(53) Personalized models of fetal brain development as biomarkers of offspring neurodevelopmental outcomes after maternal SARS-CoV-2

Maternal immune activation from viral infection in pregnancy is associated with adverse neurodevelopmental (ND) outcomes in offspring, mediated at least in part by in utero programming of fetal brain microglia resulting in aberrant synaptic pruning. We have shown the feasibility of using human cord blood mononuclear cells (CB-MNCs) to model the impact of maternal viral infection on fetal brain development.  In a large cohort, we examined associations between CB-MNC models and offspring adverse ND outcomes at 3 years of age using linked electronic health records (EHR).

Study Design: CB-MNCs were selected from 119 unvaccinated individuals enrolled in a COVID-19 biorepository (April 2020 - August 2022): 74 with maternal SARS-CoV-2 infection in pregnancy (N=38 1^st or 2^nd trimester infection, N=36 3^rd trimester infection) and 45 uninfected controls. We induced CB-MNCs to acquire microglia-like morphology and function (cord blood induced microglia or CB-iMG) using IL-34, GM-CSF, and serum withdrawal. CB-iMG phagocytosis of iPSc-derived neural synaptosomes – a proxy for microglial synaptic pruning behavior – was quantified using real-time pHrodo fluorescence-based imaging.  ND outcomes of offspring at 3 years were determined using an EHR-based algorithm to identify ICD-10 codes associated with ND in an integrated statewide health system.

Results: CB-iMG exhibited typical microglial morphology and marker expression (Fig 1A). Synaptosome phagocytosis was greater (p< 0.001) in the CB-iMG of offspring exposed to 3^rd trimester maternal SARS-CoV-2 with an adverse ND outcome (N=6) compared to exposed offspring with no ND outcome (N=30) (Fig 1B).  Among unexposed offspring (N=45), phagocytosis scores between groups were not significantly different. Phagocytosis in models derived from 1^st and 2^nd trimester infection did not differ significantly from controls.  

Conclusion:

Although additional validation is needed, personalized models using CB-MNCs available at birth show promise for identifying offspring at greatest ND risk after exposure to maternal viral infection in utero.