Molecular BioSystems View Online Dynamic Article Links Cite this: DOI: 10.1039/c0mb00242a PAPER www.rsc.org/molecularbiosystems The Haptoglobin b chain as a supportive biomarker for human lung cancersw Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A Sung-Min Kang,a Hye-Jin Sung,a Jung-Mo Ahn,a Jae-Yong Park,b Soo-Youn Lee,c Choon-Sik Parkd and Je-Yoel Cho*a Received 21st October 2010, Accepted 13th December 2010 DOI: 10.1039/c0mb00242a Haptoglobin (Hp) is produced as an acute phase reactant during inflammation, infection, malignant diseases, and several cancers. In proteomics analysis using human blood samples, the Hp peptide levels were about 3-fold higher in lung cancer patients versus normal individuals. This study is aimed at analyzing the elevation of which chain of Hp is closely related to lung cancers and can be a serum biomarker for lung cancers. In Western blot (WB) analysis, we found that the Hp b chain can be a better diagnostic biomarker for lung cancers. In the result of the Hp b chain ELISA developed by us, the concentrations of the Hp b chain in the sera increased about 4-fold in 190 lung adenocarcinoma patients versus 190 healthy controls (8.0 3.8 mg ml 1 vs. 1.9 1.2 mg ml 1). ELISA data showed that the serum levels of the Hp b chain in breast cancer (1.5 0.5 mg ml 1) and hepatocellular carcinoma (HCC) (1.4 1.0 mg ml 1) patients remained similar to those of healthy controls. Compared to lung adenocarcinoma, the Hp b chain levels in the plasma of patients with other respiratory diseases such as tuberculosis (TBC), idiopathic pulmonary fibrosis (IPF) and bronchial asthma (BA) were closer to those of healthy controls. Our data suggest that an increase of the Hp b chain can be a potential serum biomarker for lung cancers. Introduction Lung cancer is one of the leading causes of cancer-related death worldwide.1 According to a study of 5-year survival rates between 2001 and 2005 by the Ministry for Health, Welfare and Family Affairs, lung cancer (15.5%) has a very low survival rate in Korea compared to other types of cancers, including stomach cancer (56.4%), colon cancer (64.8%), breast cancer (87.3%) and thyroid cancer (98.1%). Although several screening techniques, such as X-rays and computed tomography (CT) scans, are frequently used for lung cancer diagnosis, their high cost, poor specificity, and high risk of radiation exposure are still problematic for lung cancer diagnosis.2 Lung cancer has four major histological types that can be grouped into two large categories: small-cell lung cancer a Department of Biochemistry, School of Dentistry and Brain Korea 21, Kyungpook National University, and ProtAnBio, Dongin-dong 2 Ga 101, Daegu, 700-422, South Korea. E-mail: jeycho@knu.ac.kr; Fax: +82 53-421-1418; Tel: +82 53-420-4997 b Internal Medicine, School of Medicine, Kyungpook National University Hospital, Daegu, South Korea c Department of Laboratory & Genetics, Samsung Medical Center, Sungkyunkwan University of Medicine, Seoul, South Korea d Division of Allergy and Respiratory Diseases, Department of Internal Medicine, Soonchunhyang University Hospital, Seoul, South Korea w Electronic supplementary information (ESI) available: Supplementary Table 1: Unique peptides detected in each band by proteomic analyses. Supplementary Table 2: Unique peptides detected in each chain by proteomic analyses. See DOI: 10.1039/c0mb00242a This journal is c The Royal Society of Chemistry 2011 (SCLC) and non-small-cell lung cancer (NSCLC). NSCLC, which accounts for 80% of lung cancers, consists of squamouscell carcinoma representing 20–25% of lung cancers, largecell carcinomas constituting 15–20% of lung cancers, and adenocarcinoma accounting for 30–40% of lung cancers. Adenocarcinoma has been reported to be the most prevalent histological subtype of lung carcinomas.3,4 Biomarkers, which are quantifiable measurements that distinguish between normal and abnormal conditions, include many different forms, such as physiological conditions, visual characteristics, genes, and proteins.5 Recently, there have been various attempts to improve the monitoring of lung cancer therapies using genomic and proteomic techniques. Among the many kinds of biomarkers, protein biomarkers are currently being researched in conjunction with the development of techniques in proteomics.6 Blood samples are good sources of biomarkers because proteins in blood have significant physiological functions and secreted proteins are records of physiological conditions. However, due to the existence of abundant proteins in blood, the discovery of cancer biomarkers in serum is challenging.7 Despite these difficulties, lung cancer biomarkers have been characterized, and several genes and proteins have been proposed as lung cancer-biomarker candidates: cytokeratin-19 fragment (CYFRA 21-1), carcinoembryonic antigen (CEA), cancer antigen-125 (CA-125) and neuron-specific enolase (NSE) etc. However, many of these biomarkers are not clinically used, and most still lack sensitivity and specificity.8–10 Mol. BioSyst. Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A View Online The acute-phase response is a remarkable systemic reaction to disturbances in host homeostasis caused by infection, trauma, surgery and immunological disorder. The acute-phase response stimulates the liver to secrete acute-phase proteins, which include the positive acute proteins, C-reactive protein (CRP), serum amyloid A (SAA) and tumor necrosis factor alpha (TNF-a), and the negative acute proteins, retinol binding protein (RBP), transthyretin (TTR) and transferrin. Hp is a positive acute-phase protein that is induced by the immediate host inflammatory response to stimuli primarily in hepatocytes.11 The Hp molecule is a tetrameric protein of two a and b chain dimers that are connected by disulfide bridges. These two chains originate from a common precursor protein that is proteolytically cleaved during protein synthesis. The main function of Hp is to bind free plasma hemoglobin for protection from oxidative stress and to assist in hemoglobin uptake by the hemoglobin scavenger receptor CD163;12 however, under conditions of inflammation, Hp is elevated through the activity of IL-6. Several studies have reported Hp elevation in breast cancer,13 ovarian cancer,14 pancreatic cancer,15,16 malignant lymphoma,17 urogenital tumor18 and bladder cancer.19 Another study reported that the fucosylation of Hp is increased in pancreatic cancer and ovarian cancer. In this study, we identified Hp using a LC-ESI-MS/MS analysis of crude serum samples and then verified and quantified Hp levels by WB analysis and ELISA. This analysis was done on lung adenocarcinoma sera; samples from patients with other types of solid tumors, including breast cancer and HCC; and samples from patients with respiratory diseases. Our data suggest that Hp b chain levels are increased in lung adenocarcinoma compared with other types of cancers and healthy controls. This increase of Hp b chain levels in lung cancer may be a supportive biomarker for lung cancer. Results Identification of Hp by LC-ESI-MS/MS analysis Serum samples were obtained from consenting individuals with or without lung adenocarcinoma according to the serum collection protocol described in Experimental procedures. Clinical information about the samples used in this study is summarized in Table 1. Using one-dimensional gel electrophoresis, 100 mg of protein from the pooled sera of five individuals from the healthy control and lung adenocarcinoma groups were separated. The gel was stained with Coomassie Brilliant Blue (CBB) staining solution, which showed three bands at 45, 15 and 10 kDa that had different intensities between the normal and lung adenocarcinoma samples (Fig. 1A). To identify the proteins that were elevated in the lung adenocarcinoma sample, six bands of interest were excised and in-gel digested; the tryptic peptides were then analyzed by LC-ESI-MS/MS analysis. In each band, 20–30 proteins were identified and the 1 > false positive rate (FPR) determined using a decoy database (Table 2). Hp peptides were found in the bands at 45 kDa (band no. 1), 15 kDa (band no. 2) and 10 kDa (band no. 3; Fig. 1A). Hp with 119, 91 and 42 spectral counts were detected in each band of the lung adenocarcinoma samples, whereas 32, 28 and 19 Mol. BioSyst. Table 1 Clinical data of the patients who provided plasma and serum samples Source Sample type Sex Age Serum Healthy Male Female Male Female Female Male Female Male Female Male Female Male Male Male Male Male 56.2 52.1 60.3 62.4 47.8 53.7 61.0 69.1 58.5 64.4 53.0 54.4 60.4 55.8 60.5 54.9 AdenoCA BreastCA HCC SQLC SCLC Plasma Healthy AdenoCA TBC IPF BA Total 9.3 10.4 10.3 9.6 6.4 6.0 6.7 6.4 12.4 9.0 0.0 3.3 5.2 4.2 12.3 3.0 118 83 164 75 22 16 4 36 4 39 1 50 30 50 50 50 Serum samples were obtained from the Samsung Medical Center and the Kyungpook National University Hospital, and plasma samples were obtained from the Soonchunhyang University Hospital. Abbreviations used: AdenoCA, lung adenocarcinoma; BreastCA, breast cancer; HCC, hepatocellular carcinoma; SQLC, squamous-cell lung cancer; SCLC, small-cell lung cancer; TBC, tuberculosis; IPF, idiopathic pulmonary fibrosis and BA, bronchial asthma. Data are means SD. were found in the healthy control sample bands. The average ratio of the lung adenocarcinoma to the normal samples for Hp precursor spectral counts detected by our mass analysis was 3.1 (Table 3). Hp is composed of a1, a2 and b chain polypeptides in different combinations. In our MS/MS analysis results, the major MS/MS spectrum of the Hp precursor of each band was representative of the b, a2 and a1 chains, respectively. In band no. 1, the peptide sequences SPVGVQPILNEHTFCAGMSK and SCAVAEYGVYVK were detected with 21 and 20 peptide hits, respectively, and these peptide sequences matched the Hp b chain. For band no. 2 and band no. 3, the peptide fragments LRTEGDGVYTLNDKK and AVGDKLPECEAVCGKPK were primarily detected and each matched the Hp a2 and a1 chains with 28 and 6 peptide hits, respectively (Fig. 1B). The unique sequences of Hp precursor in each band and in each chain are summarized in ESIw Tables S1 and S2, respectively, with the corresponding hit numbers. Hp proteins were identified with the coverage of 15–38% range in each band of the lung adenocarcinoma and normal samples. Fig. 1C shows the full amino acid sequences of the Hp and the tryptic peptide fragments detected in LC-MS/MS analysis are indicated with boxes. In summary, considering that the spectral count of Hp detected in the bands was much higher than that of the other proteins (data not shown), each band indicated a different chain of Hp and the key components of the elevated intensities of bands in CBB staining are three different chains of Hp. Identification of Hp in individual bands in immunoblotting As previously described, Hp is composed of three different chains, and each chain has different predicted molecular weights: 9, 19 and 29 kDa. However, the immunoblotting analysis with a polyclonal antibody (Ab) detected bands at 9, This journal is c The Royal Society of Chemistry 2011 Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A View Online Fig. 1 Identification of Hp by LC-ESI-MS/MS analysis and MS/MS spectrum of Hp fragments. (A) Sera from a total of five healthy control and five lung adenocarcinoma patients were pooled, and then 100 mg of protein of each pooled sample was separated by SDS-PAGE and stained with Coomassie Brilliant Blue. Three bands at different molecular weights that showed different intensities were excised and in-gel digested. Lane 1, pooled normal sera; Lane 2, pooled lung cancer patient sera. (B) Representative peptides of each band showing each chain of Hp. MS/MS spectrum of the peptide fragment SPVGVQPILNEHTFCAGMSK (m/z = 2171.66, charge = 2+) from band 1, which is part of the Hp b chain, with 21 peptide hits. In band 2, the Hp a2 chain fragment LRTEGDGVYTLNDKK (m/z = 1707.68, charge = 2+) was found with 28 peptide hits. The band 3 peptide AVGDKLPECEAVCGKPK (m/z = 1857.46, charge = 2+) had 6 peptide hits and represented the Hp a1 chain. (C) The figure shows the full amino acid sequence of Hp: blue letters, signal sequence; pink and green letters, a2 chain; green letters only, a1 chain; and black letters, Hp b chain. Identified peptides are in black boxes, and expected modification sites are in red letters. Table 2 The number of proteins identified by LC-MS/MS analysis from individual bands Table 3 Identification of Haptoglobin by LC-MS/MS analysis Protein Band 1 Band 2 Band 3 N LC N LC N LC Number of proteins identifieda Peptide false-positive rateb (%) 26 28 21 21 22 18 0.705 0.548 0.434 0.434 0.676 0.755 Description Peptide numbers Coverage (%) Ratioa N1 32 IPI00641737 Haptoglobin N2 precursor 28 N3 19 LC1 119 LC2 91 LC3 42 N1 26 N2 17 N3 15 LC1 38 LC2 31 LC3 33 3.71 3.25 2.21 a Ratio is LC peptide quantitative value divided by N peptide quantitative value. a Positive identification of a protein required a minimum of two peptides (95% confidence) per protein. b Calculation of the FPR is described in Experimental procedures. 19 and 45 kDa, which are different than the predicted molecular weights of each band. We hypothesized that the b chain might be N-glycosylated and thus detected at the higher molecular weight of 45 kDa. To test this hypothesis, we first searched for possible N-glycosylation motifs (NxS/T) in the This journal is c The Royal Society of Chemistry 2011 sequences and found four possible sites in the b chain; which are written in red characters in Fig. 1C. PNGase F treatment was then performed because PNGase F cleaves asparaginelinked (N-linked) oligosaccharides from glycoproteins, which deaminates asparagine to aspartic acid and leaves the oligosaccharides intact. Following the PNGase F treatment, the 45 kDa band shifted to 29 kDa, suggesting that the 45 kDa band is a glycosylated form of the b chain (Fig. 2A). Mol. BioSyst. Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A View Online Fig. 2 Elevation of the Hp b chain compared with the a2 chain in sera of lung adenocarcinoma patients compared to healthy controls. (A) Serum samples treated with PNGase F for the characterization of Hp bands were analyzed by immunoblotting with an anti-Hp polyclonal antibody. PNGase F treatment resulted in a molecular mass reduction in the Hp b chain (29 kDa) but not in the a1 and a2 chains. Lane 1, crude serum sample; Lane 2, PNGase F untreated sample; Lane 3, PNGase F treated sample. (B, C) Protein from the crude sera of 44 healthy controls, 78 lung adenocarcinoma patients, 22 breast cancer patients and 20 HCC patients was separated by SDS-PAGE, and then Hp expression was detected by immunoblotting. Representative blots of the Hp a2 chain (B) and the b chain (C) expression in each cancer patient and healthy control are shown. Each number indicates the patients’ number used for the immunoblotting. (D, E) Densitometry plot of the respective bands for the a2 chain (D) and the b chain (E). Each blot contained the same standard sample; when analyzing the results of immunoblotting by densitometry, the total pixels of each band (density area) analyzed by Scion image were normalized to the total pixels of the standard. The boxes give the upper 75% and lower 25% quartiles of the measurements with respect to the median value (horizontal line in each box) and each dot represents individual patient data, while the small open rectangle represents the mean value. The expression levels of the Hp b chain showed significant differences between the sera of lung cancer patients and healthy controls, breast cancer patients or hepatocellular carcinoma patients. Abbreviations used: AdenoCA; lung adenocarcinoma, BreastCA; breast cancer and HCC; hepatocellular carcinoma (* indicates P o 0.05). Higher increase level of the Hp b chain than the a2 chain in lung adenocarcinoma To validate the Hp expression levels, the sera of 44 healthy controls and 78 lung adenocarcinoma patients were tested to estimate the expression level of Hp by immunoblotting analysis. The a2 and b peptide chains of lung cancer samples showed higher expression levels than those of healthy controls, and the differences were great enough to be detected by WB analysis (representative data in Fig. 2B and C). We analyzed 44 healthy and 78 lung adenocarcinoma samples by WB analysis, and the densities of the a2 band and b chain bands in the WB were photographically measured by Scion image analysis. The relative protein levels were measured by densitometry analysis of the Hp a2 and b chain Mol. BioSyst. bands using the equation: (pixel number of sample band / pixel number of standard band) 100 individuals; Fig. 2D and E. To analyze the applicability for differential diagnosis, we also analyzed different types of cancers, including breast cancer and HCC. The results showed that, although the Hp a2 and b chains were both potential markers for lung cancer diagnosis, b chain levels were better indicators than a chain levels for the identification of lung adenocarcinoma. In our study, the Hp a1 chain was not clearly detectable in most of the healthy and cancer samples. The expression levels of the Hp b chain showed significant differences between lung adenocarcinomas and healthy control samples (p = 0.04299) and also between adenocarcinomas and other types of solid cancers, including breast cancer and HCC (p = 0.01941 and 0.00457; This journal is c The Royal Society of Chemistry 2011 View Online Increased expression levels of the Hp b chain in lung adenocarcinoma To better quantify the levels of the Hp b chain in the sera, we developed an ELISA using a b chain-specific Ab for detection. The specificity of the Ab was tested by WB using lung cancer serum. The Ab recognized a single specific band of the Hp b chain (Fig. 3A). Then, the sera of 190 lung adenocarcinoma Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A Fig. 2E); however, the a2 chain level was slightly increased in lung adenocarcinomas compared to healthy controls (p = 0.002), and there was not much difference between lung cancer and other types of tumors (breast cancer, p = 0.20366; HCC, p = 0.06402; Fig. 2D). Altogether, these data show that the Hp b chain is a better marker for lung adenocarcinoma diagnosis compared to the a chain. Fig. 3 ELISA showed an increase of the Hp b chain in the sera of lung cancer patients. (A) To verify the specificity of the Hp b chain antibody used for ELISA, immunoblotting was performed for crude serum of lung cancer patients, PNGase F-untreated serum of lung cancer patients and serum of lung cancer patients after PNGase F treatment. Lane 1, crude serum sample; Lane 2, PNGase F untreated sample; Lane 3, PNGase F treated sample. (B) Concentrations of the Hp b chain in the sera of 190 healthy controls and 190 lung adenocarcinoma patients were measured by ELISA. On average, the Hp b chain levels in the lung adenocarcinoma samples (8.0 mg ml 1) were 4.2-fold higher than in healthy controls (1.9 mg ml 1). Sera from patients with other histological types of lung cancers, including squamous-cell lung cancer (SQLC) and small-cell lung cancer (SCLC), were also subjected to ELISA analysis. The average value for SQLC was 8.2 mg ml 1 and the average value for SCLC was 8.9 mg ml 1. (C) When the serum samples were grouped by lung cancer stages, the Hp concentration in serum was correlated with the stages, with the exception of stage II, which had a lower number of samples. (D) The Receiver Operating Characteristic (ROC) curve of Hp b chain with the sensitivity and specificity values from the cut-off range of 2 mg ml 1. The detection of cancer samples with respect to healthy controls (* indicates P o 0.05). This journal is c The Royal Society of Chemistry 2011 Mol. BioSyst. Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A View Online patients and 190 healthy controls were analyzed by ELISA (Fig. 3B). The ELISA data showed that the average concentration of the Hp b chain was different between lung adenocarcinoma (8.0 3.8 mg ml 1) patients and healthy controls (1.9 1.2 mg ml 1). The ELISA results from lung adenocarcinoma samples were analyzed by cancer stage in order to evaluate the correlation between Hp b chain concentration and cancer progression (Fig. 3C). The Hp b chain concentration showed a gradual increase that depended on the cancer stage. The average of stage I was 6.2 2.5 mg ml 1, stage III was 9.0 4.6 mg ml 1 and stage IV was 11.8 4.9 mg ml 1; however, the data for stage II did not show correlation between stage and Hp b chain concentration unlike other stages. It seems that this is due to the small number of samples. In the ELISA analysis for lung adenocarcinoma and healthy control samples, when the cut-off was set at 2 mg ml 1, the AUC value was 0.822 and the specificity and sensitivity of the Hp b chain as a diagnostic marker were estimated to be 63.2% and 82.6%, respectively (Fig. 3D). These data suggest that the Hp b chain can be a potential biomarker for both the detection and monitoring of lung adenocarcinoma. Similar pattern of the Hp b chain increase in other histological types of lung cancer To determine whether the increase of the Hp b chain was specific for lung adenocarcinoma or common to other histological types of lung cancers, we measured the Hp b chain levels by ELISA in SQLC and SCLC samples. The average concentrations of the Hp b chain were not different from that of lung adenocarcinoma samples (8.2 3.4 mg ml 1 for SQLC; 8.9 3.6 mg ml 1 for SCLC) (Fig. 3B). Our ELISA data showed that the increase of the Hp b chain in sera was not specific to lung adenocarcinoma because the Hp b chain was also increased in other histological types of lung cancers, such as SQLC and SCLC. Comparison of the Hp b chain levels between lung adenocarcinoma patients and other cancer and respiratory disease patients ELISA analysis of the Hp b chain was also performed on sera from patients with other types of solid tumors and respiratory diseases, and these results were compared to the lung adenocarcinoma samples to test the usefulness of the Hp b chain for differential diagnosis. ELISA showed that the Hp b chain levels of the serum in breast cancer and HCC were 1.5 0.5 and 1.4 1.0 mg ml 1, respectively (Fig. 4A). The data showed that the breast cancer and HCC levels were not different from that of healthy controls. Except for one sample in HCC, all of the breast cancer and HCC samples had the Hp b chain concentrations that were lower than our cut-off value (5 mg ml 1). We also analyzed the Hp b chain by ELISA in plasma samples of patients with TBC, IPF and BA and compared the results to the plasma samples of healthy controls and lung adenocarcinoma patients. The plasma concentration of the Hp b chain was about 3-fold higher in lung adenocarcinoma patients (11.4 4.3 mg ml 1) compared to healthy controls (3.2 1.3 mg ml 1). The Hp b chain Mol. BioSyst. Fig. 4 Higher levels of the Hp b chain in the serum and plasma of lung adenocarcinoma patients relative to patients with other types of solid tumors or respiratory diseases. (A) Validation of the Hp b chain expression in other types of tumors, including 20 breast cancer and 20 hepatocellular carcinoma (HCC) patients, was performed by ELISA. When a cut-off value of 5 mg ml 1 was applied, only one HCC sample was detected above the threshold. (B) Validation of the Hp b chain expression level was also performed by ELISA in the plasma samples from healthy controls, lung adenocarcinoma (AdenoCA) patients and patients with other respiratory diseases, including tuberculosis (TBC), idiopathic pulmonary fibrosis (IPF) and bronchial asthma (BA). The concentration of the Hp b chain was significantly higher in the lung adenocarcinoma group than in healthy controls or patients with other respiratory diseases (* indicates P o 0.05). concentrations in patients with other respiratory diseases were all significantly lower than that in the lung adenocarcinoma patients: TBC (4.6 2.0 mg ml 1), IPF (5.8 2.6 mg ml 1) and BA (3.7 1.6 mg ml 1) (Fig. 4B). These results suggest that the Hp b chain might be a potential biomarker for lung cancers compared to other types of solid tumors, such as breast cancer and HCC, and, to some extent, may also be a biomarker for respiratory diseases. This journal is c The Royal Society of Chemistry 2011 View Online Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A Discussion In this study, we discovered that Hp expression was higher in the sera of lung cancer patients compared to healthy controls by LC-MS/MS analysis, and this result was validated and quantified by immunoblotting and ELISA. Among the three different chains of Hp, Hp a2 and b chains showed differences compared to healthy controls, but only the Hp b chain showed a significant difference between lung cancers and other tumors. The Hp b chain was increased not only in the sera of lung adenocarcinoma patients, but also in other histological types of lung cancers; however, the level of increase was different (4-fold higher on average) between lung cancer and other types of tumors or respiratory diseases. Hp is an acute phase-reactant tetrameric (a2b2) glycoprotein. The main function of Hp is to remove free plasma hemoglobin.20,21 However, Hp has also been shown to have angiogenic22 and antioxidant23 properties and has an important role in cell migration.24 Although Hp is mainly synthesized in the liver, the local differential expression of Hp has also been demonstrated in cancer tissues.25–27 It has been reported that the Hp a subunit is a potential biomarker in ovarian cancer and SCLC. Ye et al. identified Hp by SELDI-MS/MS and then examined the Hp a subunit expression level by WB and quantified the a chain by ELISA.28 The Hp a chain levels in 13 SCLC samples were also about 3-fold higher than in 4 healthy controls when measured by densitometry of western blots.29 In our lung cancer study, the data of the Hp a subunit showed a similar increase compared to ovarian cancer and SCLC, indicating that the Hp a chain may be similarly increased in other types of cancers. Although Hp is reportedly a potential biomarker for pancreatic and ovarian cancers,15,30 few studies have been conducted investigating Hp in relation to lung cancers. In previous reports of Hp in lung cancers, most studies have measured the total Hp without discriminating between each chain,31,32 and none of the researchers specifically measured the b chain of Hp. In this study, however, we found that the Hp b chain had a greater increase in the sera of lung cancer patients compared to the other chains, a1 and a2, and could be a better biomarker for the identification of lung cancers and for differentiating between lung cancer and other cancers such as breast cancer and hepatocellular cancers. The quantitative measurement of the Hp b chain was possible because we setup an ELISA method using a Hp b chain specific monoclonal antibody. Our data suggest that the increased level of the Hp b chain in the sera of lung cancer patients can be a supportive serum biomarker for lung cancer diagnosis and monitoring. Although our data showed that the p-values for the expression level of the Hp b chain in the respiratory diseases are not significant compared to healthy controls, the statistical analysis for the levels of the Hp b chain between lung cancer and other respiratory diseases turned out to be significant (Fig. 4B). However, there are still overlapping ranges of the Hp b chain levels between lung adenocarcinoma and other respiratory diseases. Thus, caution is needed when the Hp b chain is to be used as a potential biomarker to differentiate lung cancer patients from other respiratory disease patients. This journal is c The Royal Society of Chemistry 2011 We determined that the Hp b chain is N-glycosylated using a PNGase F treatment experiment (Fig. 2A). Our data are also supported by previous studies showing that the Hp b chain has four glycosylation sites and might exist as a glycosylated form in colon cancers.33 Another report also showed that Hp is a fucosylated protein and suggested that the increased fucosylation of Hp in cancers may indicate that cancer can induce the fucosylation of the molecule.34 An increase of fucosylated Hp in pancreatic cancer has also been reported and did not appear to be correlated with the total amount of Haptoglobin.30 One of the possible reasons why we observed a better diagnostic predicted value of the Hp b chain relative to the a chains might be due to the glycosylation characteristics of the b chain. Because the serum proteins in cancer patients are known to be further glycosylated, it is possible that the b chain is more glycosylated in lung cancer patients than in normal controls. Because the monoclonal antibody we used for the WB and ELISA can similarly detect the Hp b chain in its glycosylated and unglycosylated forms (Fig. 3A), the difference may not be caused by differences in the detection ability of the antibody for the Hp b chain. It is quite possible that the glycosylated Hp b chain in the sera of lung cancer patients might be more stable and might have a longer half-life in the serum. Thus, we would see a greater increase in the Hp b chain than the Hp a chain in the sera of lung cancer patients. For Hp to be applied to lung cancer diagnosis, the increase in the total protein amount and the fucosylated status should be combined. Conclusion We identified Hp in serum and verified the elevation of the glycosylated b chain in the lung cancer patients. Although the Hp a2 chain was slightly increased, a greater increase of the Hp b chain was found. The marked increase of the Hp b chain in lung cancers compared to other types of cancers or respiratory diseases suggests that it can be a potential serum biomarker for lung cancers. Experimental procedures Human serum samples Serum and plasma samples were obtained from patients with lung cancer, other types of solid cancers and respiratory diseases from the Samsung Medical Center (IRB No. 2008-06-007-005), the Kyungpook National University Hospital (IRB No. 74005-833) and the Soonchunhyang University Hospital, a member of the National Biobank of Korea. Informed consent was obtained from all donors. We used serum samples from 239 lung adenocarcinoma patients, 201 healthy controls, and 40 patients with other histological types of SQLC and SCLC lung cancers and other types of solid cancers, including 22 breast cancer and 20 HCC samples. We also used plasma samples from 50 healthy controls, 30 lung adenocarcinoma patients and 50 patients with the respiratory diseases TBC, IPF and BA. Serum and plasma were separated from whole blood by centrifugation within 4 h after collection and then stored at 70 1C until use. Additional information Mol. BioSyst. View Online on the samples is described in Table 1. Among samples described above, samples for WB analysis and ELISA were randomly selected and some of them were used for both analyses. with a protein that has a 95% probability, according to the Protein Prophet algorithm,38 and a minimum of 2 peptides are matched with the protein sequence, each with a 95% probability, based on the Peptide Prophet algorithm. Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A Identification by LC-ESI-MS/MS analysis Separation by SDS-PAGE and Coomassie Blue staining. Protein (100 mg) from the pooled sera of five lung adenocarcinoma patients and healthy individuals was separated by SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis). The gels were then washed three times with ddH2O for 5 min and stained with Bio-Safe Coomassie Stain solution (Coomassie G250 Stain; BioRad, Hercules, CA) for 1 h with gentle shaking at room temperature (RT). The gels were destained by incubation in ddH2O overnight. In-gel digestion. After being stained with Coomassie, the protein bands of interest were excised, sliced and in-gel digested with trypsin as previously reported.35 Briefly, six bands from lung cancer and healthy samples, which showed different intensities, were excised and subjected to in-gel digestion. The gel was sliced into six bands and destained by incubation in 75 mM ammonium bicarbonate/40% ethanol (1 : 1). The gel pieces were then shake incubated in a 5 mM dithiothreitol solution with enough volume to cover the gel. For the alkylation of proteins, the gel pieces were incubated in 300 ml 55 mM IAA solution and then dehydrated in 700 ml of 100% acetonitrile. The gel pieces were dried in a speed vacuum and treated with trypsin (Roche Applied Science) overnight at 37 1C. Tryptic peptides were eluted with 0.1% formic acid. LC-ESI-MS/MS analysis and data analysis. LC-MS/MS analysis was performed using the Thermo Finnegan ProteomeX workstation and LTQ linear IT MS (Thermo Electron, San Jose, CA, USA) equipped with an NSI source (San Jose, CA). The analysis conditions for mass spectrometry were the same as previously reported.35 MS data analyzed as previously reported. Briefly MS/MS data were searched based on the IPI human protein database (version 3.29, 69 965 entries) using the SEQUEST algorithm sorcerer 3.4 beta 2 (Sorcerer software 3.10.4, sorcerer web interface 2.2.0 r334 and Trans proteomic Pipeline 2.9.5). SEQUEST search parameters were set to a fragment ion mass tolerance of 1.00 Da and a parent ion tolerance of 1.5 Da. Oxidation of methionine and iodoacetamide modification of cysteine were specified as fixed modifications. Scaffold calculates the quantitative value number by normalizing the spectral counts across our Scaffold experiment. Protein identifications were accepted if they could be established at greater than 95.0% probability as specified by the peptide prophet algorithm36 and contained at least two identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm.37 Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. The peptide false positive (FPR) rate was calculated using the Scaffold software. For each charge state, the incorrect assignments are tabulated to calculate the FPRi = [(#Assigned Incorrect at 95% probability)/ (Total# Incorrect Assigned)] 100, with i being the charge state. The assignment is considered correct if it is associated Mol. BioSyst. Deglycosylation with peptide-N-glycosidase F (PNGase F) One microgram of glycoprotein was denatured in buffer at 100 1C for 10 min and then cooled to RT. For complete deglycosylation, the protein was incubated in reaction buffer with 1 ml PNGase F (New England BioLabs, UK) for 1 h at 37 1C. The deglycosylation of proteins was confirmed by immunoblotting. Measurement of Hp in serum samples Western blot and analysis by densitometry. Western blots (WBs) were performed as previously reported.39 Briefly, 0.5 mg of crude serum protein was separated by 15% SDS-PAGE and then transferred. The membrane was incubated with an anti-Hp Ab at a 1 : 3000 dilution (polyclonal Anti-human Hp Ab, DaKoCytomation, Denmark) for 1 h at RT, followed by incubation with a horseradish peroxidase-conjugated anti-Rabbit IgG secondary Ab at 1 : 2000 (Stressgen, BC, Canada) for 30 min at RT. The ECL Western blot analysis system (Amersham Biosciences, UK) was used to detect the immunoreactive proteins. The densitometric analysis of the thickness of bands was performed with Scion Image (Scion, Frederick, MD). An image of a calibration grid slide was measured using Scion Image and a distance of 250 mm was calculated to be equivalent to 733 pixels. Enzyme-linked immunosorbent assay (ELISA). Quantification of the Hp b chain levels in serum and plasma samples was performed using in-house Sandwich ELISA. The ELISA starter kit (Koma Biotech Inc., Korea) was used, which included coating solution, washing solution, blocking solution and color-reaction solution. Three antibodies, anti-Hp b chain monoclonal Ab (Ab frontier, Seoul, Korea), anti-Hp polyclonal Ab (Ab frontier) and anti-rabbit secondary Ab (Stressgen), were used. The anti-Hp b chain monoclonal Ab was diluted in coating buffer to 1 mg ml 1, and then 100 ml of the prepared coating solution was dispensed into each well. The plate was incubated overnight at 4 1C. After blocking with 200 ml blocking solution for 1 h at RT, serum or plasma samples were diluted 200-fold in PBS, and then 100 ml was added to each well and the samples were incubated for 1 h at RT. Purified human Hp protein (SigmaAldrich, MO, USA) was used as a standard at a concentration range of 0 ng ml 1 to 600 ng ml 1. After blocking, 100 ml of anti-Hp polyclonal Ab (1 : 20 000, Ab frontier) was added to the 96-well plates, which were then incubated for 1 h. For detection, the plates were incubated with 100 ml of goat anti-rabbit Ab (1 : 5000, Stressgen) for 1 h at RT followed by the addition of PINK-ONE TMB solution (Koma Biotech Inc.). To determine the concentration of the Hp b chain, the absorbance at 450 nm of the plates were read using an ELISA reader (Shimadzu). Statistical analysis All data are presented as the means standard deviations (SD). Statistical differences among the densitometry and ELISA results of the groups were evaluated by two-sample This journal is c The Royal Society of Chemistry 2011 View Online t-test. A statistical significance of p o 0.05 is indicated by an asterisk (*). The area under the curve (AUC) value was determined using the Panel Composer Program, provided by Yonsei Proteome Research Center (Seoul, Korea). Downloaded by Kyungpook National University on 24 January 2011 Published on 21 January 2011 on http://pubs.rsc.org | doi:10.1039/C0MB00242A Acknowledgements This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MEST) (No. 2009-0093611), by grant FPR08A2-120 of the 21C Frontier Functional Proteomics Project from the Korean Ministry of Education, Science and Technology, and by grant No. RTI04-01-01 from the Regional Technology Innovation Program of the Ministry of Knowledge Economy (MKE). References 1 D. M. Parkin, F. Bray, J. Ferlay and P. Pisani, CA–Cancer J. Clin., 2005, 55, 74–108. 2 C. I. Henschke, D. I. McCauley, D. F. Yankelevitz, D. P. Naidich, G. McGuinness, O. S. Miettinen, D. M. Libby, M. W. Pasmantier, J. Koizumi, N. K. Altorki and J. P. Smith, Lancet, 1999, 354, 99–105. 3 D. P. Carbone, Semin. Oncol., 1997, 24, 388–401. 4 E. Brambilla, W. D. Travis, T. V. Colby, B. Corrin and Y. Shimosato, Eur. Respir. J., 2001, 18, 1059–1068. 5 W. S. Dalton and S. H. Friend, Science, 2006, 312, 1165–1168. 6 L. Beretta, Nat. Methods, 2007, 4, 785–786. 7 S. M. Hanash, S. J. Pitteri and V. M. Faca, Nature, 2008, 452, 571–579. 8 W. Chi-Shing Cho, Biomed. Pharmacother., 2007, 61, 515–519. 9 R. P. Valle, C. Chavany, T. A. Zhukov and M. Jendoubi, Expert Rev. Mol. Diagn., 2003, 3, 55–67. 10 A. K. Greenberg and M. S. Lee, Curr. Opin. Pulm. Med., 2007, 13, 249–255. 11 E. Gruys, M. J. Toussaint, T. A. Niewold and S. J. Koopmans, J. Zhejiang Univ., Sci., B, 2005, 6, 1045–1056. 12 M. Kristiansen, J. H. Graversen, C. Jacobsen, O. Sonne, H. J. Hoffman, S. K. Law and S. K. Moestrup, Nature, 2001, 409, 198–201. 13 B. Hamrita, K. Chahed, M. Trimeche, C. L. Guillier, P. Hammann, A. Chaieb, S. Korbi and L. Chouchane, Clin. Chim. Acta, 2009, 404, 111–118. 14 C. Zhao, L. Annamalai, C. Guo, N. Kothandaraman, S. C. Koh, H. Zhang, A. Biswas and M. Choolani, Neoplasia, 2007, 9, 1–7. 15 E. Miyoshi and M. Nakano, Proteomics, 2008, 8, 3257–3262. 16 M. A. Firpo, D. Z. Gay, S. R. Granger, C. L. Scaife, J. A. DiSario, K. M. Boucher and S. J. Mulvihill, World J. Surg., 2009, 33, 716–722. This journal is c The Royal Society of Chemistry 2011 17 R. Epelbaum, C. Shalitin, R. Segal, C. Valansi, I. Arselan, D. Faraggi, M. Leviov, M. Ben-Shahar and N. Haim, Pathol. Oncol. Res., 1998, 4, 271–276. 18 U. Dunzendorfer, K. Jung and G. Ohlenschlager, Eur. Urol., 1980, 6, 232–236. 19 H. G. Benkmann, H. P. Hanssen, R. Ovenbeck and H. W. Goedde, Hum. Hered., 1987, 37, 290–293. 20 S. Oshiro, Y. Yajima, K. Kawamura, M. Kubota, J. Yokofujita, Y. Nishibe, M. Takahama and H. Nakajima, Chem. Pharm. Bull., 1992, 40, 1847–1851. 21 J. Wassell, Clin. Lab., 2000, 46, 547–552. 22 M. C. Cid, D. S. Grant, G. S. Hoffman, R. Auerbach, A. S. Fauci and H. K. Kleinman, J. Clin. Invest., 1993, 91, 977–985. 23 J. M. Gutteridge, Biochim. Biophys. Acta, 1987, 917, 219–223. 24 D. P. de Kleijn, M. B. Smeets, P. P. Kemmeren, S. K. Lim, B. J. Van Middelaar, E. Velema, A. Schoneveld, G. Pasterkamp and C. Borst, FASEB J., 2002, 16, 1123–1125. 25 M. B. Smeets, J. Fontijn, A. Kavelaars, G. Pasterkamp and D. P. De Kleijn, Int. J. Exp. Pathol., 2003, 84, 69–74. 26 Y. Kaneta, Y. Kagami, T. Tsunoda, R. Ohno, Y. Nakamura and T. Katagiri, Int. J. Oncol., 2003, 23, 681–691. 27 A. Nishie, K. Masuda, M. Otsubo, T. Migita, M. Tsuneyoshi, K. Kohno, T. Shuin, S. Naito, M. Ono and M. Kuwano, Clin. Cancer Res., 2001, 7, 2145–2151. 28 B. Ye, D. W. Cramer, S. J. Skates, S. P. Gygi, V. Pratomo, L. Fu, N. K. Horick, L. J. Licklider, J. O. Schorge, R. S. Berkowitz and S. C. Mok, Clin. Cancer Res., 2003, 9, 2904–2911. 29 A. Bharti, P. C. Ma, G. Maulik, R. Singh, E. Khan, A. T. Skarin and R. Salgia, Anticancer Res., 2004, 24, 1031–1038. 30 N. Okuyama, Y. Ide, M. Nakano, T. Nakagawa, K. Yamanaka, K. Moriwaki, K. Murata, H. Ohigashi, S. Yokoyama, H. Eguchi, O. Ishikawa, T. Ito, M. Kato, A. Kasahara, S. Kawano, J. Gu, N. Taniguchi and E. Miyoshi, Int. J. Cancer, 2006, 118, 2803–2808. 31 M. Kasprzyk, W. Dyszkiewicz, D. Zwarun, K. Lesniewska and K. Wiktorowicz, Pneumonol. Alergol. Pol., 2008, 76, 321–326. 32 L. F. Hoagland 4th, M. J. Campa, E. B. Gottlin, J. E. Herndon, 2nd and E. F. Patz, Jr., Cancer, 2007, 110, 2260–2268. 33 S. Y. Park, S. J. Yoon, Y. T. Jeong, J. M. Kim, J. Y. Kim, B. Bernert, T. Ullman, S. H. Itzkowitz, J. H. Kim and S. I. Hakomori, Int. J. Cancer, 2010, 126, 142–155. 34 E. Miyoshi, K. Moriwaki and T. Nakagawa, J. Biochem. (Tokyo), 2008, 143, 725–729. 35 S. H. Heo, S. J. Lee, H. M. Ryoo, J. Y. Park and J. Y. Cho, Proteomics, 2007, 7, 4292–4302. 36 A. Keller, A. I. Nesvizhskii, E. Kolker and R. Aebersold, Anal. Chem., 2002, 74, 5383–5392. 37 A. I. Nesvizhskii, A. Keller, E. Kolker and R. Aebersold, Anal. Chem., 2003, 75, 4646–4658. 38 E. L. Hendrickson, Q. Xia, T. Wang, J. A. Leigh and M. Hackett, Analyst, 2006, 131, 1335–1341. 39 B. G. Kim, H. J. Kim, H. J. Park, Y. J. Kim, W. J. Yoon, S. J. Lee, H. M. Ryoo and J. Y. Cho, Proteomics, 2006, 6, 1166–1174. Mol. BioSyst.