Identification of Human Complement Factor B as a Novel Biomarker

Jul 24, 2014 - Pancreatic cancer (PC; pancreatic ductal adenocarcinoma) is characterized by significant morbidity and mortality worldwide. Although ...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/jpr

Identification of Human Complement Factor B as a Novel Biomarker Candidate for Pancreatic Ductal Adenocarcinoma Min Jung Lee,†,‡ Keun Na,† Seul-Ki Jeong,† Jong-Sun Lim,† Sun A. Kim,§ Min-Ji Lee,∥ Si Young Song,§ Hoguen Kim,∥ William S. Hancock,⊥ and Young-Ki Paik*,†,‡ †

Yonsei Proteome Research Center and ‡Department of Integrated OMICS for Biomedical Science and Department of Biochemistry, College of Life Science and Biotechnology, Yonsei University, 50 Yonsei-ro, Sudaemoon-ku, Seoul 120-749, Korea § Department of Internal Medicine and ∥Department of Pathology, Yonsei University College of Medicine, 50-1 Yonsei Ro, Seodaemun-gu, Seoul 120-752, Korea ⊥ Barnett Institute and Department of Chemistry and Chemical Biology, Northeastern University, 360 Huntington Avenue, Boston, Massachusetts 02115, United States S Supporting Information *

ABSTRACT: Pancreatic cancer (PC; pancreatic ductal adenocarcinoma) is characterized by significant morbidity and mortality worldwide. Although carbohydrate antigen (CA) 19-9 has been known as a PC biomarker, it is not commonly used for general screening because of its low sensitivity and specificity. Therefore, there is an urgent need to develop a new biomarker for PC diagnosis in the earlier stage of cancer. To search for a novel serologic PC biomarker, we carried out an integrated proteomic analysis for a total of 185 pooled or individual plasma from healthy donors and patients with five disease groups including chronic pancreatitis (CP), PC, and other cancers (e.g., hepatocellular carcinoma, cholangiocarcinoma, and gastric cancer) and identified complement factor b (CFB) as a candidate serologic biomarker for PC diagnosis. Immunoblot analysis of CFB revealed more than two times higher expression in plasma samples from PC patients compared with plasma from individuals without PC. Immunoprecipitation coupled to mass spectrometry analysis confirmed both molecular identity and higher expression of CFB in PC samples. CFB showed distinctly higher specificity than CA 19-9 for PC against other types of digestive cancers and in discriminating PC patients from non-PC patients (p < 0.0001). In receiver operator characteristic curve analysis, CFB showed an area under curve of 0.958 (95% CI: 0.956 to 0.959) compared with 0.833 (95% CI: 0.829 to 0.837) for CA 19-9. Furthermore, the Y-index of CFB was much higher than that of CA 19-9 (71.0 vs 50.4), suggesting that CFB outperforms CA 19-9 in discriminating PC from CP and other gastrointestinal cancers. This was further supported by immunoprecipitation and qRT-PCR assays showing higher expression of CFB in PC cell lines than in normal cell lines. A combination of CFB and CA 19-9 showed markedly improved sensitivity (90.1 vs 73.1%) over that of CFB alone in the diagnosis of PC against non-PC, with similar specificity (97.2 vs 97.9%). Thus, our results identify CFB as a novel serologic PC biomarker candidate and warrant further investigation into a large-scale validation and its role in molecular mechanism of pancreatic carcinogenesis. KEYWORDS: complement factor b, carbohydrate antigen 19-9, biomarker, pancreatic cancers

■

been replaced with CA 19-9 due to its low sensitivity (54%)6 and presence in other tumor types7 during the past two decades.8 CA 19-9, a sialylated Lewis (a) antigen, was discovered in 1981 and has been known to function as adhesion in pancreatic cell and colon cell epithelium.9 Although CA 19-9 is the only currently available FDA-approved PC biomarker,10 it is also expressed in benign diseases and many types of gastrointestinal cancer11 and cannot distinguish PC from other gastrointestinal cancers or ovarian cancer.11,12 In addition to CEA and CA 19-9, other markers (e.g., CA50,

INTRODUCTION Pancreatic cancer (PC; pancreatic ductal adenocarcinoma) caused 330 000 deaths globally in 20121 and is predicted to be one of the top three cancer killers along with lung and liver cancers in 2030.2 According to the ‘Cancer Statistics 2013 Report’, PC not only progresses rapidly but also readily metastasizes to lymph nodes; therefore, early symptoms are rarely recognized and the 5-year survival rate of patients diagnosed with PC is only 1 to 3%.3,4 Although the surgical removal of tumors is a treatment option, early diagnosis remains difficult and challenging. The currently available biomarkers for PC are carcinoembryonic antigen (CEA) and carbohydrate antigen (CA) 19-9. CEA, a glycoprotein, was discovered from colon cancer in 1965 and has been used as for the first time.5 However, CEA has © 2014 American Chemical Society

Special Issue: Proteomics of Human Diseases: Pathogenesis, Diagnosis, Prognosis, and Treatment Received: March 15, 2014 Published: July 24, 2014 4878

dx.doi.org/10.1021/pr5002719 | J. Proteome Res. 2014, 13, 4878−4888

Journal of Proteome Research

Article

Research Resource Bank Program of the Korea Research Foundation of the Ministry of Science and Technology. Plasma samples were prepared according to the Human Proteome Organization reference sample collection protocol.25 This research was approved by the Institutional Review Board of Yonsei University of College of Medicine, and the diagnoses of PC were made by pathologists at the Severance Hospital of Yonsei University. Participants were classified into the following groups: healthy donors with no PC-related diseases (HD; n = 44) and patients with CP (n = 12), PC (n = 41) HCC (n = 31), CC (n = 22), and GC (n = 35). The mean age [ ± standard deviation (SD)] and gender ratio (male/female) of each group were 47 ± 15 and (10:2) for CP; 59 ± 11 and (27:14) for PC; 54 ± 10 and (24:7) for HCC; 63 ± 11 and (15:7) for CC; ansd 60 ± 12 and (26:9) for GC. Plasma samples of the same group were pooled, aliquoted, and stored at −80 °C. Detailed information about patients that provided the samples used during discovery and verification phases is provided in Supplementary Tables S1 and S2 in the Supporting Information and Table 1. The pooled samples were used for normalization and alignment purposes to minimize variation in the expression of those proteins of interest between samples.

CA242, and CA125) have also been studied as PC biomarker candidates because they often serve as indicators representing the tumor formation.13,14 Thus, there is an urgent need to develop new efficient PC biomarkers having higher specificity and sensitivity. As for other disease biomarkers, three types of clinical samples have been used in the laboratory in the search for a PC biomarker: tumor tissues, PC cell lines, and plasma from patients with PC. However, because tumor tissues of PC patients are rarely available because of the lack of a detection system for early stages of PC carcinogenesis, large screening studies on tumor tissues present great challenges.15 Although PC cell lines (such as HPAC, BXPC3, and PANC1) are useful for mechanistic studies and measuring expression of a particular targeted marker, they are not suitable for initial screening of PC biomarkers that can be directly applicable to cancer patients.10 For these reasons, plasma samples from PC patients have become the clinical specimen of choice for serologic cancer biomarker development.16 With the advent of technological developments in proteomic analysis of biofluids (e.g., plasma, urine, and tears), cancer cell lines (e.g., PANC1 cells), and various resources, several studies on biomarkers for PC have recently been performed, resulting in the identification of several candidate proteins.10,17−23 For example, Kosanam et al.17 identified laminin gamma 2(LAMC2), as a potential PC biomarker using extensive proteomic profiling and label-free quantification methods with both pancreatic tissues and plasma samples. Chen et al.22 also identified transthyretin as a serologic PC biomarker candidate using 2-D difference gel electrophoresis (2-D DIGE) coupled to matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) mass spectrometry. In addition, other proteins (e.g., DJ-1, APRIL, Plectin-1, and ULBP2)10,17−23 have also been reported as candidate protein biomarkers for PC, some of which are undergone rigorous validation processes. Nonetheless, despite such exhaustive efforts, there is still no efficient serologic biomarker candidate for PC that meets the basic criteria of high specificity and sensitivity. To identify a potential serologic biomarker for PC that could replace the current nonspecific PC biomarkers (e.g., CA 19-9, CEA), we developed an experimental strategy that employs human plasma samples from three types of patients (healthy donors (HD) and patients with chronic pancreatitis (CP) or PC) and the integrated proteomics platform (2-D DIGE, MALDI-TOF, immune-MS, and LC−MS/MS). Using this approach, we identified complement factor b (CFB), which is known to play an important role in the alternative pathway of complement system,24 as a candidate serologic PC biomarker. The diagnostic efficiency and the relative expression level of CFB were validated by comparison with those of CA 19-9 by ELISA using HD, CP, and PC plasma and PC cell lines. Here we report that CFB is a potential candidate biomarker for PC that exhibits much higher specificity than CA 19-9 in discriminating PC from not only CP but also other gastrointestinal cancers such as hepatocellular carcinoma (HCC), cholangiocarcinoma (CC), and gastric cancers (GC).

■

Table 1. Clinicopathological Information of Six Pooled Plasma Samples Obtained for Different Cancer Patients for Biomarker Discovery Discovery and Verification Set (Set 1 n = 77) gender total no.

diagnosis

healthy donor (HD) 13 chronic pancreatitis (CP) 12 pancreatic cancer (PC) 13 cholangiocarcinoma (CC) 13 gastric cancer (GC) 13 hepatocellular carcinoma 13 (HCC) Individual Verification Set

male

female

7 10 8 10 12 12

4 2 5 3 1 1

age (mean ± SD) ± ± ± ± ± ±

32 47 59 61 61 53

4 15 11 13 11 10

(Set 2 n = 185) gender

diagnosis healthy donor (HD) chronic pancreatitis (CP) pancreatic cancer (PC) cholangiocarcinoma (CC) gastric cancer (GC) hepatocellular carcinoma (HCC)

total no.

male

44 12 41 22 35 31

31 10 27 15 26 24

female age (mean ± SD) 13 2 14 7 9 7

52 47 59 63 60 54

± ± ± ± ± ±

14 15 11 11 12 10

Removal of High-Abundance Proteins from Plasma Samples

A multiple affinity removal system (MARS) (Hu-14, Agilent, Wilmington, DE) was routinely used to remove 14 major highabundance plasma proteins (e.g., albumin, transferrin, IgG, IgA, haptoglobin, antitrypsin, alpha 2-macroglobulin, fibrinogen, alpha 1-acid glycoprotein, IgM, apolipoprotein AI, apolipoprotein AII, complement C3, transthyretin).26 Chromatographic separation of the proteins by MARS was carried out using a mobile phase reagent kit according to a LC protocol provided by the manufacturer (Supplementary Table S3 in the Supporting Information). 30 μL of human plasma was diluted five-fold with 120 μL of Buffer A (Agilent), and mixtures were

EXPERIMENTAL SECTION

Clinical Specimens

All clinical samples were obtained from the archives of the Department of Pathology, Yonsei University (Seoul, Korea) and from the Liver Cancer Specimen Bank of the National 4879

dx.doi.org/10.1021/pr5002719 | J. Proteome Res. 2014, 13, 4878−4888

Journal of Proteome Research

Article

filtered through a 0.22 μm microcentrifuge filter tube by centrifugation at room temperature 16 000g for 1 min. The 150 μL of the diluted plasma was injected into the Agilent HP1100 LC system equipped with an affinity column (4.6 × 100 mm) at a flow rate of 0.125 mL/min. Flow-through fractions containing unbound proteins were collected, concentrated using a Amicon Ultra centrifugal filter unit (3K NMWL; Merck Millipore, Billerica, MA), and stored at −80 °C until use. Bound proteins were eluted from the column with Buffer B (elution buffer) at a flow rate of 1 mL/min for 7.0 min. The column was regenerated by equilibrating it with Buffer A for 11.0 min at a flow rate of 1 mL/min. The column was stored at 4 °C in a refrigerator after equilibrating with Buffer A.

Protein Identification by Mass Spectrometry

MALDI-TOF Mass Spectrometry. For MALDI-TOF, tryptic peptides were desalted and purified using a mixture of Poros R2 and Oligo R3 (Applied Biosystems, Foster City, CA), as previously described.29 The MS spectra of peptides were generated by spectrometric analysis using the 4800 MALDITOF analyzer (Applied Biosystems) in the reflectron/delayed extraction mode with an accelerating voltage of 20 kV with data summed from 500 laser pulses. The operating software used was Applied Biosystems 4000 series Data Explorer version 4.4. The T2D file obtained from 4800 MALDI-TOF was opened by Data Explorer version 4.4, and peaks were filtered according to four macro process: (1) baseline correction (peak width = 32, flexibility = 0.5, degree = 0.1), (2) noise filter/smooth: filter coefficient = 0.7, (3) spectrum peak deisotoping: adduct = H, generic formular = C6H5NO, and (4) mass calibration. (The spectrum was calibrated with the reference of tryptic autodigested peaks [m/z 842.5090, 1045.564, and 2211.1046], and monoisotopic peptide masses were obtained with Data Explorer 4.4.) A 800−4000 m/z mass range was used with 1000 shots per spectrum. At the end of the macro process, raw data were generated about centroid mass, resolution, height, and S/N ratio of each peak. These data were converted to an Excel file and used for MASCOT search. LC−MS/MS. A nano chip column (Agilent, 150 mm × 0.075 mm) was used for peptide separation. The mobile phases A and B for LC separation were 0.1% formic acid in deionized water and 0.1% formic acid in acetonitrile (ACN), respectively. The chromatography gradient was designed for a linear increase from 5% B to 8% B for 1 min, 8% B to 35% B for 19 min, 85% B for 10 min, and 5% B for 10 min. The flow rate was maintained at 400 nL/min. Product ion spectra were collected in the information-dependent acquisition (IDA) mode and analyzed by Agilent 6530 Accurate-Mass Q-TOF using continuous cycles of one full scan from 300−1500 m/z (four spectra/s) plus three product ion scans from 100−1700 m/z (two spectra/s). Precursor m/z values were selected starting with the most intense ion using a selection quadruple resolution of 4 Da.

2-D Difference Gel Electrophoresis and Image Analysis

2-D DIGE was performed as previously described.27,28 50 μg each of HD, CP, and PC plasma protein were labeled with 400 picomole Cy3, Cy5, and Cy2 fluorescent dye, respectively (GE Healthcare, Uppsala, Sweden). This internal standard was labeled with Cy2 and was run in parallel with HD, CP, and PC samples. The labeling reaction was performed in the dark on ice for 30 min and quenched with 10 mM lysine for 10 min. Labeled samples were combined, rehydrated, isoelectrically focused with 24 cm Immobiline Dry Strip pH 4−7NL (GE Healthcare) and separated in the second dimension by SDSPAGE (9−16%). The Cy2-, Cy3-, and Cy5-labeled images were scanned on a Typhoon 9400 scanner (GE Healthcare) at excitation/emission of 488/520, 530/580, and 633/670 nm, respectively. The gel images were analyzed using DeCyder 2-D analysis software v6.5.11 (GE Healthcare). Gel spot matching and statistical analysis were carried out using the Biological Variance Analysis (BVA) module. 2-D Gel Electrophoresis for the Preparative Gel

Plasma proteins from HD, CP, and PC were combined (total 1.0 mg) and diluted in 450 μL of rehydration buffer (7 M urea, 2 M thiourea, 100 mM DTT, 40 mM Tris, 4.5% CHAPS, 0.002% bromophenol blue, 2% IPG buffer). Isoelectric focusing was performed on an Ettan IPGphor (GE Healthcare) with 24 cm IPG strips (pH 4−7, GE Healthcare). After isoelectric focusing, each strip was 285 μL of isopropanol, 9.7 mL of equilibrium solution (1.5 M Tris-HCl (pH 8.8), 6 M urea, 50% glycerol, 2% SDS, 30% acrylamide), and 15.8 μL of tributyl phosphine. The equilibrated strips were transferred to 9−16% SDS-PAGE on an Ettan DALT 12 system (GE Healthcare). The preparative gel was stained with Coomassie brilliant blue G250 dye solution overnight, destained using ultrapure distilled water, and then scanned using a GS710 scanning densitometer (Bio-Rad, Hemel Hempstead, U.K.).

Data Searches for Protein Identification

MASCOT (Matrix Science, London, U.K.; version 2.2.04) was used to identify peptide sequences present in the protein sequence database NCBInr (Human). Database search criteria were as follows: (1) MALDI-TOF: NCBInr_Human_20130621 (26 617 015 sequences) and (2) LC−MS/ MS: NCBInr_Human_130324 (695 124 sequences); fixed modification, carboxyamidomethylated at cysteine residues; variable modification, oxidized at methionine residues; maximum allowed missed cleavage, 1; peptide MS tolerance (MALDI-TOF and LC−MS/MS), 100 ppm; and fragment MS tolerance (LC−MS/MS), 0.1 Da. Only peptides resulting from trypsin digests were considered.27,28 Protein scores showing greater than 66 for MALDI-TOF or individual ion scores showing greater than 42 for LC−MS/MS are regarded as significant (p < 0.05), where LC−MS/MS data have FDR < 1%.

Trypsin Digestion

2-D DIGE spots of interest for analysis were excised from the preparative gel, and the spots were transferred into each 1.5 mL tube. The spots were washed with 100 μL of distilled water; then, 50 μL of 50 mM NH4HCO3 (pH 7.8) and acetonitrile (6:4) were added to the spots and shook for in 10 min. This process was repeated at least three times until the Coomassie brilliant blue G250 dye disappeared. The supernatant was decanted, and the spots were dried in speed vacuum concentrator (LaBoGeneAps, Lynge, Denmark) for 10 min. Then, 100 ng per spot was digested with trypsin (Promega, Southampton, U.K.) in 50 mM ammonium bicarbonate and left on ice for 45 min. Spots were incubated at 37 °C for 12 h.

Western Blot Analysis

Samples were separated by 10% SDS-PAGE and transferred onto nitrocellulose membranes using an iBLOT dry blotting system (Invitrogen, Carlsbad, CA). The membranes were blocked with TBS-T buffer (20 mM Tris, 137 mM NaCl, 0.1% Tween-20, 5% skim milk, pH 7.6), incubated for 2 h with a 4880

dx.doi.org/10.1021/pr5002719 | J. Proteome Res. 2014, 13, 4878−4888

Journal of Proteome Research

Article

(ROC) curve,33 area under the ROC curve (AUC), and 95% confidence intervals (CI) were analyzed with 10-fold crossvalidations. From the optimal AUC of CFB, the diagnostic sensitivity, specificity, and Youden’s index [sensitivity (%) + specificity (%) − 100] were determined.34 The Panel Composer web tool was used for all statistical analysis.35

1:2000 dilution of mouse monoclonal [M13/12] antibody to CFB (Abcam, Cambridge, U.K.), and then incubated for 1 h with a 1:10 000 dilution of horseradish-peroxidase-conjugated goat antimouse secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Immunoreactive proteins were detected using ECL Plus Western blotting detection reagents (GE Healthcare).27,28

■

Immunoprecipitation and Band Intensity Analysis

RESULTS

Identification of CFB as a Candidate Biomarker for Pancreatic Cancer

Protein G agarose (Thermo Fisher Scientific, Rockford, IL), 1.0 mg of protein, and CFB antibody (Abcam) (2 μg antibody per 1 mg total protein) were incubated overnight at 4 °C. The beads were washed three times with PBS-T (135 mM NaCl, 31.3 mM KCl, 0.2 mM Na2HPO4, 0.5 mM KH2PO4, 0.05% Tween 20, pH 7.4), and antigens were eluted twice using 50 μL of PBS adjusted to pH 2, neutralized, and then concentrated in vacuo. Western blot analysis of the eluted samples was carried out as previously described. The CFB signals were quantified densitometrically using Image Quant TL v2005 software (GE Healthcare). Band intensity was measured for the straight line only, excluding the background and contaminated side lines.

To identify novel serologic biomarker candidates for PC, we used plasma samples obtained from HD and patients with CP or PC, as outlined in Figure 1. A typical 2-D DIGE pattern of

ELISA

Protein levels of CFB and CA 19-9 were measured by ELISA. The ELISA kit for CFB was purchased from USCNK (Wuhan, China), and the CA 19-9 kit was from Affymetrix (Santa Clara, CA). All assays were performed according to the manufacturers’ instructions. The absorbance values were read on a microplate reader (Benchmark Plus, Bio-Rad) at a wavelength of 450 nm. Cell Lines and Cell Culture

Three PC cell lines (HPAC, BXPC3, and PANC1) were from ATCC. Human pancreatic duct epithelial cell line (HPDE) was kindly provided by Dr. Ming-Sound Tsao (University of Toronto, Ontario, Canada). Cell culture media specified by ATCC for each of the three pancreatic cancer cell lines were used as follows:30 DMEM-F12 (1:1) with 10% fetal bovine serum for HPAC; RPMI 1640 with 10% fetal bovine serum for BXPC3; and DMEM with 10% fetal bovine serum for PANC1. The HPDE cell line was grown in serum-free media supplemented with bovine pituitary extract and EGF. All cells were cultured in a 37 °C incubator at 5% CO2. Cells were cultured in 100 mm culture dish at optimal seeding densities and harvested as needed for immunoprecipitation and Western blotting.

Figure 1. Schematic representation of the overall workflow.

patients’ plasma is shown in Supplementary Figure S1 in the Supporting Information. The pooled plasma samples were treated with MARS column (Hu-14), followed by 2-D DIGE analysis (Figure 2a). On average, 1782 spots in the three gels were matched (Figure 2b), among which 10 spots of PC samples showed more than 1.4 times higher abundance compared with the control groups (HD and CP) (Student’s t test, p < 0.05, Figure 2c, Table 2). From MS analysis of these 10 spots, we were able to identify five proteins that were differentially expressed between CP and PC samples (Table 2 and Supplementary Table S4 in the Supporting Information). Among these proteins, we excluded those proteins that are already known as PC-related proteins (e.g., fibrinogen gamma,36 apolipoprotein E37) and proteins involved in other cancers or diseases (e.g., E6-AP ubiquitin-protein ligase,38 haptoglobin39,40). Thus, CFB was chosen as a sole candidate protein that was differentially overexpressed in PC plasma samples (1.9 ± 0.1-fold compared with HD, 2.1 ± 0.1-fold compared with CP) (Figure 2d,e).

RT-PCR for CFB

Total RNA was isolated using easy-BLUE (iNtRON, Gyeonggi, Korea) according to the manufacturer’s protocol. cDNA was synthesized from 2 μg of total RNA using an omniscript RT kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. The sequences of the primers were as follows: CFB,5′-CAACAGAAGCGGAAGATCGTC-3′ (forward) and 5′-TATCTCCAGGTCCCGCTTCTC-3′ (reverse); 3 1 GAPDH,5′-ACCACAGTCCATGCCATCAC-3′(forward) and 5′-TCCACCACCCTGTTGCTGTA-3′ (reverse).32 PCR conditions for CFB and GAPDH were 35 cycles of denaturation at 94 °C for 1 min, annealing at 59 °C for 1 min, and primer extension at 72 °C for 1 min. Statistical Analysis

Molecular Verification of CFB by Immune-MS Analysis

To define statistical significance for differences in secretion of CFB protein among patient groups, we analyzed the relative CFB signal intensity by Mann−Whitney rank sum test, which are nonparametric statistics. Receiver operating characteristic

To verify the molecular identity of CFB detected as a PCspecific biomarker in the plasma, we subjected pooled nondepleted plasma specimens (HD [n = 13] CP [n = 12], 4881

dx.doi.org/10.1021/pr5002719 | J. Proteome Res. 2014, 13, 4878−4888

Journal of Proteome Research

Article

Figure 2. 2-D DIGE analysis of the HD, CP, PC plasma proteins and selection of CFB as a PC biomarker candidate. (a) Shown here is a type of fluorescent dye labeling of the clinical specimen for 2-D DIGE. (b) Each of three gels was labeled with CyDye DIGE Fluor Cy2 (yellow), while two samples were labeled with Cy3 (green) or Cy5 (red) minimal dyes, respectively. (c) 10 spots that were overexpressed in PC plasma are marked with circles. The gel image was analyzed by using DeCyder 2-D software. (d) CFB is indicated by arrow (nos. 1 and 2 in green) in the gel 3. (e) The 3D image results show that CFB is more abundant in PC plasma compared with HD or CP plasma. Abbreviations: HD, healthy donor; CP, chronic pancreatitis; PC, pancreatic cancer.

Figure 3. Molecular verification of CFB in pooled plasma from HD, CP, PC patients using immune-MS analysis. (a) Western blot analysis showed CFB expression in PC compared with HD and CP plasma. The lower panel shows the ratio of CFB intensity in each disease group normalized against HD values, determined from two replicate experiments. Data are presented as mean ± SD (p < 0.001). (b) MARS-depleted PC plasma (100 μg) was immunoprecipitated (IP) with antibody against CFB and subjected to SDS-PAGE. The bands marked by arrow were sliced for in-gel digestion and analyzed by MALDI-TOF. (c) Peptide sequences detected from the bands of the immunoprecipitated fraction of PC plasma were identified as CFB from the MASCOT PMF search in NCBI database. Abbreviations: HD, healthy donor; CP, chronic pancreatitis; PC, pancreatic cancer.

PC [n = 13]) to Western blot analysis using monoclonal antibody (mAb) against CFB (Abcam.). As shown in Figure 3a, CFB showed >2.0-fold higher expression in PC compared with HD and CP plasma, confirming its differential expression in PC samples. To make molecular verification of CFB, immunoprecipitation-MS (IP-MS) was performed using the mAb-captured CFB samples. As shown in Figure 3b,c, the presence of CFB

was confirmed by MALDI-TOF with a significant score and reasonable sequence coverage (>33%).

Table 2. Identification of Differentially Expressed Proteins by 2-D DIGE Analysis Coupled to LC−MS/MS or MALDI-TOFa spot no.

accession no.

protein name

score

sequence coverage

matched peptide

fold ratio (PC/HD)b

p value (PC/HD)b

fold ratio (PC/CP)

p value (PC/CP)

1c 2c 3d 4d 5d 6d 7d 8d 9d 10c

gi291922 gi291922 gi223170 gi3421149 gi4557325 gi4557325 gi178853 gi178853 gi178853 gi123508

complement factor B complement factor B fibrinogen gamma E6-AP ubiquitin-protein ligase apolipoprotein E precursor apolipoprotein E precursor apolipoprotein E apolipoprotein E apolipoprotein E haptoglobin

114 67 96 67 70 103 97 192 116 140

2% 2% 34% 80% 42% 41% 46% 56% 42% 6%

5 3 16 6 13 17 16 23 17 3

1.9 1.4 1.9 1.4 2.5 2.1 2.2 1.6 1.5 3.8

0.013 0.0015 0.022 0.014 0.00039 0.0014 0.00072 0.00037 0.019 0.00024

2.1 1.5 1.5 1.6 2.4 1.6 2.4 2.1 2..0 1.9

0.0013 0.0023 0.033 0.00011 0.0031 0.00097 0.0022 0.00054 0.01 0.0034

a

Abbreviations: HD, healthy donor; CP, chronic pancreatitis; PC, pancreatic cancer. bFold ratio and p value were estimated from biological variance analysis (BVA) mode of DeCyder software, 2-D DIGE analyzer. cIdentified only by LC−MS/MS, p value