity with breast cancer and liver cancer. Serum IPO-38 Level Validation by ELISA. On the basis of serum antibody microarray analysis, IPO-38, a kind of cell proliferation marker, was identified as an elevated biomarker. We selected it as a further validation molecule on another panel of gastric cancer. The IPO-38 mean level of 94 cases of gastric cancer was 155.097 ( 45.60 pg/mL. The IPO-38 mean level of normal control cases was 113.154 ( 26.46 pg/mL. IPO-38 level in gastric cancer group was significantly higher than that in normal control group (p ) 0.001) (Figure 2). The ROC curve and the area under curve (AUC) of IPO-38 were calculated. As

Table 3. IPO-38 Labeling Index Compared to Their Serum IPO-38 Levels in Gastric Cancer

3672

gender

age (years)

Male Male Female Female Male Female Male Female Male Female

67 54 62 46 40 51 64 66 64 58

stage (pTNM)

Lauren type

labeling index

serum IPO-38 level (pg/mL)

T3N0M0 T3N1M0 T3N1M0 T3N1M0 T3N1M1 T3N2M0 T3N1M0 T3N3M0 T2N1M0 T3N0M0

Intestinal Intestinal Diffuse Diffuse Diffuse Diffuse Diffuse Diffuse Intestinal Diffuse

31.2% 34.2% 56.7% 45.2% 77.6% 61.4% 36.2% 55.8% 27.8% 83.4%

162.301 228.383 244.683 105.85 90.694 117.634 120.695 139.667 154.628 249.485

Journal of Proteome Research • Vol. 7, No. 9, 2008

IPO-38 Is a Novel Serum Biomarker of Gastric Cancer

research articles

Figure 6. IPO-38 protein detection on gastric cancer by immunohistochemistry. Left: normal gastric mucosa shows positive only on some neck proliferating cells of gastric glands. Right: gastric cancer cells show positive staining of nuclei in scattered dots style.

Figure 7. IPO-38 protein detection on intestinal-metaplasia and mucosa dysplasia by immunohistochemistry. Left: intestinal-metaplasia showed negative staining for IPO-38 protein. Right: some of mucosa dysplasia cases showed low labeling index for IPO-38 protein by immunohistochemistry.

Figure 8. Left, Western blot of IPO-38; right, electrophoresis gel for immunoprecipitation of IPO-38. IPO-38 antigen appeared at 14 kDa position in Western blot and immunoprecipitation gel. The gel band indicated by arrow was cut off for subsequent MALDI-TOF/TOFMS identification.

result, the AUC was 0.811 ( 0.040, and significantly higher than that of null hypothesis (true area was 0.5, p < 0.001, see Figure 3). This means that the serum IPO-38 level could serve as a potential diagnostic marker for gastric cancer. To find an optimal cutoff point to dichotomize ‘diseased’ or ‘healthy’, Youden’s index was used in this study. According to Youden’s index, the optimal operating point of IPO-38 was 140.399 pg/ mL. At this cutoff point, the sensitivity was 57.4% and specificity was 90.2%. Comparison of IPO-38 Level with Classical Tumor Markers. The accuracy of IPO-38 was compared with that of classical serum tumor markers from 56 cases of gastric cancer and 15 healthy control cases with data of both IPO-38 and classical serum tumor markers. The cutoff value of IPO-38 was 140.399 pg/mL, whereas, the cutoff value of CEA, CA72–4, and CA19–9, was 10.00 ng/mL, 5.3 µg/mL, and 37 µg/mL, respectively. At these cutoff points, the sensitivities and specificities were summarized in Table 1. The ROC graphs of IPO-38 and

classical tumor markers were shown in Figure 4. The ROC curve of IPO-38 showed better than that of any classical tumor markers. Correlation of Serum IPO-38 Levels with Other Clinicopathological Parameters. Fifty-six of gastric cancers were divided into two groups according to the cutoff point of IPO38 serum levels. Of these, 31 out of 56 cases (55.4%) showed IPO-38 positive preoperation. Several clinicopathological parameters were compared between IPO-38 positive group and IPO-38 negative group (see Table 2). Serum IPO-38 level was significantly correlated with tumor size and Lauren’s histological classification. Correlation of Serum IPO-38 Levels with Survival Rate. Follow-up was performed for whole group until the end of August 2007. The result showed that IPO-38 negative patients have better survival than IPO-38 positive patients. The mean survival time of IPO-38 negative group was 18.40 ( 1.18 months. The mean survival time of IPO-38 positive group was 15.66 ( 1.41 months. Kaplan–Meier survival curves and Log-Rank statistics showed that Journal of Proteome Research • Vol. 7, No. 9, 2008 3673

research articles

Hao et al.

Figure 9. Maps of MALDI-TOF/TOF-MS for IPO-38 immunoprecipitation product. The lower map details the m/z at 901.5988. The peptide sequence was identified as LAHYNKR.

Figure 10. Western blot detection of protein extraction by IPO38 and histone H2B antibodies. Both of them showed that their molecular weights are 14 kDa.

the serum IPO-38 level can use as a prognostic biomarker for gastric cancer (Figure 5 χ2 ) 3.95, P ) 0.0467). IPO-38 Expression on Precancerous and Cancer Tissues of Stomach. The IPO-38 protein expression was examined by immunohistochemistry on 10 cases of cancer tissues with corresponding serum IPO-38 data. All 10 cases showed IPO-38 antigen expression in nuclei of proliferating cancer cells. The relation of IPO-38 labeling index on gastric cancer tissues and their serum IPO-38 levels of 10 samples was compared and summarized in Table 3. The labeling index of IPO-38 in gastric cancer tissues was between 30% and 80%, with the expressing style of scattered dots (Figure 6). Five cases showed the labeling index over 50%. The mean serum IPO-38 level was 168.43 ( 73.88 pg/mL. The mean serum IPO-38 level in the remaining five cases was 154.37 ( 47.51 pg/mL. There is no significant difference of their serum IPO-38 levels between cancer tissues with higher IPO-38 labeling index group and with lower labeling index group (P > 0.05). In addition, IPO-38 expressing status both on intestinal-metaplasia and mucosa dysplasia was also examined for another separated 10 cases, respectively. As a result, all of intestinal-metaplasia disclosed negative for IPO38 expression. Three out of 10 cases of mucosa dysplasia disclosed weak positive for IPO-38, with labeling index less than 30% (Figure 7). MALDI-TOF/TOF-MS Detection for IPO-38 Immunoprecipitation Product. The Western blot and immunoprecipitation for lysate of gastric cancer cell line KATO-III were performed at the same time by IPO-38 monoclonal antibody. 3674

Journal of Proteome Research • Vol. 7, No. 9, 2008

IPO-38 antigen appeared at 14 kDa location (Figure 8). The immunoprecipitation product from electrophoresis gel was taken for further MALDI-TOF/TOF-MS identification. Identification results showed that the immunoprecipitation product of IPO-38 was histone H2B based on search in Swiss-Prot protein database. The matched peptides occupy 47% sequence coverage with 106 score. Maps of MALDI-TOF/TOF-MS are shown in Figure 9. Verification of Correlation between IPO-38 and H2B Proteins by Western Blot. To validate the relationship between IPO-38 antigens and histone H2B, we detected protein bands of gastric cancer cell line KATO-III and SNU-16 by Western blot. As Figure 10 shows, IPO-38 and H2B were detected at the same location of 14 kDa. Co-localization Analysis of IPO-38/H2B by LSCM. Colocalization detection of IPO-38 and H2B was performed on KATO-III cell line by high-resolution observation of LSCM. The results are shown in Figure 10. IPO-38 antigen (green fluorescence) was expressed at nuclei, and was also stained by H2B antibody (red fluorescence). So, the double-stained area disclosed yellow (Figure 11).

Discussion Gastric cancer is one of the common gastrointestinal cancers. Sometimes, it is troublesome to distinguish it from colorectal cancer as well as pancreatic cancer, especially in early stage. Although some serum tumor markers such as CEA, CA19–9, and CA72–4 have been utilized for differential diagnosis, their sensitivity and specificity are not very high. Therefore, their diagnostic role remains controversial.5–9 Identifying and searching novel serological biomarker or marker group with higher specificity and sensitivity is eagerly desired. A proteomics technique is considered a very valuable method to screen biomarkers of tumor.13,14 Serum is the best protein sample representative of the whole body. Thus, serum proteomics is an area attracting intense interest as it is easily accessible and noninvasive. The serum level of some proteins can be excellent biomarkers for disease. Moreover, serum proteomics enable

IPO-38 Is a Novel Serum Biomarker of Gastric Cancer

research articles

Figure 11. Immunofluorescence detection of IPO-38 and H2B by LSCM. (A) Cell structure of KATO-III gastric cancer cell line; (B) H2B immunofluorescence staining; (C) IPO-38 immunofluorescence staining; (D) merged B and C.

parallel detection of multiple proteins in low sample volumes with a good reproducibility and high sensitivity.11 It is used to search and identify candidate biomarkers as high-throughput protein profiling.12 In this paper, we used a kind of high density antibody microarray to analyze serum protein expressing profiling of gastric cancer. As a parallel comparison, some types of other common solid cancer including colorectal cancer, hepatocellular cancer, and pancreatic cancer as well as breast cancer were included in the study. As a result, several promising biomarkers were identified as valuable biomarkers for diagnosis of gastric cancer. IPO-38 is one of the up-regulated serum proteins, which was selected for subsequent validation in this study. IPO series monoclonal antibodies were first discovered and developed by the Institute of Problems of Oncology at Ukraine.15 RPMI-1788 cell line, Daudi, and splenocytes of a patient with hairy-cell leukemia were used for immunization of BALB/c mice and IPO series antibodies were obtained.15–17 Further research indicated that IPO-38, also known as nonlineage 116 antigen (N-L116), is a nuclear antigen of proliferating cells.18–20 Hitherto, there is no systematic study on the relationship between IPO-38 proliferating marker and gastric cancer. IPO-38, as a novel serological tumor marker for gastric cancer, is first proposed in this study based on abundant experimental evidence. For instance, the mean serum level of IPO-38 in gastric cancer patients was significant higher than that in healthy control. The AUC of IPO-38 reached to 0.811. As we know, ROC curve is a useful method for evaluating clinical usefulness of a biomarker.21 The AUC of 0.5 means there is a 50–50 chance as a biomarker, which will correctly identify diseased or healthy persons. Therefore, it means that

IPO-38 can be used as a potential serum biomarker. Further, Youden index was used for measure of the cutoff point in our study, which is based on the accuracy test in clinical epidemiology.22 We found that the sensitivity and specificity of IPO38 were over 50% and 90%, respectively, which is obviously better than those of conventional tumor markers of CEA, CA19–9, and CA72–4. IPO-38 showed more accuracy for diagnostic value. The ROC curve is a useful tool for comparing the effectiveness between different biomarkers, and higher AUC represents higher diagnostic ability.23 The AUC of IPO-38 was greater than that of any other classic tumor marker. In addition, IPO-38 showed not only diagnostic value, but also predictive value for prognosis, because the negative patients for serum IPO-38 showed better survival rate than that in IPO-38 positive patients (follow-up 1–21 months, p ) 0.0467). Comparing the serum IPO-38 levels with clinicopathological factors, Intestinaltype carcinoma showed lower serum IPO-38 levels than that in diffuse-type carcinoma. It suggested that proliferating ability of intestinal-type cancer is lower than that of diffuse-type cancer.24,25 We compared the IPO-38 protein expression on cancer tissues with their serum levels. IPO-38 antigens were clearly observed in nuclei of proliferation cells. The labeling index of IPO-38 in diffuse-type gastric cancer was higher than that in intestinal-type. This phenomenon was associated with IPO-38 serum levels observed by ELISA detection. However, we do not have significant statistics because of the limited number of samples. So far, there is little information about IPO-38 antigen, except for the proliferating marker. To clarify the property of IPO-38 in nature, we performed immunoprecipitation of protein extraction from gastric cancer cells KATO-III by IPO-38 monoJournal of Proteome Research • Vol. 7, No. 9, 2008 3675

research articles clonal antibody. The immunoprecipitation product was identified by MALDI-TOF/TOF mass spectrometer. According to the search on Swiss-Prot protein database, H2B, a nuclear histone, was finally verified by our serial study, which had never been reported hitherto. This finding was confirmed by subsequent Western blot detection as well as immunofluorescence colocalization analysis of LSCM. LSCM is theoretically better than conventional microscope when using different fluorescent probes at the same time (co-localization). H2B histone was a component of nucleosomes, which include H2A, H2B, H3, and H4. Genomic DNA is wrapped around an octave of above four core histones.26 As components of nucleosomes, histones play an important role in DNA replication.27 Recently, several kinds of histone modifications during cell cycle were identified, which exhibit different dynamic changes.28,29 Upon this study, IPO-38 was proposed as a novel serum biomarker for diagnosis and predicting prognosis of gastric cancer. However, many questions remain. For instance, how does the IPO-38 protein flow into peripheral blood? It may be due to the translocation of histone by a post-translational modification mechanism, since well-known histone modifications, such as, acetylation, methylation, phosphorylation, ubiquitylation, and sumoylation, may cause downstream biological events.30–38 Some studies indicated that immunohistochemically demonstrable phosphorylated histone H3 is an excellent mitotic marker.39–41 In our immunohistochemical study, IPO38 antigens mainly appeared in nuclei of gastric cancer cell as well in small part of mucosa dysplasia cells, but not in normal gastric mucosa except for the neck proliferating cells of gastric glands, which reflected the proliferating property. It is compatible with previous report that IPO-38 antigens appeared in proliferating cells.17 We speculated that, during the DNA replication of cell cycle, the post-translational modification of H2B histone may alter the structure of the protein that is exposed as a kind of new antigen, even though it translocates to cytoplasm and blood. This will remain for further study. In summary, a novel serological biomarker, IPO-38 was found for diagnosis and predicting prognosis of gastric cancer based on antibody array screening. IPO-38 showed a higher specificity and sensitivity than conventional tumor markers. The validation study by proteomics, molecular biology, and morphology suggested that IPO-38 belongs to H2B histone, a component of nucleosomes, which may represent a kind of post-translational type of H2B histone.

Acknowledgment. We thank all members of Department of Surgery of Shanghai Ruijin Hospital for supporting of samples collection. This work was supported, in part, by grants from the Chinese National High Tech Program (863-2006AA02A402, 2006AA02Z105 and 863-2006AA02A301), the National Natural Science Foundation of China (30572127, 30770961), the Shanghai Commission of Science and Technology (05JC14013), and by Shanghai Municipal Health Bureau (05-III-005-020). Chinese Patent Application No. 200810033017.1. References (1) Landis, S. H.; Murray, T.; Bolden, S.; Wingo, P. A. Cancer statistics 1998. CA Cancer J. Clin. 1998, 48, 6–29. (2) Parkin, D. M.; Pisani, P.; Ferlay, J. Estimates of the worldwide incidence of 25 major cancers in 1990. Int. J. Cancer 1999, 80, 827– 841. (3) Yamazaki, H.; Oshima, A.; Murakami, R.; Endoh, S.; Ubukata, T. A long-term follow-up study of patients with gastric cancer detected by mass screening. Cancer 1989, 63, 613–617.

3676

Journal of Proteome Research • Vol. 7, No. 9, 2008

Hao et al. (4) Kodera, Y.; Ito, S.; Yamamura, Y.; Mochizuki, Y.; Fujiwara, M.; Hibi, K.; Ito, K.; Akiyama, S.; Nakao, A. Follow-up surveillance for recurrence after curative gastric cancer surgery lacks survival benefit. Ann. Surg. Oncol. 2003, 10, 898–902. (5) Kodama, I.; Koufuji, K.; Kawabata, S.; Tetsu, S.; Tsuji, Y.; Takeda, J.; Kakegawa, T. The clinical efficacy of CA 72–4 as serum marker for gastric cancer in comparison with CA19–9 and CEA. Int. Surg. 1995, 80 (1), 45–48. (6) Ohuchi, N.; Takahashi, K.; Matoba, N.; Sato, T.; Taira, Y.; Sakai, N.; Masuda, M.; Mori, S. Comparison of serum assays for TAG72, CA19–9 and CEA in gastrointestinal carcinoma patients. Jpn. J. Clin. Oncol. 1989, 19 (3), 242–248. (7) Ychou, M.; Duffour, J.; Kramar, A.; Gourgou, S.; Grenier, J. Clinical significance and prognostic value of CA72–4 compared with CEA and CA19–9 in patients with gastric cancer. Dis. Markers 2000, 16 (3–4), 105–110. (8) Lai, I. R.; Lee, W. J.; Huang, M. T.; Lin, H. H. Comparison of serum CA72–4, CEA, TPA, CA19–9 and CA125 levels in gastric cancer patients and correlation with recurrence. Hepatogastroenterology 2002, 49 (46), 1157–1160. (9) Carpelan-Holmstrom, M.; Louhimo, J.; Stenman, U. H.; Alfthan, H.; Haglund, C. CEA, CA 19–9 and CA 72–4 improve the diagnostic accuracy in gastrointestinal cancers. Anticancer Res. 2002, 22, 2311–2316. (10) Somiari, R. I.; Somiari, S.; Russell, S.; Shriver, C. D. Proteomics of breast carcinoma. J. Chromatogr., B 2005, 815 (1–2), 215–225. (11) Haab, B. B.; Dunham, M. J.; Brown, P. O. Protein microarrays for highly parallel detection and quantitation of specific proteins and antibodies in complex solutions. Genome Biol. 2001, 2 (2), RESEARCH0004.1–0004.13. (12) Miller, J. C.; Zhou, H.; Kwekel, J.; Cavallo, R.; Burke, J.; Butler, E. B.; Teh, B. S.; Haab, B. B. Antibody microarray profiling of human prostate cancer sera: antibody screening and identification of potential biomarkers. Proteomics 2003, 3 (1), 56–63. (13) Kawada, N. Cancer serum proteomics in gastroenterology. Gastroenterology 2006, 130 (6), 1917–1919. (14) Ren, H.; Du, N.; Liu, G.; Hu, H. T.; Tian, W.; Deng, Z. P.; Shi, J. S. Analysis of variabilities of serum proteomic spectra in patients with gastric cancer before and after operation. World J. Gastroenterol. 2006, 12 (17), 2789–2792. (15) Pinchuk, V. G.; Sidorenko, S. P.; Vetrova, E. P.; Berdova, A. G.; Shlapatskaia, L. N. Monoclonal antibodies to RPMI-1788 lymphoblastoid line cells. Eksp. Onkol. 1986, 8 (6), 41-46; [article in russian]. (16) Sidorenko, S. P.; Vetrova, E. P.; Iurchenko, O. V.; Shlapatskaia, L. N.; Berdova, A. G.; Elenskaia, A. M.; Bal’shin, M. D.; Gluzman, D. F. Monoclonal antibodies of the IPO series in studying and diagnosing malignant lymphoproliferative diseases. Gematol. Transfuziol. 1990, 35 (4), 19-22; [article in russian]. (17) Sidorenko, S. P.; Vetrova, E. P.; Yurchenko, O. V.; Berdova, A. G.; Shlapatskaya, L. N.; Gluzman, D. F. Monoclonal antibodies of IPO series against B cell differentiation antigens in leukemia and lymphoma immunophenotyping. Neoplasma 1992, 39 (1), 3–9. (18) Mikhalap, S. V.; Lopez, F.; Shlapatskaya, L. N. Monoclonal antibody IPO-38 recognizes a novel nuclear antigen of proliferating cells. In Leucocyte Typing VI; Kishimoto, T., Eds.; Garland Publishing, Inc.: New York, NY, 1997; pp 609–610. (19) Mikhalap, S. V.; Shlapatskaya, L. N.; Berdova, A. G.; Yurchenko, O. V.; Lopes, F.; Lukyanova, N. Y.; Sidorenko, S. P. Monoclonal antibody IPO-38 in evaluation of proliferative activity of tumor cells. Exp. Oncol. 2000, 22 (Suppl. 2), 36–38. (20) Thosaporn, W.; Iamaroon, A.; Pongsiriwet, S.; Ng, K. H. A comparative study of epithelial cell proliferation between the odontogenic keratocyst, orthokeratinized odontogenic cyst, dentigerous cyst, and ameloblastoma. Oral. Dis. 2004, 10 (1), 22–26. (21) Perkins, N. J.; Schisterman, E. F. The inconsistency of “optimal” cutpoints obtained using two criteria based on the receiver operating characteristic curve. Am. J. Epidemiol. 2006, 163 (7), 670– 675. (22) Schisterman, E. F.; Perkins, N. J.; Liu, A; Bondell, H. Optimal cutpoint and its corresponding Youden index to discriminate individuals using pooled blood samples. Epidemiology 2005, 16, 73– 81. (23) Schisterman, E. F.; Faraggi, D.; Reiser, B. Adjusting the generalized ROC curve for covariates. Stat. Med. 2004, 23 (21), 3319–3331. (24) Broll, R.; Mahlke, C.; Best, R.; Schimmelpenning, H; Strik, M. W.; Schiedeck, T; Bruch, H. P.; Duchrow, M. Assessment of the proliferation index in gastric carcinomas with the monoclonal antibody MIB 1. J. Cancer Res. Clin. Oncol. 1998, 124 (1), 49–54. (25) Davessar, K.; Pezzullo, J. C.; Kessimian, N.; Hale, J. H.; Jauregui, H. O. Gastric adenocarcinoma: prognostic significance of several

IPO-38 Is a Novel Serum Biomarker of Gastric Cancer

research articles

pathologic parameters and histologic classifications. Hum. Pathol. 1990, 21 (3), 325–332. Luger, K.; Mader, A. W.; Richmond, R. K.; Aergent, D. F.; Richmond, T. J. Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature 1997, 389, 251–260. Jackson, V.; Chalkley, R. Histone segregation on replicating chromatin. Biochemistry 1985, 24 (24), 6930–6938. Bonenfant, D.; Towbin, H.; Coulot, M.; Schindler, P.; Mueller, D. R.; van Oostrum, J. Analysis of dynamic changes in post-translational modifications of human histones during cell cycle by massspectrometry. Mol. Cell. Proteomics 2007, 6, 1917–1932. Beck, H. C.; Nielsen, E. C.; Matthiesen, R.; Jensen, L. H.; Sehested, M.; Finn, P.; Grauslund, M.; Hansen, A. M.; Jensen, O. N. Quantitative proteomic analysis of post-translational modifications of human histones. Mol. Cell. Proteomics 2006, 5 (7), 1314–1325. Kuo, M. H.; Allis, C. D. Roles of histone acetyltransferases and deacetylases in gene regulation. BioEssays 1998, 20 (8), 615–626. Cheung, P.; Allis, C. D.; Sassone-Corsi, P. Signaling to chromatin through histone modifications. Cell 2000, 103 (2), 263–271. Kouzarides, T. Histone methylation in transcriptional control. Curr. Opin. Genet. Dev. 2002, 12 (2), 198–209. Robzyk, K.; Recht, J.; Osley, M. A. Rad6-dependent ubiquitination of histone H2B in yeast. Science 2000, 287 (5452), 501–504. Nathan, D.; Ingvarsdottir, K.; Sterner, D. E.; Bylebyl, G. R.; Dokmanovic, M.; Dorsey, J. A.; Whelan, K. A.; Krsmanovic, M.; Lane, W. S.; Meluh, P. B.; Johnson, E. S.; Berger, S. L. Histone sumoylation is a negative regulator in Saccharomyces cerevisiae

and shows dynamic interplay with positive-acting histone modifications. Genes Dev. 2006, 20 (8), 966–976. Nathan, D.; Sterner, D. E.; Berger, S. L. Histone modifications: Now summoning sumoylation. Proc. Natl. Acad. Sci. U.S.A. 2003, 100 (23), 13118–13120. Strahl, B. D.; Allis, C. D. The language of covalent histone modifications. Nature 2000, 403, 41–45. Cheung, P.; Allis, C. D.; Sassone-Corsi, P. Signaling to chromatin through histone modifications. Cell 2000, 103, 263–271. Jenuwein, T.; Allis, C. D. Translating the histone code. Science 2001, 293, 1074–1080. Altheim, B. A.; Schultz, M. C. Histone modification governs the cell cycle regulation of a replication-independent chromatin assembly pathway in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. U.S.A. 1999, 96 (4)), 1345–1350. Brenner, R. M.; Slayden, O. D.; Rodgers, W. H.; Critchley, H. O.; Carroll, R.; Nie, X. J.; Mah, K. Immunocytochemical assessment of mitotic activity with an antibody to phosphorylated histone H3 in the macaque and human endometrium. Hum. Reprod. 2003, 18, 1185–1193. Davidson, E. J.; Morris, L. S.; Scott, I. S.; Rushbrook, S. M.; Bird, K.; Laskey, R. A.; Wilson, G. E.; Kitchener, H. C.; Coleman, N.; Stern, P. L. Minichromosome maintenance (Mcm) proteins, cyclin B1 and D1, phosphohistone H3 and in situ DNA replication for functional analysis of vulval intraepithelial neoplasia. Br. J. Cancer 2003, 88, 257–262.

(26)

(27) (28)

(29)

(30) (31) (32) (33) (34)

(35) (36) (37) (38) (39)

(40)

(41)

PR700638K

Journal of Proteome Research • Vol. 7, No. 9, 2008 3677