Plasma Proteome Profiles Associated with Early Development of Lung Injury in Extremely Preterm Infants

Am J Respir Cell Mol Biol. 2024 Jul 25. doi: 10.1165/rcmb.2024-0034OC. Online ahead of print.

Abstract

The biological mediators which initiate lung injury in extremely preterm infants during early postnatal life remain largely unidentified, limiting opportunities for early treatment and diagnosis. This exploratory study used SWATH-mass spectrometry to identify bronchopulmonary dysplasia (BPD)-specific changes in protein abundance in plasma samples obtained in the first 72 hours of life from extremely preterm infants and bioinformatic analysis to identify BPD-related biological categories and pathways. Lasty, binary logistic regression analysis was used to test the BPD predictive potential of a base model alone (gestational age, birth weight, sex) and with the protein biomarker added, with bootstrap resampling used to internally validate protein predictors and adjust for overoptimism. We observed disturbance of key processes including coagulation, complement activation, development and extracellular matrix organisation in the first days of life in extremely preterm infants who were later diagnosed with BPD. In the BPD prediction analysis, 49 plasma proteins were identified which when each singularly was combined with birth characteristics had a C-index of 0.65-0.84 (optimism-adjusted C-index) suggesting predictive potential for BPD outcomes. Taken together, this study demonstrates that alterations in plasma proteins can be detected from 4 hours of age in extremely preterm infants who later develop BPD and that protein biomarkers when combined with three birth characteristics have the potential to predict BPD development within the first 72 hours of life.

Keywords: lung injury phenotype; prediction modelling; preterm lung injury; proteomics.