|
Abstract INTRODUCTION: Reliable specific biomarkers are essential for the improvement of diagnosis and accelerated drug discovery. Biofluids like plasma and serum are unique sources of potential biomarkers, capable of revealing an organism’s status. However, despite major efforts and significant investment, only a limited number of new biomarkers stand up to the scrutiny of validation via the analysis of large sample cohorts. Serum extraction is a facile procedure capable of delivering a high consistency of processed samples independent of origin and operator. However, variables such as the collection tube material, blood clotting time, centrifugation speed and temperature etc., can have detrimental effects on the concentration of so-called "contaminating proteins" derived from blood cells.
OBJECTIVE and METHODS: Here we describe a set of novel fast and reproducible low-flow LC-MS/MS methods suitable for label-free quantification of proteins in crude serum samples using the UltiMate 3000 RSLCnano system coupled to a Q Exactive HF-X HRAM MS. The high reproducibility of these methods permitted the study of the impact of both technical (run-to-run) and biological variation. Alterations in serum protein abundance was observed that are likely caused by artifacts in sample preparation corresponding to serum from four different serum sources.
RESULTS: The methods afforded throughputs of 180, 100, 60, 30 and 24 samples per day, respectively. The shortest permitted the profiling of more than 140 proteins and identification of 1340 peptide groups whilst the longest method, comprising a 60 min total cycle time, enabled the profiling of more than 250 proteins based on 2900 peptide groups. 233 proteins (90%) were quantified with CVs below 15%. Analysis of biological replicates revealed that 90% of the quantified proteins had a sample to sample concentration variation below 45%. This variation increased when data for serum samples obtained from 4 different sources were combined. However, the concentration variation of the majority of proteins remained below 60%. Among the proteins that showed a high concentration variation (> 4 times) in serum obtained from different sources, were major components of blood cells, e.g. hemoglobin subunits, platelet proteins, or common proteomics sample preparation contaminants (Keratins).
CONCLUSION: These data demonstrate how robust and precise quantitative low-flow LC-MS/MS serum proteome profiling combined with careful experimental planning may help to avoid identification of false positive biomarker candidates. However, deeper proteome profiling of serum is required to identify biologically relevant drug- or disease-specific biomarkers.
DISCLAIMER: for research use only. |