A major challenge of proteome research is detecting disease biomarkers in biological fluids. Disease markers, described by Adkins as "proteins that undergo a change in concentration or state in association with a biological process or disease," can be key factors for early diagnosis, monitoring response to therapy, and detection of relapse of most types of cancers as well as of other diseases . However, disease biomarkers are usually present at relatively low concentrations (ng/mL or less). Serum and plasma offer particularly promising resources for biomarker discovery because collection of these samples is minimally invasive and the blood is thought to contain the majority of protein constituents found in the body [2-4].
The complex nature of serum and plasma, and the presence of a modest number of proteins at mg/mL levels, e.g., 0.1-401 mg/mL, make detection of low-abundance disease biomarkers very challenging. Although protein-rich plasma and serum have been used as diagnostic tools for decades, there are still fewer than 1000 distinct proteins identified, and only a small portion of these known proteins have been shown to have diagnostic potential [5,6]. Traditional 2-D gel methods are most commonly used for most quantitative proteome analyses. However, when applied to analysis of serum or plasma, sample load capacity of 2-D gels is severely limited by the presence of high-abundance proteins in addition to other well-known limitations. Consequently, prefractionation methods as well as alternative non2-D methods are being used to divide proteomes into smaller subsets to identify as many proteins, or patterns of proteins, as possible and detect low-abundance disease biomarkers [3, 7, 8].
A highly promising first step for most analysis strategies of serum or plasma is to deplete as many of the major proteins as possible. A range of methods to deplete high-abundance proteins have been evaluated in the past, specifically Cibacron blue, - a chlorotriazine dye which has a high affinity for albumin [9-11], - as well as Protein A or G to deplete immunoglobulins [12-13]. While dye-based kits bind the majority of albumin, they often also bind a large number of nonspecific proteins, resulting in potential losses. This nonspecific binding probably includes both proteins that bind to the dye as well as some minor proteins that bind albumin. Ciba-cron blue and other dye-based methods are known to bind proteins with nucleated binding domains as well as via ionic and hydrophobic interactions [9-11]. In addition, the buffers typically used with these kits are unlikely to dissociate minor proteins that are complexed with albumin .
Conversely, Protein A and G may not bind all of the immunoglobulin subgroups, thereby leaving a portion of these very heterogeneous abundant proteins behind. Individual antibody methods have proven to be more specific in depleting targeted proteins and give more complete removal of abundant proteins. Monoclonal antibodies are a promising choice for their high specificity, but they may not recognize all forms of the targeted protein, including proteolytic fragments and PTM forms of the antigen . Polyclonal antibodies, on the other hand, are more likely to deplete multiple structural forms of a protein. Ideally, for biomarker discovery it is desirable to deplete as many high-abundance proteins as possible while minimizing incidental losses of nontargeted proteins. Thus, a recently developed depletion method that mixes six high-specificity polyclonal antibodies to rapidly and efficiently deplete multiple proteins in a single purification step is particularly promising . A commercial version of this method, multiple affinity removal system (MARS), recently became available and was used for these experiments.
In this study, we compared depletion of six abundant human blood proteins using a polyclonal HPLC column with: older Cibacron blue/Protein A or G depletion methods, two prototype antibody spin columns, and no depletion. The most critical considerations for major protein depletion are the extent to which unde-sired nonspecific losses of proteins occur during major protein depletion, and the potential positive impact that major protein depletion has on detection of lower abundance proteins. Results using normal human serum and plasma show that the HPLC column containing polyclonal antibodies to six abundant human proteins can efficiently and reproducibly deplete about 85% of the total protein. Although this depletion allows larger amounts ofserum or plasma to be analyzed, the next most abundant proteins subsequently interfere with detection of very low-abundance proteins by masking major regions of the gel. In addition, most blood proteins are structurally heterogeneous due to physiological and/or artifactual proteolysis, and varying degrees of PTMs. As a result, most proteins are separated into many spots on 2-D gels, making detection more difficult. Hence, non2-D gel analysis methods are more likely to detect low-abundance proteins.
Was this article helpful?