Mass spectrometry-based proteomics of human cerebrospinal fluid. Biomarker discovery and verification in multiple sclerosis
MetadataVis full innførsel
The main focus of this study was to use mass spectrometry-based proteomics to study protein abundance in cerebrospinal fluid (CSF) to reveal proteins that could serve as biomarker candidates in multiple sclerosis (MScl). By combining a CSF pooling strategy and label-free relative quantification we discovered 65 proteins of differential abundance between MScl patients and controls. A selection of 17 biomarker candidates was further subjected to two independent verification steps: using stable isotope dimethyl labeling coupled to Accurate Inclusion Mass Screening (dimethyl-AIMS) and Stable Isotope Dilution Selected Reaction Monitoring (SID-SRM) for targeted quantification. The SID-SRM study included a larger patient cohort of 125 cases and controls. To our knowledge, this is the first report of a larger SRM verification study for biomarker candidates in MScl. The most interesting results from the biomarker discovery and verification study were the significantly decreased abundance of Apolipoprotein D, Cystatin C, Kallikrein-6 and Alpha-1-acid glycoprotein 1 in MScl patients compared to controls. Furthermore, we performed a comprehensive characterization of the normal CSF proteome. We identified 18, 807 peptide mapping to 1987 proteins by applying immuno-affinity depletion and SDS-PAGE for enhanced proteome coverage. We obtain a comprehensive set of reference proteins that could further be used for investigations in MScl. The experiment gave us the opportunity to examine the size distribution on the SDS-PAGE gel of a selection of biomarker candidate proteins in normal CSF, with the aim to reveal potential protein variants (isoforms, truncation products and proteolytic processed products). We hypothesized that the identification of non-tryptic peptides could indicate truncation products of proteins in CSF. 13 and nine non-tryptic peptides were identified for the biomarker candidates Cystatin C and Secretogranin-1, respectively. This information had immediate utility for investigation in MScl. Based on the observed spread in size distribution of biomarker candidate proteins in normal CSF, we aimed to obtain quantitative information of proteins present in both high and low mass fractions on the gel and further target these proteins to obtain an abundance ratio between RRMS (patients) and OIND (controls), and investigate if these ratios differed for the same protein. The most striking observation was the opposite regulation level in CSF of Secretogranin-1.