Long-lasting alterations to DNA methylation and ncRNAs could underlie the effects of fetal alcohol exposure in mice

Abstract

Fetal alcohol spectrum disorders (FASDs) are characterized by life-long changes in gene expression, neurodevelopment and behavior. What mechanisms initiate and maintain these changes are not known, but current research suggests a role for alcohol-induced epigenetic changes. In this study we assessed alterations to adult mouse brain tissue by assaying DNA cytosine methylation and small noncoding RNA (ncRNA) expression, specifically the microRNA (miRNA) and small nucleolar RNA (snoRNA) subtypes. We found long-lasting alterations in DNA methylation as a result of fetal alcohol exposure, specifically in the imprinted regions of the genome harboring ncRNAs and sequences interacting with regulatory proteins. A large number of major nodes from the identified networks, such as Pten signaling, contained transcriptional repressor CTCF-binding sites in their promoters, illustrating the functional consequences of alcohol-induced changes to DNA methylation. Next, we assessed ncRNA expression using two independent array platforms and quantitative PCR. The results identified 34 genes that are targeted by the deregulated miRNAs. Of these, four (Pten, Nmnat1, Slitrk2 and Otx2) were viewed as being crucial in the context of FASDs given their roles in the brain. Furthermore, ≈ 20% of the altered ncRNAs mapped to three imprinted regions (Snrpn-Ube3a, Dlk1-Dio3 and Sfmbt2) that showed differential methylation and have been previously implicated in neurodevelopmental disorders. The findings of this study help to expand on the mechanisms behind the long-lasting changes in the brain transcriptome of FASD individuals. The observed changes could contribute to the initiation and maintenance of the long-lasting effect of alcohol.

Fig. 1.

Hierarchical clustering analysis of differential DNA methylation enrichment peaks from arrays examining the effect of continuous preference drinking on B6 male brains. To compare differentially enriched regions between ethanol-exposed and control mice, the normalized log₂-ratio scan values were averaged and then used to calculate the M′ value [M′=Average(log₂ MeDIPE/InputE) – Average(log₂ MeDIPC/InputC)] for each probe. NimbleScan sliding-window peak-finding algorithm was run on this data to find the differential enrichment peaks (DEPs).

Fig. 2.

The functional significance of altered CTCF-binding-site methylation after FAE. (A,B) Gene promoters (‘P’) with CTCF sites showing increased (A) and decreased (B) methylation after FAE. (C) Schematic of H19/IGF2 imprinting regulation and the effects of FAE. The black rectangle represents the H19 ICR, white lollipops represent unmethylated DNA, and black lollipops represent methylated DNA. On the wild-type locus, the ICR exhibits paternal-specific methylation and contains binding sites for CTCF. On the maternal allele, CTCF binds to the ICR and blocks the Igf2 promoter from accessing the 3′ shared enhancers (E). On the paternal allele, the ICR is methylated, and H19 transcription is repressed. Because CTCF binding is methylation sensitive, the ICR cannot act as an insulator on the paternal allele, allowing Igf2 expression to be driven from the enhancer. Our results suggest that, in the case of FAE, imprinting is deregulated owing to increased methylation in the CTCF-binding site, which causes the maternal allele to exhibit paternal imprinting marks.

Fig. 3.

Gene Mania network analysis of significantly differentially methylated genes containing CTCF-binding sites, from the ‘Behavior, Neurological Disease, and Psychological Disorders’ network.

Fig. 4.

Heat maps of miRNA expression generated using hierarchical clustering of the four FAE paradigms. Trimester 1 injection model, trimester 2 injection model, trimester 3 injection model and continuous preference drinking model (ANOVA P-value of 0.05 and minimum 1.2-fold cut-off).

Fig. 5.

Venn diagram of common and unique differentially expressed miRNAs identified by four FAE models. Continuous preference drinking; trimester 1, binge treatment at GD8 and GD11; trimester 2, binge treatment at GD14 and GD16; trimester 3, binge treatment at PD4 and PD7. (ANOVA P-value of 0.05 and minimum 1.2-fold cut-off.)

Fig. 6.

A bar graph depicting the quantitation of mmu-mir-679-5p expression in control and fetal alcohol-exposed (CPD) adult brains. The y-axis depicts the relative mir-679 expression normalized to snoRNA 202, expressed as a mean ± s.e.m. of both biological (n=6) and technical (n=3) replicates. *P<0.05.

Fig. 7.

Summary of select observed epigenetic associations. White lollipops indicate an absence of DNA cytosine methylation, whereas black lollipops indicate its presence. Green arrows indicate an increase of either methylation or expression, whereas red arrows indicate a decrease.