Proteomic analysis identifies MMP-9, DJ-1 and A1BG as overexpressed proteins in pancreatic juice from pancreatic ductal adenocarcinoma patients

The study was approved by the Ethical Commitee of Shandong Academy of Medical Science. Fresh pancreatic juice samples were collected with informed consent from 9 PDAC patients and 9 cancer-free controls undergoing endoscopic retrograde cholangiopancreatography (ERCP) in Qilu Hospital and Qianfoshan Hospital of Shandong University. Clinical data of the patients included were summarized in Table 1. After collection, the pancreatic juice samples were centrifuged at 10,000 rpm for 20 min at 4°C and supernatant of each was aliquoted and stored at -80°C until used.

Pancreatic juice samples were first precipitated with acetone. Briefly, 1.2 mL cold acetone (Fluca) was added to 300 μL pancreatic juice and kept at -20°C for 2 hours, then centrifuged at 13,000 rpm for 10 min at 4°C. Supernatant was discarded and pellet was dissolved in 500 μL of lysis buffer containing 30 mM Tris, 8 M urea, 2 M thiourea, 2% 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS), 18 mM dithiothreitol (DTT) and 0.5% IPG Buffer (GE Healthcare). The mixture was then centrifuged at 12,000 rpm for 10 min at 4°C, and the supernatant was collected and stored at -80°C. Protein concentration was determined using a 2-D Quant Kit according to the manufacturer's instruction (GE Healthcare).

Cancerous or control pancreatic juice protein extracts were first pooled for traditional 2-DE. Two hundred micrograms of each pooled protein sample was diluted in 450 μL rehydration buffer (7 M urea, 2 M thiourea, 4% CHAPS, 0.5% IPG buffer, 0.002% bromophenol blue). Isoelectric focusing was performed on Ettan IPGphor (GE Healthcare) with 24 cm IPG strips (pH 3–10 NL, GE Healthcare). The IPG strips were first rehydrated at 30 V for 12 hours, then focused at 500 V for 1 hour, 1,000 V for 1 hour, and maintained at 8,000 V until a total of 65,000 Vhr was arrived. After isoelectric focusing, the strips were equilibrated with 0.375 M Tris-HCl (pH 8.8), 6 M urea, 20% glycerol, 2% sodium dodecyl sulfate (SDS), and 0.2% bromophenol blue. IEF strips were first treated with 130 mM DTT for 10 min, followed by 135 mM iodoacetamide for 10 min with constant shaking. The equilibrated strips were transferred to 12.5% SDS polyacrylamide gel electrophoresis (SDS-PAGE) on Ettan DALT twelve system (GE Healthcare) with constant power (0.2 W/gel, 1 hour; 1.7 W/gel, 4.5 hours) at 20°C. All gels were stained with Coomassie blue R350 (GE Healthcare), and scanned using a PowerLook 2100 XL scanner system (Umax USA).

Fifty micrograms of each of cancerous and cancer-free control pancreatic juice protein extracts was minimally labeled with 400 pmol Cy3 or 400 pmol Cy5 fluorescent dye (GE Healthcare). An internal standard pool generated by combining equal amounts of extracts from all the samples was labeled with 400 pmol Cy2. The labeling reaction was carried out in the dark on ice for 30 min, and quenched with 10 mM lysine for 10 min. 2-DE was performed as described above, except that low-fluorescent glass plates were used. The Cy2, Cy3, and Cy5-labeled images were acquired on a Typhoon Trio scanner (GE Healthcare) at the excitation/emission of 488/520 nm, 532/580 nm and 633/670 nm, respectively.

Spots of interest were excised from gels stained by Coomassie Blue R350, and were digested with sequencing grade modified porcine trypsin (Promega) as described previously [7]. Subsequent protein identification was carried out on the ABI 4700 Proteomic Analyzer MALDI-TOF-MS/MS mass spectrometry (Applied Biosystems, USA) on a reflective mode. Peptide mass fingerprint (PMF) was acquired between 800–3500 Da. The strongest five peaks from PMF were selected to obtain MS/MS spectra. PMF and MS/MS data were then searched against a human subset of the Swiss-Prot database using GPS explorer software (Applied Biosystems, USA).

Thirty micrograms of pancreatic juice protein extracts from 6 PDAC and 6 cancer-free controls were used for SDS-PAGE. The separated proteins were transferred to nitrocellulose membranes (GE Healthcare). The membranes were blocked for 1 hour and incubated with anti-DJ-1 antibody (1:1000 dilutions, MBL international corporation, MA), anti-MMP-9 antibody (1:1000 dilutions, Santa Cruz Biotechnology) and anti-A1BG antibody (1:1000 dilutions, Aviva System Biology, San Diego, CA) overnight at 4°C. After peroxidase-conjugated secondary antibody was added, proteins were detected using an ECL Plus kit (GE Healthcare).

To study the tissue compartment for DJ-1, MMP-9 and A1BG, a tissue array (Chaoying Biotechnology co. Xian, China) containing 51 cases of PDAC and 8 adjacent normal pancreas tissues was used in immunohistochemical analysis. Clinicopathological data of the tissue array were summarized (Table 2). Paraffin-embedded tissue array slides were processed for antigen retrieval using microwave heating in citrate buffer, and immunohistochemically stained with the same antibodies used in Western blot analysis at 1:100 dilution. Visualization was performed using ABC kit according to the manufacturer's recommendation (Zhongshan Bio co., Beijing, China).

Serum from 8 PDAC patients, 9 chronic pancreatitis patients and 8 healthy controls not included in the proteomic analysis were collected for MMP-9 ELISA analysis. Serum levels of human MMP-9 were analyzed with a commercially available kit (Biotrak) according to the manufacturer's recommendation. Serum samples were diluted at 1:100; all assays were done in duplicate. The sensitivity limit of MMP-9 ELISA was 0.08 ng/ml.

DIGE images were analyzed by ImageMaster 6.0 DIGE-enable software (GE Healthcare). The best internal standard image was assigned as the master reference. The protein spots on the remaining internal standard images were matched to the master reference to ensure that the same protein spots were compared between gels. Spot intensity was normalized by dividing each Cy3 or Cy5 spot volume with the corresponding Cy2 (internal standard) spot volume. Statistical analyses were performed using SPSS 13.0 software. The relationships with DJ-1, MMP-9 and A1BG expression and clinicopathological parameters were analyzed using Chi-square or Fisher's exact tests. The differences of MMP-9 levels in serum among the various groups were analyzed using independent-samples t-test. p < 0.05 was considered to be statically significant.

Given the limited amount of pancreatic juice sample available, we first made sample pools of all cancerous and cancer-free control pancreatic juice protein extracts separately to identify the differently expressed proteins by traditional 2-DE. Each pool was repeated three times. Coomassie blue R350 staining was applied to visualize the protein spots, because of its compatibility with protein identification by MS. Moreover, we carried out DIGE on each individual sample to verify differential protein expression found by the traditional 2-DE. Most spot changes in DIGE analysis were consistent with those found in Coomassie blue stained 2-DE analysis (Figure 1). Twenty four proteins with more than two-fold expression change between pancreatic cancer juice and cancer-free pancreatic juice were identified (Table 3). Fourteen proteins were significantly up-regulated and ten proteins were significantly down-regulated in pancreatic cancer juice compared with the cancer-free controls (Figure 2). Three up-regulated proteins MMP-9, DJ-1 and A1BG were further confirmed by Western blot and immunohistochemistry.

As seen in Figure 3, compared with cancer-free controls, increased MMP-9, DJ-1 and A1BG expression in cancerous pancreatic juice were detected by Western blot. In addition, cancerous pancreatic juice with MMP-9 expression evinced both 92 kDa and 82 kDa bands, corresponding to the latent and activated forms of MMP-9, respectively.

The expressions of MMP-9, DJ-1 and A1BG were confirmed by immunohistochemistry in 51 pancreatic cancer and 8 normal pancreas samples that were not included in the proteomic experiment (Table 4; Figure 4). MMP-9 was detected in the malignant ductal epithelia in 82.4% of PDAC tissues, but was barely detectable in normal pancreas. In addition, 9 cases of PDAC demonstrated strong MMP-9 expression in the stroma.

In normal pancreas, DJ-1 was negatively or weakly expressed in duct epithelium, acinar cells and islet cells. In PDAC, 72.5% demonstrated DJ-1 over-expression in cancer cells. In addition, DJ-1 over-expression was related to the differentiation of PDAC. DJ-1 positive stain was observed in 7/9 well differentiated tumors, 16/19 moderately differentiated tumors, and 14/23 poorly differentiated tumors, respectively. There was a significant difference of DJ-1 over-expression between moderately differentiated PDAC with poorly differentiated PDAC (p < 0.05).

No positive A1BG staining was detected in all the normal pancreas tissues. A1BG was over-expressed in the cytoplasma of malignant epithelia in 86.3% of pancreatic cancer tissues, significantly higher than that in normal pancreas tissues (p < 0.01).

As shown in Figure 5, serum MMP-9 levels were significantly higher in patients with pancreatic cancer (median = 255.14 ng/ml; quartile range, 125.43 ng/ml) than those with chronic pancreatitis (median = 210.22 ng/ml; quartile range, 12.48 ng/ml; p = 0.009) and normal controls (median = 203.77 ng/ml; quartile range, 17.04 ng/ml; p = 0.027).