miR-483-5p offsets functional and behavioural effects of stress in male mice through synapse-targeted repression of Pgap2 in the basolateral amygdala

Abstract

Severe psychological trauma triggers genetic, biochemical and morphological changes in amygdala neurons, which underpin the development of stress-induced behavioural abnormalities, such as high levels of anxiety. miRNAs are small, non-coding RNA fragments that orchestrate complex neuronal responses by simultaneous transcriptional/translational repression of multiple target genes. Here we show that miR-483-5p in the amygdala of male mice counterbalances the structural, functional and behavioural consequences of stress to promote a reduction in anxiety-like behaviour. Upon stress, miR-483-5p is upregulated in the synaptic compartment of amygdala neurons and directly represses three stress-associated genes: Pgap2, Gpx3 and Macf1. Upregulation of miR-483-5p leads to selective contraction of distal parts of the dendritic arbour and conversion of immature filopodia into mature, mushroom-like dendritic spines. Consistent with its role in reducing the stress response, upregulation of miR-483-5p in the basolateral amygdala produces a reduction in anxiety-like behaviour. Stress-induced neuromorphological and behavioural effects of miR-483-5p can be recapitulated by shRNA mediated suppression of Pgap2 and prevented by simultaneous overexpression of miR-483-5p-resistant Pgap2. Our results demonstrate that miR-483-5p is sufficient to confer a reduction in anxiety-like behaviour and point to miR-483-5p-mediated repression of Pgap2 as a critical cellular event offsetting the functional and behavioural consequences of psychological stress.

Fig. 1. Regulation of miRNAs in mice by restraint stress in the basolateral amygdala.

a C57BL/6J mice were subjected to a 6-h of restraint stress or control, stress-naïve animals were left undisturbed in their home cages. Mice were then anaesthetised and (b) miRNA was isolated from dissected basolateral amygdalae (c). The resulting miRNA was then hybridised onto microarrays to examine miRNAs’ expression levels. d Bioinformatics analysis of the microarray data revealed upregulation of miR-1192, miR-1224, miR-1892, miR-1894-3p and miR-483-5p in the basolateral amygdala in mice after stress as highlighted by red dotted line in the volcano plot. e Pre- (red rectangles) and post-stress (blue rectangles) expression levels identified miR-483-5p as the most significantly elevated miRNA. Data are presented on a box and Whiskers plot where the box extends from 25th to 75th percentile, and whiskers range from minimum and maximum value, with the centre denoting the median value; n = 4 microarrays where each array was hybridised with pulled mRNA of 3 individual animals (12 animals in total); statistical significance is assessed by two-sided unpaired t test, *p < 0.05, **p < 0.01, ***p < 0.001 vs control mice (f) To verify the microarray results, miRNA isolated from basolateral amygdalae of mice subjected to 6-h immobilisation was analysed by quantitative RT-PCR using miRNA-specific probes. qRT-PCR confirmed that miR-483-5p was among the miRNAs most prominently induced by stress in the basolateral amygdala in mice. Data are presented as dot plots with mean ± SEM; n = 3 animals per group; statistical significance is assessed by two-sided unpaired t test, **p < 0.01, ***p < 0.001 vs control mice. Source data and the exact p-values are provided in a Source Data file.

Fig. 2. miR-483-5p is expressed in amygdala neurons in mice upon restraint stress.

Fluorescence in situ hybridisation (FISH) targeting miR-483-5p (green) and immunohistochemistry for neuronal marker NeuN (a, red) and synaptic marker HOMER (a, purple) or PGAP2 (b, purple) revealed miR-483-5p expression on amygdala neurons following restraint stress. The arrows indicate the colocalization of the miR-483-5p with HOMER (a) or PGAP2 (b). Scale bar = 200 μm (upper panels) and 5 μm (lower panels). Representative images, experiments were performed independently with similar results on 3 animals. LA—lateral amygdala, BLA—basolateral amygdala, CEA—central amygdala.

Fig. 3. miR-483-5p is induced by stress in mice in the juxtasynaptic compartment to control dendritic arborisation and spine morphology.

a Mice (stress-naïve or subjected to 6-h immobilisation stress) were anaesthetised, their basolateral amygdalae were dissected and the cytosol and synaptosomal fractions were isolated. miRNAs were extracted from both fractions, reverse transcribed, and their expression levels analysed by quantitative RT-PCR. The study revealed that stress elevated the expression of miR-483-5p predominantly in the synaptosomal fraction. Data are presented as dot plots with mean ± SEM; n = 4 animals per group; statistical significance is assessed by 2-way ANOVA with Tukey’s multiple comparisons test as a post-hoc test, **p < 0.01 vs an indicated group. b Sholl analysis of neuronal structure revealed that overexpression of miR-483-5p (mimicking its stress-induced upregulation) in primary mouse amygdala neurons, caused shrinkage of the distal parts of the dendritic arbour (130–160 μm from the soma). Data are presented as mean ± SEM, n_neurons (scrambled) = 21; n_neurons (miR-483-5p) = 34 neurons; statistical significance is assessed by mixed effect model with 2-step linear procedure of Benjamin, Krieger and Yekuteli for individual comparisons, **p < 0.01, *p < 0.05 vs scrambled. c Representative images of primary amygdala neurons transfected with either the miR-483-5p or scrambled control vector. Scale bar = 50 μm. d Overexpression of miR-483-5p in primary mouse amygdala neurons caused an increase in the proportion of mature, mushroom-like spines, with a simultaneous decrease in the immature, filopodia-like protrusions. Data are presented on a violin plot where the centre denotes median value; n_neurons (scrambled) = 20 and N_regions (scrambled) = 59; n_neurons (miR-483-5p) = 21 neurons and N_regions (miR-483-5p) = 54, statistical significance is assessed by 2-way ANOVA with Bonferonni’s correction as a post-hoc test, ****p < 0.0001 vs scrambled. e Representative images of dendritic spines in primary mouse amygdala neurons transfected with either the miR-483-5p or scrambled control vector. Scale bar = 5 μm. f The density of dendritic spines in primary mouse amygdala neurons was not affected by the overexpression of miR-483-5p. Data are presented on a violin plot where the centre denotes median value, statistical significance is assessed by two-sided unpaired t test, n_neurons (scrambled) = 20 and N_regions (scrambled) = 59; n_neurons (miR-483-5p) = 21 neurons and N_regions (miR-483-5p) = 54. g Mice were first injected with lentiviruses carrying miR-483-5p-EGFP, Pgap2-shRNA-EGFP or Pgap2^miR-483-5p-EGFP (Pgap2 resistant to miR-483-5p) into basolateral amygdala, and following 6-h immobilisation dendritic spines were visualised using Dil staining. Expression of both miR-483-5p and Pgap2-shRNA resulted in increased proportion of mature mushroom-shaped spines, while no change to a population of mushroom spines were found following expression of miR-483-5p-resistant Pgap2 when compared to the control. N = 3 animals per group, 1.6–5.5 mm of the dendrites per group, n = 338–4767 spines per group, n_regions (scrambled) = 50, n_regions (miR-483-5p) =45, n_regions (Pgap2) = 78. Expression of both miR-483-5p and Pgap2-shRNA lead to about 10% increase of dendritic spine density. Data are presented on a violin plot where the centre denotes median value; statistical significance is assessed by 2- (graphs 1 and 3) or 1-way ANOVA (graph 2) or by two-sided t test (graph 4), ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 vs an indicated group. h Representative images of Dil stained neurons. Scale bar = 2 μm. Source data and the exact p-values are provided in a Source Data file.

Fig. 4. Identification and characterisation of stress-related genes in the amygdala.

a Bioinformatics database cross-comparison identified 12 genes expressed in the amygdala (the EIMMo database), which are also regulated in response to stress (the AmiGo search engine) as potential targets of miR-483-5p (the TargetScan algorithm). b To experimentally identify genuine stress-related miR-483-5p targets, Neuro2a cells were transfected with the miR-483-5p-expressing vector (or scrambled control vector) and gene expression levels following dexamethasone were measured by quantitative RT-PCR (see also Supplementary Fig. 1). The expression of Pgap2, Gpx3 and Macf1 mRNAs was suppressed by miR-483-5p, confirming these genes as genuine miR-483-5p targets. Data are presented as dot plots with mean ± SEM; n = 5 animals per group; statistical significance is assessed by 2-way ANOVA with Šídák’s multiple comparisons test as a post-hoc test, ****p < 0.0001, ***p < 0.001, **p < 0.01 vs an indicated group. Source data and the exact p-values are provided in a Source Data file.

Fig. 5. miR-483-5p controls set of target gene expression in the juxtasynaptic compartment of amygdala neurons.

a C57/BL/6J mice were subjected to 6-h immobilisation (the stress-naïve group remained in their home cages), anaesthetised, and their basolateral amygdalae dissected rapidly on ice from coronal brain sections (dissection boundaries shown as red dotted lines) and the cytosol and synaptosomal fractions were isolated. mRNAs were extracted from both fractions, reverse transcribed, and Pgap2, Gpx3 and Macf1 gene expression was measured by quantitative RT-PCR. b qRT-PCR revealed 50–60% stress-induced decrease in the expression of Pgap2 and Macf1 in the whole-cell amygdala homogenate. In contrast only the expression of Pgap2 was suppressed in the synaptosomal fraction. Data are presented as dot plots with mean ± SEM; n = 3–6 animals per group; statistical significance is assessed by 2-way ANOVA with Šídák’s multiple comparisons test as a post-hoc test, *p < 0.05 vs an indicated group. Source data and the exact p-values are provided in a Source Data file. c Immunohistochemistry revealed that PGAP2, MACF1 and GPX3 proteins are present in the neurons in the direct vicinity of the cell membrane. Representative images, experiments were performed independently with similar results on 3 animals. Scale bar = 20 μm.

Fig. 6. miR-483-5p controls the expression of Pgap2, Gpx3 and Macf1 genes directly through binding to 3’ UTR regions.

a Alignment of miR-483-5p sequence (red) with 3’UTRs of mouse Pgap2, Gpx3 and Macf1 mRNA sequences (violet, green and blue, respectively). Complementary nucleotides within miR-483-5p and the 3’UTRs of its target mRNAs are connected with vertical lines. Arrowheads point to the nucleotides within the 3’UTR seed match sequences mutated in order to disrupt miR-483-5p binding. b To perform in vitro luciferase assay the 3’UTR of Pgap2, Gpx3 and Macf1 mRNA were inserted into the pmiRGLO Dual-Luciferase miRNA-target expression vector, downstream of the firefly luciferase reporter gene. Neuro2a cells were double-transfected with the above vector together with the miR-483-5p-expressing (or scrambled control) vector, and the intensity of the luminescence produced by firefly luciferase was measured. The signal was normalised to the Renilla luciferase luminescence. In all three cases miR-483-5p expression reduced the Firefly luciferase activity, confirming that miR-483-5p suppressed Pgap2, Gpx3 and Macf1. Mutating the seed match sequences of the 3’UTR (arrowheads in a) abolished miR-483-5p-mediated suppression, confirming that miR-483-5p affects expression of all three genes directly. Data are presented as dot plots with mean ± SEM; n = 4–6 animals per group; statistical significance is assessed by 2-way ANOVA with Šídák’s multiple comparisons test as a post-hoc test, **p < 0.01, *p < 0.05 vs an indicated group. Source data and the exact p-values are provided in a Source Data file. c The sequence of miR-483-5p from mouse, rat, man and guinea pig. d miR-483-5p target sequences located in 3’UTRs of the mouse, rat, human, chimp and rabbit Pgap2 genes. Conserved nucleotides marked in red.

Fig. 7. Expression of miR-483-5p in the BLA in mice promotes anxiolysis through repression of Pgap2 gene.

a Lentiviral particles containing the miR-483-5p sequence (pLenti-UbC-EGFP-mmu-miR-483-5p), or the control scrambled sequence, were injected bilaterally into the basolateral amygdalae (BLA) of mice. b Representative BLA injection site visualised with reporter protein GFP immunohistochemistry. Injection site was verified for all animals—see Supplementary Fig. 3 for the locations of injection sites. c After three weeks of recovery anxiety-like behaviour was tested in the elevated plus-maze. d Overexpression of miR-483-5p in the BLA led to decreased anxiety-like behaviour in mice compared to control animals as measured by the decrease in an anxiety index, a ratio between total arm entries to open arm entries. e The total number of arm entries (a measure of total locomotor activity) was not affected by the expression of miR-483-5p. f Representative elevated-plus maze movements trajectories of mice injected with either miR-483-5p-containing or control lentiviral particles. g shRNA-mediated suppression of Pgap2 (pLenti-H1-shRNA-Rsv(GFP-Bsd vector) in the BLA led to decreased levels of anxiety, thus recapitulating the behavioural effect of miR-483-5p. h Similarly to miR-483-5p, the total number of arm entries was not affected by the suppression of the Pgap2 gene. i Representative elevated-plus maze movement trajectories of mice injected with lentiviruses containing either the Pgap2-supressing or control sequences. Data are presented as dot plots with mean ± SEM; n (scrambled) = 8, n (miR-483-5p & Pgap2 shRNA) = 7 animals per group; statistical significance is assessed by unpaired two-sided t test, *p < 0.05 vs control (scrambled) group. Source data and the exact p-values are provided in a Source Data file.

Fig. 8. Expression of miR-483-5p in the BLA prevents anxiogenesis through downregulation of Pgap2 gene.

Lentiviral particles containing the miR-483-5p, Pgap2^miR-483-5p or the control scrambled sequence, were injected bilaterally into the basolateral amygdalae of mice. Following three weeks animals were subjected to 6-h immobilisation and next day tested on the elevated plus-maze. a Overexpression of miR-483-5p in the BLA prevented development of the anxiety-like behaviour when compared with animals expressing scrambled, control sequence while co-expression of miR-resistant form of Pgap2 with miR-483-5p reinstated anxiety-like behaviour to the level of control group. b The total number of arm entries (a measure of total locomotor activity) was not affected by the expression of miR-483-5p or Pgap2^miR-483-5p. c Representative elevated-plus maze movement trajectories of mice injected with lentiviral particles. d shRNA-mediated suppression of Pgap2 in the BLA led to decreased levels of anxiety-like behaviour in mice, thus recapitulating the behavioural effect of miR-483-5p. e Similarly to miR-483-5p, the total number of arm entries was not affected by the suppression of the Pgap2 gene. f Representative movement trajectories of mice injected with lentiviruses containing either the Pgap2-supressing or control sequences. Data are presented as dot plots with mean ± SEM; n (scrambled & Pgap2^miR-483-5p) = 7, n (miR-483-5p & Pgap2 shRNA) = 8 animals per group; statistical significance is assessed by Kruskal–Wallis test with Dunn’s multiple comparison test (a) or by a 1-way ANOVA with Bonferroni’s multiple comparisons test (b) or by an unpaired two-sided t test (d, e), ***p < 0.001, *p < 0.05 vs control group. Source data and the exact p-values are provided in a Source Data file.