Septal miR-132-3p and miR-124-3p expression is dynamically altered following social fear acquisition and extinction

Since miR-132 has frequently been reported as a dynamic regulator of neuronal plasticity, learning, and memory, we investigated its time-dependent expression within the septum of SFC+ and SFC− mice, 30 min, 90 min, 180 min, and 24 h after social fear acquisition, or 30 min, 90 min, and 180 min after social fear extinction (Fig. 1D and Supplementary Fig. S1E). All SFC+ mice showed similar social fear learning, as they received an equal number of CS-US-pairings (foot shock when sniffing the conspecific) during  acquisition (Supplementary Table S1). During extinction, SFC+ and SFC− mice showed the expected percentage of social investigation (Supplementary Table S2).

In septal tissue of SFC+ mice, miR-132-3p was found to be upregulated 90 min after acquisition and was not altered 180 min after acquisition compared to respective SFC− mice (Fig. 1D). In response to extinction, miR-132-3p levels were significantly reduced 180 min after extinction independent of the animal´s conditioning status but remained unchanged at all other time-points in comparison to the respective post-acquisition levels. 30 min after acquisition and extinction as well as 24 h after acquisition, no changes in miR-132-3p were detected in SFC+ compared to SFC− mice. In sum, this revealed a specific temporal dynamic of miR-132-3p transcript level in response to SFC.

Additionally, septal miR-124-3p levels were found to be upregulated in SFC+ compared to SFC− mice 90 min after acquisition (Supplementary Fig. S1B, F). Moreover, miR-124-3p transcript levels in SFC− mice were increased 90 min after extinction compared to respective SFC− post-acquisition levels. However, septal miR-124-3p transcript levels remained unchanged in all groups at all earlier (30 min) or later (180 min, 24 h) time points.

Various forms of social encounters stimulate the central release of OXT and subsequent OXT receptor (OXTR)-mediated signaling [32, 33], and higher levels of intra-septal OXT release were associated with elevated investigation levels [6], which might be at least partly mediated by miRNAs. To investigate whether miR-132-3p or miR-124-3p transcription is altered in response to repeated social encounters, SFC− and SFC+ mice were exposed to either the extinction protocol (3 non-social and 6 social stimuli) or to nine non-social stimuli only. Independent of the conditioning status, neither septal miR-132-3p (Fig. 1E) nor miR-124-3p (Supplementary Fig. S1C) levels were altered 90 min post-exposure. To further analyze whether these miRNAs are altered after OXTR activation, septal miRNA transcript levels were assessed 90 min and 180 min after icv infusion of OXT or Veh. Here, no significant alterations of septal miR-132-3p (Fig. 1F) and miR-124-3p (Supplementary Fig. S1D) were found.

Transcript levels of miR-132-3p are known to increase in response to neuronal activation [17]. Therefore, we assessed the cellular activation within the LS in response to SFC acquisition or extinction via c-Fos immunohistochemistry. SFC+ mice received equal CS-US pairings (Supplementary Table S1) during acquisition and displayed the expected initially reduced social investigation during extinction compared to SFC− mice (Fig. 1G). c-Fos levels were increased 90 min after acquisition in SFC+ mice compared to respective SFC− mice, whereas no alterations in response to social fear conditioning were observed after extinction (Fig. 1H-I). Whether the observed increase in septal miR-132-3p reflects local cellular activation needs further investigation.

In summary, the expression of both miR-132-3p and miR-124-3p in the septum is dynamically altered in response to SFC. As our preliminary data indicate an upregulation of miR-132-3p specifically within the hypothalamic paraventricular nucleus by icv OXT infusion in male rats [34], we decided to further decipher the functional involvement of miR-132-3p in SFC-related behavior and OXT-mediated social fear reversal.

Septal miR-132-3p inhibition impairs, whereas its overexpression facilitates extinction of social fear

The dynamic alterations of septal miR-132-3p in response to SFC (Fig. 1D and Supplementary Fig. S1E) prompted us to further analyze the involvement of septal miR-132-3p in social fear acquisition and extinction using LNA-induced inhibition or AVV-induced overexpression of miR-132-3p (Fig. 2A). Septal LNA and AAV infusion sites, as well as viral expression, were validated by fluorescence microscopy (Fig. 2B).

Fig. 2: Inhibition of septal miR-132-3p impairs, whereas its overexpression facilitates extinction of social fear.

A Schematic representation of septal treatments: local miR-132-3p inhibition was achieved by septal infusion of a locked nucleic acid (LNA) complementary to miR-132-3p (Inh-LNA; 0.5 nmol) or a scrambled sequence as control (Scr-LNA; 0.5 nmol). miR-132 overexpression (132-OE) was achieved by septal infusion of an adeno-associated virus (AAV; 4.9 × 1012 GC/ml) or control AAV (Ctrl-OE). Infusions were performed 48 h (LNA) or 3 weeks (AAV) prior to social fear acquisition. B Representative immunofluorescent verification of septal microinfusion placement and adequate intracellular localization of LNA and expression of AAV. LNA and AAV infusions into the lateral septum resulted in cellular transfection of the entire septum. Scale bar represents 100 µm; LV: lateral ventricle. C Number of CS-US pairings presented to Inh-LNA and Scr-LNA-infused conditioned (SFC+) mice during acquisition of social fear. D Percentage of time investigating the presented 3 non-social and 6 social stimuli during social fear extinction of Inh-LNA and Scr-LNA-treated SFC+ and unconditioned (SFC−) mice. *p < 0.05, (*) s2: p = 0.054, s5: p = 0.066 Inh-LNA/SFC+ vs Scr-LNA/SFC+. E Percentage of investigation time of 6 social stimuli presented during social fear recall. F Number of CS-US pairings presented to 132-OE or Ctrl-OE AAV-infused SFC+ mice during acquisition of social fear. G Percentage of nvestigation time of non-social and social stimuli presented during social fear extinction of 132-OE and Ctrl-OE SFC+ and SFC− mice. **p < 0.01, *p < 0.05 132-OE/SFC+ vs Ctrl-OE/SFC+. H Percentage of investigation time of 6 social stimuli during social fear recall. Data represent mean ± SEM; n = 8–12/group.

During acquisition, neither miR-132-3p inhibition (Inh-LNA; Fig. 2C) nor overexpression of miR-132-3p (132-OE; Fig. 2F) within the septum altered the number of CS-US pairings in SFC+ mice in comparison to their controls (LNA: Scr-LNA; AAV: Ctrl-OE). During extinction, no treatment effect was found on the investigation of the three non-social stimuli (Fig. 2D, G). In contrast, septal miR-132 inhibition differentially delayed extinction of social fear in SFC+ mice. First, Inh-LNA-treated SFC+ animals showed lower investigation of social stimuli 2 (by trend), 4, and 5 (by trend) compared to Scr-LNA-infused SFC+ mice (Fig. 2D). Second, Inh-LNA/SFC+ mice still showed lower social investigation times of stimuli 1 to 5 compared to Inh-LNA/SFC− mice, indicating persistent social fear, whereas Scr-LNA/SFC+ controls displayed lower investigation only of stimuli 1 to 3 when compared to Scr-LNA/SFC− mice (for statistical details see Supplementary Table S6).

In support of the functional involvement of miR-132-3p in extinction, AAV-induced overexpression of septal miR-132 (132-OE; Fig. 2G) facilitated fear extinction reflected by (i) higher investigation times of 132-OE-infused SFC+ mice during exposure to social stimulus 4 and 5 compared to Ctrl-OE-infused SFC+ animals, and (ii) reduced social investigation only during social stimuli 1–2 in 132-OE/SFC+ compared to 132-OE/SFC− mice, whereas Ctrl-OE/SFC+ mice continued to avoid social stimuli 4 and 5 when compared to Ctrl-OE/SFC− mice (for statistical details see Supplementary Table S6).

Importantly, neither inhibition nor overexpression of septal miR-132 altered social investigation of SFC− mice indicating that septal miR-132 is not involved in general social behavior. During recall, LNA-mediated inhibition (Fig. 2E) and AAV-mediated overexpression (Fig. 2H) of septal miR-132 did not affect investigation times in all groups. Moreover, neither general anxiety-related behavior nor locomotion assessed in the OF/NO test was affected by miR-132-3p inhibition or overexpression (Supplementary Fig. S2). These data indicate that septal miR-132 contributes to the facilitation of social fear extinction.

Hypothalamic miR-132-3p inhibition has been shown to control water homeostasis and body weight gain [35]. However, we did not find alterations in body weight gain 24 h, 48 h, or 3 weeks after bilateral infusion of LNA or AAV into the LS (Supplementary Fig. S2I–J).

Septal miR-132-3p is required for the OXT-induced reversal of social fear

Previous studies reported that OXT signaling within the LS promotes social fear extinction in male [6] and female [3] SFC+ mice. To analyze the involvement of miR-132-3p in OXT-mediated reversal of social fear, septal miR-132-3p of SFC+ mice was inhibited by Inh-LNA infusion and combined with local OXT infusion performed 10 min before extinction resulting in the following groups: Inh-LNA/OXT, Inh-LNA/Veh, Scr-LNA/OXT, and Scr-LNA/Veh (Fig. 3A).

Fig. 3: Septal miR-132-3p is required for the reversal of social fear mediated by oxytocin (OXT) receptor signaling within the lateral septum.

A Schematic representation of septal treatments: 48 h prior to acquisition of social fear, mice were bilaterally infused with either a miR-132-3p inhibitor (Inh-LNA; 0.5 nmol) or scrambled (Scr-LNA; 0.5 nmol) locked nucleic acid (LNA; light gray) into the septum. 10 min prior to extinction of social fear they were locally infused (dark gray) with OXT (5 ng/0.2 µL/hemisphere) or vehicle (Veh; Ringer solution). Additionally, selective downregulation of miR-132 in OXT receptor-expressing neurons was achieved in OXT receptor (OXTR)-Cre mice, which were infused with a pre-miR-132 shRNA (sh132; 9.4 × 1013 GC/ml) or scrambled (shScr; 9.4 × 1013 GC/ml) adeno-associated virus into the septum 3 weeks prior to social fear acquisition. B Number of CS-US pairings during acquisition of social fear presented to conditioned (SFC+) Inh-LNA or Scr-LNA-treated mice that were further infused with OXT or Veh into the LS. C Percentage of investigation time of 3 non-social and 6 social stimuli during social fear extinction in Inh-LNA and Scr-LNA SFC+ mice locally treated with OXT or Veh. **p < 0.01; *p < 0.05 Inh-LNA/OXT vs Scr-LNA/OXT; (#) p = 0.061 Inh-LNA/Veh vs Scr-LNA/Veh. D Percentage of investigation time of  6 social stimuli during social fear recall. **p < 0.01, *p < 0.05 Inh-LNA/OXT vs Scr-LNA/OXT; #p < 0.05 Inh-LNA/Veh vs Scr-LNA/Veh. E Representative immunofluorescent images of septal sh132 or shScr microinfusion placement. Viral shRNA expression in OXTR-Cre mice was turned on using a SICO vector, whereas GFP expression is turned off in the presence of Cre, while mCherry expression remains constitutive. AAV infusions into the lateral septum resulted in neuronal infection of the entire septum. Scale bar represents 100 µm; LV lateral ventricle. F Number of CS-US pairings presented to SFC+ sh132 or shScr-treated mice during social fear acquisition. G Investigation time of non-social and social stimuli during social fear extinction of SFC+ and unconditioned (SFC−) mice treated with septal sh132 or shScr. *p < 0.05 sh132/SFC+ vs shScr/SFC+. H Investigation time of social stimuli during social fear recall. Data represent mean ± SEM; n = 6–12/group.

In concordance with the previous experiment (Fig. 2C–E), septal miR-132-3p inhibition in SFC+ mice neither influenced acquisition of social fear nor the investigation of the 3 non-social stimuli during extinction (Fig. 3B, C). In support of our hypothesis that septal miR-132-3p inhibition prevents the social fear-reversing effect of OXT, significantly lower investigation times were displayed by Inh-LNA/OXT-treated compared to Scr-LNA/OXT-treated SFC+ mice during exposure to social stimuli 1, 3, 4, and 6 during extinction. Additionally, Inh-LNA/Veh treatment resulted in a decreased investigation time of social stimulus 1 compared to Scr-LNA/Veh. During recall (day 3), Inh-LNA/OXT-treated SFC+ mice showed a reduced investigation of social stimuli 2, 3, 4, and 6 compared to Scr-LNA/OXT-treated animals, whereas Inh-LNA/Veh treatment resulted in an overall decreased investigation of social stimuli 1 to 6, when compared to Scr-LNA/Veh-treated animals (Fig. 3D).

In this experiment, however, we could not confirm the reversing effect of OXT per se on extinction [6] in Scr-LNA/OXT-treated SFC+ mice, likely due to the essential pre-treatment septal surgeries and microinfusions. These procedures, together with a reduced post-surgical recovery period of only 72 h resulted in an overall decrease of social investigation times during extinction and recall independent of the treatment (e.g., Scr-LNA/Veh: approx. 60% (Fig. 3C) vs Scr-LNA/SFC+ : approx. 80% (Fig. 2D)). Moreover, lesions of septal tissue, i.e., via cannulation, is known to affect social behavior [36].

To further confirm the involvement of miR-132 in OXTR-mediated facilitation of social fear extinction, pre-miR-132, and concomitantly also mature miR-132-3p, was selectively down-regulated within septal OXTR-expressing neurons of OXTR-Cre mice using an AAV expressing a floxed shRNA against miR-132(sh132; Fig. 3A, E). pre-miR-132 knockdown in septal OXTR-expressing neurons neither affected the number of CS-US pairings during acquisition (Fig. 3F) nor investigation times of non-social stimuli independent of the conditioning status, nor social investigation times in SFC− mice (Fig. 3G). However, pre-miR-132 knockdown, specifically in septal OXTR-expressing neurons, delayed the extinction of social fear in SFC+ mice. This was reflected by (1) reduced investigation of social stimuli 4 and 5 compared with shScr/SFC+ mice, and (2) lower investigation of stimulus 5 compared with sh132/SFC− mice, whereas SFC+ mice treated with shScr showed reduced investigation of stimuli 1 to 4 only (shScr/SFC+ vs shScr/SFC−). This illustrated delayed extinction of social fear by knockdown of miR-132-3p in septal OXTR expressing neurons (for statistical details see Supplementary Table S6). During recall, no differences between the groups were found (Fig. 3H).

Taken together, septal miR-132-3p is significantly involved in signaling downstream of the OXTR and required for OXTR-mediated reversal of social fear, as its inhibition partly prevented the social fear-reversing effect of OXT, and its knockdown selectively in OXTR-expressing neurons impaired social fear extinction.

Septal miR-132-3p influences social fear extinction by regulation of growth differentiation factor 5 (Gdf-5)

To identify targets of septal miR-132-3p, which potentially mediate the observed effects of miR-132-3p on social fear extinction, Ago2-IP analysis was performed 48 h after septal infusion of Inh-LNA or Scr-LNA. In theory, only mRNAs, which are regulated by miRNAs, are enriched after Ago2-IP, when compared to the respective input samples since Ago proteins bind to miRNAs and tightly associate with the target mRNAs. Furthermore, mRNAs regulated by miR-132-3p are not enriched or even de-riched in Inh-LNA samples after Ago2-IP since the inhibitor binds to miR-132-3p, thereby dissociating miRNA-Ago complexes from miR-132-3p-specific target mRNAs (Fig. 4A). As expected, a principal component analysis (PCA) revealed the highest variation between input and Ago2-IP samples, confirming that Ago2 binds to distinct mRNAs during degradation (Supplementary Fig. S4A). Furthermore, the PCA provided evidence that the composition of mRNAs bound by Ago2 varies between Inh-LNA or Scr-LNA treatment, and the transcript abundance of 164 genes differed between both IP conditions (Supplementary Fig. 4B, used cutoff: p < 0.05; Supplementary Tables S12–S14). Further k-means clustering of the differentially enriched genes revealed three clusters (1, 4, 10), including 44 mRNAs, which were enriched in Scr-LNA-IP samples compared to the respective input sample and downregulated in Inh-LNA compared to Scr-LNA samples (Supplementary Fig. 4B). Additionally, 52 genes (clusters 3, 5, 6, 7, 8) were downregulated after Inh-LNA treatment compared to Scr-LNA, but not enriched after Scr-LNA-IP compared to its input. Selected septal mRNAs of these clusters were further analyzed 90 min after acquisition or extinction in SFC− and SFC+ mice using a customized RT2 Profiler PCR Array (Qiagen; Supplementary Fig. 4C; Supplementary Table S4). SFC+ mice included in this analysis received equal CS-US pairings (Supplementary Table S4) during acquisition and showed initially low social investigation time during extinction, as expectced (Supplementary Fig. 4G). From all analyzed target mRNAs, including all annotated GDF-family members, Gdf-5 intensity was found to be increased in the Ago2-IP analysis after septal Inh-LNA treatment (Fig. 4B, C; Supplementary Fig. 4E, F; Supplementary Table S3). Further analysis by miRNA target prediction software (miRDB [37], miRMap [38], miRWalk2 [39], and TargetScan Mouse 7.1 [40]) revealed Gdf-5 as a putative target with an 8mer binding site within the miR-132-3p seed sequence. In confirmation of previous results (Fig. 1D), septal miR-132-3p was increased 90 min after acquisition in SFC+ mice compared to SFC− mice, whereas septal Gdf-5 mRNA levels were decreased by trend (Fig. 4D). Analysis of septal GDF-5 protein levels revealed no alteration in the dimeric pro-form and monomeric mature form in response to acquisition or extinction (Supplementary Fig. 4H). However, the ratio of mature GDF-5 to pro-GDF-5 was decreased in SFC+ mice 90 min after acquisition compared to respective SFC− animals (Fig. 4E), showing reduced septal mature GDF-5 in proportion to pro-GDF-5. This downregulation of septal GDF-5 might result from increased miR-132-3p in SFC+ mice after acquisition. Details of the behavior displayed by these mice during social fear acquisition and extinction are found in Supplementary Table S5.

Fig. 4: Septal miR-132-3p influences social fear extinction by regulation of growth differentiation factor 5 (Gdf-5).

A Schematic representation of the performed Argonaute-RNA-co-immunoprecipitation (Ago2-IP) microarray analysis of septal tissue of mice infused with the miR-132-3p inhibitor (Inh-LNA; 0.5 nmol) or scrambled control (Scr-LNA; 0.5 nmol) locked nucleic acid (LNA). Target mRNAs regulated by miR-132-3p are not enriched or even de-riched in Inh-LNA samples after Ago2-IP, whereas Scr-LNA infusion is supposed to result in enrichment of miR-132-3p target mRNAs. B Heat-map showing the log fold-change to mean intensity of relevant genes after microarray analysis of Ago2-IP samples. Only predicted targets of miR-132-3p (congruent in MiRDB, miRMap, miRWalk2, and TargetScan Mouse 7.1), which show higher intensities in Scr-LNA compared to input samples are shown. C Volcano plot depicting the fold-change of conditioned (SFC+) mice after acquisition (Acq) of social fear (normalized to unconditioned mice (SFC−)/Acq) in dependence of the p-value of all genes detected via PCR Array analysis. D Relative miR-132-3p and Gdf-5 RNA expression in SFC− and SFC+ mice 90 min after social fear Acq or extinction (Ext); *p < 0.05. E Ratio of septal mature GDF-5 to pro-GDF-5 protein level in SFC− and SFC+ mice 90 min after Acq or Ext of social fear; *p < 0.05. F Relative Hmga2 and Gdf-5 mRNA level in Neuro-2a neuroblastoma cells 48 h after transfection with 3 nM of a negative control (neg ctrl), positive control (let-7c; pos ctrl), or miR-132-3p (132-3p Inh) inhibitor. *p < 0.05 vs all groups (Hmga2) or vs pos ctrl (Gdf-5). G Relative Gdf-5 level in Neuro-2a neuroblastoma cells transfected with neg ctrl or 132-3p Inh for 48 h and further stimulated with vehicle (Veh; growth medium) or oxytocin (OXT; 250 nmol) for 90 min. *p < 0.05; **p < 0.01. H Investigation time of non-social and social stimuli during social fear extinction in SFC+ mice locally infused with Veh (0.2 µl Ringer solution/hemisphere) or GDF-5 (0.05 µg/0.2 µl/hemisphere) 30 min prior to extinction. *p < 0.05; (*) p < 0.07 GDF-5 vs Veh. I In summary, in response to SFC, septal miR-132-3p was increased, which in turn downregulated Gdf-5. Septal upregulation of miR-132-3p, leading to decreased Gdf-5, resulted in high social fear, whereas inhibition of miR-132-3p, resulting in increased Gdf-5, led to low social fear. Moreover, miR-132-3p inhibition was found to prevent OXT-induced reversal of social fear. E, H data represent mean ± SEM; B n = 1–2/group; C–E, H n = 6–11/group; F, G n = 8–10/group.

To further prove whether miR-132-3p mediates the regulation of Gdf-5, Neuro-2a cells were transfected with either a miR-132-3p inhibitor (132-3p Inh) or negative (neg ctrl) or positive control (pos ctrl) inhibitor. Confirming the successful transfection, we revealed decreased miR-132-3p levels in 132-3p Inh cells compared to neg ctrl treatment (Supplementary Fig. S4I). Hmga2, which is a predicted (TargetScan Mouse 8.0) and validated [41] target of miR-132-3p, was increased after 132-3p Inh treatment compared to control inhibitors (Fig. 4F). Importantly, Gdf-5 was increased in 132-3p Inh-treated cells compared to positive control inhibitor-treated cells, suggesting that miR-132-3p indeed downregulates Gdf-5 expression. Further stimulation of 132-3p Inh-transfected Neuro-2a cells with OXT increased Gdf-5 levels in 132-3p Inh/OXT treated cells compared to neg ctrl/OXT treatment (Fig. 4G), suggesting that OXT leads to increased miR-132-3p, which regulated Gdf-5 expression. Additionally, OXT treatment decreased Gdf-5 levels in miR-132-3p neg ctrl-treated cells (neg ctrl/OXT vs neg ctrl/Veh), suggesting that miR-132-3p mediates the stimulatory effect of OXT on Gdf-5 expression.

Overexpression of miR-132-3p using a miR-132-3p mimic and corresponding negative and positive controls resulted in aberrant overexpression of miR-132-3p in Neuro-2a cells (Supplementary Fig. S4I), which might induce various counterbalancing regulatory effects. Therefore, analysis of Gdf-5 and effects of OXT treatment were not analyzed after miR-132-3p overexpression in Neuro-2a cells.

To further assess whether GDF-5 regulated by miR-132-3p is involved in the observed behavioral effect of septal miR-132-3p in social fear extinction, GDF-5 [42] or Veh was applied into the LS of SFC+ mice 30 min prior to extinction of SFC+ mice. The number of CS-US pairings during acquisition did not differ prior to treatment (Supplementary Fig. S4J). During extinction, no alterations in non-social investigation times were detected, whereas investigation times of social stimuli 2 and 3 (by trend) were reduced after local GDF-5 treatment indicating impaired extinction (Fig. 4H). The recall remained unaffected by local GDF-5 treatment (Supplementary Fig. S4K). These data further strengthen the suggestion of septal miR-132-3p mediating its effect on social fear extinction by post-transcriptional regulation of septal Gdf-5 abundance. However, viral Cre-dependent GDF-5 overexpression in septal OXTR-positive neurons did not alter acquisition, extinction, and recall in SFC+ mice compared to eGFP overexpression (Supplementary Fig. S3), suggesting that additional neuronal populations or target mRNAs might be involved in mediating the effect of miR-132-3p on social fear extinction.

Summarizing these molecular and behavioral analyses, we suggest septal miR-132-3p to influence social fear extinction via regulation of Gdf-5.