Hippocampal transcriptome analysis following maternal separation implicates altered RNA processing in a mouse model of fetal alcohol spectrum disorder

Abstract

Background: Fetal alcohol spectrum disorders (FASD) are common, seen in 1-5% of the population in the USA and Canada. Children diagnosed with FASD are not likely to remain with their biological parents, facing early maternal separation and foster placements throughout childhood.

Methods: We model FASD in mice via prenatal alcohol exposure and further induce early life stress through maternal separation. We use RNA-seq followed by clustering of expression profiles through weighted gene co-expression network analysis (WGCNA) to analyze transcriptomic changes that result from the treatments. We use reverse transcription qPCR to validate these changes in the mouse hippocampus.

Results: We report an association between adult hippocampal gene expression and prenatal ethanol exposure followed by postnatal separation stress that is related to behavioral changes. Expression profile clustering using WGCNA identifies a set of transcripts, module 19, associated with anxiety-like behavior (r = 0.79, p = 0.002) as well as treatment group (r = 0.68, p = 0.015). Genes in this module are overrepresented by genes involved in transcriptional regulation and other pathways related to neurodevelopment. Interestingly, one member of this module, Polr2a, polymerase (RNA) II (DNA directed) polypeptide A, is downregulated by the combination of prenatal ethanol and postnatal stress in an RNA-Seq experiment and qPCR validation (q = 2e-12, p = 0.004, respectively).

Conclusions: Together, transcriptional control in the hippocampus is implicated as a potential underlying mechanism leading to anxiety-like behavior via environmental insults. Further research is required to elucidate the mechanism involved and use this insight towards early diagnosis and amelioration strategies involving children born with FASD.

Keywords: Fetal alcohol spectrum disorder; Gene expression; Hippocampus; Maternal separation; Prenatal alcohol; WGCNA.

Fig. 1

Module formation and trait association using weighted gene co-expression network analysis (WGCNA). Hierarchical clustering dendrogram of module eigengenes with the dissimilarity of eigengenes given by 1-(eigengene correlation) and module-trait correlation heatmap for 11 traits: prenatal or postnatal treatment, experimental group (group), Barnes maze learning score (learningscore), weight at postnatal day 21 (p21weight), open field test activity (OFTactivity), distance (OFTdistance), latency to enter the center (OFTlatency), number of center entries (OFTentries), home cage activity (HCactivity), and number of rears (HCrears), with positive (red) and negative (blue) correlations, with the number of significantly (p < 0.05) positively or negatively correlated modules with each trait indicated

Fig. 2

Differential gene expression between groups as detected by sleuth. Volcano plots indicating effect size (beta value) and significance (p < 0.01 by color) for each transcript in a ethanol, b stress, and c ethanol + stress groups compared to control treatment with nominally significant (p < 0.001), annotated genes labeled. d Venn diagram of overlapping transcripts differentially expressed in each treatment as compared to controls (p < 0.01). e Volcano plot indicating effect size (beta value) and significance for transcripts with an expanded scale and significant (q < 0.05) annotated genes labeled for ethanol + stress compared to control, where dashed line indicates the axis cutoff of 6 presented in c

Fig. 3

Transcript-specific differential gene expression following ethanol + stress treatment. Transcript abundance in transcripts per million (tpm ± inferential variance) as detected by sleuth from RNA-Seq for aEsrrb and bPolr2a, ***q value < 0.05. c The relative quantity of Polr2a is decreased 1.24-fold with postnatal stress alone and 1.59-fold when mice were prenatally exposed to ethanol and postnatal stress, as detected by reverse transcription qPCR, *p < 0.05, **p < 0.01. Open circles represent biological replicates that were not used in the RNA-Seq experiment, while closed circles represent samples included in the RNA-Seq experiment and serve as technical replicates