212 - Metabolome-Wide Association Study of Bronchopulmonary Dysplasia in Extremely Premature Infants

Publication Number: 212.5674

THAYBETH I. MALAVE-MENDEZ, University of California, San Francisco, SAN FRANCISCO, CA, United States; Miguel A`. Guardado, UCSF Benioff Children's Hospital San Francisco, San Francisco, CA, United States; Cheryl J. Chapin, UC San Francisco, SAN FRANCISCO, CA, United States; Ryan Hernandez, UCSF, San Francisco, CA, United States; Philip Ballard, UCSF Benioff Children's Hospital San Francisco, Bodega Bay, CA, United States; Dara G. Torgerson, University of California San Francisco, San Francisco, CA, United States

Background: Bronchopulmonary dysplasia (BPD) is a common outcome in extremely low gestational age newborns (ELGANs), yet its pathogenesis remains poorly understood. Metabolic dysregulation has been attributed to the development of BPD. However, there is a need for larger population-based studies with multiple time point assessments to further characterize metabolic changes that are associated with BPD.

Objective: Identify early changes in the urinary metabolome and metabolic pathways that associate with a diagnosis of BPD at 36 weeks postmenstrual age.

Design/Methods: Untargeted metabolomic profiling was performed using UHPLC/MS/MS on urine samples collected at two time points between postnatal days 6-14, time point 1 (TP1), and 23-30, time point 2 (TP2), for 178 BPD cases and 136 BPD controls from two independent studies of ELGANs born < 29 weeks gestational age (TOLSURF, and PROP) (Table). Logistic regression was used to compare metabolite levels at each time point between BPD cases and BPD controls adjusting for self-reported maternal race, gestational age, birth weight, nutritional source (enteral nutrition (EN) vs. total parenteral nutrition(TPN)), and use of corticosteroids. Comparisons were performed within each study and then combined using a weighted meta-analysis. Pathway enrichment was evaluated using a Fisher’s Exact test among characterized metabolites found to associate with BPD at p< 0.05.

Results: Following quality control, we identified 872 and 909 urinary metabolites that belonged to 88 and 91 different metabolic sub-pathways at TP1 and TP2, respectively. At TP1, we identified 24 metabolites that varied in level by BPD status at p < 0.05 that were enriched in the lactoyl amino acid sub-pathway (p = 0.0002, FDR = 0.017). At TP2, the urinary levels of 56 metabolites varied by BPD status at p < 0.05 that were enriched in fatty acid metabolism (acylcarnitine, medium chain sub-pathway) (p = 0.001, FDR = 0.087). A total of 2 metabolites were consistently at higher levels in infants who did not develop BPD at both time points at p< 0.05 (3-hydroxykynurenine and an uncharacterized metabolite). At TP2, 93% of the metabolites that varied by BPD status were elevated in infants that did not develop BPD (52 out of 56 metabolites [p=5.8x10-7], Figure 1). Sub-analyses by nutritional sources at TP2 revealed 61% and 100% of the differential metabolites having higher levels in No BPD infants on TPN and EN, respectively (Figure 1).

Conclusion(s): Our findings indicate metabolic differences in lactoyl amino acids and in fatty acid metabolism in infants that develop BPD, and an overall reduction in metabolism.

Characteristics of infants from TOLSURF and PROP included in metabolomic studies by BPD status at 36 weeks postmenstrual age
Quantitative measurements were compared using a t-test, and categorical data were compared using a Fisher’s Exact test.

Volcano plots of the results of the metabolome-wide association studies using logistic regression for the identification of differential urinary metabolites that vary by BPD status at two distinct time points.
The dashed red line represents p < 0.05 and the dashed black line divides metabolites higher in infants that developed BPD (log2(OR) > 0) from those higher in infants that did not develop BPD (log2(OR) < 0). Each individual metabolite is represented as a dot colored by super pathway. These plots present that there are multiple metabolites whose levels vary significantly between infants with and without BDP early after birth. Additionally, it shows that the dysregulation at time point 2 is even more pronounced when infants are stratified by nutrition source, with most of the differing metabolites being at lower levels in infants who develop BPD.