Fear memory regulation by the cAMP signaling pathway as an index of reexperiencing symptoms in posttraumatic stress disorder

Abstract

Posttraumatic stress disorder (PTSD) is a psychiatric disorder associated with traumatic memory, yet its etiology remains unclear. Reexperiencing symptoms are specific to PTSD compared to other anxiety-related disorders. Importantly, reexperiencing can be mimicked by retrieval-related events of fear memory in animal models of traumatic memory. Recent studies revealed candidate PTSD-associated genes that were related to the cyclic adenosine monophosphate (cAMP) signaling pathway. Here, we demonstrate the tight linkage between facilitated cAMP signaling and PTSD by analyzing loss- and gain-of-cAMP signaling effects on fear memory in mice and the transcriptomes of fear memory-activated mice and female PTSD patients with reexperiencing symptoms. Pharmacological and optogenetic upregulation or downregulation of cAMP signaling transduction enhanced or impaired, respectively, the retrieval and subsequent maintenance of fear memory in mice. In line with these observations, integrative mouse and human transcriptome analysis revealed the reduced mRNA expression of phosphodiesterase 4B (PDE4B), an enzyme that degrades cAMP, in the peripheral blood of PTSD patients showing more severe reexperiencing symptoms and the mouse hippocampus after fear memory retrieval. Importantly, more severe reexperiencing symptoms and lower PDE4B mRNA levels were correlated with decreased DNA methylation of a locus within PDE4B, suggesting the involvement of methylation in the mechanism of PTSD. These findings raise the possibility that the facilitation of cAMP signaling mediating the downregulation of PDE4B expression enhances traumatic memory, thereby playing a key role in the reexperiencing symptoms of PTSD patients as a functional index of these symptoms.

Fig. 1. Effects of the systemic injection of rolipram or NB001 on fear memory retrieval and fate.

Effects of the systemic injection of rolipram (ROL) (A) or NB001 (B) on contextual fear memory. A Vehicle (VEH), n = 10; ROL, n = 11. (B), VEH, n = 10; NB001, n = 11. Effects of the systemic injection of ROL (C) or NB001 (D) on IA memory. C VEH, n = 10; ROL, n = 10. (D), VEH, n = 10; NB001, n = 10. *p < 0.05, **p < 0.01, post hoc Bonferroni’s test; #p < 0.05, paired t-test. Error bars indicate SEM.

Fig. 2. Effects of the optogenetic manipulation of cAMP levels in the dorsal hippocampus on fear memory retrieval and fate.

A, B Effects of increased cAMP levels in the dorsal hippocampus by 473-nm blue light stimulation on contextual fear memory. A Schematic illustration of virus injection and mGFP expression in the dorsal hippocampus (upper images). Photostimulation of photoactivatable adenylyl cyclase (bPAC; lower image). B Effects of optogenetic increased cAMP levels in the dorsal hippocampus on contextual fear memory. GFP, n = 12; no stimulation (No Stim.), n = 9; optical stimulation (Optical Stim.), n = 11. C, D Effects of decreased cAMP levels in the dorsal hippocampus by 473-nm blue light stimulation on contextual fear memory. C Schematic illustration of virus injection and GFP expression in the dorsal hippocampus (upper images). Photostimulation of light-activated phosphodiesterase (LAPD; lower image). D Effects of optogenetic decreased cAMP levels in the dorsal hippocampus on contextual fear memory. GFP, n = 11; No Stim., n = 11; Optical Stim., n = 11. *p < 0.05; post hoc Bonferroni’s test. Error bars indicate SEM.

Fig. 3. Mouse fear memory retrieval-associated transcriptome changes and human PTSD reexperiencing-associated transcriptome changes.

A IA memory retrieval rapidly induced gene expression changes in the mouse dorsal hippocampus. Volcano plot showing 3997 upregulated (log₂ fold change > 0.38, p < 0.05, red) and 3140 downregulated (log₂ fold change < −0.38, p < 0.05, blue) genes in the dorsal hippocampus of the reactivated group compared to the non-reactivated group. B Relationship between human PTSD reexperiencing-associated expression changes and mouse fear memory retrieval-associated expression changes. All of the 16,721 genes analyzed in both humans (i.e., the female PTSD patients) and mice are included in this figure. The x-axis represents the -log₁₀(p) of correlations between human PTSD reexperiencing symptoms and blood gene expression levels measured by microarray. The y-axis represents the -log₁₀(p) of expression change after fear memory retrieval in mouse hippocampus measured by RNA-sequencing. Blue dots indicate those genes that show the same (consistent) direction of change in expression (i.e., either upregulation or downregulation) between humans and mice, while orange dots indicate those genes that show the opposite (inconsistent) direction of change in expression between humans and mice. C, D Consistency between human PTSD reexperiencing-associated expression changes and mouse fear memory retrieval-associated expression changes. C Pie chart showing the proportion of consistently vs. inconsistently regulated genes between human and mouse data for the 413 overlapping genes. The four (i.e., 2 × 2) patterns by upregulation and downregulation for humans and mice are distinguished. D Relationship between human PTSD reexperiencing-associated expression changes and mouse fear memory retrieval-associated expression changes for the 413 genes. The x-axis represents standardized correlation coefficients between human PTSD reexperiencing symptoms and blood gene expression levels. The y-axis represents standardized fold change after fear memory retrieval in mouse hippocampus.

Fig. 4. Integrative human and mouse transcriptome analysis.

A Flow diagram of integrative human and mouse transcriptome analysis. The top left and right arms describe the flow of human and mouse transcriptome data analysis, respectively. The bottom two boxes indicate further selection steps by integrating the human and mouse data. The human data were examined at the probe level, that is, different microarray probes within the same gene were distinguished. B List of the 15 genes overlapping across all three datasets. In this figure, gene symbols are represented using human gene nomenclature, i.e., all letters are italicized and in upper-case (e.g., PDE4B).

Fig. 5. Relationship of PDE4B expression levels with PTSD diagnosis, symptoms, and DNA methylation levels in humans.

A, B, C Relationship of PDE4B expression levels with PTSD diagnosis and associated symptoms in humans. Plots in purple indicate PTSD patients and those in white indicate healthy controls. A Scatterplot showing the correlation between PDE4B expression levels and reexperiencing symptoms in PTSD (n = 31). B Combined dot- and box-plot showing PDE4B expression levels in PTSD patients (n = 32) compared to healthy controls (n = 16). *p = 0.031. C Scatterplot showing the correlation between PDE4B expression levels measured by microarray and trait anxiety symptoms assessed with the State-Trait Anxiety Inventory-Trait scores in the total sample (n = 48). D, E, F Mediation model in which PDE4B expression levels mediate the relationship between PDE4B cg14227435 methylation status and PTSD reexperiencing symptoms in patients. D Schema of the mediation model tested. The independent variable (X) was cg14227435 methylation level. The dependent variable (Y) was reexperiencing symptom severity. The mediator variable (M) was PDE4B mRNA expression levels. E Scatterplot showing the correlation between cg14227435 methylation levels and PDE4B expression levels in patients (n = 32). F Scatterplot showing the correlation between cg14227435 methylation levels and reexperiencing symptom severity in patients (n = 31).

Fig. 6. Changes of mRNA levels in the murine dorsal hippocampus and peripheral blood following IA memory retrieval.

A Experimental design. B Changes of mRNA levels in the dorsal hippocampus. Non-reactivated (NR), n = 6–10; 30 min after reactivation (React-30), n = 6–10. C Changes of mRNA levels in peripheral blood. NR, n = 8–10; React-30, n = 8–11; 90 min after reactivation (React-90), n = 8–10. *p < 0.05, post hoc Bonferroni’s test; #p < 0.05, Student’s t-test. Error bars indicate SEM.