Early Multiomic Signals of Severe Bronchopulmonary Dysplasia in Extremely Low Birthweight Premature Infants

Le résumé fourni par la source

Abstract Background: Biomarker-based estimation of the risk of developing severe bronchopulmonary dysplasia (SevBPD) near birth may facilitate personalized therapy in extremely low birthweight (ELBW) newborns. Methods: We collected blood samples within 72 hours of birth in a prospective observational cohort of 302 ELBWs and performed whole exome sequencing (WES), mRNA and miRNA transcriptomics, and proteomics to identify multiomic markers of early SevBPD. We integrated these omics with advanced systems biology and novel network analyses to gain insight into predictors of SevBPD or resilience to disease (NoBPD). Results: We identified multiomic predictors of risk or resilience to SevBPD, including extensive alterations in the immune response, extracellular matrix integrity, and antioxidant processes that may underlie early BPD and predicted disease development weeks before it could be clinically diagnosed (Fig 1). WES identified significant differences in the mucin genes MUC19 and MUC4 in infants with SevBPD. Important mRNA transcriptomic differences included upregulated genes related to immune maturation and inflammation (CEBPE, AZUL1, BPI, PF4, CAMP, DEFA3, and LY6GF6), including several antimicrobial defensins. Let-7b-5p, miR-96-5p, miR-218-5p, and miR-3200-3p were the most differentially expressed small RNAs. Plasma proteomics identified both general markers of inflammation, such as C-reactive protein, alpha-1-acid glycoproteins, and LPS-binding protein. Integrated multiomic factor analysis (MOFA2) identified that mRNA differences in NPIPB4, TRIM58, and RPL29P12 drove latent factor 1 (70.1% variance, Fig 3). MiRNA differences, including in let-7, shaped latent factor 2 (66.2% variance). We also developed machine learning models to integrate the multiomic data with clinical characteristics and identified several mRNAs and miRNAs, most notably miR-4732-5p, that outperformed established clinical predictors. Conclusions: Integrated multiomics in peripheral blood soon after birth distinguish infants who will later develop BPD from their disease-resistant peers. Of note, the miRNA let-7 is the strongest biomarker of BPD we have identified to date. Our study paves the way for biomarker-based early targeted therapies to treat lung disease in newborns and generates a rich dataset for further mechanistic research on BPD. Fig. 1. Study Overview. (A) We collected central venous blood samples from 300 extremely low birthweight premature infants within 72 hours of birth, then performed whole exome sequencing, small RNA profiling, transcriptomics, or plasma proteomics. (B) We used a systems biology approach with network analyses and machine learning to integrate the different omic datasets at multiple levels.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.