Maternal immune activation alters behavior in adult offspring, with subtle changes in the cortical transcriptome and epigenome

Abstract

Maternal immune activation during prenatal development, including treatment with the viral RNA mimic, polyriboinosinic-polyribocytidilic acid (poly IC), serves as a widely used animal model to induce behavioral deficits reminiscent of schizophrenia and related disease. Here, we report that massive cytokine activation after a single dose of poly IC in the prenatal period is associated with lasting working memory deficits in adult offspring. To explore whether dysregulated gene expression in cerebral cortex, contributes to cognitive dysfunction, we profiled the cortical transcriptome, and in addition, mapped the genome-wide distribution of trimethylated histone H3-lysine 4 (H3K4me3), an epigenetic mark sharply regulated at the 5' end of transcriptional units. However, deep sequencing-based H3K4me3 mapping and mRNA profiling by microarray did not reveal significant alterations in mature cerebral cortex after poly IC exposure at embryonic days E17.5 or E12.5. At a small set of genes (including suppressor of cytokine signaling Socs3), H3K4me3 was sensitive to activation of cytokine signaling in primary cultures from fetal forebrain but adult cortex of saline- and poly IC-exposed mice did not show significant differences. A limited set of transcription start sites (TSS), including Disrupted-in-Schizophrenia 1 (Disc1), a schizophrenia risk gene often implicated in gene-environment interaction models, showed altered H3K4me3 after prenatal poly IC but none of these differences survived after correcting for multiple comparisons. We conclude that prenatal poly IC is associated with cognitive deficits later in life, but without robust alterations in epigenetic regulation of gene expression in the cerebral cortex.

Figure 1

(A) IL-6, (B) TNF-a, (C) IFN-g and (D) IL-1b protein levels in maternal serum 3 hours following poly(I:C) (black bars) or saline (white bars) on embryonic day 12.5 (E12.5) and E17.5 (n = 3 mice/group) group). (E) IL-6 in fetal brain 3 hours after poly(I:C) injection (n=3–4 litters/group) * = P <0.05 (two tailed t-test).

Figure 2. T-maze performance and working memory is impaired in adult E17.5 poly IC mice

(A) Adult mice from mothers injected with poly(I:C) on E17.5 demonstrate significantly fewer numbers of T-maze alternations compared to saline on both days 1 and 2 of testing (saline n = 28 [10 litters]; poly(I:C) n = 34 [12 litters]). (B) Adult mice from mothers injected with poly(I:C) on E12.5 do not demonstrate a significantly different number of alternation on day 1 or day 2 of T-maze (saline n = 51 [9 litters]; poly(I:C) n = 27 [8 litters]). * = P < 0.05 (two-tailed t-test).

Figure 3. H3K4me3 ChIP-Seq of adult mouse cortex yields a high percentage of uniquely mappable reads and proximal peaks

(A) A high percentage (87%) of reads from H3K4me3 libraries generated from cortex of adult mice are uniquely mappable to mouse genome (version mm9) for both saline and poly(I:C). A smaller percentage (9%) map to multiple locations and 4% are not mappable. (B) The majority of peaks called by MACS are positioned within 2Kb from an annotated transcription start site (TSS); additionally, 5% of peaks are positioned 2–10Kb from the next TSS, and 16–17% are further removed from a TSS. N = 6 mice/group.

Figure 4. Adult E17.5 poly IC offspring show subtle, non-significant H3K4me3 alterations in cortex

The heatmap depicts the 36 genes for which DESeq detected significantly different numbers of tags between groups within 500bp from the transcription start site of the corresponding gene. None of the p-values survived correction with the false discovery rate. Fold-change compared to control group (N=6 animals from different litters, 3 male and 3 female/group as indicated) (range): 1.25 to 1.61-fold for 17 genes with increased H3K4me3 in poly IC animals;-1.26 to -2.47-fold for six genes with decreased H3K4me3 in poly IC animals.

Figure 5. Histone methylation peaks at the Disc1 locus

(A) University of California Santa Cruz (UCSC) genome browser track for approximately 60 Kb region on chromosome 8, surrounding Disc1 TSS as indicated. (B,C) Distribution of sequence tags from anti-H3Kme3 ChIP-seq from cerebral cortex of (B) saline and (C) poly IC exposed animal. Notice sharp peaks at TSS of Disc1 and of adjacent Tsnax gene.

Figure 6. 12 hour IL-6 treatment alters H3K4me3 levels at many gene transcription start sites in rat forebrain culture

(A) Following 12 hrs of IL-6 treatment, 33 genes demonstrated increased H3K4me3 levels ≥ 1.5-fold (149 ≥ 1.2-fold); 38 genes showed decreased levels ≤-1.5-fold (60 ≤-1.2-fold). The heatmap illustrates all IL-6 induced histone methylation changes ≥ 1.5-fold (red = increase relative to saline; blue = decrease). (B) The H3K4me3 profile from the UCSC genome browser for the Socs3 gene, which demonstrated the most robust and significant increase in methylation following 12 hrs of IL-6 treatment (fold-change = 3.2; p = 1.94⁻²⁰). Note the atypical pattern of H3K4me3, which encompasses the entire Socs3 gene and surrounding region.