Background: Proteomic-based discovery of biomarkers for disease has recently come under scrutiny for a variety of issues; one prominent issue is the lack of orthogonal validation for biomarkers following discovery. Validation by ELISA or Western blot requires the use of antibodies, which for many potential biomarkers are under-characterized and may lead to misleading or inconclusive results. Gelsolin is one such biomarker candidate in HIV-associated neurocognitive disorders.
Methods: Samples from human (plasma and CSF), monkey (plasma), monocyte-derived macrophage (supernatants), and commercial gelsolin (recombinant and purified) were quantitated using Western blot assay and a variety of anti-gelsolin antibodies. Plasma and CSF was used for immunoaffinity purification of gelsolin which was identified in eight bands by tandem mass spectrometry.
Results: Immunoreactivity of gelsolin within samples and between antibodies varied greatly. In several instances, multiple bands were identified (corresponding to different gelsolin forms) by one antibody, but not identified by another. Moreover, in some instances immunoreactivity depended on the source of gelsolin, e.g. plasma or CSF. Additionally, some smaller forms of gelsolin were identified by mass spectrometry but not by any antibody. Recombinant gelsolin was used as reference sample.
Conclusions: Orthogonal validation using specific monoclonal or polyclonal antibodies may reject biomarker candidates from further studies based on misleading or even false quantitation of those proteins, which circulate in various forms in body fluids.