MicroRNA as repressors of stress-induced anxiety: the case of amygdalar miR-34

Abstract

The etiology and pathophysiology of anxiety and mood disorders is linked to inappropriate regulation of the central stress response. To determine whether microRNAs have a functional role in the regulation of the stress response, we inactivated microRNA processing by a lentiviral-induced local ablation of the Dicer gene in the central amygdala (CeA) of adult mice. CeA Dicer ablation induced a robust increase in anxiety-like behavior, whereas manipulated neurons survive and appear to exhibit normal gross morphology in the time period examined. We also observed that acute stress in wild-type mice induced a differential expression profile of microRNAs in the amygdala. Bioinformatic analysis identified putative gene targets for these stress-responsive microRNAs, some of which are known to be associated with stress. One of the prominent stress-induced microRNAs found in this screen, miR-34c, was further confirmed to be upregulated after acute and chronic stressful challenge and downregulated in Dicer ablated cells. Lentivirally mediated overexpression of miR34c specifically within the adult CeA induced anxiolytic behavior after challenge. Of particular interest, one of the miR-34c targets is the stress-related corticotropin releasing factor receptor type 1 (CRFR1) mRNA, regulated via a single evolutionary conserved seed complementary site on its 3' UTR. Additional in vitro studies demonstrated that miR-34c reduces the responsiveness of cells to CRF in neuronal cells endogenously expressing CRFR1. Our results suggest a physiological role for microRNAs in regulating the central stress response and position them as potential targets for treatment of stress-related disorders.

Figures

Figure 1.

In vivo depletion of Dicer in mice central amygdala leads to an increase in anxiety-like behavior. A, Schematic representation of site-specific deletion of Dicer in the CeA using conditional Dicer mouse model and a lentiviral-mediated Cre expression. Lentiviruses expressing sdCre were injected into the CeA of mice carrying both floxed-dicer and floxed-transcriptional “stop”–YFP alleles at the age 9–10 weeks. Infected cells in the amygdala undergo Cre-mediated recombination of loxP sites, leading to the ablation of Dicer gene and the induction of YFP expression. In the dark/light transfer test, two-tailed Student's t test revealed that CeA–DCR–KO mice (n = 7) spent significantly less time in light (B), visited the illuminated chamber significantly fewer times (C), and traveled a significantly shorter distance in the light chamber compared with CeA–GFP controls (D) (n = 7). E, F, Representative traces of CeA–GFP control (E) and CeA–DCR–KO (F) activity in the DLT as recorded by the VideoMot system (TSE Systems). In the open-field test, two-tailed Student's t test revealed that CeA–DCR–KO mice (n = 6) spent significantly less time in the center of the OF arena (G) and had a tendency toward longer latency to enter the center compared with CeA–GFP controls (H) (n = 6), whereas total distance traveled in the OF arena was not affected (I). J, K, Representative traces of CeA–GFP control (J) and CeA–DCR–KO (K) activity in the OF arena as recorded by the VideoMot system (TSE Systems). L, Home-cage locomotion is similar for CeA–DCR–KO and CeA–GFP mice. Insets represent average light- or dark-phase locomotion over 4 d of testing (repeated-measures two-way ANOVA; dark, p = 0.91; light, p = 0.45). All behavioral tests were administered 2–7 weeks after viral injection. M, Plasma corticosterone levels as measured by ELISA under basal conditions (0) and 30, 90, and 150 min after acute restraint stress (n = 4 and 5, Cre and control, respectively). DCR, Dicer. Bars represent mean ± SEM. *p < 0.05, #p < 0.1.

Figure 2.

sdCre-infected neurons in vivo and in vitro. A, Schematic representation of site of delivery adapted from Paxinos and Franklin digital mouse brain atlas. B, Enlargement of the amygdala region corresponding to the injection site at 8 weeks after lentiviral infection. C, A representative microscope image of virally infected CeA, after behavioral studies. YFP expression was assessed using immunohistochemistry for GFP. D–G, CeA adult neurons, depleted of Dicer-processed miRNAs in vivo. Neuronal identity of infected cells was assessed using immunohistochemistry for NeuN (red). D, Confocal image of neurons stained for both GFP and NeuN (arrows) at the infection site. Some non-infected neurons, negative for GFP (#), and infected non-neuronal cells, negative for NeuN (star), can also be identified in the field. E–G, An isolated, single DCR–KO neuron (found at the periphery of injection) stained both for NeuN (E) and GFP (F). G, Merge photomicrograph of E and F. H, Cortical primary neurons derived from E18 mouse embryos carrying both conditional Dicer alleles (Dcrflox/flox) and conditional R26R–YFP reporter alleles, after sdCre infection, express YFP and appear to exhibit intact gross morphology and synaptic associations 5 d after sdCre introduction. I, Semiquantitative RT-PCR for Dicer or control (S16) mRNA derived from uninfected primary neurons (None) or primary neurons infected with sdCre-expressing lentiviruses (Cre), 5 d after infection. J, Densitometry analysis for Dicer expression normalized to S16 expression, presented as ratio compared with non-infected control (None). CeC/CeL/CeM, Central amygdaloid nucleus, central, lateral, and medial divisions, respectively.

Figure 3.

CeA neurons survive 8 weeks after sdCre expression. A–C, Representative infection sites for GFP (B) and LV–sdCre (C), immunostained for GFAP. D, GFAP immunoreactivity quantification in the region represented by a dashed rectangle in A–C. E–G, Region of dense sdCre infection immunostained for GFP (E) and GFAP (F). G, Overlay of E and F. H–M, Representative infection sites for LV–sdCre injection (H–J) and LV-GFP injection (K–M). N, Quantification of cell density (Hoechst-positive nuclei/tissue volume) in the region infected by sdCre and GFP.

Figure 4.

Differential expression of miRNAs in the amygdala 90 min after acute stress. A, Agilent array results. B, Affymetrix array results. Normalized values are depicted as log2 ratio (stress vs control) of spot intensity plotted against average intensities across conditions (n = 2, 2). The intensity of each miRNA was calculated as the average normalized intensity across biological repeats. miR-100, miR-15a, miR-15b, miR-34a, miR-34c, and miR-92a are indicated in red. miR-124, a well-established neuronal marker not affected by the stress protocol, is indicated in white. C, qRT-PCR for selected miRNA. *p < 0.05; #p < 0.1. D, Stress-response-related predicted targets of miRNAs upregulated by stress. One hundred fifty-seven genes are common to EIMMo search for upregulated miRs and AmiGO “response to stress” annotation. Venn diagram created using Venny (http://bioinfogp.cnb.csic.es/tools/venny/index.html).

Figure 5.

miR-34c overexpression in the CeA. A, RT-qPCR showing miR-34c upregulation 90 min after acute restraint stress in a group of mixed background mouse strain (n = 2 and 2). B, RT-qPCR showing miR-34c upregulation 2 weeks following 10 d of daily social defeat stress (n = 15 and 6). C, miR-34c downregulation in the DCR–KO amygdala primary culture. D, Schematic representation of the lentiviral expression plasmids used for the production of LV–miR-34c and LV–EGFP. E, Illustration of LV–miR-34c and LV–EGFP injection sites. Red circles represent the punch area used for RNA extraction. F, RT-qPCR verification of miR-34c overexpression in the amygdala 7 d after viral injection of LV–miR-34c relative to LV–EGFP control. Expression is presented as fold change relative to control levels (n = 6 and 4 punches, respectively). G, Left, Schematic representation of site of delivery adapted from Paxinos and Franklin digital mouse brain atlas. Middle, Enlargement of the amygdala region corresponding to the injection site. Right, A representative microscope image of virally infected CeA, after behavioral studies. GFP expression was assessed using immunohistochemistry. RRE, Rev-responsive element; cPPT, central polypurine tract; WPRE, Woodchuck hepatitis posttranscriptional regulatory element. *p < 0.05.

Figure 6.

Over-expression of miR-34c in mice central amygdala has anxiolytic properties. A–D, In the dark/light transfer test, CeA–EGFP controls show indices of increased anxiety 24 h after acute restraint stress, whereas CeA–miR-34c–OE mice behavior remains unaffected after stress (repeated-measures two-way ANOVA, asterisks represent post hoc Student's t test comparisons). At 24 h after acute restraint stress, CeA–miR-34c–OE mice (n = 6) spent more time in light (A), visited the illuminated chamber more (B), traveled a significantly longer distance in the light chamber (C), and showed a longer latency to enter the light chamber compared with CeA–EGFP controls (D) (n = 6). E, In the EPM test, CeA–miR-34c–OE mice traveled a significantly larger percentage of their overall distance in the open arms and spent more time in the open arms compared with CeA–EGFP controls. F, CeA–miR-34c–OE present normal home-cage locomotion compared with CeA–EGFP mice. Insets represent average light- or dark-phase locomotion over 2.5 d of testing. U, Unstressed; S, 24 h after stress. All behavioral tests were administered 2–7 weeks after viral injection. Bars represent mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.005; #p < 0.1.

Figure 7.

miR-34c regulates the expression of corticotropin releasing factor receptor type 1 via one seed match on its 3′ UTR. A, CRFR1 3′ UTR is 995 bases long and carries one conserved miR seed match. Interspecies alignment of a 70 base sequence including the miR-34 seed match demonstrates its conservation among mammals (adapted from www.targetscan.org). B, Illustration of intact (top) and mutant (bottom) PsiCHECK-2–CRFR1-3′ UTR luciferase constructs. C, In vitro validation of miR-34c overexpression after transfection of HEK293T cells with pEGFP–miR-34c. D, E, Luciferase activity measured in HEK293T cells cotransfected with miR-34c–EGFP overexpressing, miR-34a-EGFP overexpressing, or GFP-expressing plasmid and a luciferase reporter plasmid carrying either an intact (D) or mutant (D, E) form of CRFR1-3′ UTR. Renilla luciferase activity was normalized by firefly luciferase expression levels and presented as percentage of activity achieved by the mutant form of CRFR1-3′ UTR at the presence of miR-34c overexpression. CRFR1-3′ UTRmut, Mutated from of CRFR1 3′ UTR lacking the seed match for miR-34; UTR, untranslated region; CMVp, cytomegalovirus promoter; pA, SV40 polyadenylation signal. **p < 0.01.

Figure 8.

miR-34 functionally regulates CRF responsiveness of N2a cells that endogenously express CRFR1. A, RT-PCR for CRFR1, CRFR2, and HPRT on RNA produced from N2a murine neuroblastoma cell line (N2a), HT22 murine hippocampal cell line (HT22), and mouse muscle biopsies (Muscle). B, RT-qPCR for CRFR1 in N2a cells or amygdala punches, normalized by HPRT levels. Expression is presented as fold change relative to CRFR1 levels in the amygdala. Levels of CRFR1 mRNA in HT22 used as negative control are undetectable after 40 PCR cycles. N.D, Nondetected. C, In vitro example of miR-34c overexpression in N2a cells after stable infection with LV-E/Syn–miR-34c, measured by qRT-PCR. D, Illustration of the cAMP–RE–firefly luciferase construct. E, F, Luciferase activity measured in N2a cells stably infected with either LV-E/Syn–EGFP (E, F, gray bars), LV-E/Syn–miR-34c (E, white bars), or LV-E/Syn–miR-34a (F, white bars) and transfected with a firefly luciferase reporter plasmid driven by a cAMP–RE. A Renilla luciferase vector served as normalizer. Firefly luciferase activity was normalized by Renilla luciferase expression levels and presented as fold change relative to basal luciferase activity (BSA treatment). cAMP–RE, A fragment of the EVX1 gene, which contains a potent cAMP response element. **p < 0.01, ***p < 0.005.