Promising Diagnostic Biomarkers for Primary Biliary Cirrhosis Identified With Magnetic Beads and MALDI-TOF-MS.код для вставкиСкачать
THE ANATOMICAL RECORD 292:455–460 (2009) Promising Diagnostic Biomarkers for Primary Biliary Cirrhosis Identiﬁed With Magnetic Beads and MALDI-TOF-MS YONG-ZHE LI,1* CHAO-JUN HU,1 XIAO-MEI LENG,1 GUAN-FEI ZHAO,1 NING LI,2 AND YANG XU2* 1 Department of Rheumatology, Peking Union Medical College Hospital, Peking Union Medical College & Chinese Academy of Medical Sciences, Beijing, China 2 Center for Clinical Laboratory Development, Peking Union Medical College & Chinese Academy of Medical Sciences, Beijing, China ABSTRACT (PBC) is not a rare disease worldwide. Most patients are diagnosed at the advanced stage, primarily because there are not yet any valid biomarkers available for early diagnosis. Useful biomarkers are absolutely necessary for early detection of PBC. Fortunately, the use of MALDITOF-MS and pattern recognition software has been successful in ﬁnding speciﬁc markers for the early detection of the disease. To screen for potential protein biomarkers in the serum for diagnosing PBC, MALDITOF-MS combined with magnetic beads and pattern recognition software was used to investigate 119 serum samples from 44 patients with PBC, 32 controls with other hepatic disease, and 43 healthy controls. A total of 69 discriminant m/z peaks were identiﬁed as being associated with PBC. Of them, the m/z peaks at 3445, 4260, 8133, and 16,290 were used to construct a model for the diagnosis of PBC. This diagnostic model can distinguish PBC from non-PBC controls with a sensitivity of 93.3% and a speciﬁcity of 95.1%. In our blind test, it demonstrated good sensitivity and speciﬁcity: 92.9% and 82.4%, respectively. These results indicate that useful serum biomarkers for PBC can be discovered by MALDI-TOF-MS combined with the use of magnetic beads and pattern recognition software. The pattern of multiple markers provides a powerful and reliable diagnostic method for PBC with high sensitivity and speciﬁcity. Anat Rec, 292:455–460, 2009. Ó 2009 Wiley-Liss, Inc. Key words: primary biliary cirrhosis; biomarker; magnetic beads; MALDI-TOF-MS Primary biliary cirrhosis (PBC) is a chronic autoimmune cholestatic liver disease characterized by progressive, nonsuppurative inﬂammation, and the destruction of small and medium-sized bile ducts (Locke et al., 1996; Kaplan and Gershwin, 2005). Recently, studies have suggested that PBC is not a rare disease worldwide (Metcalf and James, 1997), especially asymptomatic PBC, which is probably more prevalent in the Asian population than Grant sponsor: National Natural Science Foundation of China (NSFC); Grant numbers: 30640084, 30471617. Yong-zhe Li and Chao-jun Hu contributed equally to this work. Development, Peking Union Medical College & Chinese Academy of Medical Sciences, Beijing 102206, China and Center for Organelle Proteomics of Diseases, Zhejiang University School of Basic Medical Science, Hangzhou, China. Fax: 186-1067016735. E–mail: email@example.com *Correspondence to: Yong-zhe Li, Department of Rheumatology, Peking Union Medical College Hospital, Peking Union Medical College & Chinese Academy of Medical Sciences, Beijing 100730, China. Fax: 186-10-88068795. E-mail: firstname.lastname@example.org or Yang Xu, Center for Clinical Laboratory DOI 10.1002/ar.20870 Published online in Wiley InterScience (www.interscience.wiley. com). Ó 2009 WILEY-LISS, INC. Received 14 April 2008; Accepted 25 November 2008 456 LI ET AL. thought previously. Research has shown that PBC is a signiﬁcant cause of morbidity and mortality in Chinese patients (Wong et al., 2005). Currently, the most reliable procedure for diagnosing PBC is histopathology of a liver biopsy (Heathcote, 2000). Unfortunately, the characteristic histopathological changes that are associated with PBC cannot be observed in all tissue samples, and the patients are generally reluctant to accept this invasive procedure (Zeniya et al., 2005). Therefore, high concentrations of anti-mitochondrial antibodies (AMA) in the serum have served as a hallmark for diagnosing PBC (Leung et al., 1997; Neuberger, 1997). AMA is neither a characteristic nor a speciﬁc marker for PBC, however, and can be frequently detected in patients with other medical conditions, such as infectious liver disease (Toda et al., 1997; Miyakawa et al., 2006). In general, over 10% of patients with PBC have an undetectable level of serum AMA (Kaplan, 1996; Sakauchi et al., 2006). Frequently, patients with PBC are in the terminal stages of their disease when they are diagnosed. A new method for diagnosing the early stage of PBC is still an unmet need in clinical practice. Proteomics has been shown to be a promising method for the early detection of cancer, neuropathic disease, infectious disease, and rheumatic diseases (Shiwa et al., 2003; Dotzlaw et al., 2004; De Seny et al., 2005; Kang et al., 2005; Agranoff et al., 2006). In this study, we use proteomic approaches to identify relevant biomarkers that could replace invasive and nonspeciﬁc tests for the early diagnosis of PBC. The goal of this study was to screen for potential protein biomarkers in serum for the early diagnosis of PBC using MALDI-TOF-MS combined with magnetic beads and pattern recognition software. PATIENTS AND METHODS Patients One hundred nineteen serum samples were collected from 44 patients with PBC, 32 from patients with liver diseases other than PBC (10 were autoimmune hepatitis patients; 10 were hepatitis B patients; nine were hepatocirrhosis patients; one was an alcoholic hepatitis patient; one was a drug-induced liver disorder patient; and one was a primary sclerotic cholangitis patient), and 43 were healthy volunteers. Blood samples were obtained with informed consent from all subjects. The protocol was approved by the review board of Peking Union Medical College Hospital. The 44 patients with PBC were clinically diagnosed according to generally accepted criteria (Heathcote, 2000). The demographic data of the PBC patient group are shown in Table 1. The blood samples were collected in 4 mL BD Vacutainers without anticoagulation and allowed to clot at room temperature for up to 1 hr; the samples were then centrifuged at 48C for 5 min at 1000 rpm (1000g). The sera were frozen and stored at 2808C for future analysis. Serum Pretreatment and Binding to Magnetic Beads Serum samples were pretreated with WCX magnetic beads (SEDTM) (Beijing SED Science and Technology, Inc.). Brieﬂy, 10 mL of each serum sample was mixed with 20 mL of U9 solution (9 mol/L urea, 2% CHAPS) in a 0.5 mL centrifuge-tube. After incubating for 30 min at TABLE 1. Demographics of the PBC patients and control groups Sample type Normal PBC Hepatic disease Number of samples Male/ female Age range Mean age 43 44 32 27/16 3/41 17/15 25–78 16–74 18–75 38.6 610.9 54.1 6 11.3 50.6 613.1 48C, the samples were diluted 40-fold with the addition of 370 mL of binding buffer (50 mmol/L sodium acetate, pH 4.0). Then, 50 mL of WCX magnetic beads were added to a PCR tube, and the tubes were placed in a magnet separator for 1 min. The supernatant was removed carefully using a pipette. The magnetic beads were then washed twice with 100 mL binding buffer. Diluted serum sample in a volume of 100 mL was added and mixed with the activated magnetic beads carefully by pipetting up and down several times; the samples were incubated for 1 hr at 48C and then washed twice with 100 mL binding buffer. Following binding and washing, the bound proteins were eluted from the magnetic beads using 10 mL of 0.5% (vol/vol) triﬂuoroacetic acid (TFA). Then, 5 mL of the eluted sample was diluted in the ratio of 1:2 with 5 mL of SPA (50% CAN 1 0.5% TFA). One microliter of the resulting mixture was aspirated and spotted onto an 8-spot pre-structured sample chip (Au-chip, Ciphergen). After air-drying for 5 min at room temperature, protein crystals on the chip were detected by MALDI-TOF-MS (Ciphergen, PBS IIc). Statistical Analysis Data were collected by averaging 80 laser shots per spot with an intensity setting of 205 and a detector sensitivity setting of 8. The spectra from all samples were initially analyzed using Biomarker Wizard Version 3.1 with the following processes: the 4091 mass peak in serum was used to normalize dimensions; the baseline was subtracted; the peaks were automatically detected and clustered (First 5, Min Peak 10%, Cluster Mass 0.3%, Second Pass 2); statistics on the sample group was performed. Model Construction The data were analyzed and a model was constructed using Biomarker Patterns Software (BPS) Version 5.0. The ﬁnal model was setup as: Relative cost: 0.362; Method 5 0; Advanced 5 10, 1; Testing 5 21; and Costs 5 1:1. Data Score The examiner was blind to the data used to test the diagnostic model. Forty-eight samples were tested in total, including 14 patients with PBC, 14 patients with other hepatic diseases, and 20 healthy controls. RESULTS Optimization of the Experimental Conditions and Evaluation of the Reproducibility Reproducibility was evaluated with four mixed serum samples from the healthy controls of blood type O (two women and two men). The mixed serum samples were 457 PROMISING DIAGNOSTIC BIOMARKERS FOR PBC Fig. 1. An 8-spot reproducibility test showed good reproducibility. The CV of all the selected mass peaks was below 10%. TABLE 2. The 69 discriminating m/z peaks among PBC, hepatic disease controls, and normal controls m/z 3088 2543 8133a 7628 8600 5477 31842 3935 2759 4281 4300 7967 3955 22798 a P m/z P m/z P m/z P m/z P 2.70E–09 1.3E–08 2.2E–07 2.7E–07 3E–07 1.6E-05 2.1E–05 4.2E–05 4.7E–05 7.1E–05 7.9E–05 0.0001 0.0002 0.0002 15874 11485 4068 8071 15608 5803 11697 6846 16290a 16489 3375 9280 2869 16077 0.0002 0.0002 0.0002 0.0003 0.0003 0.0004 0.0006 0.0007 0.0008 0.0009 0.001 0.001 0.0013 0.0018 5855 3774 8290 4382 28026 4645 9487 23386 5246 14016 5634 9794 14103 7405 0.0023 0.0024 0.0024 0.0027 0.0027 0.0027 0.0027 0.0033 0.0037 0.0042 0.0051 0.0056 0.0056 0.0066 5315 43917 46614 8554 3681 28790 7761 3445a 25099 4528 43219 24015 4676 7049 0.0075 0.0101 0.0111 0.0114 0.0122 0.0124 0.0125 0.0135 0.0151 0.0189 0.0194 0.0235 0.0236 0.0273 3397 3887 9606 32550 5214 5384 7554 4260a 47633 1498 9399 9116 5525 0.03 0.03 0.0333 0.0351 0.0383 0.0414 0.0419 0.0434 0.0441 0.0452 0.463 0.474 0.0487 Selected to construct a model for diagnosis of PBC. spotted on an eight spot Au-chip. The CV was under 10% for all the selected mass peaks (Fig. 1). Protein Fingerprint Analysis of Serum Samples in Patients with PBC, Other Hepatic Disease Controls, and Healthy Controls The protein proﬁle of the serum samples from the 44 patients with PBC, the 32 patients with other hepatic diseases, and the 43 healthy controls were extracted by magnetic beads and examined by MALDI-TOF-MS. The data were analyzed by Biomarker Wizard Version 3.1; 69 m/z peaks were found to discriminate the patients with PBC, patients with other hepatic diseases, and normal controls (Table 2). Identiﬁcation of Biomarker Pattern and Construction of Diagnostic Model The 69 m/z peaks that were distinct among patients with PBC, patients with other hepatic diseases, and healthy controls were analyzed by BPS Version 5.0 to 458 LI ET AL. Fig. 2. The representative m/z peaks at 3445 (a), 4260 (b), 8133 (c), and 16290 (d) in the PBC, hepatic disease control (HD), and normal control (Nor) samples. TABLE 3. Serum sample used in the training and test groups Immunology Normal PBC Hepatic disease control Fig. 3. TABLE 4. Classiﬁcation of the data in the diagnostic model for PBC Training group Test group Total Group Cases N N correct-classed Pct accurate 23 30 18 20 14 14 43 44 32 PBC Control 30 41 28 39 93.33 95.12 The decision trees of diagnostic model for PBC. identify biomarker patterns. The m/z peaks at 3445, 4260, 8133, and 16290 Da of 71 objects (Fig. 2, Table 3) were selected as the best markers to construct a diagnostic model for PBC. The decision tree is shown in Fig. 3, and the characteristics of the diagnostic model are shown in Table 4 and Fig. 4. Fig. 4. ROC curve of the diagnostic model for class PBC. Note: The data used to construct the diagnostic model included 71 samples: 30 PBC, 18 hepatic disease controls, and 23 healthy controls. Test of the Diagnostic Model for PBC in a Blind Test Forty-eight samples, including 14 from patients with PBC, 14 from patients with other hepatic diseases, and 459 PROMISING DIAGNOSTIC BIOMARKERS FOR PBC 20 healthy controls (Table 3) were used to test the PBC diagnostic model. The results are shown in Table 5. Discrimination of m/z Peaks Between Patients with PBC with Positive AMA-M2 and Patients with PBC with Negative AMA-M2 The BPS identiﬁed ﬁve m/z peaks that were able to discriminate AMA-M2 positive and AMA-M2 negative patients with PBC. The results are shown in Fig. 5 and Table 6. DISCUSSION Cholestatic disorders are not uncommon and should be differentiated from other hepatic diseases (Kim et al., 2000). There are currently no diagnostic tests available that can meet this goal, however. At present, liver histological presentations are used for diagnosis. Biochemical and serological tests can only identify 50%–60% of PBC cases in the early stage of the disease course. The early diagnosis of PBC remains a technical challenge (He et al., 2006). In recent years, as proteomics research has developed, more and more proteomic technologies have emerged and have begun to be used in clinical research (Baumeister, 2005; Sinz et al., 2005; Chung et al., 2007). The search for new biomarkers for complicated diseases has become more and more successful with the use of new high throughput proteomic techniques (Choi et al., 2002; Hayman and Przyborski, 2004). Biomarkers, especially patterns of multiple biomarkers, are more reliable and powerful for diagnosis, differential diagnosis, and treatment guidance than the current techniques. Serum pro- TABLE 5. The predictions for PBC made by the diagnosis model in a blind test Group Cases N N correct-classed Pct accurate PBC Control 14 34 13 28 92.86 82.35 teome analysis by MALDI-TOF-MS has provided new information about peptides and proteins that have the potential to be surrogate markers. As magnetic beads can provide a large absorbable surface, they have much more potential to adsorb peptides and proteins from serum than the ProteinChip surface alone (Villanueva et al., 2004; Whiteaker et al., 2007). The combination of magnetic beads and MALDI-TOF-MS can allow us to enrich and discover more proteins that have low abundance in sera. In our study, the protein ﬁngerprints in sera of 44 subjects with PBC and 75 controls were analyzed with WCX magnetic beads and MALDI-TOF-MS. Sixty-nine potential protein biomarkers were found, and a diagnostic model for PBC was constructed by pattern recognition software using the m/z peaks of protein biomarkers 3445, 4260, 8133, and 16290. In a blind test on 48 samples, the diagnostic model had a sensitivity of 92.9% and a speciﬁcity of 82.4%. This result shows that the detection of serum proteins by WCX magnetic beads, MALDI-TOF-MS, and pattern recognition software can be a powerful tool to distinguish PBC from other hepatic diseases and healthy controls. The identiﬁcation of novel biomarkers for PBC may increase our understanding of the progression and pathophysiology of the disease on a molecular level. Because PBC is a multi-factorial liver disease, using a combination of multiple markers should be a more reliable and powerful way to diagnose the disease with higher speciﬁcity and sensitivity than using a single TABLE 6. The ﬁve m/z peaks that discriminate between PBC patients with positive AMA-M2 and PBC patients with negative AMA-M2 m/z P Mean-M21 Mean-M22 4068 3682 3935 4472 8133 0.008 0.026 0.026 0.026 0.028 0.94 1.18 1.46 6.97 5.86 1.49 1.66 2.04 9.03 7.78 Fig. 5. The representative m/z peaks at 3682, 3935, 4068, 4472, and 8133 in patients with PBC with positive AMA-M2, and patients with PBC with negative AMA-M2. 460 LI ET AL. marker alone. In our study, the selection of 3445, 4260, 8133, and 16290 m/z peaks by BPS as a biomarker pattern for PBC could separate patients with PBC from controls with a high conﬁdence level. The selection of multi-markers was achieved by the BPS (Wiesner, 2004). One limitation of our study is that only 30 samples from patients with PBC were included in the building of the diagnostic model, and only 14 PBC sera samples were used for the blind test. Therefore, it is still an open question whether these potential biomarkers are really speciﬁc to PBC. A larger sample with more differential diagnosis, such as other rheumatisms, inﬂammatory diseases, and other organ speciﬁc diseases is warranted for further investigation. In conclusion, using a small number of PBC samples, we have shown that MALDI-TOF-MS combined with magnetic beads and pattern recognition software can distinguish PBC from other hepatic diseases and healthy controls. In particular, we have found biomarker patterns at 3445, 4260, 8133, and 16290 m/z peaks and have constructed a diagnostic model for PBC that has good sensitivity and speciﬁcity. LITERATURE CITED Agranoff D, Fernandez-Reyes D, Papadopoulos MC, Rojas SA, Herbster M, Loosemore A, Tarelli E, Sheldon J, Schwenk A, Pollok R, Rayner CF, Krishna S. 2006. Identiﬁcation of diagnostic markers for tuberculosis by proteomic ﬁngerprinting of serum. Lancet 368:1012–1021. Baumeister W. 2005. From proteomic inventory to architecture. FEBS Lett 579:933–937. Choi J, Malakowsky CA, Talent JM, Conrad CC, Gracy RW. 2002. Identiﬁcation of oxidized plasma proteins in Alzheimer’s disease. Biochem Biophys Res Commun 293:1566–1570. Chung CH, Levy S, Chaurand P, Carbone DP. 2007. Genomics and proteomics: emerging technologies in clinical cancer research. Crit Rev Oncol Hematol 61:1–25. De Seny D, Fillet M, Meuwis MA, Geurts P, Lutteri L, Ribbens C, Bours V, Wehenkel L, Piette J, Malaise M, Merville MP. 2005. Discovery of new rheumatoid arthritis biomarkers using the surface-enhanced laser desorption/ionization time-of-ﬂight mass spectrometry ProteinChip approach. Arthritis Rheum 52:3801–3812. Dotzlaw H, Schulz M, Eggert M, Neeck G. 2004. A pattern of protein expression in peripheral blood mononuclear cells distinguishes rheumatoid arthritis patients from healthy individuals. Biochim Biophys Acta 1696:121–129. Hayman MW, Przyborski SA. 2004. Proteomic identiﬁcation of biomarkers expressed by human pluripotent stem cells. Biochem Biophys Res Commun 316:918–923. He XS, Ansari AA, Ridgway WM, Coppel RL, Gershwin ME. 2006. New insights to the immunopathology and autoimmune responses in primary biliary cirrhosis. Cell Immunol 239:1–13. Heathcote EJ. 2000. Management of primary biliary cirrhosis, management of primary biliary cirrhosis. The American Association for the study of liver diseases practice guidelines. Hepatology 31:1005–1013. Kang X, Xu Y, Wu X, Liang Y, Wang C, Guo J, Wang Y, Chen M, Wu D, Wang Y, Bi S, Qiu Y, Lu P, Cheng J, Xiao B, Hu L, Gao X, Liu J, Wang Y, Song Y, Zhang L, Suo F, Chen T, Huang Z, Zhao Y, Lu H, Pan C, Tang H. 2005. Proteomic ﬁngerprints for potential application to early diagnosis of severe acute respiratory syndrome. Clin Chem 51:56–64. Kaplan MM. 1996. Primary biliary cirrhosis. N Engl J Med 335:1570–1580. Kaplan MM, Gershwin ME. 2005. Primary biliary cirrhosis. N Engl J Med 353:1261–1273. Kim WR, Ludwig J, Lindor KD. 2000. Variant forms of cholestatic diseases involving small bile ducts in adults. Am J Gastroenterol 95:1130–1138. Leung PS, Coppel RL, Ansari A, Munoz S, Gershwin ME. 1997. Antimitochondrial antibodies in primary biliary cirrhosis. Semin Liver Dis 17:61–69. Locke GR, III, Therneau TM, Ludwig J, Dickson ER, Lindor KD. 1996. Time course of histological progression in primary biliary cirrhosis. Hepatology 23:52–56. Metcalf J, James O. 1997. The geoepidemiology of primary biliary cirrhosis. Semin Liver Dis 17:13–22. Miyakawa H, Tanaka A, Selmi C, Hosoya N, Mataki N, Kikuchi K, Kato T, Arai J, Goto T, Gershwin ME. 2006. Serum reactivity against bacterial pyruvate dehydrogenase: increasing the speciﬁcity of anti-mitochondrial antibodies for the diagnosis of primary biliary cirrhosis. Clin Dev Immunol 13:289–294. Neuberger J. 1997. Primary biliary cirrhosis. Lancet 350:875–879. Sakauchi F, Mori M, Zeniya M, Toda G. 2006. Antimitochondrial antibody negative primary biliary cirrhosis in Japan: utilization of clinical data when patients applied to receive public ﬁnancial aid. J Epidemiol 16:30–34. Shiwa M, Nishimura Y, Wakatabe R, Fukawa A, Arikuni H, Ota H, Kato Y, Yamori T. 2003. Rapid discovery and identiﬁcation of a tissue-speciﬁc tumor biomarker from 39 human cancer cell lines using the SELDI ProteinChip platform. Biochem Biophys Res Commun 309:18–25. Sinz A, Kalkhof S, Ihling C. 2005. Mapping protein interfaces by a trifunctional cross-linker combined with MALDI-TOF and ESIFTICR mass spectrometry. J Am Soc Mass Spectrom 16:1921– 1931. Toda G, Zeniya M, Watanabe F, Imawari M, Kiyosawa K, Nishioka M, Tsuji T, Omata M. 1997. Present status of autoimmune hepatitis in Japan—correlating the characteristics with international criteria in an area with a high rate of HCV infection. Japanese National Study Group of Autoimmune Hepatitis. J Hepatol 26:1207–1212. Villanueva J, Philip J, Entenberg D, Chaparro CA, Tanwar MK, Holland EC, Tempst P. 2004. Serum peptide proﬁling by magnetic particle-assisted, automated sample processing and MALDI-TOF mass spectrometry. Anal Chem 76:1560–1570. Whiteaker JR, Zhao L, Zhang HY, Feng LC, Piening BD, Anderson L, Paulovich AG. 2007. Antibody-based enrichment of peptides on magnetic beads for mass-spectrometry-based quantiﬁcation of serum biomarkers. Anal Biochem 362:44–54. Wiesner A. 2004. Detection of tumor markers with ProteinChip technology. Curr Pharm Biotechnol 5:45–67. Wong GL, Hui AY, Wong VW, Chan FK, Sung JJ, Chan HL. 2005. A retrospective study on clinical features and prognostic factors of biopsy-proven primary biliary cirrhosis in Chinese patients. Am J Gastroenterol 100:2205–2211. Zeniya M, Watanabe F, Morizane T, Shibata M, Maeyama S, Kage M, Nakanuma Y, Toda G. 2005. Diagnosing clinical subsets of autoimmune liver diseases based on a multivariable model. J Gastroenterol 40:1148–1154.