Unraveling the Complex Proteome for Biomarker Discovery in Gastrointestinal and Liver Diseases

Abstract

Proteomics holds great promise to transform clinical medicine with novel diagnostics and therapeutics. The completion of the sequencing of the human genome in 2003 was a massive, 13-year undertaking. In comparison, the human proteome is orders of magnitude more complex, and unraveling and mining the human proteome is still an emerging field. Proteomics is the study of the proteome utilizing technologies that allow for large-scale and/or high-throughput analysis. The field encompasses not only the identification of direct translational products from gene transcripts, but also identifying posttranslational modifications including cleaved fragments, quantitative analysis of protein levels, localization within the cell and tissues, protein activity assays, and protein–protein interactions. The dynamic nature of biological systems adds additional complexity as protein properties evolve under different conditions.Approach to Proteomic AnalysisKey components in proteomic analysis are the sample(s), the technologies used, and information processing. The importance of experimental design and proper selection of samples for analysis cannot be overemphasized. This is particularly the case for human samples, as heterogeneity between subjects or patients introduces potential bias, and sample acquisition, processing, and storage must also be carefully designed and controlled to avoid false discovery.1Rai A.J. Gelfand C.A. Haywood B.C. Warunek D.J. Yi J. Schuchard M.D. Mehigh R.J. Cockrill S.L. Scott G.B. Tammen H. Schulz-Knappe P. Speicher D.W. Vitzthum F. Haab B.B. Siest G. Chan D.W. HUPO Plasma Proteome Project specimen collection and handling: towards the standardization of parameters for plasma proteome samples.Proteomics. 2005; 5: 3262-3277Google Scholar Proteomic technologies are rapidly advancing and newer technologies such as protein microarrays have increasing visibility.2Liotta L.A. Espina V. Mehta A.I. Calvert V. Rosenblatt K. Geho D. Munson P.J. Young L. Wulfkuhle J. Petricoin 3rd, E.F. Protein microarrays: meeting analytical challenges for clinical applications.Cancer Cell. 2003; 3: 317-325Google Scholar, 3Wulfkuhle J.D. Liotta L.A. Petricoin E.F. Proteomic applications for the early detection of cancer.Nat Rev Cancer. 2003; 3: 267-275Google Scholar Currently, the major technology for proteomic discovery relies on the use of mass spectrometry. Various mass spectrometry methods and instrumentation are utilized, each with distinctive advantages.4Koomen J. Hawke D. Kobayashi R. Developing an understanding of proteomics: an introduction to biological mass spectrometry.Cancer Invest. 2005; 23: 47-59Google Scholar One approach that has been popularized called surfaced enhanced laser desorption/ionization relies not on identification of proteins, but rather on the recognition of specific spectral patterns in a sample. Initial excitement using this approach for cancer screening was significantly dampened once issues of sensitivity and reproducibility were raised owing to limitations of this particular technology.5Ransohoff D.F. Lessons from controversy: ovarian cancer screening and serum proteomics.J Natl Cancer Inst. 2005; 97: 315-319Google Scholar More advanced mass spectrometry instrumentation allows substantial resolution, depth of analysis, and confidence in protein identification. Additional bioinformatics and other knowledge are required to analyze the proteins identified to determine biological relevance and formulate hypotheses.6States D.J. Omenn G.S. Blackwell T.W. Fermin D. Eng J. Speicher D.W. Hanash S.M. Challenges in deriving high-confidence protein identifications from data gathered by a HUPO plasma proteome collaborative study.Nat Biotechnol. 2006; 24: 333-338Google Scholar It follows therefore that proteomic research requires a multidisciplinary team with expertise across disciplines that include protein biochemistry, mass spectrometry, bioinformatics, statistics, and clinical medicine working closely together to avoid the many pitfalls in the data acquisition and analysis process.Sources of Specimens for Proteomic AnalysisThere are numerous strategies that may be implemented for biomarker discovery using proteomics (Figure 1). These include tumor and control tissue analysis to identify candidates such as secreted proteins and proteins shed from the cell surface, which may be detected in circulation as biomarkers. Such a strategy requires that efficient assays be available to test candidates in serum or plasma. Other strategies include the use of tumor cells and cell lines to identify such candidates in the case of cancer. The most direct but challenging path to biomarker identification is the analysis of serum and plasma based on well-designed studies of diagnostic or prediagnostic serum or plasma and appropriately matched controls. Mouse models of cancer potentially represent an efficient means for uncovering diagnostic markers as genetic alterations associated with human tumors can be engineered in mice. In addition, defined stages of tumor development, breeding conditions, and blood sampling can all be controlled and standardized to limit heterogeneity. The concept that plasma from engineered mouse models of cancer contains tumor-derived proteins that may be relevant as candidate markers for human cancer is attractive but largely untested.7Hingorani S.R. Petricoin E.F. Maitra A. Rajapakse V. King C. Jacobetz M.A. Ross S. Conrads T.P. Veenstra T.D. Hitt B.A. Kawaguchi Y. Johann D. Liotta L.A. Crawford H.C. Putt M.E. Jacks T. Wright C.V. Hruban R.H. Lowy A.M. Tuveson D.A. Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse.Cancer Cell. 2003; 4: 437-450Google ScholarProteomics for Biomarker DiscoveryThe discovery of disease biomarkers utilizing proteomics, especially for various cancers, has been a major area of proteomic research. The National Cancer Institute (NCI) through its Early Detection Network (EDRN) has spearheaded a multicenter effort for early detection of cancer using biomarkers.8Verma M. Srivastava S. New cancer biomarkers deriving from NCI early detection research.Recent Results Cancer Res. 2003; 163 (discussion 264–266): 72-84Google Scholar Several promising proteomic approaches for discovery have been utilized. Protein microarrays allow for high-throughput analysis for biomarker discovery. One application of protein microarrays is to coat glass slides with proteins, cell lysates, or antibodies and then incubate with a complementary biological sample (Figure 2A). Investigators have used antibodies to coat glass slides and then incubate serum from cancer and control patients to identify differences. A recent study utilized recombinant antibody microarrays to identify a protein expression signature in patients with gastric adenocarcinoma in the setting of Helicobacter pylori infection.9Ellmark P. Ingvarsson J. Carlsson A. Lundin S.B. Wingren C. Borrebaeck C.A. Identification of protein expression signatures associated with H. pylori infection and gastric adenocarcinoma using recombinant antibody microarrays.Mol Cell Proteomics. 2006; 5: 1638-1646Google Scholar Using proteins as the capture agent, NCI-60 cancer cell line lysates have previously been spotted on a protein microarray and then incubated with commercially available antibodies. From this analysis, 2 promising pathologic markers were identified that could distinguish colon from ovarian adenocarcinoma.10Nishizuka S. Charboneau L. Young L. Major S. Reinhold W.C. Waltham M. Kouros-Mehr H. Bussey K.J. Lee J.K. Espina V. Munson P.J. Petricoin 3rd, E. Liotta L.A. Weinstein J.N. Proteomic profiling of the NCI-60 cancer cell lines using new high-density reverse-phase lysate microarrays.Proc Natl Acad Sci U S A. 2003; 100: 14229-14234Google Scholar A promising area of protein microarray research is the identification of autoantibodies in human serum. This approach utilizes protein or lysate spotted arrays incubated with human serum from cancer and control patients. Microarrays that contain natural proteins derived from tumor cells have the potential to substantially accelerate the discovery of tumor antigens via a molecular signature for immune responses directed against protein targets in different types of cancer. In a study of colon cancer, microarrays printed with 1760 distinct protein fractions, prepared from the LoVo colon adenocarcinoma cell line, were hybridized with individual sera.11Nam M.J. Madoz-Gurpide J. Wang H. Lescure P. Schmalbach C.E. Zhao R. Misek D.E. Kuick R. Brenner D.E. Hanash S.M. Molecular profiling of the immune response in colon cancer using protein microarrays: occurrence of autoantibodies to ubiquitin C-terminal hydrolase L3.Proteomics. 2003; 3: 2108-2115Google Scholar A fraction that exhibited immunoglobulin G–based reactivity with 9/15 colon cancer sera was found to contain ubiquitin C-terminal hydrolase L3 (UCH-L3) by tandem mass spectrometry (ESI-Q-TOF). The highest levels of UCH-L3 mRNA among the 329 tumors of different types analyzed by DNA microarrays were found in colon tumors. Independent validation by Western blotting demonstrated that UCH-L3 antibodies existed in 19/43 sera from patients with colon cancer, and in 0/54 sera from subjects with lung cancer, colon adenoma, or otherwise healthy controls. These data point to the utility of microarrays printed with natural proteins for the identification of tumor antigens that have induced an antibody response in patients with specific cancersFigure 2Promising proteomic approaches. (A) Protein microarrays coated either with antibodies or proteins. Red circles represent fluorescent tags that can be detected using a scanner, which are represented by white and red spots on the scanned image. (B) Representative mass ion spectra of a peptide. The heavy labeled peptide is from cancer patient serum and the light labeled peptide is from normal patient serum. Ratio is calculated using total ion intensity. (C) Example of fractionation and isotope labeling of proteins in pancreatic cancer versus control serum. Each row represents a single protein and the columns represent protein fractions based on charge state and hydrophobic properties. Each colored spot identifies the presence of a peptide for that protein in that fraction. Red represents up-regulated in cancer, green is down-regulated in cancer, and yellow is no dysregulation compared with control.View Large Image Figure ViewerDownload (PPT)Another advance in proteomics has been the use of isotope labeling for protein quantitative analysis. The basic principle is to label one sample with a “heavy” isotope and another sample with a “light” isotope and then use mass spectrometry to detect the relative abundance of labeled proteins between the 2 samples (Figure 2B). Isotope labeling has been used to detect relative protein concentration differences in pancreatic cancer. Using isotope-code affinity tag technology, the proteome of human pancreatic cancer tissue and juice has been analyzed.12Chen R. Pan S. Yi E.C. Donohoe S. Bronner M.P. Potter J.D. Goodlett D.R. Aebersold R. Brentnall T.A. Quantitative proteomic profiling of pancreatic cancer juice.Proteomics. 2006; 6: 3871-3879Google Scholar, 13Chen R. Yi E.C. Donohoe S. Pan S. Eng J. Cooke K. Crispin D.A. Lane Z. Goodlett D.R. Bronner M.P. Aebersold R. Brentnall T.A. Pancreatic cancer proteome: the proteins that underlie invasion, metastasis, and immunologic escape.Gastroenterology. 2005; 129: 1187-1197Abstract Full Text Full Text PDF Scopus (164) Google Scholar Differentially regulated proteins were found to be involved in functions encompassing angiogenesis, invasion, extracellular matrix destruction, and immunologic escape. Further validation of over-expressed proteins in these studies was performed using Western blot or immunohistochemistry techniques.One of the challenges to proteomic analysis has been the wide dynamic range of protein concentration. In human blood, protein concentration can span ≥10 orders of magnitude.14Anderson N.L. Anderson N.G. The human plasma proteome: history, character, and diagnostic prospects.Mol Cell Proteomics. 2002; 1: 845-867Google Scholar Extensive multidimensional fractionation of serum or plasma can be used to identify low abundance proteins, which may be enriched in biomarkers as they are less likely to represent acute phase reactants or classical blood proteins.6States D.J. Omenn G.S. Blackwell T.W. Fermin D. Eng J. Speicher D.W. Hanash S.M. Challenges in deriving high-confidence protein identifications from data gathered by a HUPO plasma proteome collaborative study.Nat Biotechnol. 2006; 24: 333-338Google Scholar Coupling isotope labeling with multidimensional fractionation is attractive. In our laboratory, we have taken such an approach to identify promising blood-based biomarkers for gastrointestinal cancers across 7 orders of magnitude of protein concentration. Multidimensional separation not only provides quantitative information, but can provide information on protein isoforms and cleavage products as elution patterns of proteins can be analyzed across the fractionation scheme (Figure 2C).From Discovery to ValidationAlthough proteomic analyses to date have identified differentially regulated proteins in disease versus control samples, validation of discovery findings remains a challenge. The discovery phase of potential biomarkers is just the beginning toward establishing a clinically useful test. The EDRN has set forth specific criteria for biomarker development for the early detection of cancer.15Pepe M.S. Etzioni R. Feng Z. Potter J.D. Thompson M.L. Thornquist M. Winget M. Yasui Y. Phases of biomarker development for early detection of cancer.J Natl Cancer Inst. 2001; 93: 1054-1061Google Scholar Indeed, these criteria can also be applied to nonmalignant diseases. The 5 phases of biomarker development start with preclinical exploratory studies and end with large-scale studies that show the impact of screening with biomarkers on decreasing disease burden at the population level (Figure 3). Epidemiology and biostatistics become increasingly important across the 5 phases.Figure 3Guidelines for biomarker development as set forth by EDRN.View Large Image Figure ViewerDownload (PPT)One of the obstacles to biomarker development is the availability and adequacy of testing platforms. A diagnostic test should ideally be one that is simple to run, have high throughput, and be of low cost. Although mass spectrometry is commonly used for biomarker discovery, it is not suitable for widespread diagnostic testing. Within the proteomics field, there has been interest in developing novel multiplexing assays that can analyze many proteins with only a small amount of biological sample. However, such technologies depend on the availability of high-quality antibodies or other affinity reagents, which unfortunately are lacking for many potential protein biomarkers discovered. The use of protein isoforms, cleaved products, or peptides as biomarkers require specifically designed antibodies.Future DirectionsIndividual expertise can improve the various proteomic technology elements, but multidisciplinary research teams are essential for biomarker discovery and validation. Although there has been great interest in proteomic efforts for various cancers, there are many nonmalignant diseases within gastroenterology and hepatology that would benefit from proteomic analysis. Inflammatory bowel disease (IBD), irritable bowel syndrome, and fatty liver represent diseases for which many questions remain that proteomics may help answer. Within IBD, better markers are needed to distinguish between Crohn’s disease and ulcerative colitis and to predict or diagnose those who will flare or go on to need surgery. A marker to identify colonic dysplasia would avoid the need for annual surveillance colonoscopies that are taxing both to the patient as well as the endoscopist. Proteomics is an emerging yet quite promising field and the technologies available today should be applied for discovery that will lead to the introduction of novel diagnostics and therapeutics to improve patient outcomes. Proteomics holds great promise to transform clinical medicine with novel diagnostics and therapeutics. The completion of the sequencing of the human genome in 2003 was a massive, 13-year undertaking. In comparison, the human proteome is orders of magnitude more complex, and unraveling and mining the human proteome is still an emerging field. Proteomics is the study of the proteome utilizing technologies that allow for large-scale and/or high-throughput analysis. The field encompasses not only the identification of direct translational products from gene transcripts, but also identifying posttranslational modifications including cleaved fragments, quantitative analysis of protein levels, localization within the cell and tissues, protein activity assays, and protein–protein interactions. The dynamic nature of biological systems adds additional complexity as protein properties evolve under different conditions. Approach to Proteomic AnalysisKey components in proteomic analysis are the sample(s), the technologies used, and information processing. The importance of experimental design and proper selection of samples for analysis cannot be overemphasized. This is particularly the case for human samples, as heterogeneity between subjects or patients introduces potential bias, and sample acquisition, processing, and storage must also be carefully designed and controlled to avoid false discovery.1Rai A.J. Gelfand C.A. Haywood B.C. Warunek D.J. Yi J. Schuchard M.D. Mehigh R.J. Cockrill S.L. Scott G.B. Tammen H. Schulz-Knappe P. Speicher D.W. Vitzthum F. Haab B.B. Siest G. Chan D.W. HUPO Plasma Proteome Project specimen collection and handling: towards the standardization of parameters for plasma proteome samples.Proteomics. 2005; 5: 3262-3277Google Scholar Proteomic technologies are rapidly advancing and newer technologies such as protein microarrays have increasing visibility.2Liotta L.A. Espina V. Mehta A.I. Calvert V. Rosenblatt K. Geho D. Munson P.J. Young L. Wulfkuhle J. Petricoin 3rd, E.F. Protein microarrays: meeting analytical challenges for clinical applications.Cancer Cell. 2003; 3: 317-325Google Scholar, 3Wulfkuhle J.D. Liotta L.A. Petricoin E.F. Proteomic applications for the early detection of cancer.Nat Rev Cancer. 2003; 3: 267-275Google Scholar Currently, the major technology for proteomic discovery relies on the use of mass spectrometry. Various mass spectrometry methods and instrumentation are utilized, each with distinctive advantages.4Koomen J. Hawke D. Kobayashi R. Developing an understanding of proteomics: an introduction to biological mass spectrometry.Cancer Invest. 2005; 23: 47-59Google Scholar One approach that has been popularized called surfaced enhanced laser desorption/ionization relies not on identification of proteins, but rather on the recognition of specific spectral patterns in a sample. Initial excitement using this approach for cancer screening was significantly dampened once issues of sensitivity and reproducibility were raised owing to limitations of this particular technology.5Ransohoff D.F. Lessons from controversy: ovarian cancer screening and serum proteomics.J Natl Cancer Inst. 2005; 97: 315-319Google Scholar More advanced mass spectrometry instrumentation allows substantial resolution, depth of analysis, and confidence in protein identification. Additional bioinformatics and other knowledge are required to analyze the proteins identified to determine biological relevance and formulate hypotheses.6States D.J. Omenn G.S. Blackwell T.W. Fermin D. Eng J. Speicher D.W. Hanash S.M. Challenges in deriving high-confidence protein identifications from data gathered by a HUPO plasma proteome collaborative study.Nat Biotechnol. 2006; 24: 333-338Google Scholar It follows therefore that proteomic research requires a multidisciplinary team with expertise across disciplines that include protein biochemistry, mass spectrometry, bioinformatics, statistics, and clinical medicine working closely together to avoid the many pitfalls in the data acquisition and analysis process.Sources of Specimens for Proteomic AnalysisThere are numerous strategies that may be implemented for biomarker discovery using proteomics (Figure 1). These include tumor and control tissue analysis to identify candidates such as secreted proteins and proteins shed from the cell surface, which may be detected in circulation as biomarkers. Such a strategy requires that efficient assays be available to test candidates in serum or plasma. Other strategies include the use of tumor cells and cell lines to identify such candidates in the case of cancer. The most direct but challenging path to biomarker identification is the analysis of serum and plasma based on well-designed studies of diagnostic or prediagnostic serum or plasma and appropriately matched controls. Mouse models of cancer potentially represent an efficient means for uncovering diagnostic markers as genetic alterations associated with human tumors can be engineered in mice. In addition, defined stages of tumor development, breeding conditions, and blood sampling can all be controlled and standardized to limit heterogeneity. The concept that plasma from engineered mouse models of cancer contains tumor-derived proteins that may be relevant as candidate markers for human cancer is attractive but largely untested.7Hingorani S.R. Petricoin E.F. Maitra A. Rajapakse V. King C. Jacobetz M.A. Ross S. Conrads T.P. Veenstra T.D. Hitt B.A. Kawaguchi Y. Johann D. Liotta L.A. Crawford H.C. Putt M.E. Jacks T. Wright C.V. Hruban R.H. Lowy A.M. Tuveson D.A. Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse.Cancer Cell. 2003; 4: 437-450Google ScholarProteomics for Biomarker DiscoveryThe discovery of disease biomarkers utilizing proteomics, especially for various cancers, has been a major area of proteomic research. The National Cancer Institute (NCI) through its Early Detection Network (EDRN) has spearheaded a multicenter effort for early detection of cancer using biomarkers.8Verma M. Srivastava S. New cancer biomarkers deriving from NCI early detection research.Recent Results Cancer Res. 2003; 163 (discussion 264–266): 72-84Google Scholar Several promising proteomic approaches for discovery have been utilized. Protein microarrays allow for high-throughput analysis for biomarker discovery. One application of protein microarrays is to coat glass slides with proteins, cell lysates, or antibodies and then incubate with a complementary biological sample (Figure 2A). Investigators have used antibodies to coat glass slides and then incubate serum from cancer and control patients to identify differences. A recent study utilized recombinant antibody microarrays to identify a protein expression signature in patients with gastric adenocarcinoma in the setting of Helicobacter pylori infection.9Ellmark P. Ingvarsson J. Carlsson A. Lundin S.B. Wingren C. Borrebaeck C.A. Identification of protein expression signatures associated with H. pylori infection and gastric adenocarcinoma using recombinant antibody microarrays.Mol Cell Proteomics. 2006; 5: 1638-1646Google Scholar Using proteins as the capture agent, NCI-60 cancer cell line lysates have previously been spotted on a protein microarray and then incubated with commercially available antibodies. From this analysis, 2 promising pathologic markers were identified that could distinguish colon from ovarian adenocarcinoma.10Nishizuka S. Charboneau L. Young L. Major S. Reinhold W.C. Waltham M. Kouros-Mehr H. Bussey K.J. Lee J.K. Espina V. Munson P.J. Petricoin 3rd, E. Liotta L.A. Weinstein J.N. Proteomic profiling of the NCI-60 cancer cell lines using new high-density reverse-phase lysate microarrays.Proc Natl Acad Sci U S A. 2003; 100: 14229-14234Google Scholar A promising area of protein microarray research is the identification of autoantibodies in human serum. This approach utilizes protein or lysate spotted arrays incubated with human serum from cancer and control patients. Microarrays that contain natural proteins derived from tumor cells have the potential to substantially accelerate the discovery of tumor antigens via a molecular signature for immune responses directed against protein targets in different types of cancer. In a study of colon cancer, microarrays printed with 1760 distinct protein fractions, prepared from the LoVo colon adenocarcinoma cell line, were hybridized with individual sera.11Nam M.J. Madoz-Gurpide J. Wang H. Lescure P. Schmalbach C.E. Zhao R. Misek D.E. Kuick R. Brenner D.E. Hanash S.M. Molecular profiling of the immune response in colon cancer using protein microarrays: occurrence of autoantibodies to ubiquitin C-terminal hydrolase L3.Proteomics. 2003; 3: 2108-2115Google Scholar A fraction that exhibited immunoglobulin G–based reactivity with 9/15 colon cancer sera was found to contain ubiquitin C-terminal hydrolase L3 (UCH-L3) by tandem mass spectrometry (ESI-Q-TOF). The highest levels of UCH-L3 mRNA among the 329 tumors of different types analyzed by DNA microarrays were found in colon tumors. Independent validation by Western blotting demonstrated that UCH-L3 antibodies existed in 19/43 sera from patients with colon cancer, and in 0/54 sera from subjects with lung cancer, colon adenoma, or otherwise healthy controls. These data point to the utility of microarrays printed with natural proteins for the identification of tumor antigens that have induced an antibody response in patients with specific cancersAnother advance in proteomics has been the use of isotope labeling for protein quantitative analysis. The basic principle is to label one sample with a “heavy” isotope and another sample with a “light” isotope and then use mass spectrometry to detect the relative abundance of labeled proteins between the 2 samples (Figure 2B). Isotope labeling has been used to detect relative protein concentration differences in pancreatic cancer. Using isotope-code affinity tag technology, the proteome of human pancreatic cancer tissue and juice has been analyzed.12Chen R. Pan S. Yi E.C. Donohoe S. Bronner M.P. Potter J.D. Goodlett D.R. Aebersold R. Brentnall T.A. Quantitative proteomic profiling of pancreatic cancer juice.Proteomics. 2006; 6: 3871-3879Google Scholar, 13Chen R. Yi E.C. Donohoe S. Pan S. Eng J. Cooke K. Crispin D.A. Lane Z. Goodlett D.R. Bronner M.P. Aebersold R. Brentnall T.A. Pancreatic cancer proteome: the proteins that underlie invasion, metastasis, and immunologic escape.Gastroenterology. 2005; 129: 1187-1197Abstract Full Text Full Text PDF Scopus (164) Google Scholar Differentially regulated proteins were found to be involved in functions encompassing angiogenesis, invasion, extracellular matrix destruction, and immunologic escape. Further validation of over-expressed proteins in these studies was performed using Western blot or immunohistochemistry techniques.One of the challenges to proteomic analysis has been the wide dynamic range of protein concentration. In human blood, protein concentration can span ≥10 orders of magnitude.14Anderson N.L. Anderson N.G. The human plasma proteome: history, character, and diagnostic prospects.Mol Cell Proteomics. 2002; 1: 845-867Google Scholar Extensive multidimensional fractionation of serum or plasma can be used to identify low abundance proteins, which may be enriched in biomarkers as they are less likely to represent acute phase reactants or classical blood proteins.6States D.J. Omenn G.S. Blackwell T.W. Fermin D. Eng J. Speicher D.W. Hanash S.M. Challenges in deriving high-confidence protein identifications from data gathered by a HUPO plasma proteome collaborative study.Nat Biotechnol. 2006; 24: 333-338Google Scholar Coupling isotope labeling with multidimensional fractionation is attractive. In our laboratory, we have taken such an approach to identify promising blood-based biomarkers for gastrointestinal cancers across 7 orders of magnitude of protein concentration. Multidimensional separation not only provides quantitative information, but can provide information on protein isoforms and cleavage products as elution patterns of proteins can be analyzed across the fractionation scheme (Figure 2C).From Discovery to ValidationAlthough proteomic analyses to date have identified differentially regulated proteins in disease versus control samples, validation of discovery findings remains a challenge. The discovery phase of potential biomarkers is just the beginning toward establishing a clinically useful test. The EDRN has set forth specific criteria for biomarker development for the early detection of cancer.15Pepe M.S. Etzioni R. Feng Z. Potter J.D. Thompson M.L. Thornquist M. Winget M. Yasui Y. Phases of biomarker development for early detection of cancer.J Natl Cancer Inst. 2001; 93: 1054-1061Google Scholar Indeed, these criteria can also be applied to nonmalignant diseases. The 5 phases of biomarker development start with preclinical exploratory studies and end with large-scale studies that show the impact of screening with biomarkers on decreasing disease burden at the population level (Figure 3). Epidemiology and biostatistics become increasingly important across the 5 phases.Figure 3Guidelines for biomarker development as set forth by EDRN.View Large Image Figure ViewerDownload (PPT)One of the obstacles to biomarker development is the availability and adequacy of testing platforms. A diagnostic test should ideally be one that is simple to run, have high throughput, and be of low cost. Although mass spectrom