Proteomics is the study of the identity, properties, and functions of proteins. A proteome is the total set of proteins in a cell or organism, for example, just as a genome is the total set of genes in a cell or organism. Proteomics is more complicated than its sister science of genomics for a number of reasons. Proteins are gene products that outnumber genes many-fold; in humans, for example, some 20,000 to 25,000 protein-coding genes may produce more than 1 million different protein species. Many genes can produce more than one version of the proteins they encode. And, unlike the genome of an organism, which remains relatively stable over time, the protein composition changes constantly through new production, elimination, and modification of proteins in response to internal and external stimuli. Thus, the proteome can vary widely among individuals based on such factors as age, gender, diet, exercise, and disease.
Protein Signatures
Cancer cells can secrete specific proteins or protein fragments into the blood and other body fluids (urine, saliva, sweat). The hope for proteomics as it relates to cancer is to identify protein 'signatures'-sets or patterns of proteins-in readily accessible fluids that can provide information about risk, presence, and disease progression. One hope is that assays based on protein signatures will be developed that allow screening for early detection of cancer or cancer risk. Another is that assays can be used to predict likelihood of response to particular treatments, detect response, anticipate adverse effects, and screen for early signs of disease recurrence.
In short, success in some or all of these initiatives would greatly advance the ability to tailor cancer treatment to individual patients. (For more information on personalizing care through proteomics, see the Education Session, Proteomics Pathway Analysis for Tumor Classification and Therapy, being presented on June 8 at the 2010 ASCO Annual Meeting.)
Monitoring for cancer-related proteins has been a part of clinical practice for some time-eg, in the form of screening for ovarian cancer using the protein CA-125 and screening for prostate cancer using the protein prostate-specific antigen (PSA). Relatively high rates of false-positive and/or false-negative results mark the use of these and other secreted single proteins as screening biomarkers. For example, elevated blood levels of CA-125 are found in only 50% to 60% of women with early-stage ovarian cancer, but elevated levels are also found in noncancer conditions, such as pregnancy and endometriosis. (Watch for abstract 5061, being presented on June 5 at the ASCO Annual Meeting, for further information on the use of proteomics to diagnose ovarian cancer.) Similarly, only 25% to 30% of men with elevated PSA are found to have prostate cancer on biopsy, and 15% of men with PSA levels in the normal range are found to have prostate cancer. Evidence exists that using multiple proteins in protein signatures may greatly increase accuracy in cancer detection.
Key Techniques
The two main approaches in identifying candidate proteins/signatures for use in cancer screening consist of protein identification and pattern recognition. Both require high-powered computing and bioinformatics systems to process and analyze the enormous amount of data produced in proteomics studies (Fig. 1). Techniques for protein identification include the use of antibodies that bind to proteins thought or known to be overexpressed in particular cancers. Antibodies in microarray wells are exposed to fluid containing the proteins (eg, blood, urine). The proteins that bind to the antibodies can then be detected by fluorescence microscopy, with the amounts in samples from cancer patients being compared with amounts from samples from those without cancer. Another technique is gel electrophoresis, in which proteins are separated according to mass and electric charge; those that are more common in samples from cancer patients are identified by enzyme-linked immunosorbent assay (ELISA), which uses antibodies to identify the proteins.
A primary technology in pattern recognition studies is mass spectrometry (MS), which determines the masses and relative quantities of all proteins in a sample. The protein profiles or signatures produced by this technique can differentiate samples from patients with cancer from samples from those without cancer, but the technique does not identify the proteins constituting the signature.
Chromatography plus tandem MS can be used to identify the individual proteins in an MS protein signature. Identifying the individual proteins is important to avoid incorrect characterization of a signature as intrinsically representing cancer-related differences, when it might actually reflect differences in the way cancer samples and noncancer samples are collected, stored, and processed.
Challenges to Proteomics Progress
The latter problem highlights one of the many difficulties facing proteomics initiatives. Proteins are much more likely to be altered during isolation, storage, and handling than are DNA or other molecules of interest, jeopardizing the accuracy of findings and the conclusions based on them. In response to this problem, researchers and organizations are working to develop standards and best practice guidelines for collecting and processing patient samples.
Another difficulty is the limited sensitivity of mass spectrometers. Although these devices can detect proteins that are 100 to 10,000 times less common than other proteins in a sample, proteins produced by cancer cells are often present in amounts that may evade detection by MS. Work is ongoing to improve MS sensitivity. In addition, the development of antibodies or other molecules (collectively, affinity reagents) that accurately identify proteins is an expensive and time-consuming process that lags behind the detection of proteins and protein signatures used to differentiate cancer from noncancer specimens.
Promising results have been achieved in identifying protein signatures for many cancers. Due in part to differences among laboratories in methods of processing specimens and analyzing proteins, many of these findings have been difficult to replicate with currently available shared technology. Promising signatures must be validated in patient populations other than those in which they were initially derived and findings must be replicable by different laboratories before large-scale trials can accurately assess the potential clinical utility of these signatures.
NCI Initiatives
The National Cancer Institute provides major support for proteomics and its potential to advance cancer research and treatment in the form of intramural and extramural initiatives (see box). The Antibody Characterization Program develops antibodies and other affinity reagents. The Clinical Proteomic Technologies for Cancer initiative comprises: the Clinical Proteomic Technology Assessment for Cancer program, a private/public collaboration assessing MS and affinity capture platforms; the Advanced Proteomic Platforms and Computational Sciences initiative, devoted to the development of new technologies and computational approaches for analysis and sharing of data; and the Proteomic Reagents and Resources Core, a source for affordable, well-characterized, and validated reagents.
The Early Detection Research Network is an extensive collaboration of researchers developing and testing promising biomarkers/technologies for early detection of cancer. Proteomics-based research projects have produced a number of promising candidate markers for lung, pancreas, and prostate cancers. The Biomedical Proteomics Program is devoted to identification and characterization of protein signatures of cancer cells/tissues and the application of technologies directly to diagnosis, monitoring of side effects of treatment, and treatment improvements.
Proteomics Information: Resources
National Cancer Institute
NCI FactSheet: Proteomics and Cancer: http://www.cancer.gov/cancertopics/factsheet/detection/proteomics
Clinical Proteomic Technologies for Cancer (CPTC) initiative: http://proteomics.cancer.gov and http://proteomics.cancer.gov/library/primer.asp
Clinical Proteomic Technology Assessment for Cancer (CPTAC) network: http://proteomics.cancer.gov/programs/CPTAC/
Other Resources
Yurong G, Fu Z, Van Eyk JE: A proteomic primer for the clinician. Proc Am Thorac Soc 4:9-17, 2007.
Hanash SM, Pitteri SJ, Faca VM: Mining the plasma proteome for cancer biomarkers. Nature 452:571-579, 2008.
Tonack S, Aspinall-O'Dea M, Neoptolemos JP, et al: Pancreatic cancer: Proteomic approaches to a challenging disease. Pancreatology 9:567-576, 2009.
Wong SCC, Chan CML, Ma BBY, et al: Advanced proteomic technologies for cancer biomarker discovery. Exp Rev Proteomics 6:123-134, 2009.
Larkin SET, Bashar Z, Taylor MG, et al: Proteomics in prostate cancer biomarker discovery. Exp Rev Proteomics 7:93-102, 2010.
Gast M-CW, Schellens JHM, Beijnen JH: Clinical proteomics in breast cancer: A review. Breast Cancer Res Treat 116:17-29, 2009.
