Serious neonatal morbidities are associated with differences in DNA methylation among very preterm infants

BACKGROUND: Infants born very preterm are more likely to experience neonatal morbidities compared to their term peers. Variations in DNA methylation (DNAm) associated with these morbidities may yield novel information about the processes impacted by these morbidities. METHODS: This study included 532 infants born < 30 weeks gestation, participating in the Neonatal Neurobehavior and Outcomes in Very Preterm Infants study. We used a neonatal morbidity risk score, which was an additive index of the number of morbidities experienced during the NICU stay, including bronchopulmonary dysplasia (BPD), severe brain injury, serious neonatal infections, and severe retinopathy of prematurity. DNA was collected from buccal cells at discharge from the NICU, and DNAm was measured using the Illumina MethylationEPIC. We tested for differential methylation in association with the neonatal morbidity risk score then tested for differentially methylated regions (DMRs) and overrepresentation of biological pathways. RESULTS: We identified ten differentially methylated CpGs (α Bonferroni-adjusted for 706,278 tests) that were associated with increasing neonatal morbidity risk scores at three intergenic regions and at HPS4, SRRD, FGFR1OP, TNS3, TMEM266, LRRC3B, ZNF780A, and TENM2. These mostly followed dose–response patterns, for 8 CpGs increasing DNAm associated with increased numbers of morbidities, while for 2 CpGs the risk score was associated with decreasing DNAm. BPD was the most substantial contributor to differential methylation. We also identified seven potential DMRs and over-representation of genes involved in Wnt signaling; however, these results were not significant after Bonferroni adjustment for multiple testing. CONCLUSIONS: Neonatal DNAm, within genes involved in fibroblast growth factor activities, cellular invasion and migration, and neuronal signaling and development, are sensitive to the neonatal health complications of prematurity. We hypothesize that these epigenetic features may be representative of an integrated marker of neonatal health and development and are promising candidates to integrate with clinical information for studying developmental impairments in childhood.