NINDS Clinical Proteomics Unit

National Institute Of Neurological Disorders And Stroke

Investigators

Linked publications, trials & patents

Abstract

The NINDS Clinical Proteomics Unit (CPU) continues a long-term collaboration with Dr. Avindra Nath and colleagues in the Section of Infections of the Nervous System (SINS) to quantify differences in mitochondrial protein expression in patients with Chronic Fatigue Syndrome (CFS). Nath and colleagues produced mitochondrial respiration data from patient samples suggesting impaired mitochondrial function may play a role in CFS. In 2019, CPU conducted a pilot isobaric mass tagging study, using PBMCs as an inflammatory model, to quantify changes in the mitochondrial proteome directly from whole cell lysates. We were able to quantify a few thousand proteins, only 16% of these proteins were linked with mitochondria-related Gene Ontology (GO) terms. We used this mass tagging strategy to quantify differences in the mitochondrial proteome of 3 healthy and 3 CFS patients. Weighted gene co-expression network (WGCNA) analysis suggested several distinct modules of proteins with correlated expression, both up and down in CF. Submission of lists of proteins in each module for STRING analysis showed many of the proteins had known functional interactions and GO analysis showed mitochondria-relatedness of a subset of each module of proteins. In 2020, we sought to expand the number of control vs CF samples to twenty by analyzing them as 2 groups of 10. We hypothesized that the 2 groups of 10 patient samples could be joined together by including a pool of all 20 samples as a common reference channel for normalization. We used two 11-plex isobaric mass tagging kits to label whole cell PBMC lysates from 8 healthy and 12 chronic fatigue patients taken as two groups of 4 control vs 6 CF with the pooled sample as the common 11th sample in each group. We used the 2D LC-MS3 methods developed previously to collect three sets of triplicate analyses. Nine replicate analyses each contained 12 high pH RPLC fractions from group A&B, for a total of 216 raw files collected over 18 days. MaxQuant, a robust bioinformatics software package designed for large, high-resolution datasets, was used to identify protein constituents and quantify expression their levels using peptide level data. Detailed inspection of the peptide and Protein Groups reports for each triplicate analysis showed high precision of quantitation within single multiplexed experiments. CPU is currently collaborating with Drs. Fahad Almsned and Kory Johnson, NINDS Bioinformatics Section, for help with statistical analysis using MSstats TMT, an R-based statistics package for protein significance analysis in proteomics experiments with TMT labeling. Principal component analysis of replicates showed strong batch effects and we are currently exploring normalization strategies.