Quest for a new generation of biomarkers using quantitative proteomics

Advances in proteomics have constantly altered our understanding of cell biology and biochemistry by providing new approaches and techniques to identify complex proteomes, protein-protein interactions and post-translational modifications. Additionally, proteomic approaches are believed to have enormous potential for discovery of disease biomarkers that can provide diagnostic, prognostic and therapeutic targets and address important problems in clinical and translational research. Unfortunately, to date the development of new assays based on biomarkers discovered by proteomics has been unsuccessful mostly due to low sensitivity and specificity of various candidate biomarkers.

Within the proteomics discovery phase, quest for clinically relevant protein biomarkers has presented a number of candidates that need further verification and validation to establish their proper usefulness. Such requirements need clinically orientated studies which are complex and time consuming and currently, there is no accepted process of validation similar to therapeutical clinical trials. Therefore, it is not easy to recognise those biomarkers that have outstanding potential. Hence, there is a need for a new strategy of validation process that could provide information on assay development and subsequently impact clinical decisions leading to an improved patient outcome. More importantly, there is a need for new biomarker discovery strategies that use reliable fractionation and quantitative analytical tools and discovery fluids and tissues that are reflective of clinical needs and concerns. Proteomics technologies are constantly being improved to increase throughput, sensitivity and peptide/protein quantification. Such a trend is well reflected by rapid advancement in mass spectrometry instrumentation and related studies. This innovation is expected to help discover and more precisely allow better selection of biomarkers with probability of higher selectivity and specificity. Several quantitative proteomic methodologies are aimed at studying differential expression of proteins in body fluids and tissues to pin-point candidate biomarkers which may allow early stage detection of a disease and help to improve the cure rates as well as overall survival time of the patient.

Additionally, discovery of any prognostic biomarkers may help to identify those patients who may likely recur with a particular indication and who will need adjuvant treatments. Companion diagnostic assays may help to identify patients who are likely to respond to a particular treatment and predictive biomarkers may provide an early assessment of the patient’s response to that particular therapy. New generation of additional protein biomarkers may help to identify patients who will experience a severe adverse event and who should not be considered for that drug treatment and in other patients help to define the appropriate dosage for a drug. Because the protemic discovery phase is a quick and easy step, the improvements of proteomic tools significantly facilitate more effective and targeted validation and assay development process. Hence, well designed follow-up studies in clinical environment may yield positive answers to critical questions in relation to the therapy.

Clinicians, drug discovery scientists, pathologists, researchers and technologists may have different definitions for a biomarker. Generally, a biomarker may be defined as a biological substance that can be used to detect a disease, measure its progression or evaluate effects of a drug treatment. More importantly, it should be readily accessible, provide sufficient sensitivity and specificity to accurately distinguish between true positives, false positives, and false negatives outcomes when tested with a view to providing a clinical benefits to the patient. In broader terms, there may be several biomarker types including: Translation biomarker in preclinical and clinical settings, Disease biomarker that relates to a clinical outcome or measure of disease, Efficacy biomarker that reflects a beneficial effect of a given treatment, Staging biomarker that may distinguish between different stages of a chronic disorder, Surrogate biomarker that is regarded as a valid substitute for a clinical outcomes measure, Toxicity biomarker that reports a toxicological effect of a drug on an in vitro or in vivo system, Mechanism biomarker that reports a downstream effect of a drug and Target biomarker that reports interaction of the drug with its target1. A variety of methods have been applied to identify new biomarkers. Unlike the techniques focusing on DNA or mRNA, the approaches characterising proteins and considering protein biomarkers have the advantage that the proteins are directly responsible for the function and biological phenotype. Additionally, current improvements in quantitative proteomics facilite the quest for the new generation of protein biomarkers that should reveal a number of concepts including mechanistic correlations of the biomarkers, biomarker accessibility and stability in body fluids, baseline levels of the biomarkers in control samples, and potential interferences in the chemical matrix of the measurement2.

Many diseases are caused by disturbance in genetic balance, gene transcription and regulation which is associated with perturbations in cellular signalling and metabolic pathways. These changes result in altered protein composition and cellular phenotypes. Consequently, the proteins are shed into extracellular fluid including tissue interstitial fluids and blood plasma. Whilst tissue interstitial fluids are in direct contact with tissue/cells via transfer of molecules, the composition of blood plasma results from the communication with tissue interstitial fluids. On the other hand, the blood plasma influences the composition of other body fluids. It is important to realise that relative concentration of biomarkers is highest in tissue interstitial fluids which in turn drain into lymph and lymph from different tissues finally merge and drain into blood. As a result, the final concentration of biomarkers in blood is significantly lower compared to interstitial fluids. Hence, various body fluids represent a more or less enriched source of biomarkers which can be understood in context of plasma and preferably analysed using proteomic tools3. Because plasma is easily available following non-invasive blood taking, many studies aimed in identifying disease biomarkers have been performed directly on blood plasma samples4.

The first generation of quantitative tools have been used in the past to study differential expression of proteins in body fluids, cells and tissues. They have included classical 2-DE coupled to mass spectrometry for identification of protein profiles. The major bottleneck of this approach was the preference for analysis of high abundant proteins which in turn resulted in identification of candidate biomarkers which were not selective due to low sensitivity and specificity. Additionally, classical 2-DE based approach suffered from very low dynamic range of quantification using various staining procedures and not very high level of reproducibility. The use of fluorescence protein labelling and run of the mixture of two different samples in one gel known as difference in gel electrophoresis (DIGE) has provided researchers with a ‘handle’ to look closely at the biomarkers of interest but did not completely solve the problems associated with gel-based techniques as well as protein sample complexity or high dynamic range of protein abundances in contrast to rather low dynamic range of protein detection and quantification5.

Hence, newer technologies were needed if the field of biomarker discovery was to expand. The problems appear to be overcome by modern mass-spectrometry based quantification methods that have gained popularity over the last five years. The determination of protein abundances in samples is performed using stable isotope labelling techniques including stable isotope labelling by amino acids in cell culture (SILAC) or isotope-coded protein labels (ICPL) prior protein fractionation. In addition, an improved approach using amine-reactive isobaric tags for relative and absolute quantitation (iTRAQ) has been developed with potential usefulness. This technique is based on chemically tagging the N-terminus of peptides generated from protein digests that have been isolated from cells cultured under different experimental conditions6,7. A specific way to introduce isotope label into peptide is the use of trypsin-catalysed incorporation of 18O during protein digestion. The incorporated labels create specific mass tags that can be recognised by mass spectrometry and at the same time provide the basis for quantification. Unlike relative quantification, absolute quantification (AQUA) can be achieved using internal standard by adding known quantity of stable isotope-labeled standard peptide to a protein digest and subsequent comparison of mass spectrometric signals of standard peptide and corresponding endogenous peptide in digest. The specific modification of AQUA is a method called multiple reaction monitoring (MRM) in which the mass spectrometer monitors not only intact peptide mass but also one or more specific fragmentations of those peptides. The combination of retention time, peptide mass and fragment mass eliminates ambiguities in peptide assignment and allows for higher range of quantification. Development of hybrid methods coupling MRM assay with enrichment of proteins by immuno-depletion or enrichment of the peptides by antibody capture and further improvements may be capable of extending MRM method to full dynamic range of plasma. With this view in mind, MRM measurement of plasma peptide may provide a rapid and specific assay platform for biomarker validation8,9.

Due to problems posed by isotope labelling methods such as removal of contaminating reagents, sample loss, costs and reproducibility, label-free methodologies for quantitation were introduced. Quantitation by spectral counting of MS/MS spectra of peptide derived from a protein and comparison of spectral counts of a protein between different MS runs provided relative quantitation. One of the disadvantages of such technology was the reliance on software tools used for summation of peptide spectra, normalisation of data and removal of background noise9. Other group of label–free method is presented by protein microarrays. Each array can contain hundreds or thousands of immobilised proteins or antibodies and the specificity of the protein or antibody is crutial. They are versatile tools for examining large numbers samples for biomarkers in various environments such as drug development and clinical studies. A major bottleneck in the biomarker development process exists in validation of candidate markers emerging from genomic and proteomic screens. Protein and tissue microarray technologies provide a high-throughput platform for various investigations including pathological investigation, thus allowing analysis of several hundreds of samples simultaneously. Moreover, the advent of digital slide technology has afforded an unparalleled opportunity for archiving of valuable histopathological and immunohistochemical specimens. Such digital images offer a unique opportunity to provide a paradigm shift in relation to interpretation of such data via the use of automated image analysis algorithms10.

The application of a wide range of proteomics methods to the discovery of new protein biomarkers should serve to identify early markers of disease as well as predict and monitor patient response to therapeutic intervention. Examples of this include early detection of various cancers including breast11, pancreatic12 and prostate13. These combined processes could make the most effective use of existing drugs and therapeutic regimes and the development of new effective therapeutic strategies14. In addition, the continued development of protein biomarkers to predict and monitor drug toxicity is likely to have a major impact on the drug development process. It is increasingly apparent that the validation of biomarker expression and qualification of utility is of critical importance and remains challenging.

Proteomic biomarkers are increasingly used as key decision tools in therapeutic drug development. The promise of these biomarkers lies in the potential to collect information about candidate drugs earlier in the development allowing shortening of timelines to reach a critical decision point. Safety and toxicity markers may help stop programs prior to costly pivotal studies, whereas dose estimates and patient stratification may reduce the size and duration of such studies15. The ability to improve the early detection and clinical management of a disease represents a significant goal for clinical diagnostics. The ability to identify novel biomarkers that are associated with that disease; to define the appropriate biomarkers and assays that may be useful for routine laboratory use; and to validate these biomarkers for clinical applications represents a significant challenge in molecular medicine today16. Biomarker validation studies are best accomplished in prospective clinical trials performed outside of the research environment.

The last five years has witnessed significant growth in quantitative proteomic technologies including mass spectrometry-based techniques and protein microarrays. Mass spectrometry-based methods appears to be more suited for discovery and identification of new biomarkers because no previous knowledge of the protein sample is required except for protein sequence database. On the contrary, protein or tissue microarrays may be applied rather for validation of a new biomarker with good potential for subsequent clinical application. The major distinction of proteomic approaches is the capability to analyse body fluids, cells or tissues at the level of hundreds of proteins in a single run, while current clinical tests usually monitor only one or a few proteins or analytes. Although the recent development of proteomics tools is very impressive and giving us prospects for a new generation of suitable and hopefully applicable protein biomarkers, it is necessary to realise that we are still far from clinical quantitative proteomics. There are numbers of issues that need to be addressed during the biomarker discovery phase including reproducibility of experimental set up, detection limits and development of appropriate statistical approaches in order to properly evaluate data and obtain correct biological interpretation. It is no doubt that human body fluid proteome analyses belong currently to the very active areas of biomarker research. Furthermore, cancer disease biomarkers appear to be very promising and parallel a combination of high sensitivity/low specificity biomarkers with the less sensitive/highly specific ones may represent preferred and productive approach in a clinical setting. Encompassing all the above criteria, there appears to be a significant hope in pursuing the new generation of biomarkers using quantitative proteomics for clinical benefits.

References

1. Boguslavsky, J. Biomarkers as checkpoints. Drug Disc Dev. 2004, 7: 26-33.
2. Hurko, O. Understanding the strategic importance of biomarkers for the discovery and early development phases. Drug Disc World. 2006, 7: 63-74.
3. Ahn, S-M., Simpson, R.J. Body fluids proteomics: Prospects for biomarker discovery. Proteomics Clin. Appl. 2007, 1: 1004- 1015.
4. Hu, S., Loo, J., Wong, D.T. Human body fluid proteome analysis. Proteomics 2006, 6: 6326-6353.
5. Meyer, H.E., Stühler, K. High-performance proteomics as a tool in biomarker discovery. Practical Proteomics 2007, 1: 18-26.
6. Ross, P.L., Huang, Y.N. et al. Multiplexed protein quantitation in Saccharomyces cerevisiae using amine-reactive isobaric tagging reagents. Mol. Cell. Proteomics 2004, 3: 1154-1169.
7. Skalnikova H., Rehulka, P. et al. Relative quantitation of proteins fractionated by the ProteomeLab PF 2D system using isobaric tags for relative and absolute quantitation (iTRAQ). Anal. Bioanal. Chem. 2007, 389:1639-1645.
8. Hu, Z., Hood, L., Tian, Q. Quantitative proteomic approaches for biomarker discovery. Proteomics Clin. Appl. 2007, 1:1036-1041.
9. Bantscheff, M., Schirle, M., Sweetman, G., Rik, J., Kuster, B. Quantitative mass spectrometry in proteomics: a critical review. Anal. Bioanal. Chem. 2007, 389: 1017-1031.
10. Pollard H.B., Srivastava, M. et al. Proteomics Clin. Appl. 2007, 1: 934-952.
11. Metodiev M., Alldrige, L. Phosphoproteomics: A possible route to novel biomarkers of breast cancer. Proteomics Clin. Appl. 2008, 2: 181-194.
12. Aspinall-O’Dea, M., Costello, E., The pancreatic cancer proteome – recent advances and future promise. Proteomics Clin. Appl. 2007, 1:1066-1079.
13. Hellström, M., Jonmarker, S. et al. Proteomics in clinical prostate research. Proteomics Clin. Appl. 2007, 1:1058-1067.
14. Tjalsma, H., Schaeps R. M J., et. Al. Immunoproteomics: From biomarker discovery to diagnostic applications. Proteomics Clin. Appl. 2008, 2:167-180.
15. Harris, L., Fritsche, H. et al. American Society of Clinical Oncology 2007. Update of Recommendations for the Use of Tumor Markers in Breast Cancer. J. Clin. Oncol., 2007, 25: 5287-5312.
16. Hanash, S.M., Pitteri S.J., Faca V.M. Mining the plasma proteome for cancer biomarkers. Nature, 2008, 452: 571-579.