Abstract
Introduction: Reliable biomarkers could aid antibiotic decisions in paediatric LRTI. The proteomic interrogation of airway secretions, which may better reflect the local inflammatory environ, could provide them
Methods: Protein was extracted from BAL (Bronchoalveolar Lavage) supernatant (from non-bronchoscopic BAL) using a TrisHCL/SDS based method. A mBCA assay measured total protein concentration and extracts ran on total protein gel for quality control. Extracts were pooled due to low sample volume and underwent chemical labelled (iTRAQ & TMT) LC/MS Proteomics. Proteins were identified using UniProt database with analysis in R & IPA
Biomarker selection was by (1) ranking eligible proteins by the size of fold change & significance, then (2) promoting proteins that showed similar characteristics in parallel experiments run with archived samples from murine models of LRTI and consistent good diagnostic performance (AUC ≥0.8) identified at systematic review
Results: BAL from 10 children with LRTI (6 bacterial & 4 viral determined by expert review panel) and 6 controls was obtained. 1176 proteins identified. 975 proteins in the bacterial & 889 in the viral groups were differentially expressed compared to control. 500 proteins differentiated bacterial and viral LRTI. Bioinformatics identified top networks in both groups related to terms acute phase response, LXR/RXR activation, ER2 signalling & inflammation. The final biomarker panel consisted of: HP, AGT, PIGR & AX2
Conclusions: This novel minimally invasive BAL and proteomic methodology has identified HP, AGT, PIGR & AX2 as potential biomarkers to differentiate viral and bacterial LRTI within the paediatric BAL proteome.