Transcriptome analysis of rat dorsal hippocampal CA1 after an early life seizure induced by kainic acid

Abstract

Seizures that occur during early development are associated with adverse neurodevelopmental outcomes. Causation and mechanisms are currently under investigation. Induction of an early life seizure by kainic acid (KA) in immature rats on post-natal day (P) 7 results in behavioral changes in the adult rat that reflect social and intellectual deficits without overt cellular damage. Our previous work also demonstrated increased expression of CA1 hippocampal long-term potentiation (LTP) and reduced desensitization of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-type ionotropic glutamate receptors (AMPA-R) one week following a kainic acid induced seizure (KA-ELS). Here we used RNA sequencing (RNAseq) of mRNA from dorsal hippocampal CA1 to probe changes in mRNA levels one week following KA-ELS as a means to investigate the mechanisms for these functional changes. Ingenuity pathway analysis (IPA) confirmed our previous results by predicting an up-regulation of the synaptic LTP pathway. Differential gene expression results revealed significant differences in 7 gene isoforms. Additional assessments included AMPA-R splice variants and adenosine deaminase acting on RNA 2 (ADAR2) editing sites as a means to determine the mechanism for reduced AMPA-R desensitization. Splice variant analysis demonstrated that KA-ELS result in a small, but significant decrease in the "flop" isoform of Gria3, and editing site analysis revealed significant changes in the editing of a kainate receptor subunit, Grik2, and a serotonin receptor, Htr2c. While these specific changes may not account for altered AMPA-R desensitization, the differences indicate that KA-ELS alters gene expression in the hippocampal CA1 one week after the insult.

Keywords: ADAR2; Development; Kainate; RNAseq; Seizure; Splice-variant.

Figure 1.

Distribution of Isoform Expression Data. Differential expression of gene isoforms was determined using a multifactor design for littermate pairing using DESeq2(Love et al. 2016). Shown is the distribution of the expression of gene isoforms with significant gene isoforms are labeled. Significance values were calculated using a Wald test for treatment effects, and the Benjamini-Hochberg method for false discovery rate was used to adjust p-values for multiple testing.

Figure 2.

Ingenuity Pathway Analysis. A. Top canonical signaling pathways with relevance to learning and memory, and the gene (ovals) overlap between pathways (rectangles). B. Top canonical signaling pathways with relevance to stress pathways, and gene overlap between pathways. Figure was generated using Cytoscape(Shannon et al. 2003). Genes are colored based on the number of pathways in the figure that they are associated with (5 = yellow; 4 = green; 3 = orange; 2 = blue; 1= fuschia).

Figure 3.

ADAR2 editing site differences between control and KA-ELS rats. Genome aligned reads and the GATK pipeline for variant calling and filtering(Piskol et al. 2013) were used to generate variant calling files (vcf) for each sample. The chromosome locations of ADAR2 editing sites were determined and located within each sample vcf. Percent editing was determined based on the read counts for alternate bases at each chromosome location of interest. Blcap Y2C (Con 35.6 ± 0.44, KA-ELS 35.3 ± 1.3, p = 0.83), Blcap Q5R (Con 22.8 ± 0.61, KA-ELS 23.1 ± 0.90, p = 0.77), Flna (Con 43.3 ± 2.54, KA-ELS 47.0 ± 3.0, p = 0.367), Gabra3 I342M (Con 69.3 ± 2.7, KA-ELS 73.0 ± 1.9, p = 0.274), Gria2 Q607R (Con 99.646 ± 0.09, KA-ELS 99.651 ± 0.09, p = 0.972), Gria2 R764G (Con 62.0 ± 1.2, KA-ELS 63.0 ± 1.2, p = 0.525), Gria3 R769G (Con 84.6 ±0.44, KA-ELS 85.7 ± 0.56, p = 0.162), Grik1 Q636R (Con 48.85 ± 3.6, KA-ELS 43.8 ± 5.8, p = 0.461), Grik2 I567V (Con 62.8 ± 2.5, KA-ELS 61.3 ± 3.7, p = 0.732), Grik2 Y571C (Con 72.7 ± 1.33, KA-ELS 78.9 ± 1.48; p = 0.008), Grik2 Q621R (Con72.7 ± 2.0, KA-ELS 70.0 ± 3.6, p = 0.516), Htr2c B (Con 81.5 ± 2.6, KA-ELS 79.8 ± 4.2, p = 0.725), Htr2c C (Con 73.4 ± 3.7, KA-ELS 68.6 ± 5.7, p = 0.5), and Htr2c D (I161V) (Con 62.2 ± 3.02, KA-ELS 79.0 ± 4.73; p = 0.01) Statistics presented are mean ± SEM, and two tailed t-test).

Figure 4.

Gria Flip/Flop alternative splicing. A. Gria3 transcripts showing the location of the flip and flop exons. Example read distributions surrounding the flip and flop exons of Gria3 from two control and two KA-ELS samples. B. Gria3 is approximately 90% flip, and there is a small but significant increase in the percentage of the subunits that are flip in the KA-ELS rats (Con 91.6 ± 0.75% flip, KA-ELS 93.9 ± 0.64% flip, p = 0.04, two-tailed t-test). C. Gria1 is approximately 75% flip, with no significant difference (p = 0.843, two-tailed t-test) between control (73.6 ± 0.01% flip) and KA-ELS rats (74.0 ± 0.01% flip). Gria2 is approximately 60% flip, with no significant difference (p = 0.474, t-tailed t-test) between control (62.4 ± 0.01% flip) and KA-ELS (60.7 ± 0.02% flip). Gria4 is approximately 50% flip with no significant difference (p = 0.308, two-tailed t-test) between control (52 ± 2.6 % flip) and KA-ELS rats (55.6 ± 2.3 % flip).