Mapping prenatal predictors and neurobehavioral outcomes of an epigenetic marker of neonatal inflammation - A longitudinal population-based study
Anna Suleri,Nicole Creasey,Esther Walton,Ryan Muetzel,Janine F Felix,Liesbeth Duijts,Veerle Bergink,Charlotte A M Cecil
DOI: https://doi.org/10.1016/j.bbi.2024.08.053
2024-08-29
Abstract:Background: DNA methylation levels at specific sites can be used to proxy C-reactive protein (CRP) levels, providing a potentially more stable and accurate indicator of sustained inflammation and associated health risk. However, its use has been primarily limited to adults or preterm infants, and little is known about determinants for - or offspring outcomes of - elevated levels of this epigenetic proxy in cord blood. The aim of this study was to comprehensively map prenatal predictors and long-term neurobehavioral outcomes of neonatal inflammation, as assessed with an epigenetic proxy of inflammation in cord blood, in the general pediatric population. Methods: Our study was embedded in the prospective population-based Generation R Study (n = 2,394). We created a methylation profile score of CRP (MPS-CRP) in cord blood as a marker of neonatal inflammation and validated it against serum CRP levels in mothers during pregnancy, as well as offspring at birth and in childhood. We then examined (i) which factors (previously associated with sustained inflammation) explain variability in MPS-CRP at birth, including a wide range of prenatal lifestyle and clinical conditions, pro-inflammatory exposures, as well as child genetic liability to elevated CRP levels; and (ii) whether MPS-CRP at birth associates with child neurobehavioral outcomes, including global structural MRI and DTI measures (child mean age 10 and 14 years) as well as psychiatric symptoms over time (Child Behavioral Checklist, at mean age 1.5, 3, 6, 10 and 14 years). Results: MPS-CRP at birth was validated with serum CRP in cord blood (cut-off > 1 mg/L) (AUC = 0.72). Prenatal lifestyle pro-inflammatory factors explained a small part (i.e., < 5%) of the variance in the MPS-CRP at birth. No other prenatal predictor or the polygenic score of CRP in the child explained significant variance in the MPS-CRP at birth. The MPS-CRP at birth prospectively associated with a reduction in global fractional anisotropy over time on mainly a nominal threshold (β = -0.014, SE = 0.007, p = 0.032), as well as showing nominal associations with structural differences (amygdala [(β = 0.016, SE = 0.006, p = 0.010], cerebellum [(β = -0.007, SE = 0.003, p = 0.036]). However, no associations with child psychiatric symptoms were observed. Conclusion: Prenatal exposure to lifestyle-related pro-inflammatory factors was the only prenatal predictor that accounted for some of the individual variability in MPS-CRP levels at birth. Further, while the MPS-CRP prospectively associated with white matter alterations over time, no associations were observed at the behavioral level. Thus, the relevance and potential utility of using epigenetic data as a marker of neonatal inflammation in the general population remain unclear. In the future, the use of epigenetic proxies for a wider range of immune markers may lend further insights into the relationship between neonatal inflammation and adverse neurodevelopment within the general pediatric population.