Cytokine changes associated with the maternal immune activation (MIA) model of autism: A penalized regression approach

Abstract

Maternal immune activation (MIA) during pregnancy induces a cytokine storm that alters neurodevelopment and behavior in the progeny. In humans, MIA increases the odds of developing neuropsychiatric disorders such as autism spectrum disorder (ASD). In mice, MIA can be induced by injecting the viral mimic polyinosinic:polycytidylic acid (poly(I:C)) to pregnant dams. Although the murine model of MIA has been extensively studied, it is not clear whether MIA results in cytokine changes in the progeny at early postnatal stages. Further, the murine model of MIA suffers from a lack of reproducibility and high inter-individual variability. Multivariable (MV) statistical analysis is widely used in human studies to control for confounders and covariates such as sex, age and exposure to environmental factors. We therefore reasoned that animal studies in general and studies on the MIA model in particular could benefit from MV analyses to account for complex phenotype interactions and high inter-individual variability. Here, we used MV statistical analysis to identify cytokines associated with MIA after adjustment for covariates. Besides confirming the association between previously described variables and MIA, we identified new cytokines that could play a role in behavioural alterations in the progeny during the early postnatal period.

Fig 1. Methodological workflow for the multivariable analysis.

Fig 2. Correlation between numerical variables.

Heatmap of pairwise Spearman's rank r correlation coefficients between variable pairs. Spearman’s r coefficients are color-coded and proportional to dot area. Only significant correlations are shown (p < 0.05).