ABSTRACT
Biomarkers of cancer, such as carcinoembryonic antigen (CEA) for colorectal cancer and prostate-specific antigen (PSA) for prostate cancer, have a wide range of potential clinical utilities including early detection, disease monitoring, and prediction of treatment response. Discovery of novel tumor markers is an area of active research in oncology utilizing rapidly developing proteomic and genomic biotechnologies. Since proteins are closer to cell function than DNA and mRNA, it is logical to consider proteins (and peptides) as potential biomarkers. Measuring proteins is a complex task, however. First, proteins are dynamic. They are changed by interactions, posttranslational modifications, and perturbations from the environment. Thus, measurements are, strictly speaking, specific to cells, time, and environment. Second, in contrast to the digital nature of DNA and mRNA, proteins have three-dimensional structures and do not hybridize to their complements like DNA and mRNA. This makes the physical measurement of proteins and especially their quantification challenging. Third, proteins exist naturally in a mixture (e.g., in human serum), and some proteins (e.g., albumin and transferin) dominate with respect to abundance in the mixture. This necessitates the separation of proteins in the mixture, especially if potential biomarkers are not as abundant as many other proteins in the mixture.
