Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Apr 26;113(17):4830-5.
doi: 10.1073/pnas.1524857113. Epub 2016 Apr 12.

Stress-induced Gene Expression and Behavior Are Controlled by DNA Methylation and Methyl Donor Availability in the Dentate Gyrus

Affiliations
Free PMC article

Stress-induced Gene Expression and Behavior Are Controlled by DNA Methylation and Methyl Donor Availability in the Dentate Gyrus

Emily A Saunderson et al. Proc Natl Acad Sci U S A. .
Free PMC article

Abstract

Stressful events evoke long-term changes in behavioral responses; however, the underlying mechanisms in the brain are not well understood. Previous work has shown that epigenetic changes and immediate-early gene (IEG) induction in stress-activated dentate gyrus (DG) granule neurons play a crucial role in these behavioral responses. Here, we show that an acute stressful challenge [i.e., forced swimming (FS)] results in DNA demethylation at specific CpG (5'-cytosine-phosphate-guanine-3') sites close to the c-Fos (FBJ murine osteosarcoma viral oncogene homolog) transcriptional start site and within the gene promoter region of Egr-1 (early growth response protein 1) specifically in the DG. Administration of the (endogenous) methyl donor S-adenosyl methionine (SAM) did not affect CpG methylation and IEG gene expression at baseline. However, administration of SAM before the FS challenge resulted in an enhanced CpG methylation at the IEG loci and suppression of IEG induction specifically in the DG and an impaired behavioral immobility response 24 h later. The stressor also specifically increased the expression of the de novo DNA methyltransferase Dnmt3a [DNA (cytosine-5-)-methyltransferase 3 alpha] in this hippocampus region. Moreover, stress resulted in an increased association of Dnmt3a enzyme with the affected CpG loci within the IEG genes. No effects of SAM were observed on stress-evoked histone modifications, including H3S10p-K14ac (histone H3, phosphorylated serine 10 and acetylated lysine-14), H3K4me3 (histone H3, trimethylated lysine-4), H3K9me3 (histone H3, trimethylated lysine-9), and H3K27me3 (histone H3, trimethylated lysine-27). We conclude that the DNA methylation status of IEGs plays a crucial role in FS-induced IEG induction in DG granule neurons and associated behavioral responses. In addition, the concentration of available methyl donor, possibly in conjunction with Dnmt3a, is critical for the responsiveness of dentate neurons to environmental stimuli in terms of gene expression and behavior.

Keywords: DNA methylation; behavior; hippocampus; immediate-early gene; stress.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
The effect of SAM on FS-induced behavior. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before FS (15 min, 25 °C) and 24 h later were forced to swim again under the same conditions. The graphs show the climbing, swimming, and immobility behavior scored in 10-s bins during the first 5 min of the initial test (A) and retest (B). Data are shown as the mean behavioral score (mean ± SEM, n = 8–9). *P < 0.05 compared with the respective vehicle-treated group; &P = 0.072 compared with the respective vehicle-treated group. For more information on statistical analyses in Figs. 1–7, see SI Statistics Information to Main Manuscript Figs. 1–7.
Fig. 2.
Fig. 2.
The effect of SAM on FS-evoked c-Fos and Egr-1 induction in the DG. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before FS (15 min, 25 °C) and killed 60 min after the start of the challenge (FS60). The baseline (BL) groups were killed 90 min after injection. The graphs show the number of c-Fos+ and Egr-1+ neurons in the whole DG within a 50-µm section (A and B, respectively). Data are shown as the average number of c-Fos+ or Egr-1+ neurons from three 50-μm-thick coronal brain slices per animal (mean ± SEM, n = 5–6). *P < 0.05 compared with the respective BL group; $P < 0.05 compared with the respective vehicle/FS60 group.
Fig. S1.
Fig. S1.
Representative images of c-Fos immunostaining in the dentate gyrus of SAM-treated, forced swim-challenged rats. Rats were s.c. injected with vehicle or SAM (100 mg/kg) and either killed under baseline conditions (Control) or 60 min after the start of a 15-min forced swim (FS) challenge. (AD) Representative images of the four treatment conditions with (Left) low-magnification images and (Right) higher magnification images of the indicated rectangles depicted in the respective Left images. The black arrows indicate examples of positively stained dentate gyrus neuronal nuclei. GCL, granular cell layer.
Fig. S2.
Fig. S2.
The effect of SAM on forced swimming-evoked c-Fos and Egr-1 induction in the dentate gyrus. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before forced swimming (15 min, 25 °C) and killed 60 min after the start of the challenge (FS60). The baseline (BL) groups were killed 90 min after injection. The graphs show the number of c-Fos+ and Egr-1+ neurons split between the dorsal and the ventral blade (A and B, respectively). Representative images of c-Fos+ (C) and Egr-1+ (D) neurons in the dorsal blade of the dentate gyrus under baseline (BL) or stress conditions (FS60) are shown. GCL, granular cell layer; H, hilus; ML, molecular layer. Data in A and B are shown as the average number of c-Fos+ or Egr-1+ neurons from three 50-μm-thick coronal brain slices per animal (mean ± SEM, n = 5–6). (A) Three-way ANOVA; effect of SAM, F(1,36) = 19, P < 0.0001; effect of stress, F(1,36) = 130, P < 0.0001; effect of location, F(1,36) = 600, P < 0.0001; interaction SAM × stress, F(1,36) = 34, P < 0.0001; interaction of SAM × location, F(1,36) = 15, P < 0.0001; interaction stress × location, F(1,36) = 80, P < 0.0001; interaction SAM × stress × location, F(1,36) = 14, P < 0.01. (B) Three-way ANOVA: effect of SAM, F(1,36) = 9.4, P < 0.01; effect of stress, F(1,36) = 6.3, P < 0.05; effect of location, F(1,36) = 170, P < 0.0001; interaction SAM × stress, F(1,36) = 8.1, P < 0.01; interaction of SAM × location, F(1,36) = 5.5, P < 0.05; interaction stress x location, F(1,36) = 8.6, P < 0.01; interaction SAM × stress × location, F(1,36) = 8.4, P < 0.01. Bonferroni-corrected post hoc test with contrasts: *P < 0.05 compared with the respective BL group; $P < 0.05 compared with the respective vehicle/FS60 group; #P < 0.05 compared with the respective ventral blade group.
Fig. S3.
Fig. S3.
The effect of SAM on forced swimming-induced c-Fos and Egr-1 induction in the CA1 and CA3. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before forced swimming (15 min, 25 °C) and killed 60 min after the start of the challenge (FS60). The baseline (BL) groups were killed 90 min after the injection. The graphs show the number of c-Fos+ and Egr-1+ neurons in the CA1 (A and B, respectively) and the CA3 (C and D, respectively). Data are shown as an average number of c-Fos+ or Egr-1+ neurons from two 50-μm-thick coronal brain slices per animal (mean ± SEM, n = 3–4). Statistical analysis (two-way ANOVA): (A) effect of SAM, F(1,12) = 0.91, P = 0.36; effect of stress, F(1,12) = 43, P < 0.0001; interaction SAM x stress, F(1,12) = 1.1, P = 0.31. (B) Effect of SAM, F(1,12) = 0.24, P = 0.63; effect of stress, F(1,12) = 5.1, P < 0.05; interaction SAM x stress, F(1,12) = 0.12, P = 0.73. (C) Effect of SAM, F(1,11) = 0.61, P = 0.45; effect of stress, F(1,11) = 1.6, P = 0.23; interaction SAM x stress, F(1,11) = 0.042, P = 0.84. (D) Effect of SAM, F(1,14) = 0.082, P = 0.78; effect of stress, F(1,14) = 2.6, P = 0.13; interaction SAM x stress, F(1,14) = 0.34, P = 0.57. Bonferroni-corrected post hoc test with contrasts: *P < 0.05 compared with the respective BL group.
Fig. 3.
Fig. 3.
The effect of SAM on H3S10p-K14ac formation in the DG after FS. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before FS (15 min, 25 °C) and killed at FS60. The BL groups were killed 90 min after the injection. The graphs show the number of H3S10p-K14ac+ neurons in the DG. Data are shown as an average number of H3S10p-K14ac+ neurons from three 50-μm-thick coronal brain slices per animal (mean ± SEM, n = 4–6). *P < 0.05 compared with the respective BL group.
Fig. S4.
Fig. S4.
The effect of SAM on H3S10p-K14ac formation in the dentate gyrus after forced swimming. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before forced swimming (15 min, 25 °C) and killed at FS60. The BL groups were killed 90 min after the injection. The graphs show the number of H3S10p-K14ac+ neurons split between the dorsal and the ventral blade (A). Representative images of H3S10p-K14ac+ (B) neurons in the dorsal blade of the dentate gyrus under baseline (BL) or stress conditions (FS60) are shown. GCL, granular cell layer; H, hilus; ML, molecular layer. Data in A are shown as an average number of H3S10p-K14ac+ neurons from three 50-μm-thick coronal brain slices per animal (mean ± SEM, n = 4–6). (A) Three-way ANOVA: effect of SAM, F(1,32) = 0.33, P = 0.57; effect of stress, F(1,32) = 28, P < 0.0001; effect of location, F(1,32) = 50, P < 0.0001; interaction SAM x stress, F(1,32) = 0.0049, P = 0.95; interaction of SAM x location, F(1,32) = 0.22, P = 0.65; interaction stress x location, F(1,32) = 11, P < 0.01; interaction SAM x stress x location, F(1,32) = 0.053, P = 0.82. Bonferroni-corrected post hoc test with contrasts: *P < 0.05 compared with the respective BL group; #P < 0.05 compared with the respective ventral blade group.
Fig. 4.
Fig. 4.
FS-induced CpG-specific DNA methylation changes in the c-Fos promoter region. Rats were killed immediately (BL group) or subjected to FS (15 min, 25 °C) and killed at FS60. The location of CpGs within areas 1 and 2 with respect to the rat c-Fos gene are shown in Fig. S5. The graphs show DNA methylation changes at CpGs in area 1 and area 2 in the DG. Data are shown as percentage methylation (mean ± SEM, n = 3–6). *P < 0.05; &, P < 0.1, compared with the respective BL group.
Fig. S5.
Fig. S5.
Forced swimming-induced CpG-specific DNA methylation changes in the c-Fos promoter region in the CA regions of the hippocampus. Rats were killed immediately (BL group) or subjected to forced swimming (15 min, 25 °C) and killed at FS60. The location of CpGs within areas 1 and 2 with respect to the rat c-Fos gene are shown in A. The graph shows DNA methylation changes at CpGs in area 1 and area 2 in the CA regions of the dorsal hippocampus (B). Data are shown as percentage methylation (mean ± SEM, n = 3–6). (B) Area 1: effect of CpG number, F(6,54) = 28, P < 0.0001; effect of stress, F(1,54) = 0.030, P = 0.87; interaction CpG number x stress, F(6,54) = 0.44, P = 0.85. Area 2: effect of CpG number, F(6,60) = 43, P < 0.0001; effect of stress, F(1,60) = 0.73, P = 0.41; interaction CpG number x stress, F(6,60) = 1.2, P = 0.30.
Fig. S6.
Fig. S6.
Forced swimming-induced CpG-specific DNA methylation changes in the Egr-1 promoter region in the CA regions of the hippocampus. Rats were killed immediately (BL group) or subjected to forced swimming (15 min, 25 °C) and killed at FS60. The location of CpGs within areas A and B with respect to the rat Egr-1 gene is shown in A. The graphs show DNA methylation changes at CpGs in area A and area B in the CA regions of the dorsal hippocampus (B). Data are shown as percentage methylation (mean ± SEM, n = 5–6). *P < 0.05; &P < 0.1, compared with the respective BL group. Statistical analysis (two-way ANOVA with repeated measures): (B) area A: effect of CpG number, F(16,160) = 71, P < 0.0001; effect of stress, F(1,160) = 0.32, P = 0.58; interaction CpG number x stress, F(16,160) = 0.63, P = 0.85. Area B: effect of CpG number, F(7,56) = 360, P < 0.0001; effect of stress, F(1,56) = 2.0, P = 0.20; interaction CpG number x stress, F(7,56) = 1.6, P = 0.16.
Fig. 5.
Fig. 5.
FS-induced CpG-specific DNA methylation changes in the Egr-1 promoter region. Rats were killed immediately (BL group) or subjected to FS (15 min, 25 °C) and killed at FS60. The location of CpGs within areas A and B with respect to the rat Egr-1 gene are shown in Fig. S6. The graphs show DNA methylation changes at CpGs in area A and area B in the DG. Data are shown as percentage methylation (mean ± SEM, n = 5–6). *P < 0.05; &P < 0.1, compared with the respective BL group.
Fig. 6.
Fig. 6.
The effect of SAM treatment on FS-induced DNA methylation changes at CpGs within the c-Fos UTR in the DG. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before FS (15 min, 25 °C) and killed at FS60. The BL groups were killed 90 min after the injection. The graph shows methylation of CpGs in area 2 of the c-Fos UTR in the DG. Data are shown as percentage methylation (mean ± SEM, n = 4–6). *P < 0.05 compared with the respective vehicle/FS60 group; $P < 0.05 compared with the respective SAM/BL group; +P = 0.076 compared with the respective SAM/BL group; &P = 0.076 compared with the respective vehicle/FS60 group.
Fig. 7.
Fig. 7.
The effect of SAM treatment on FS-induced DNA methylation changes at CpGs within the Egr-1 gene promoter in the DG. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before FS (15 min, 25 °C) and killed at FS60. The BL groups were killed 90 min after the injection. The location of CpGs within area A with respect to the Egr-1 gene is shown in Fig. S6. The graph shows methylation of CpGs in area A of the Egr-1 gene promoter in the DG. Data are shown as percentage methylation (mean ± SEM, n = 4–6). *P < 0.05 compared with the respective vehicle/FS60 group; $P < 0.05 compared with the respective SAM/BL group.
Fig. S7.
Fig. S7.
The effect of SAM treatment and forced swimming on CpG methylation in the c-Fos UTR and Egr-1 gene promoter in the CA regions of the hippocampus. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before forced swimming (15 min, 25 °C) and killed at FS60. The BL groups were killed 90 min after the injection. The graphs show methylation of CpGs in area 2 of the c-Fos gene promoter (A) and area A of the Egr-1 gene promoter (B) from the CA regions. Data are shown as percentage methylation (mean ± SEM, n = 4–6). Statistical analysis (three-way ANOVA): (A) Effect of SAM, F(1,134) = 3.6, P = 0.060; effect of stress, F(1,134) = 18, P < 0.0001; effect of CpG number, F(6,134) = 150, P < 0.0001; interaction SAM x stress, F(1,134) = 28, P < 0.0001; interaction SAM x CpG number, F(5,134) = 0.89, P = 0.49; interaction stress x CpG number, F(6,134) = 1.2, P = 0.31; interaction SAM x stress x CpG number, F(5,134) = 0.42, P = 0.84. (B) Effect of SAM, F(1,280) = 1.9, P = 0.17; effect of stress, F(1,280) = 6.2, P < 0.05; effect of CpG number, F(16,280) = 110, P < 0.0001; interaction SAM x stress, F(1,280) = 0.0, P = 1.0; interaction SAM x CpG number, F(11,280) = 0.47, P = 0.92; interaction stress x CpG number, F(16,280) = 0.83, P = 0.66; interaction SAM x stress x CpG number, F(11,280) = 0.60, P = 0.83. Bonferroni-corrected post hoc test with contrasts: *P < 0.05 compared with the respective vehicle/BL group; $P < 0.05 compared with the respective SAM/BL group.
Fig. S8.
Fig. S8.
The enrichment of H3K4me3, H3K9me3, and H3K27me3 at the c-Fos and Egr-1 promoters after SAM treatment and forced swimming. Rats were given one injection of vehicle or SAM (100 mg/kg, s.c.) 30 min before forced swimming (15 min, 25 °C) and killed at 60 min after the start of the stressor (FS60). The BL groups were killed 90 min after the injection. The graphs show the enrichment of H3K4me3 (A), H3K9me3 (C), and H3K27me3 (E) at the c-Fos gene promoter, and H3K4me3 (B), H3K9me3 (D) and H3K27me3 (F) at the Egr-1 gene promoter. Data are shown as a ratio of target DNA after immunoprecipitation (Bound) compared with target DNA in the Input sample (A and B, mean ± SEM, n = 3; CF, mean and range, n = 2). Statistical analysis (two-way ANOVA): (A) Effect of SAM, F(1,8) = 0.036, P = 0.85; effect of stress, F(1,8) = 0.85, P = 0.38; interaction SAM x stress, F(1,8) = 0.044, P = 0.84. (B) Effect of SAM, F(1,8) = 0.42, P = 0.54; effect of stress, F(1,8) = 0.92, P = 0.36; interaction SAM x stress, F(1,8) = 0.63, P = 0.45.
Fig. 8.
Fig. 8.
Effect of FS on Dnmt3a mRNA expression in the DG and CA regions of the hippocampus. Rats were killed immediately (BL group) or subjected to FS (15 min, 25 °C) and killed immediately (FS15), 30 min (FS30), 60 min (FS60), or 180 min (FS180) after the start of the challenge. The graphs show Dnmt3a mRNA expression in the DG (A) and the CA regions (B) of the hippocampus. Data are shown as relative mRNA copy number standardized to the expression of the housekeeping genes Hprt1 and Ywhaz (mean ± SEM, n = 8–9). Statistical analysis (one-way ANOVA): (A) F(5,38) = 3.0, P < 0.05; (B) F(5,39) = 0.97, P = 0.43. Dunnett’s post hoc test: *P < 0.05 compared with the BL group.
Fig. S9.
Fig. S9.
Effect of forced swimming on Dnmt3b, Dnmt1, and Tet1 mRNA expression in the dentate gyrus and CA regions of the hippocampus. Rats were killed immediately (BL group) or subjected to forced swimming (15 min, 25 °C) and killed immediately (FS15), 30 min (FS30), 60 min (FS60), or 180 min (FS180) after the start of the challenge. The graphs show Dnmt3b, Dnmt1, and Tet1 expression in the dentate gyrus (A, C, and E) and the CA regions (B, D, and F) of the hippocampus. Data are shown as relative mRNA copy number calculated using the Pfaffl method of analysis, standardized to the expression of the housekeeping genes Hprt1 and Ywhaz (mean ± SEM, n = 8–9). Statistical analysis (one-way ANOVA): (A) F(4,38) = 2.0, P = 0.12; (B) F(4,39) = 0.71, P = 0.59; (C) F(4,38) = 0.82, P = 0.52; (D) F(4,39) = 1.4, P = 0.24; (E) F(4, 38) = 1.71, P = 0.17; (F) F(4, 37) = 0.75, P = 0.57.
Fig. 9.
Fig. 9.
Association of Dnmt3a, Dnmt1, and Tet1 with c-Fos and Egr-1 gene loci after FS. Rats were killed under baseline conditions or at 60 min after the start of a 15-min FS session (FS60). ChIP for Dnmt3a, Dnmt3b, and Tet1 was conducted on hippocampus tissue, followed by quantitative PCR (qPCR) for the c-Fos (A) and Egr-1 (B) loci studied for DNA methylation changes after SAM and FS. Data are expressed as the enrichment of the respective enzymes at the loci at FS60 relative to the enrichment in the baseline situation (mean ± SEM, n = 4). *P < 0.01, Student’s t test.

Similar articles

See all similar articles

Cited by 26 articles

See all "Cited by" articles

Publication types

MeSH terms

Substances

LinkOut - more resources

Feedback