Rulin Zhang 1 Lisa Barker 1 Deborah Pinchev 2 John Marshall 1 Michèle Rasamoelisolo 1 Chris Smith 1 Peter Kupchak 1 Inga Kireeva 1 Leslee Ingratta 2 George Jackowski 1, 3 1 Syn-X Pharma, Toronto, Ontario, Canada 2 Department of Biochemistry, University of Guelph, Guelph, Ontario, Canada 3 Department of Laboratory Medicine and Pathology, University of Toronto, Toronto, Ontario, Canada Mining biomarkers in human sera using proteomic tools One of the major difficulties in mining low abundance biomarkers from serum or plasma is due to the fact that a small number of proteins such as albumin, a2-macro- globulin, transferrin, and immunoglobulins, may represent as much as 80% of the total serum protein. The large quantity of these proteins makes it difficult to identify low abundance proteins in serum using traditional 2-dimensional electrophoresis. We recently used a combination of multidimensional liquid chromatography and gel elec- trophoresis coupled to matrix-assisted laser desorption/ionization-quadrupole-time of flight and Ion Trap liquid chromatography-tandem mass spectrometry to identify pro- tein markers in sera of Alzheimer’s disease (AD), insulin resistance/type-2 diabetes (IR/ D2), and congestive heart failure (CHF) patients. We identified 8 proteins that exhibit higher levels in control sera and 36 proteins that exhibit higher levels in disease sera. For example, haptoglobin and hemoglobin are elevated in sera of AD, IR/D2, and CHF patients. The levels of several other proteins including fibrinogen and its fragments, alpha 2-macroglobulin, transthyretin, pro-platelet basic protein, protease inhibitors clade A and C, as well as proteins involved in the classical complement pathway such as complement C3, C4, and C1 inhibitor, were found to differ between IR/D2 and con- trol sera. The sera levels of proteins, such as the 10 kDa subunit of vitronectin, alpha 1-acid glycoprotein, apolipoprotein B100, fragment of factor H, and histidine-rich glyco- protein were observed to be different between AD and controls. The differences observed in these biomarker candidates were confirmed by Western blot and the en- zyme-linked immunosorbent assay. The biological meaning of the proteomic changes in the disease states and the potential use of these changes as diagnostic tools or for therapeutic intervention will be discussed. Keywords: Biomarkers / Body fluids / Diagnostic / Mass Spectrometry PRO 0495 1 Introduction Proteomics has recently emerged as a new field of protein science. This field is dominated by separation technolo- gies including traditional 2-DE and the alternative multidi- mensional liquid chromatography (MDLC) to separate proteins, followed by the analysis of these proteins by MS [1–3]. These technologies, in combination with bioin- formatics, are powerful tools for protein identification and characterization [4–8]. These tools have been used to cre- ate and compare protein profiles of normal and diseased body fluids such as serum [8–10], cerebrospinal fluid [11– 14], and urine [15–17]. Since blood has direct contact with most of the tissues in the human body, pathological changes are likely to be reflected in proteomic changes in serum. Thus, biomarkers discovered in serum may form the basis of noninvasive and simple diagnostic tests or be candidates for therapeutic intervention. However, it is difficult to identify low abundance proteins in serum [18]. One of the major difficulties in mining low abundance biomarkers from serum or plasma is due to the fact that a small number of proteins, including albumin, a2-macro- globulin, transferrin, and immunoglobulins, may represent about 80% of total serum protein. These proteins provide a significant background and make it difficult to identify low abundance proteins in serum using the traditional 2-DE approach. We recently used a combination of MDLC and gel electrophoresis coupled to MS to identify protein markers in sera of Alzheimer’s Disease (AD), insulin resis- tance/type-2 diabetes (IR/D2), and congestive heart fail- ure (CHF) patients. Sera samples were fractionated using one or two columns and further separated on 1-D or 2-D gels. Bands or spots that were consistently different be- tween disease samples and controls were isolated, in-gel trypsin digested, and identified by MS analysis and data- Correspondence: Dr. Rulin Zhang, Director of Proteomics, Syn- X Pharma, 1 Marmac Drive, Toronto, Ontario, Canada M9W 1E7 E-mail: rzhang@synxpharma.com Fax: +1-416-798-3447 Abbreviations: AD, Alzheimer’s Disease; CHF , congestive heart failure; D2, type-2 diabetes; Hb, hemoglobin; IR, insulin resis- tance; MDLC, multidimensional liquid chromatography; ROC, receiver operating characteristic 244 Proteomics 2004, 4, 244–256  2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.de DOI 10.1002/pmic.200300495