Transcriptome of the Krushinsky-Molodkina Audiogenic Rat Strain and Identification of Possible Audiogenic Epilepsy-Associated Genes

Abstract

Audiogenic epilepsy (AE), inherent to several rodent strains is widely studied as a model of generalized convulsive epilepsy. The molecular mechanisms that determine the manifestation of AE are not well understood. In the present work, we compared transcriptomes from the corpora quadrigemina in the midbrain zone, which are crucial for AE development, to identify genes associated with the AE phenotype. Three rat strains without sound exposure were compared: Krushinsky-Molodkina (KM) strain (100% AE-prone); Wistar outbred rat strain (non-AE prone) and "0" strain (partially AE-prone), selected from F2 KM × Wistar hybrids for their lack of AE. The findings showed that the KM strain gene expression profile exhibited a number of characteristics that differed from those of the Wistar and "0" strain profiles. In particular, the KM rats showed increased expression of a number of genes involved in the positive regulation of the MAPK signaling cascade and genes involved in the positive regulation of apoptotic processes. Another characteristic of the KM strain which differed from that of the Wistar and "0" rats was a multi-fold increase in the expression level of the Ttr gene and a significant decrease in the expression of the Msh3 gene. Decreased expression of a number of oxidative phosphorylation-related genes and a few other genes was also identified in the KM strain. Our data confirm the complex multigenic nature of AE inheritance in rodents. A comparison with data obtained from other independently selected AE-prone rodent strains suggests some common causes for the formation of the audiogenic phenotype.

Keywords: KM rat strain; MAPK signaling cascade; Msh3; Ttr; Wistar rats; audiogenic epilepsy; transcriptomic analysis (RNA-seq).

FIGURE 1

(A) The sequential procedure for deriving the Krushinsky-Molodkina (KM) and “0” strains from the Wistar strain. (B) Venn diagram depicting differentially expressed genes (DEGs) between KM samples and those from W and “0” rats, which exhibit distinct genotypes. Genes with differential expression, p < 0.05, and expression changes greater than 1.5-fold (either up- or downregulated) were used for enrichment analysis. Enriched Gene Ontology (GO) terms are presented in colored text. The font size is proportional to the number of genes in a term (see the “font size legend”). Font color indicates the enrichment test p-value (see the “font color legend”).

FIGURE 2

Differential expression profiles of genes participating in the most affected KEGG pathways in the KM, W, and “0” rats. For each KEGG term, genes were identified (irrespective of p-value) and sorted by decreasing log2-fold change. Red-to-blue sub-plots illustrate expression level changes (log-scale, with values decreasing across the plot). The log2 expression level fold change (LogFC) range extends from −1 (i.e., 2-fold downregulation; blue) to +1 (i.e., 2-fold overexpression; red). The cell border indicates the gene set enrichment (GSEA) p-value for a pathway. A red border indicates that a KEGG pathway is enriched with upregulated genes; a blue border indicates that a KEGG pathway is enriched with downregulated genes. The p-value was calculated for up- and downregulated genes using Fisher’s exact test.

FIGURE 3

The expression of MAPK pathway and ERK1/2 genes was alleviated in “0” rats. (A) Differential expression of genes participating in the MAPK pathway (rno04010). Gene expression changes are shown in color (log2-scale; LogFC; red – upregulation, blue – downregulation, gray – no change, white background – the absence of expression). Each gene block is divided into two sections (the left block shows differential expression between KM and W rats; right – between “0” and KM samples). Genes that demonstrate opposite changes between two pairs of comparisons (“compensated” genes) are shown in green font. (B) Scatter plot depicting the pairwise comparison of log2-fold changes between the KM vs W and “0” vs KM groups. R – Spearman’s correlation coefficient. (C) Western blot analysis of phosphorylated and unphosphorylated forms of ERK1/2 (top) and the calculation of the optical density of the signals (bottom). For semiquantitative estimation, the optical density of phosphorylated ERK1/2 was subjected to two levels of normalization: first between samples and second on the unphosphorylated form of ERK1/2. *p < 0.05, Kruskal-Wallis test with post hoc analysis.

FIGURE 4

The expression levels of most genes in the “0” rat samples tended to be reverted to the levels in the W rats. (A) The log2-fold change of gene expression that shows the significant difference in the expression levels between the KM and W rat transcriptomes (X-axis) and between the “0” and KM rat transcriptomes (Y-axis). Positive values correspond to activation in the KM or “0” rats. One point represents one gene, and only genes exhibiting significant changes (p < 0.01, QLF test) are shown. Genes with differences in expression between the KM and W transcriptomes are shown in red, and those with differences only between the “0” and KM transcriptomes are shown in black. (B) Dependence of the compensation coefficient (Y-axis) on the log2-fold change in the KM and W rat comparison (X-axis). The compensation coefficient was calculated as a percentage of mean (logCPM) KM – mean (logCPM) 0)/(mean (logCPM) KM – mean (logCPM) W. (C) Heat map of the top 50 compensated genes (>50 of the compensation coefficient value) displayed in line from the highest p-value at the top to the lowest p-value at the bottom. (D) Heat map of the top 50 genes exhibiting the least or no compensatory expression (<50 of compensation coefficient value) displayed in line from the highest p-value at the top to the lowest p-value at the bottom. The expression values of the genes presented in panels (C,D) are Z-transformed.

FIGURE 5

qPCR validation of gene expression changes initially observed through RNA-seq analysis. Expression levels determined by the 2^–dCt equation, including normalization of the genes of interest levels to a housekeeping gene (Ywhaz) for each group of samples. Expression levels were estimated for no fewer than 6 rats. *p < 0.05, **p < 0.01, Kruskal-Wallis test with post hoc analysis.