Urinary Proteomic Biomarkers in Coronary Artery Disease

Urinary proteomics is emerging as a powerful non-invasive tool for diagnosis and monitoring of variety of human diseases. We tested whether signatures of urinary polypeptides can contribute to the existing biomarkers for coronary artery disease (CAD). We examined a total of 359 urine samples from 88 patients with severe CAD and 282 controls. Spot urine was analyzed using capillary electrophoresis on-line coupled to ESI-TOF-MS enabling characterization of more than 1000 polypeptides per sample. In a first step a “training set” for biomarker definition was created. Multiple biomarker patterns clearly distinguished healthy controls from CAD patients, and we extracted 15 peptides that define a characteristic CAD signature panel. In a second step, the ability of the CAD-specific panel to predict the presence of CAD was evaluated in a blinded study using a “test set.” The signature panel showed sensitivity of 98% (95% confidence interval, 88.7–99.6) and 83% specificity (95% confidence interval, 51.6–97.4). Furthermore the peptide pattern significantly changed toward the healthy signature correlating with the level of physical activity after therapeutic intervention. Our results show that urinary proteomics can identify CAD patients with high confidence and might also play a role in monitoring the effects of therapeutic interventions. The workflow is amenable to clinical routine testing suggesting that non-invasive proteomics analysis can become a valuable addition to other biomarkers used in cardiovascular risk assessment. Urinary proteomics is emerging as a powerful non-invasive tool for diagnosis and monitoring of variety of human diseases. We tested whether signatures of urinary polypeptides can contribute to the existing biomarkers for coronary artery disease (CAD). We examined a total of 359 urine samples from 88 patients with severe CAD and 282 controls. Spot urine was analyzed using capillary electrophoresis on-line coupled to ESI-TOF-MS enabling characterization of more than 1000 polypeptides per sample. In a first step a “training set” for biomarker definition was created. Multiple biomarker patterns clearly distinguished healthy controls from CAD patients, and we extracted 15 peptides that define a characteristic CAD signature panel. In a second step, the ability of the CAD-specific panel to predict the presence of CAD was evaluated in a blinded study using a “test set.” The signature panel showed sensitivity of 98% (95% confidence interval, 88.7–99.6) and 83% specificity (95% confidence interval, 51.6–97.4). Furthermore the peptide pattern significantly changed toward the healthy signature correlating with the level of physical activity after therapeutic intervention. Our results show that urinary proteomics can identify CAD patients with high confidence and might also play a role in monitoring the effects of therapeutic interventions. The workflow is amenable to clinical routine testing suggesting that non-invasive proteomics analysis can become a valuable addition to other biomarkers used in cardiovascular risk assessment. Coronary artery disease (CAD) 1The abbreviations used are: CAD, coronary artery disease; CE, capillary electrophoresis; ETD, electron transfer dissociation; ROC, receiver operating characteristic; AUC, area under the ROC curve; CI, confidence interval; DTA, .dta file; LDL, low density lipoprotein; HDL, high density lipoprotein. 1The abbreviations used are: CAD, coronary artery disease; CE, capillary electrophoresis; ETD, electron transfer dissociation; ROC, receiver operating characteristic; AUC, area under the ROC curve; CI, confidence interval; DTA, .dta file; LDL, low density lipoprotein; HDL, high density lipoprotein. is a leading cause of morbidity and mortality worldwide. The underlying molecular causes are still largely unknown but are likely to involve alterations in gene and protein expression (1McGregor E. Dunn M.J. Proteomics of the heart: unraveling disease.Circ. Res. 2006; 98: 309-321Crossref PubMed Scopus (119) Google Scholar). Despite multiple clinical, electrographic, and biochemical characteristics, there are subgroups of patients who progress to severe, life-threatening CAD without many symptoms and signs (2Fazzini P.F. Prati P.L. Rovelli F. Antoniucci D. Menghini F. Seccareccia F. Menotti A. Epidemiology of silent myocardial ischemia in asymptomatic middle-aged men (the Eccis Project).Am. J. Cardiol. 1993; 72: 1383-1388Abstract Full Text PDF PubMed Scopus (61) Google Scholar). For example, patients with type 2 diabetes and the elderly frequently suffer from silent myocardial infarctions with significantly increased risk of complications (3Scognamiglio R. Negut C. Ramondo A. Tiengo A. Avogaro A. Detection of coronary artery disease in asymptomatic patients with type 2 diabetes mellitus.J. Am. Coll. Cardiol. 2006; 47: 65-71Crossref PubMed Scopus (189) Google Scholar). Early diagnosis of CAD in its presymptomatic stage would allow for better, targeted, and hence more effective primary prevention as compared with current clinical recommendations. Proteomics is increasingly used to examine dynamic changes in protein expression providing new insights into cellular processes. Moreover proteomics analyses have already resulted in the identification of clinically useful biomarkers and can assist in diagnosis and disease staging (1McGregor E. Dunn M.J. Proteomics of the heart: unraveling disease.Circ. Res. 2006; 98: 309-321Crossref PubMed Scopus (119) Google Scholar, 4Hanash S. Disease proteomics.Nature. 2003; 422: 226-232Crossref PubMed Scopus (861) Google Scholar, 5Kolch W. Neususs C. Pelzing M. Mischak H. Capillary electrophoresis-mass spectrometry as a powerful tool in clinical diagnosis and biomarker discovery.Mass Spectrom. Rev. 2005; 24: 959-977Crossref PubMed Scopus (256) Google Scholar). Substances contained in body fluids hold an abundance of information and can be used as a dynamic and concurrent gauge for monitoring the well-being of an organism. Urine presents a rich source of information related to the functioning of many internal organs (5Kolch W. Neususs C. Pelzing M. Mischak H. Capillary electrophoresis-mass spectrometry as a powerful tool in clinical diagnosis and biomarker discovery.Mass Spectrom. Rev. 2005; 24: 959-977Crossref PubMed Scopus (256) Google Scholar, 6Hewitt S.M. Dear J. Star R.A. Discovery of protein biomarkers for renal diseases.J. Am. Soc. Nephrol. 2004; 15: 1677-1689Crossref PubMed Scopus (240) Google Scholar, 7O'Riordan E. Goligorsky M.S. Emerging studies of the urinary proteome: the end of the beginning?.Curr. Opin. Nephrol. Hypertens. 2005; 14: 579-585Crossref PubMed Scopus (35) Google Scholar), and the appearance of certain proteins in the blood stream may result in their appearance in the urine either in the intact form or as peptide fragments. The protein and peptide composition of the urine is determined by the function of the glomerular filtration apparatus, proximal tubular absorption of ultrafiltered proteins, and the capacity of the brush border and lysosomal proteolytic machinery to degrade filtered proteins (8D'Amico G. Bazzi C. Pathophysiology of proteinuria.Kidney Int. 2003; 63: 809-825Abstract Full Text Full Text PDF PubMed Scopus (367) Google Scholar). Therefore, detection of one or several proteins or polypeptides may provide a signature of a particular pathological process (7O'Riordan E. Goligorsky M.S. Emerging studies of the urinary proteome: the end of the beginning?.Curr. Opin. Nephrol. Hypertens. 2005; 14: 579-585Crossref PubMed Scopus (35) Google Scholar). We hypothesized that proteomics analysis of urine should yield a panel of biomarker peptides useful as additional tools for the diagnosis and monitoring of CAD. Furthermore we aimed to obtain sequences of biomarkers of the CAD signature panel to gain insight into pathogenetic mechanisms and facilitate comparison with currently used biochemical markers. Capillary electrophoresis on-line coupled to electrospray ionization-time-of-flight mass spectrometry (CE-ESI-TOF MS) seems ideally suited for this purpose because of its non-invasive nature, high resolution, and amenability for future adaptation to clinical laboratory analysis (5Kolch W. Neususs C. Pelzing M. Mischak H. Capillary electrophoresis-mass spectrometry as a powerful tool in clinical diagnosis and biomarker discovery.Mass Spectrom. Rev. 2005; 24: 959-977Crossref PubMed Scopus (256) Google Scholar). We enrolled 88 patients with CAD confirmed by coronary angiography. Patients were recruited at the Western Infirmary, Glasgow, UK. Eighty-two of the 88 patients were reassessed after a further 14 weeks, one patient died, and five patients refused to participate at a follow-up examination. At both assessments blood and midstream spot urine samples were collected. Thirty-two subjects with no history of angina, CAD, or peripheral artery disease who were recruited from a local health club and from surgical wards at Gartnavel General Hospital, Glasgow, UK served as controls. Plasma total cholesterol, low density lipoproteins, high density lipoproteins, triglycerides, high sensitivity C-reactive protein, and serum creatinine were assessed using standard biochemical methods. The modification of diet in renal disease formula was used for the estimation of glomerular filtration rate in study participants (9National Kidney FoundationK/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification.Am. J. Kidney Dis. 2002; 39: S1-S266PubMed Google Scholar). Vascular stiffness was assessed by two methods. First, pulse contour analysis of the diastolic pressure decay was used to estimate large (C1) and small artery compliance (C2; HDI/Pulse Wave CR2000, HDI Inc., Eagan, MN) based on a three-element Windkessel model (10Cohn J.N. Finkelstein S. McVeigh G. Morgan D. LeMay L. Robinson J. Mock J. Noninvasive pulse wave analysis for the early detection of vascular disease.Hypertension. 1995; 26: 503-508Crossref PubMed Scopus (424) Google Scholar). Second, the augmentation index of the central aorta was derived from the radial pulse waveform using a generalized transfer function (SphygmoCor pulse wave analysis system, AtCor Medical, West Ryde, New South Wales, Australia) (11Pauca A.L. O'Rourke M.F. Kon N.D. Prospective evaluation of a method for estimating ascending aortic pressure from the radial artery pressure waveform.Hypertension. 2001; 38: 932-937Crossref PubMed Scopus (971) Google Scholar). The augmentation index was calculated from the ratio of the pulse pressure at the second systolic peak to that at the first systolic peak. The study was approved by the West Glasgow Ethics Committee, and all subjects gave written informed consent. This study was designed according to current guidelines for studies on clinical proteomics (12Mischak H. Apweiler R. Banks R. Conaway M. Coon J. Dominiczak A. Ehrich J. Fliser D. Girolami M. Hermjakob H. Hochstrasser D. Jankowski J. Julian B.A. Kolch W. Massy Z.A. Neusuess C. Novak J. Peter K. Rossing K. Schanstra J. Semmes O.J. Theodorescu D. Thongboonkerd V. Weissinger E.M. Van Eyk J.E. Yamamoto T. Clinical proteomics: a need to define the field and to begin to set adequate standards.Proteomics Clin. Appl. 2007; 1: 148-156Crossref PubMed Scopus (280) Google Scholar) and the minimum information about proteomics experiments (MIAPE) (13Taylor C.F. Paton N.W. Lilley K.S. Binz P.A. Julian Jr., R.K. Jones A.R. Zhu W. Apweiler R. Aebersold R. Deutsch E.W. Dunn M.J. Heck A.J. Leitner A. Macht M. Mann M. Martens L. Neubert T.A. Patterson S.D. Ping, P, Seymour S.L. Souda P. Tsugita A. Vandekerckhove J. Vondriska T.M. Whitelegge J.P. Wilkins M.R. Xenarios I. Yates III, J.R. Hermjakob H. The minimum information about a proteomics experiment (MIAPE).Nat. Biotechnol. 2007; 25: 887-893Crossref PubMed Scopus (580) Google Scholar). To exclude the effect of medication, 17 paired urine samples from age- and sex-matched patients with hypertension and type 2 diabetes, but without albuminuria, before and 12 weeks after commencing treatment with the angiotensin-converting enzyme inhibitor ramipril (5–10 mg once daily) were evaluated. To rule out center specific bias, samples from 233 new appointees at the University of Hannover who were free of self-reported illness were also analyzed. Detailed characteristics of all patients and controls are shown in Table I with additional data on 233 healthy university recruits and 18 ramipril patients shown in Table II.Table IDemographics and clinical dataControls (n = 32)CAD, base line (n = 77)CAD, follow-up (n = 82)pSex, male/female21/956/2159/230.91Age (yr)54 ± 1361 ± 11 †62 ± 11N/ABMI (kg/m2)25.3 ± 3.126.2 ± 4.826.5 ± 4.60.38Smokers, active/stopped/none3/9/2014/38/25ap < 0.05.13/42/270.93Diabetic, n0670.84Statin therapy, n075bp < 0.001.770.93Systolic blood pressure (mm Hg)123 ± 12132 ± 20cp < 0.01.133 ± 190.42Diastolic blood pressure (mm Hg)76 ± 776 ± 974 ± 90.15Total cholesterol (mmol/liter)5.4 ± 0.93.9 ± 0.8bp < 0.001.3.8 ± 0.80.24LDL cholesterol (mmol/liter)3.2 ± 0.71.9 ± 0.7bp < 0.001.1.8 ± 0.80.13HDL cholesterol (mmol/liter)1.5 ± 0.41.2 ± 0.3cp < 0.01.1.3 ± 0.30.12Triglycerides (mmol/liter)1.3 (1.0;2.7)1.5 (1.8; 2.2)1.4 (1.1; 2.1)0.48C-reactive protein (mg/liter)1.3 (0.3;2.4)2.6 (1.0; 6.3)cp < 0.01.1.2 (0.6; 2.2)<0.001Creatinine (μmol/liter)90 ± 991 ± 2198 ± 230.03eGFR (ml/min/1.73m2)75 ± 1075 ± 969 ± 14<0.001AI (%)26.4 ± 11.732.1 ± 10.4ap < 0.05.30.5 ± 9.40.18C1 (ml/mm Hg × 10)14.0 ± 3.911.8 ± 4.3cp < 0.01.13.3 ± 4.30.01C2 (ml/mm Hg × 100)5.7 ± 3.93.8 ± 2.8cp < 0.01.3.6 ± 1.60.50a p < 0.05.b p < 0.001.c p < 0.01. Open table in a new tab Table IICharacteristics of the training and the test setsTraining setTest setCADaIn all depicted parameters there were no differences between CAD patients and controls in the training set and test set, respectively.ControlsaIn all depicted parameters there were no differences between CAD patients and controls in the training set and test set, respectively.Ramipril samplesHannover samplesCADaIn all depicted parameters there were no differences between CAD patients and controls in the training set and test set, respectively.ControlsaIn all depicted parameters there were no differences between CAD patients and controls in the training set and test set, respectively.Total number of patients3020182324712Sex, male/female22/814/614/4101/13734/137/5Age (yr)62 ± 1154 ± 1359 ± 1134 ± 1161 ± 1254 ± 12BMI (kg/m2)25.9 ± 3.825.7 ± 3.530.5 ± 5.026.5 ± 5.324.8 ± 2.2Smoker, yes/no5/252/184/148/391/11Diabetic, n101850Systolic blood pressure (mm Hg)133 ± 19123 ± 11151 ± 13132 ± 20124 ± 13Diastolic blood pressure (mm Hg)75 ± 1176 ± 786 ± 1076 ± 875 ± 8Total cholesterol (mmol/liter)3.9 ± 0.85.2 ± 1.05.5 ± 1.33.9 ± 0.95.6 ± 0.6LDL cholesterol (mmol/liter)1.9 ± 0.63.0 ± 0.83.4 ± 1.01.9 ± 0.83.4 ± 0.6HDL cholesterol (mmol/liter)1.2 ± 0.31.4 ± 0.41.3 ± 0.31.2 ± 0.31.6 ± 0.4Triglycerides (mmol/liter)1.6 (1.2; 2.4)1.3 (1.1; 2.5)1.6 (1.0; 2.7)1.4 (1.2; 1.9)1.3 (1.0; 1.6)a In all depicted parameters there were no differences between CAD patients and controls in the training set and test set, respectively. Open table in a new tab At follow-up examination, self-reported physical activity was assessed. The physical activity was graded into two categories: no regular physical activity (patients mainly confined indoors) or low grade physical activities like walking on flat terrain or non-strenuous gardening and a very active group with hiking, biking, and golfing several times a week. This classification was independently validated by physiotherapists who categorized patients’ activity levels according to clinical data (r = 0.379, p = 0.006). Furthermore physiotherapists performed an incremental shuttle walk test (14Singh S.J. Morgan M.D. Hardman A.E. Rowe C. Bardsley P.A. Comparison of oxygen uptake during a conventional treadmill test and the shuttle walking test in chronic airflow limitation.Eur. Respir. J. 1994; 7: 2016-2020PubMed Google Scholar) after base-line examination in 52 of the 88 patients. There was a significant correlation between metabolic equivalent obtained by the test and self-reported activity (r = 0.399, p = 0.003). After collection, all the spot urine samples were frozen at −80 °C until analysis. For proteomics analysis samples were prepared as described previously (15Theodorescu D. Wittke S. Ross M.M. Walden M. Conaway M. Just I. Mischak H. Frierson H.F. Discovery and validation of new protein biomarkers for urothelial cancer: a prospective analysis.Lancet Oncol. 2006; 7: 230-240Abstract Full Text Full Text PDF PubMed Scopus (366) Google Scholar, 16Weissinger E.M. Wittke S. Kaiser T. Haller H. Bartel S. Krebs R. Golovko I. Rupprecht H.D. Haubitz M. Hecker H. Mischak H. Fliser D. Proteomic patterns established with capillary electrophoresis and mass spectrometry for diagnostic purposes.Kidney Int. 2004; 65: 2426-2434Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar). CE-MS analysis was performed as described previously (15Theodorescu D. Wittke S. Ross M.M. Walden M. Conaway M. Just I. Mischak H. Frierson H.F. Discovery and validation of new protein biomarkers for urothelial cancer: a prospective analysis.Lancet Oncol. 2006; 7: 230-240Abstract Full Text Full Text PDF PubMed Scopus (366) Google Scholar, 16Weissinger E.M. Wittke S. Kaiser T. Haller H. Bartel S. Krebs R. Golovko I. Rupprecht H.D. Haubitz M. Hecker H. Mischak H. Fliser D. Proteomic patterns established with capillary electrophoresis and mass spectrometry for diagnostic purposes.Kidney Int. 2004; 65: 2426-2434Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar, 17Wittke S. Mischak H. Walden M. Kolch W. Radler T. Wiedemann K. Discovery of biomarkers in human urine and cerebrospinal fluid by capillary electrophoresis coupled to mass spectrometry: towards new diagnostic and therapeutic approaches.Electrophoresis. 2005; 26: 1476-1487Crossref PubMed Scopus (128) Google Scholar) using a P/ACE MDQ capillary electrophoresis system (Beckman Coulter, Fullerton, CA) on-line coupled to a Micro-TOF MS instrument (Bruker Daltonics, Bremen, Germany). The performance of the sample preparation procedure as well as the analytical performance of the instrumental setup was evaluated (5Kolch W. Neususs C. Pelzing M. Mischak H. Capillary electrophoresis-mass spectrometry as a powerful tool in clinical diagnosis and biomarker discovery.Mass Spectrom. Rev. 2005; 24: 959-977Crossref PubMed Scopus (256) Google Scholar, 18Wittke S. Fliser D. Haubitz M. Bartel S. Krebs R. Hausadel F. Hillmann M. Golovko I. Koester P. Haller H. Kaiser T. Mischak H. Weissinger E.M. Determination of peptides and proteins in human urine with capillary electrophoresis-mass spectrometry, a suitable tool for the establishment of new diagnostic markers.J. Chromatogr. A. 2003; 1013: 173-181Crossref PubMed Scopus (177) Google Scholar). The average recovery of the sample preparation procedure is ∼85% with a detection limit of ∼1 fmol. The monoisotopic mass signals could be resolved for z ≤ 6. The mass accuracy of the CE-TOF-MS method was determined to be <25 ppm for monoisotopic resolution and <100 ppm for unresolved peaks (z > 6). The precision of the analytical method was determined by assessing (a) the reproducibility achieved for repeated measurement of the same aliquot and (b) the reproducibility achieved for repeated preparation and measurement of the same urine sample (5Kolch W. Neususs C. Pelzing M. Mischak H. Capillary electrophoresis-mass spectrometry as a powerful tool in clinical diagnosis and biomarker discovery.Mass Spectrom. Rev. 2005; 24: 959-977Crossref PubMed Scopus (256) Google Scholar, 18Wittke S. Fliser D. Haubitz M. Bartel S. Krebs R. Hausadel F. Hillmann M. Golovko I. Koester P. Haller H. Kaiser T. Mischak H. Weissinger E.M. Determination of peptides and proteins in human urine with capillary electrophoresis-mass spectrometry, a suitable tool for the establishment of new diagnostic markers.J. Chromatogr. A. 2003; 1013: 173-181Crossref PubMed Scopus (177) Google Scholar). The 200 most abundant polypeptides were detected with a rate of 98%. The performance of the analytical system over time was assessed with consecutive measurements of the same aliquot over a period of 24 h. No significant loss of polypeptides was observed implying the stability of the CE-MS setup, the postpreparative stability of the urine samples at 4 °C, and their resistance to e.g. oxidizing processes or precipitation (5Kolch W. Neususs C. Pelzing M. Mischak H. Capillary electrophoresis-mass spectrometry as a powerful tool in clinical diagnosis and biomarker discovery.Mass Spectrom. Rev. 2005; 24: 959-977Crossref PubMed Scopus (256) Google Scholar, 18Wittke S. Fliser D. Haubitz M. Bartel S. Krebs R. Hausadel F. Hillmann M. Golovko I. Koester P. Haller H. Kaiser T. Mischak H. Weissinger E.M. Determination of peptides and proteins in human urine with capillary electrophoresis-mass spectrometry, a suitable tool for the establishment of new diagnostic markers.J. Chromatogr. A. 2003; 1013: 173-181Crossref PubMed Scopus (177) Google Scholar). Data processing and cluster analysis were performed as described previously (15Theodorescu D. Wittke S. Ross M.M. Walden M. Conaway M. Just I. Mischak H. Frierson H.F. Discovery and validation of new protein biomarkers for urothelial cancer: a prospective analysis.Lancet Oncol. 2006; 7: 230-240Abstract Full Text Full Text PDF PubMed Scopus (366) Google Scholar, 19Theodorescu D. Fliser D. Wittke S. Mischak H. Krebs R. Walden M. Ross M. Eltze E. Bettendorf O. Wulfing C. Semjonow A. Pilot study of capillary electrophoresis coupled to mass spectrometry as a tool to define potential prostate cancer biomarkers in urine.Electrophoresis. 2005; 26: 2797-2808Crossref PubMed Scopus (135) Google Scholar). Only signals observed in a minimum of three consecutive spectra with a minimum signal-to-noise ratio of 4 were considered. Mass spectral ion peaks representing identical molecules at different charge states were deconvoluted into single masses using either the distance between resolved isotope peaks of the ion or according to conjugated signals for unresolved isotope peaks (MosaiquesVisu software (16Weissinger E.M. Wittke S. Kaiser T. Haller H. Bartel S. Krebs R. Golovko I. Rupprecht H.D. Haubitz M. Hecker H. Mischak H. Fliser D. Proteomic patterns established with capillary electrophoresis and mass spectrometry for diagnostic purposes.Kidney Int. 2004; 65: 2426-2434Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar, 18Wittke S. Fliser D. Haubitz M. Bartel S. Krebs R. Hausadel F. Hillmann M. Golovko I. Koester P. Haller H. Kaiser T. Mischak H. Weissinger E.M. Determination of peptides and proteins in human urine with capillary electrophoresis-mass spectrometry, a suitable tool for the establishment of new diagnostic markers.J. Chromatogr. A. 2003; 1013: 173-181Crossref PubMed Scopus (177) Google Scholar)). In addition, migration time and ion signal intensity (amplitude) were normalized using internal polypeptide standards (15Theodorescu D. Wittke S. Ross M.M. Walden M. Conaway M. Just I. Mischak H. Frierson H.F. Discovery and validation of new protein biomarkers for urothelial cancer: a prospective analysis.Lancet Oncol. 2006; 7: 230-240Abstract Full Text Full Text PDF PubMed Scopus (366) Google Scholar). The resulting peak list characterizes each polypeptide by its molecular mass (kDa), normalized migration time (min), and normalized signal intensity. All detected polypeptides were deposited, matched, and annotated in a Microsoft SQL (structured query language) database, allowing further analysis and comparison of multiple samples (patient groups). Polypeptides within different samples were considered identical if the mass deviation was less than 100 ppm and the migration time deviation was less than 3%. CE-MS data of all individual samples can be accessed in the supplemental table. For biomarker panel definition, we used polypeptides that were found in more than 75% of the urine samples in at least one of the different groups of the training set (e.g. CAD or healthy controls). Polypeptides fulfilling this criterion were further evaluated using receiver operating characteristic (ROC) statistics (20DeLeo J. Receiver operating characteristics laboratory (ROCLAB): software for developing decision strategies that account for uncertainty.in: Second International Symposium on Uncertainty Modeling and Analysis: Proceedings April 25–28, 1993 University of Maryland College Park, Maryland (Isuma ‘93) Institute of Electrical and Electronics Engineers. IEEE Computer Society, Washington, D. C.1993: 318-325Crossref Scopus (180) Google Scholar). The amplitude distribution of the CE-MS data of polypeptides present in the samples was used as the ROC variable, and the affiliation to a diagnostic group (i.e. CAD or healthy control) was used as the classification variable. The obtained area under the ROC curve (AUC) value of the analysis of a given polypeptide was interpreted as a measure of its discriminatory potential. An initial list of potential marker candidates was further refined using the Mann-Whitney test with p ≤ 0.05 as the significance level. Model establishment and sample classification were performed by using a linear classifier algorithm according to F = ∑i ci log Ai with F as classification factor, ci as classification coefficient, and Ai as amplitude of the CE-MS signal of the marker i. The algorithm generates a classification model based on polypeptides that are best suited to discriminate between two defined sample groups. Models consist of fewer biomarkers than samples to avoid overfitting of models. The probability to have CAD at a given classification factor F taking into account the related probability for a negative diagnosis was calculated according to Equation 1. PCAD=11+Xwith X=S.D.CADS.D.HCe(FCADmean-F)22S.D.CAD2-(FHCmean-F)22.S.D.HC2(Eq. 1) For the first phase of the study a training set was established. The training set consisted of 50 urine samples from randomly selected subjects, 30 CAD patients, and 20 control subjects, respectively. The first step of biomarker selection led to a set of 187 potential CAD-specific polypeptides. In a second step, these preselected polypeptides were compared with 233 urine samples from healthy volunteers from different centers to eliminate polypeptides that may show center-specific bias. To exclude the effect of medication on constituting markers, an additional control group of patients before (n = 15) and after (n = 17) 12-week treatment with the angiotensin-converting enzyme ramipril was used to refine the selected polypeptides. Polypeptides that showed up/down-regulation of CE-MS signal intensity in direct comparison of both groups and in addition a uniform behavior in pairwise comparison in the majority of patients were considered as medication artifacts and eliminated. Subsequently the established pattern of 15 polypeptides was evaluated in a blinded assessment of 59 urine samples: 47 samples from patients with CAD and 12 samples from healthy controls. All samples were examined using the CAD panel. Seventy-six urine samples from follow-up examination were also evaluated using the CAD panel. Peptide sequencing was performed using an LTQ-Orbitrap™ hybrid mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) equipped with a Dionex Ultimate 3000 nanoflow system and a nanoelectrospray ion source. Peptide separation took place on a 5-μm C18 nanocolumn (NanoSeparations, Nieuwkoop, Netherlands) in a precolumn setup using a flow rate of 5 μl/min followed by a flow of 250 nl/min and a linear gradient (60 min) from 2 to 50% MeCN in H2O (0.1% formic acid). The mass spectrometer was operated in data-dependent mode to automatically switch between MS and MS/MS acquisition. Survey full-scan MS spectra (from m/z 300 to 2000) were acquired in the Orbitrap with resolution R = 60,000 at m/z 400 (target value of 500,000 charges in the linear ion trap). The most intense ions were sequentially isolated for fragmentation in the linear ion trap using collisionally induced dissociation and the detection took place either in the linear ion trap (parallel mode; target value 10,000) or in the Orbitrap (target value of 500,000). Orbitrap MS/MS were acquired with resolution R = 15,000 at m/z 400. General mass spectrometric conditions were: electrospray voltage, 1.6 kV; no sheath and auxiliary gas flow; ion transfer tube temperature, 225 °C; collision gas pressure, 1.3 millitorrs; normalized collision energy, 32% for MS/MS. The ion selection threshold was 500 counts for MS/MS. Further analysis was performed using instruments with electron transfer dissociation (ETD) capability (21Coon J.J. Shabanowitz J. Hunt D.F. Syka J.E. Electron transfer dissociation of peptide anions.J. Am. Soc. Mass Spectrom. 2005; 16: 880-882Crossref PubMed Scopus (209) Google Scholar, 22Good D.M. Coon J.J. Advancing proteomics with ion/ion chemistry.BioTechniques. 2006; 40: 783-789Crossref PubMed Scopus (46) Google Scholar, 23Syka J.E. Coon J.J. Schroeder M.J. Shabanowitz J. Hunt D.F. Peptide and protein sequence analysis by electron transfer dissociation mass spectrometry.Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 9528-9533Crossref PubMed Scopus (1999) Google Scholar). Upon arrival, samples were resuspended (50 μl of 100 mm acetic acid) and bomb-loaded onto a 360 × 75-μm microcapillary precolumn that was connected to a 360 × 50-μm analytical column with a ∼ 1-μm tip pulled with a laser puller (both columns were packed in-house with ∼5–8 cm of C18 resin). Peptides were separated by nano-reversed phase HPLC (Agilent 1100; flow split by a tee to ∼100 nl/min) and introduced into either an ETD-enabled LTQ quadrupole linear