Genomic and epigenomic responses to chronic stress involve miRNA-mediated programming

Abstract

Stress represents a critical influence on motor system function and has been shown to impair movement performance. We hypothesized that stress-induced motor impairments are due to brain-specific changes in miRNA and protein-encoding gene expression. Here we show a causal link between stress-induced motor impairment and associated genetic and epigenetic responses in relevant central motor areas in a rat model. Exposure to two weeks of mild restraint stress altered the expression of 39 genes and nine miRNAs in the cerebellum. In line with persistent behavioural impairments, some changes in gene and miRNA expression were resistant to recovery from stress. Interestingly, stress up-regulated the expression of Adipoq and prolactin receptor mRNAs in the cerebellum. Stress also altered the expression of Prlr, miR-186, and miR-709 in hippocampus and prefrontal cortex. In addition, our findings demonstrate that miR-186 targets the gene Eps15. Furthermore, we found an age-dependent increase in EphrinB3 and GabaA4 receptors. These data show that even mild stress results in substantial genomic and epigenomic changes involving miRNA expression and associated gene targets in the motor system. These findings suggest a central role of miRNA-regulated gene expression in the stress response and in associated neurological function.

Figure 1. A: Body weight growth curves.

The time course of body weight gain in animals undergoing two weeks of stress (gray) and non-stress control animals (black). Note that stress moderately diminished the average rate of weight gain (mean ± SD). B: Body weight (means ± SD) after two weeks of stress (“Stress”) and after two weeks of stress followed by two weeks of recovery (“Recovery”).

Figure 2. Concentration of plasma corticosterone (means ± SD; ng/ml) in control and stress animals as measured on the first day of stress, last day of stress and after two weeks of recovery from stress (“Recovery”).

Asterisks represent statistical significance (*** p<0.001).

Figure 3. Skilled reaching performance in male rats.

A, B: Series of photographs illustrating forelimb and digit movements of a rat grasping a food pellet. C: skilled reaching success in rats before (baseline), during and after two weeks of stress (mean ± SEM). Note that stress diminished skilled reaching performance after acute (day 1) and chronic (day 14) of daily stress treatment. Reaching success did not return to baseline levels within 14 days of recovery. Asterisks represent statistical significance (*** p<0.001).

Figure 4. mRNA microarray expression analysis in animals after two weeks of stress (2WSTRESS), recovery from stress (4WSTRESS) and appropriate controls for each time point (2WCONTROL, 4WCONTROL).

Genes with a 2-fold difference and a p-value of p<0.05 are shown in logarithmic scale (log₂). Genes that were changed are represented as dark-red diamonds, in top-left (down-regulated) and top-right (up-regulated) parts of each figure. 2S_2C – groups of 2WSTRESS vs. 2WCONTROL, 4C_2C – 4WCONTROL vs. 2WCONROL, 4S_2S – 4WSTRESS vs. 2WSTRESS, 4S_2C – 4WSTRESS vs. 2WCONTROL.

Figure 5. sq-RT-PCR analysis.

A: prolactin receptor (Prlr) gene; B: Adipoq gene. C: ephrin B3 receptor (Efnb3) gene; D: GABA (A) receptor 4 (Gabra4) gene. E: GAPDH. Data are represented as an average of three animals per group. Asterisks represent statistical significance (* p<0.05; ** p<0.01; *** p<0.001). Error bars represent standard deviation of the mean. Photographs below bars represent corresponding PCR fragments in duplicates for each animal for three animals per group.

Figure 6. qRT-PCR analysis of Prlr (A, B) and Adipoq (C, D) expression level.

Data are represented as a normalized relative fold change to control. Asterisks represent statistical significance (* p<0.01; *** p<0.001). Control animals are represented in black, stressed animals in grey. For more details see (Tables S3, S4, S5, S6). It should be noted that the level of Adipoq expression in hippocampus is very low (C(t) values >40 and beyond detection), which results in a high fold change that is not significant.

Figure 7. Functional annotation clustering.

Green - corresponding gene-term association positively reported; black - corresponding gene-term association not reported yet. A: cluster of four genes: Tcf21, Msx1, Adipoq, Cited, which were grouped based on their involvement in positive regulation of the macromolecule metabolic process. B: cluster of three genes: Prlr, Otc, Adipoq, which were grouped based on their involvement in protein complex assembly. C: cluster of four receptors: Prlr, Osmr, Itgb6, Ssta6. D: cluster of three genes: Cldn3, Itgb6, Cdh3, which were grouped based on their involvement in cell adhesion.

Figure 8. Analysis of miRNA expression in the rat cerebellum after two weeks of stress compared to controls.

The microarray heatmap demonstrates the log₂ ratio of miRNA signal difference between control and stress samples. Up-regulated miRNAs are shown in red, while down-regulated miRNAs are shown in green. The first three columns on each figure represent the level of expression in control animals, while the three last columns represent the level of corresponding miRNA expression in stress animals.

Figure 9. qRT-PCR analysis of miR-186 (A,C,E) and miR-709 (B,D,F) expression level in cerebellum (A,B), hippocampus (C,D), and prefrontal cortex (E,F).

Data are represented as a normalized relative fold change to control. Asterisks represent statistical significance (*p<0.1; **p<0.05; ***p<0.001). Control animals are represented in black, stressed animals in grey. For more details see (Tables S7, S8, S9, S10, S11, S12).

Figure 10. A: Dose-dependent inhibition of Eps15 expression in the Luciferase Assay after transfection of HEK-293 cells with miR-186.

B: Dose-dependent inhibition of the Nab1 gene in the Luciferase Assay after transfection of MCF7 cells with miR-709. C: Luciferase Assay with pFN4 (3′UTR Eps15) and miR-186. The first bar demonstrates relative level of luciferase activity after transfection of MCF-7 cells with pFN4 only. Second bar: pFN4 + miR-186. Third bar: pFN4 mut + miR-186. Asterisks represent statistical significance (p<0.001). D: Luciferase Assay with pFN7 (3′UTR Nab1) and miR-709. The first bar demonstrates relative level of luciferase activity after transfection of MCF-7 cells with pFN7 only. Second bar: pFN7 + miR-709. Third bar: pFN7 mut + miR-709. Bars represent the normalized average of relative luciferase units.

Figure 11. Hypothetical scheme of the stress response in the brain.

A: Changes in PRLR expression. B: Changes in miRNA expression. We hypothesize that restraint stress causes a different response in cerebellum and other brain regions, such as hippocampus and prefrontal cortex. There might be an immediate response in the cerebellum reflected by an increase in the level of prolactin receptors after two weeks of stress. After the recovery from stress the expression of PRLR returns to normal in cerebellum, but is down-regulated in hippocampus and up-regulated in prefrontal cortex. Negative correlation between prolactin and its receptors suggests that levels of prolactin in the cerebellum may decrease after stress. Alternatively, changes in expression levels of PRLR may occur in response to changes of other hormones or cytokines that bind to PRLR. Similarily, immediate responses in the cerebellum could be reflected in a decrease of miR-709 levels after two weeks of stress. After recovery from stress the expression of miR-709 is up-regulated in hippocampus and prefrontal cortex.

Figure 12. Time course of the experimental manipulations.

24 animals were randomly assigned to one of the following experimental groups: two weeks of daily restraint stress (2WSTRESS, n = 6), two weeks naive controls (2WCONTROL, n = 6), two weeks of daily restraint stress + two weeks of recovery from stress (4WSTRESS, n = 6), four weeks naive controls (4WCONTROL, n = 6). Animals were sacrificed immediately after stress (2 weeks), or after two weeks of recovery from stress (4 weeks).