Oral Presentation HUPO 2019 - 18th Human Proteome Organization World Congress

The sweet separation between bacterial and viral infections by glycopeptide profiling (#112)

Esther Willems 1 , Jolein Gloerich 1 , Anouk Suppers 1 , Ronand de Groot 1 , Alain J. van Gool 1 , Hans J.C.T. Wessels 1 , Marien I. de Jonge 1
  1. Radboudumc, Nijmegen, GELDERLAND, Netherlands

Background

Accurate treatment of febrile children is an important challenge faced by healthcare providers worldwide. Current diagnostic tests, based on C-reactive protein or procalcitonin, are unable to unambiguously distinguish life-threatening bacterial infections from self-resolving viral infections. This often results in empirical treatment with broad-spectrum antibiotics. Due to the emergence of antibiotic resistance there is a strong demand for improved diagnostics. Several studies have been conducted to identify better protein biomarkers but thus far failed to enhance diagnostics. The vast majority of proteins are glycosylated, which offers attractive possibilities for biomarker research since aberrant glycomics signatures have been associated with many genetic and acquired human diseases. Here, we applied glycopeptide profiling to identify site-specific glycosylation changes in response to bacterial and/or viral infection.

 

Methods

Blood plasma samples from 92 febrile pediatric patients, undergoing either a bacterial or viral infection, and 43 samples from healthy individuals were subjected to tryptic digestion and subsequent glycopeptide enrichment. Intact glycopeptides were analyzed by reversed phase LC-MS/MS (maXis plus; Bruker Daltonics). Data pre-processing, integration, and analysis was performed in MATLAB (MathWorks).

 

Results

We detected 3682 unique glycopeptide features that were consistently detected in at least 75% of the samples from any sample class. A library of glycan- and peptide-moiety identifications from a previous glycopeptide profiling study were mapped onto the feature matrix to infer glycosylation signatures for multiple sites in 37 different proteins, each carrying between 1 to 20 different glycan forms. By means of Partial Least Squares – Discriminant Analysis we obtained a subset of differential glycopeptides, which enabled a clear discrimination between sample classes.

 

Conclusion

The selected subset of differential glycopeptide features was able to adequately separate bacterial from viral infected patients. Current work is focused on interpretation of the glycosylation changes in response to disease and implementation of these results into novel diagnostics.