Acute and persistent neuroendocrine and behavioral alterations after social fear conditioning in adolescent male mice
Anna Bludau, Melanie Kabas, Rohit Menon, Inga D. Neumann
TL;DR
This study shows that social trauma in adolescent male mice leads to lasting changes in stress and social behavior, affecting both short-term and long-term responses.
Contribution
The study reveals acute and persistent neuroendocrine and behavioral effects of social trauma in adolescent mice.
Findings
Social fear conditioning in early adolescent male mice leads to persistent social avoidance into adulthood.
SFC in early adolescence alters corticosterone and oxytocin responses to social stressors.
Oxytocin system dysfunction is observed after social trauma, both acutely and long-term.
Abstract
Adolescence is a critical developmental period with heightened stress susceptibility. Traumatic experiences during this phase are highly predictive of future affective disorders, such as social anxiety disorder (SAD), which may manifest during early adolescence. Social avoidance, a major symptom of SAD, can be robustly generated in adult male and female mice using the social fear conditioning (SFC) paradigm. Using the SFC paradigm in adolescent mice, we analyze behavioral and neuroendocrine responses after adolescent social trauma. Here, we demonstrate that social fear elicited by SFC in early adolescent (EA) male mice (SFC + EA/29d) persists until adulthood (SFC + EA/57d). We further compared neuroendocrine responses to a heterotypic (elevated platform) or homotypic (exposure to a conspecific) stressor after SFC performed either in EA (SFC + EA/29d, SFC + EA/57d) or adulthood (SFC +…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4| Behavior | Plasma CORT | Plasma OXT | |||||||
|---|---|---|---|---|---|---|---|---|---|
| SFC | Effect of conditioning | Effect of stressor | Effect of conditioning | Effect of stressor | |||||
| SFC+/−EA/29d | SFC+: social avoidance and successful extinction | BAS | N/C | SFC−EA/29d | EPF > BAS | BAS | N/C | SFC−EA/29d | N/C |
| EPF | SFC+ < SFC− | N/C | EPF | N/C | N/C | ||||
| SOC | SFC+ > SFC− | EPF > SOC | SOC | N/C | N/C | ||||
| SFC+EA/29d | EPF > BAS | SFC+EA/29d | N/C | ||||||
| SOC > BAS | N/C | ||||||||
| EPF > SOC | N/C | ||||||||
| SFC+/−EA/57d | SFC+: social avoidance and successful extinction | BAS | N/C | SFC−EA/57d | EPF > BAS | BAS | N/C | SFC−EA/57d | N/C |
| EPF | N/C | SOC > BAS | EPF | N/C | SOC > BAS | ||||
| SOC | N/C | N/C | SOC | SFC+ < SFC− | EPF < SOC | ||||
| SFC+EA/57d | EPF > BAS | SFC+EA/57d | N/C | ||||||
| SOC > BAS | N/C | ||||||||
| N/C | N/C | ||||||||
| SFC+/‐AD |
| BAS | N/C | SFC−AD | EPF > BAS | BAS | N/C | SFC−AD | N/C |
| EPF | N/C | SOC > BAS | EPF | N/C | SOC > BAS | ||||
| SOC | N/C | EPF > SOC | SOC | SFC+ < SFC− | EPF < SOC | ||||
| SFC+AD | EPF > BAS | SFC+AD | N/C | ||||||
| N/C | N/C | ||||||||
| EPF > SOC | N/C | ||||||||
- —Deutsche Forschungsgemeinschaft10.13039/501100001659
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsStress Responses and Cortisol · Neurotransmitter Receptor Influence on Behavior · Memory and Neural Mechanisms
INTRODUCTION
1
Adolescence is a critical developmental period characterized by profound neuroanatomical plasticity, elevated stress sensitivity, and dynamic adaptations in socio‐emotional behavior. Adolescent rodents display increased social activity,1 risk‐taking,2 and impulsivity3 as opposed to other age groups. Traumatic experiences during adolescence are highly predictive of the development of various affective disorders,4 such as social anxiety disorder (SAD). Major symptoms of SAD include intense fear and avoidance of social situations. With a lifetime prevalence of 8%–15%,5 SAD is deemed a major health concern. The onset of SAD typically occurs in early adolescence (EA) and tends to persist unless treated effectively. Acute or repeated exposure to severe traumatic events, such as social trauma, or stress during adolescence, a period of heightened brain plasticity, has a severe impact on behavioral and neuroendocrine development.6 For example, exposure of rats and mice to chronic non‐social or social stressors during or around adolescence exerts long‐lasting effects on social, emotional, and cognitive behaviors. Some of these consequences have been linked to reduced hippocampal neurogenesis and neuroplasticity,7 and to alterations of hypothalamic–pituitary–adrenal (HPA) axis functions, including increased cellular activity within the paraventricular nucleus (PVN) of the hypothalamus.8, 9
Stress hormones, such as corticosterone (CORT), secreted in response to threatening stimuli and traumatic experiences, are known to promote fear generalization10 and the retention of contextual fear.11 CORT was also shown to improve the extinction of fear memory in adult rodents.12 During adolescence, however, a high variability in the stress response of the HPA axis of rodents has generally been observed, with increased,13 decreased,14 or unaltered15 plasma CORT responses dependent on the characteristics and severity of the applied chronic non‐social stress.
The neuropeptide oxytocin (OXT), which essentially promotes various social behaviors, also exerts stress‐protective effects mostly studied in adulthood.16, 17, 18 The distribution and quantity of OXT receptor‐expressing cells within distinct brain regions are dynamically shifted during development, with adolescence being a critical remodeling period.19, 20, 21 These developmental changes in the OXT system are suggested to contribute to age‐specific adaptations in socio‐emotional behavior.6, 16, 22, 23 Interestingly, intranasal application of synthetic OXT has been suggested to mitigate isolation‐induced SAD in adolescent patients.24 However, the neurobiological mechanisms underlying the etiology of SAD in adolescent subjects remain elusive.
Social fear conditioning (SFC) is an established mouse model of SAD that has been shown to robustly evoke social avoidance in adolescent22 and adult male25 and female26 mice by application of a mild foot shock when in direct contact with a conspecific. Subsequently, mice are repeatedly exposed to social stimuli during SFC extinction training, thereby shifting their behavior from social avoidance to social approach and investigation. However, it remains elusive whether social avoidance elicited by SFC during EA persists until adulthood (AD). Also, to what extent SFC exposure in EA (SFC^+^EA) or in AD (SFC^+^AD) differentially affects HPA axis and OXT responses to either a heterotypic or homotypic stressor is unknown.
Here, we studied whether SFC^+^EA is capable of eliciting robust social fear, which may even persist until AD. We further compared the HPA axis and OXT responses to an acute 5‐min exposure to either a heterotypic (elevated platform; EPF) or homotypic (unknown conspecific) stressor in SFC^+^EA and SFC^+^AD male mice.
MATERIALS AND METHODS
2
Animals and husbandry
2.1
Male CD1 mice (EA: 27–29 days, AD: 55–57 days at the start of experiment; Charles River Laboratories, Germany) were maintained under standard laboratory conditions (12/12 h light/dark cycle; lights on at 07:00 am, 21–23°C, 55%–60% humidity) with food and water available ad libitum. Initially, after arrival from the supplier, mice were group‐housed in groups of four for 3 days and single‐housed in observation cages (30 cm × 25 cm × 35 cm) for 2 days prior to acquisition of social fear until the end of the experiment. Weight‐, age‐, and sex‐matched conspecifics were used as social stimuli. All experiments were conducted between 08:00 am and 12:00 pm in accordance with the Guide for the Care and Use of Laboratory Animals by the National Institutes of Health, Bethesda, USA, approved by the government of Unterfranken and performed according to international guidelines on the ethical use of animals and ARRIVE guidelines.27
Social fear conditioning (SFC)
2.2
The SFC paradigm was performed with a computerized fear conditioning system (TSE System GmbH, Germany) as previously described.22, 25, 26, 28 Briefly, during acquisition of social fear (Figure 1A), mice were habituated to the chamber for 30 s and subsequently exposed to a non‐social stimulus (small empty wire‐mesh cage) followed by a social stimulus (age and weight‐matched conspecific placed in a wire‐mesh cage; conditioned stimulus: CS). Unconditioned mice (SFC^−^) were allowed to freely explore the social stimulus, whereas conditioned mice (SFC^+^) received a mild foot shock (0.7 mA, approx. 1 s; unconditioned stimulus: US) when in direct contact with the social stimulus.
Schematic representation of the experimental protocol. (A) During acquisition of social fear, early adolescent (EA, 28d of age) or adult (AD, 56d of age) mice were habituated to the chamber for 30 s and subsequently exposed to a non‐social stimulus (empty wire‐mesh cage) for 3 min followed by a social stimulus (conspecific in wire‐mesh cage). Social fear‐conditioned (SFC+) mice received a mild foot shock when investigating and sniffing an unknown same‐sex, same‐age conspecific, whereas unconditioned (SFC−) mice were allowed to freely investigate the social stimulus. (B) To analyze acute (29d old; 1 day after acquisition) and persistent (57d old; 28 days after acquisition) social avoidance elicited by SFC in EA, SFC+ and SFC− mice were consecutively exposed to 3 non‐social stimuli (wire‐mesh cages) and 6 different social stimuli (wire‐mesh cages containing an unknown conspecific) for 3 min each with a 3‐min inter‐stimulus interval during social fear extinction training. On the subsequent day during recall, mice were exposed to 6 unknown social stimuli. Hence, the following groups were analyzed: SFC−EA/29d, SFC+EA/29d, SFC−EA/57d, and SFC+EA/57d. (C) A separate cohort of mice was exposed to a 5‐min heterotypic (elevated platform; EPF) or homotypic (social stimulus) stressor or remained undisturbed in the home cage (basal) 1 day (SFC+/−EA/29d; adulthood SFC+/‐AD) or 28 days (SFC+/−EA/57d) after acquisition of social fear either in EA or AD. Plasma hormone levels for corticosterone (CORT), and oxytocin (OXT), OXT+ brain cells, as well as relative organ weights (thymus, adrenals, spleen) were analyzed, and splenocytes were quantified 10 or 30 min after termination of the stressor. Here, the stressor response of the following groups was analyzed: SFC−EA/29d, SFC+EA/29d, SFC−EA/57d, SFC+EA/57d, SFC−AD and SFC+AD. Created with BioRender (www.BioRender.com).
To evaluate the persistence of social avoidance after SFC in EA, mice were exposed to 3 non‐social and 6 social stimuli either on the subsequent day (EA/29d) or 28 days later (EA/57d) during social fear extinction training (Figure 1B). During extinction recall on the following day (either 30d or 58d), mice were exposed to 6 social stimuli. The following groups were analyzed: SFC^−^EA/29d, SFC^+^EA/29d, SFC^−^EA/57d, and SFC^+^EA/57d.
Social investigation behavior was manually scored by an observer blind to treatment and is shown as percentage of time spent in close interaction with and sniffing at the respective stimulus.
As the behavioral effects of SFC in adult male mice are well known,22, 25, 26, 28 we included separate groups of adult mice (SFC^−^AD, SFC^+^AD) only after SFC acquisition to assess subsequent CORT and OXT responses to stressor exposure (Figure 1C).
Stressor exposure
2.3
To compare HPA and OXT responses to either a heterotypic or homotypic stressor after SFC in EA or in AD, mice were exposed to either the EPF or a social stressor (conspecific) for 5 min or were left undisturbed (basal; Figure 1C) either 1 day (EA: 29d; AD: 57d) or 4 weeks (EA: 57d) after acquisition of social fear (Figure 1A). The following groups were analyzed: SFC^+^EA/29d, SFC^−^EA/29d, SFC^+^EA/57d, SFC^−^EA/57d, SFC^+^AD, and SFC^−^AD.
EPF (heterotypic stressor)
2.3.1
Mice were individually placed on the EPF (diameter 18 cm; elevation 75 cm; 160 lux) for 5 min without behavioral monitoring and were returned to their home cage thereafter.29
Social (homotypic stressor)
2.3.2
Mice were exposed to a novel social stimulus (within a small wire‐mesh cage) placed into their home cage for 5 min. Their behavior was videotaped. The social investigation behavior was manually scored by an observer blind to treatment and is shown as percentage of time spent in direct social contact.
Trunk blood sampling
2.4
Mice were sacrificed by decapitation under CO_2_ anesthesia either 10 min or 30 min after termination of acute stressor exposure. Trunk blood was collected in EDTA‐coated tubes (Sarstedt, Germany) on ice within 2 min after removal from their home cage, and centrifuged at 4°C (5000 rpm, 10 min). Plasma was stored at −20°C until assayed.
Quantification of OXT
- cells
2.5
To quantify OXT‐producing (OXT^+^) neurons in the hypothalamic PVN and supraoptic nuclei SON of SFC^−/+^EA/29d and SFC^−/+^EA/57d male mice, animals were deeply anaesthetized with an intraperitoneal injection of ketamine (10%; 1 mL/kg; WDT, Garbsen, Germany) and xylazine (2%; 0.5 mL/kg; Serumwerk Bernburg AG, Bernburg, Germany) 90 min after 5‐min social stressor exposure. Subsequently, animals were transcardially perfused using 0.01 M PBS (pH 7.4) followed by a 4% paraformaldehyde solution (PFA; pH 7.4; in 0.01 M PBS). Brains were extracted, post‐fixed overnight in 4% PFA, cryo‐protected in 30% sucrose in 0.01 M PBS for 2–3 days at 4°C and finally snap‐frozen in N‐methylbuthane (Sigma Aldrich, Deisenhofen, Germany) on dry ice. Thirty or 40‐μm coronal brain sections were cryo‐cut (Cryostat CM 3050S; Leica, Wetzlar, Germany) and stained for OXT using immunofluorescent methods. Briefly, non‐specific binding was blocked using 0.1% triton X‐100 and 2% normal goat serum (NGS; Vector Laboratories, Burlingame, USA) in 0.01 M PBS for 1 h. Brain sections were incubated overnight at 4°C with the primary antibody mouse anti‐OXT‐neurophysin1 (PS‐38, 1:1000 in 0.01 M PBS, 0.1% Triton X‐100 and 2% NGS; kindly provided by Dr. Harold Gainer, NIH, Bethesda, USA) and subsequently incubated for 2 h at room temperature with the secondary antibody (1:1000; goat anti‐mouse Alexa Fluor™ 488, A11001, Thermo Fisher Scientific, Waltham, USA) in 2% NGS/0.01 M PBS. Slices were mounted onto SuperFrost object slides using ROTI®Mount FluorCare mounting medium with DAPI (Carl Roth GmbH & Co KG, Karlsruhe, Germany). Tissue images of the PVN and SON were acquired using a Leica Thunder Tissue Imager microscope with 20× objective lens (z‐stack with 1 μm steps), OXT (GFP: ex: 450–490, dc: 495, em: 500–550; exposure: 120 ms, illumination: 30%) and DAPI (DAPI: ex: 325–375, dc: 400, em: 435–485; exposure: 140 ms, illumination: 30%). Thunder automated computational clearance (large volume; medium: vectashield) was performed. Exported z‐stack pictures (lossless compression; .tiff) were compiled to 2D (>z‐projection > maximum intensity for OXT); (>z‐projection > average intensity for DAPI) and the OXT channel was edited (>subtract background > rolling ball radius 75 pixels and > enhance contrast > saturated pixels 0.15%) using ImageJ. The selection of the regions of interest (PVN or SON) as well as quantification of OXT^+^ cells were performed manually using ImageJ. DAPI channel images were further edited in Cellprofiler using the method EnhanceOrSupressFeatures (suppress; 14) and DAPI^+^ cells were counted automatically within the annotated regions using the StarDist model (dsb2018_heavy_augment.pb; normalizePercentiles (1, 99); threshold (0.3); pixelSize (0.6); cellExpansion (4)) in QuPath. For each animal, the mean value of OXT^+^/DAPI^+^ cells of 2–3 images from PVN and 4–6 images from SON were evaluated.
Determination of organ weights and splenocyte number
2.6
After decapitation, the left and right adrenal glands, thymus, and spleen were removed, pruned from fat, and weighed separately. Splenocytes were isolated and counted as previously described.30
Plasma hormone analysis
2.7
Plasma CORT concentrations were analyzed using a commercially available ELISA (analytical sensitivity <1.631 nmol/L, intra‐assay and inter‐assay coefficients of variation ≤6.35%; IBL International, Germany). Plasma OXT was analyzed after an extraction procedure by radioimmunoassay (Riagnosis, Germany) as described.31
Statistical analysis
2.8
For statistical analysis (SPSS 29; IBM) data were tested for normal distribution using the Kolmogorov–Smirnov test. Mixed model ANOVA followed by Bonferroni post hoc (ns1‐ns3 and s1‐s6 separately) was used to analyze extinction and recall of social fear: factor conditioning × stimulus. Geisser–Greenhouse correction was applied when sphericity was violated (tested by Mauchly test). Parametric two‐way analysis of variance (ANOVA; factor conditioning × stressor) followed by Bonferroni post hoc was performed to analyze plasma hormone levels within the three age groups (EA/29d, EA/57d, and AD). Parametric student's t‐test was performed to analyze social investigation during social exposure and as indicated in Table S1. Pearson's correlation was used to analyze relations between plasma hormone levels and social investigation. Statistical significance was accepted at p < .05. Statistical outliers were calculated by “mean ± 2 × standard deviation” and have been removed from analysis. Detailed reports for significant statistical analysis and group sizes are available in Tables S1 and S2. Graphs were plotted using Prism 10 (GraphPad).
RESULTS
3
A summary of all behavioral and endocrine data can be found in Table 1.
Social fear conditioning during early adolescence results in acute and persistent social avoidance
3.1
To analyze whether SFC during EA, i.e., at the age of 28d, results not only in acute, but also in a persistent expression of social fear, SFC^+^EA and SFC^−^EA mice were subjected to extinction training either on the next day (SFC^+/−^EA/29d) or 4 weeks later (SFC^+/−^EA/57d) (Figure 1A,B). SFC^+^EA/29d and SFC^+^EA/57d mice received a similar number of CS‐US pairings during acquisition of social fear (SFC^+^EA/29d: 1.85 ± 0.19; SFC^+^EA/57d: 1.87 ± 0.13). During social fear extinction training, SFC^+^EA/29d and SFC^+^EA/57d mice showed lower investigation times especially of the first social stimuli (both SFC^+^EA/29d and SFC^+^EA/57d: s1–s3 p < .04 vs. SFC^−^EA/29d and SFC^−^EA/57d, respectively; Figure 2B,E and Table S1 for detailed statistics), representing social fear. Reflecting the successful extinction of social fear, both SFC^+^EA/29d and SFC^+^EA/57d mice showed a gradual increase in social investigation time of the six presented social stimuli. Independent of the period between acquisition and extinction, the success of social fear extinction was also reflected by overall high social investigation times of SFC^+^EA/29d and SFC^+^EA/57d mice on the subsequent day during recall (Figure 2C,F). The investigation times of the three non‐social stimuli presented prior to extinction training did not differ between SFC^+^ and SFC^−^ groups (Figure 2B,E), which illustrates the absence of general fear after SFC. In summary, SFC in adolescence results in robust social fear, which persists until adulthood.
*Social fear conditioning (SFC) in early adolescence (EA) leads to acute and persistent social avoidance in male mice. (A) Schematic representation of acute effects of SFC in EA (age 28d on day of acquisition). EA mice were conditioned (SFC+) or remained unconditioned (SFC−) during social fear acquisition and underwent extinction (age 29d) and recall (age 30d) on the subsequent days (groups SFC+/−EA/29d). Investigation time of the 3 presented non‐social and 6 presented social stimuli during (B) social fear extinction training, and (C) recall in SFC+/−EA/29d mice. (D) Schematic representation of persistent effects of SFC in EA (SFC+/−EA/57d; age 28d on day of acquisition). Mice underwent extinction (age 57d) and recall (age 58d) in adulthood. Investigation time during (E) social fear extinction training and (F) recall in SFC+/−EA/57d mice. p < .05 SFC+ vs. SFC−. Data represents mean ± SEM; n = 7‐13/group.
Social trauma prevents CORT hypo‐response to social stress exposure in early adolescence
3.2
To test whether acquisition of social fear, either during early adolescence or in adulthood, alters HPA axis responses to a heterotypic (EPF) or homotypic (social) stressor, plasma CORT levels were analyzed 10 min (and 30 min, see Figure S2) after 5‐min stressor exposure, either 1 day (SFC^+^EA/29d, SFC^−^EA/29d) or 28 days (SFC^+^EA/57d, SFC^−^EA/57d) after acquisition during early adolescence, or 1 day after acquisition in adulthood (SFC^+^AD, SFC^−^AD) (Figure 3A). In confirmation of the previous experiments (Figure 2),25 all SFC^+^ mice (i.e., SFC^+^EA/29d, SFC^+^EA/57d, and SFC^+^AD) spent a lower percentage of time investigating the social stimulus compared to their respective SFC^−^ controls (all age groups p < .003; Figure 3B), reflecting social fear during the 5‐min social exposure time.
*Social investigation during 5‐min exposure to a social stimulus, and plasma corticosterone (CORT) concentrations after 5‐min exposure to a non‐social or social stimulus in adolescent (EA) and adult (AD) social fear‐conditioned (SFC+) and unconditioned (SFC−) mice. (A) Male mice underwent social fear acquisition in EA (age 28d) or AD (age 56). On the subsequent day (EA/29d and AD) or 4 weeks later (EA/57d) they were either exposed to a 5‐min heterotypic (elevated platform; EPF) or homotypic (a single conspecific; SOC) stressor or remained undisturbed (BAS). Plasma CORT levels were obtained 10 min post stressor exposure. (B) Percent of social investigation time of SFC+/−EA/29d, SFC+/−EA/57d, and SFC+/‐AD mice during 5‐min exposure to a single SOC stressor. (C) Plasma CORT under BAS conditions and 10 min after exposure to the EPF or SOC stressor in SFC+/−EA/29d, SFC+/−EA/57d, and SFC+/‐AD mice. (D) Correlation of investigation time of the social stimulus and plasma CORT levels in SFC+/−EA/29d mice. *p < .05, *p < .01 SFC+ vs. SFC−, # p < .05, ## p < .01 stress vs. respective basal; ++ p < .01 SOC vs. respective EPF. Data represents mean + SEM; n = 6–8/group.
Independent of SFC, increased plasma CORT concentrations were observed after exposure to the EPF compared to respective basal values in all age groups (for all comparisons p < .005, see Table S1 for detailed statistics; Figure 3C). However, an increased HPA axis response to a social stimulus was only found in adult mice, which acquired social fear during adolescence, independent of SFC, in SFC^+^EA/57d (p < .001) and SFC^−^EA/57d (p = .019), and upon adult SFC in SFC^−^AD (p = .001) when compared to basal values. In contrast, in EA, a rise in plasma CORT in response to the social stimulus was only found in SFC^+^EA/29d (p = .023), but not in SFC^−^EA/29d mice, indicating a physiological HPA axis hypo‐response to social stressors in EA, which becomes dysregulated after acquisition of social fear. In support, comparison of stress‐induced plasma CORT concentrations between SFC^+^ and SFC^−^ mice revealed increased levels in response to the social stimulus in SFC^+^EA/29d mice (p = .034 vs. SFC^−^EA/29d), whereas exposure to the EPF resulted in lower levels in SFC^+^EA/29d mice (p = .001 vs. SFC^−^EA/29d), indicating a stressor‐specific shift in HPA responsiveness after SFC^+^EA.
An SFC^+^‐induced shift in HPA axis responses was not found in adult mice (EA/57d or AD). Importantly, only plasma CORT concentrations of EA/29d, but not of EA/57d or AD mice, negatively correlated with social investigation times (R ^2^ = 0.482; p = .012; Figure 3D), confirming higher plasma CORT and social fear in SFC^+^EA/29d mice compared to SFC^−^EA/29d mice. Thus, the presence of a conspecific seems to be highly stressful in socially traumatized EA mice, which normalizes by adulthood. Independent of SFC, comparison of HPA axis responses to EPF versus social exposure revealed generally lower CORT responses to the social stimulus in EA mice (SFC^+^EA/29d and SFC^−^EA/29d both p < .001 vs. EPF) and in AD mice (SFC^+^AD p = .006, SFC^−^AD p = .001). However, in SFC^+/−^EA/57d mice we did not reveal such lower plasma CORT response to a social stressor compared to the EPF.
Similar SFC^+^‐induced alterations in social investigation and in the patterns of stress‐induced CORT concentrations were found in the 30 min post‐stressor EA/29d, EA/57d, and AD groups (Figure S2B,D).
In summary, social trauma in EA induced a CORT hyper‐response to a social, but not heterotypic, stressor seen in early adolescence, which becomes normalized until adulthood.
Social trauma in early adolescence impairs OXT response to social stress in adult mice
3.3
We further analyzed plasma OXT concentrations 10 min (and 30 min, see Figure S2) after exposure to either a heterotypic or homotypic stressor in SFC^+^ and SFC^−^ mice of the EA/29d, EA/57d, and AD groups (Figure 3A).
Independent of conditioning, exposure to the EPF did not alter plasma OXT levels compared to basal levels in all groups. However, plasma OXT levels rose in response to the social challenge, but only in unconditioned adult mice (SFC^−^EA/57d and SFC^−^AD p < .001 vs. respective basal; Figure 4A). In contrast, in adolescent mice, the OXT system seems still unresponsive to a social challenge, with unchanged OXT concentrations seen in SFC^−^EA/29d mice. Importantly, the increased OXT response to a social challenge found in adult SFC^−^ controls was completely prevented by social fear acquisition, that is, in SFC^+^EA/57d and SFC^+^AD mice, with lower OXT levels found in comparison to respective SFC^−^EA/57d and SFC^−^AD groups (both p < .001). Further, a positive correlation of plasma OXT levels and social investigation time was only found in EA/57d (R ^2^ = 0.783; p < .001; Figure 4B), but not in EA/29d or AD mice.
*Plasma oxytocin (OXT) concentrations of adolescent (EA) and adult (AD) social fear‐conditioned (SFC+) and unconditioned (SFC−) mice after 5‐min stressor exposure. Male mice underwent SFC in EA or AD and were either exposed to a 5‐min heterotypic (elevated platform; EPF) or homotypic (conspecific; SOC) stressor or remained undisturbed (BAS) on the subsequent day (EA/29d and AD) or 4 weeks later (EA/57d). Plasma OXT levels were obtained 10 min post stressor exposure. (A) Plasma OXT under BAS conditions or 10 min after exposure to the EPF or SOC stressor in SFC+/−EA/29d, SFC+/−EA/57d, and SFC+/‐AD mice. **p < .01 SFC+ vs. SFC−; ## p < .01 stress vs. respective basal; ++ p < .01 SOC vs. respective EPF. (B) Correlation of plasma OXT levels and social investigation time of SFC+/−EA/57d mice. (C) Percentage of OXT+/DAPI+ cells within the paraventricular (PVN) and supraoptic (SON) nuclei of the hypothalamus of SFC+/−EA/29d and SFC+/−EA/57d mice. (D) Representative immunofluorescent image of OXT+/DAPI+ cells within the PVN and SON of SFC+/−EA/29d and SFC+/−EA/57d mice. Scale bar: 100 μm. *p < .01 EA57d vs. EA29d. Data represents mean + SEM; n = 5–9/group.
In support of the specific OXT hypo‐response to the social stimulus after SFC^+^, higher OXT levels were only found in SFC^−^EA/57d and SFC^−^AD (both p < .001 vs. EPF), but not in SFC^+^EA/57d and SFC^+^AD mice after social versus non‐social exposure (EPF). Similar SFC‐induced alterations in social investigation and in the patterns of stress‐induced OXT concentrations with a positive correlation of plasma OXT and social investigation time only in EA/57d mice were found in the 30 min post‐stressor EA/29d, EA/57d, and AD groups (Figure S2B,D).
Further analysis of OXT^+^ cells revealed a lower percentage of OXT^+^/DAPI^+^ cells within the PVN, but not SON, in EA/57d mice compared to EA/29d animals irrespective of the conditioning status of the animal (p = .008), illustrating an age‐dependent difference in the quantity of OXT^+^ cells (Figure 4C,D).
Taken together, social trauma either in early adolescence or in adulthood prevented peripheral OXT release in response to a social challenge in adulthood. Moreover, we revealed a blunted plasma OXT response to social exposure in early adolescent mice independent of SFC.
No effects of SFC and acute stressor exposure on organ weights
3.4
We further assessed the relative adrenal and spleen weight of SFC^+^ and SFC^−^ mice of the EA/29d, EA/57d, and AD groups 10 min (Figure S1A,B) and relative thymus weight and splenocyte quantity 30 min (Figure S2F–H) after EPF or social stressor exposure and of their respective unstressed controls. Neither SFC nor acute stressor exposure altered these parameters in all groups of mice. However, independent of SFC, adult mice (EA/57d, AD) showed lower relative thymus und spleen weights compared to EA/29d mice, confirming known physiological adaptations of lymphatic organs in AD (Figure S2I).
DISCUSSION
4
In this study, we show that SFC successfully induces social avoidance in male EA mice, which can be long‐lasting, i.e., persists for at least 4 weeks, and can be extinguished by repeated social exposure during extinction training performed either on the next day during EA or in AD. Furthermore, we describe developmental differences in the responsiveness of the HPA axis and the OXT system between unconditioned (SFC^−^) EA and AD mice. Specifically, we found a CORT hyper‐response to a non‐social stressor (EPF), but a CORT hypo‐response to a social challenge in EA, in comparison to AD mice. Exposure to social trauma in EA (SFC^+^EA) resulted in a CORT hyper‐response specifically to a social stimulus 1 day later, that is, in SFC^+^EA/29d mice, an effect which disappeared in SFC^+^AD mice. Regarding the responsiveness of the OXT system, we found elevated plasma OXT concentrations in SFC^−^AD, but not SFC^−^EA mice in response to a social challenge. This robust rise in plasma OXT of adult SFC^−^ mice was found to be completely abolished by social fear acquisition performed either in EA or AD, that is, in SFC^+^EA/57d and in SFC^+^AD mice. Thus, our data provide a comprehensive view of the neuroendocrine (CORT and OXT) responses to social and non‐social stressors across different stages of development, which are severely affected by social trauma.
SFC induces social avoidance in EA and AD mice
4.1
In our study, exposure to the SFC paradigm during early adolescence induced robust acute and persistent social fear in mice. Social fear extinction training performed either 1 day or 28 days after acquisition successfully reversed SFC‐induced social fear (Figure 2). These results extend our previous study performed in adolescent mice of two strains differing in stress susceptibility.22
Studies assessing neurobiological mechanisms underlying associative learning in juvenile and adolescent rodents have mostly focused on non‐social forms of conditioned fear and described dramatic developmental changes in fear learning. Specifically, in juvenile rodents, contextual fear expression and extinction recruit only the amygdala, making juveniles unable to develop fear associations, whereas engagement of the hippocampus in addition to the amygdala in adults enables them to generate robust fear memories.32, 33 During adolescence, the hippocampus and medial prefrontal cortex (mPFC) are additionally recruited, leading to an age‐specific expression and extinction of fear.32 Similarly, human studies also revealed blunted fear extinction learning in adolescent compared to adult individuals,33 which has been suggested to originate in a shift in the amygdala‐PFC connectivity.34 However, EA appears to be a unique developmental period, wherein expression and extinction of contextual fear are temporarily impaired before recovering in AD.32 In contrast, we find intact expression and extinction of social fear in EA/29d male mice (Figure 2B,C), with fear expression persisting until adulthood as shown in EA/57d mice (Figure 2D,E).
Social trauma experienced either in adolescence or adulthood has been associated with SAD and exerts long‐lasting effects.35 In this context, adolescent humans have been shown to exhibit a heightened sensitivity toward social trauma compared to adults.36 Here we show that SFC during EA resulted in social avoidance behavior that persists until adulthood, confirming the long‐lasting effects of social trauma during puberty. In rodents, social isolation in early life or adulthood is known to increase social avoidance in a novel environment, increase anxiety‐related behavior, and alter stress responses.37, 38 Interestingly, the 4‐week social isolation in SFC^+/−^EA/57d mice did neither intensify social avoidance behavior nor facilitate extinction of social fear compared to EA/29d and adult mice (Figure 2).22, 25, 28, 39 This suggests the interference of unknown mechanisms that prevent avoidance, exacerbating increased social motivation effects of social isolation on long‐lasting SFC‐induced social fear.
Different neuronal circuits and molecular mechanisms are involved in the processing of social vs. non‐social fear. In this context, the amygdala is the principal brain region shaping fear responses of any kind, and across all age groups.40 In contrast, age‐dependent morphological and neurochemical adaptations have been observed within social fear‐related brain regions, such as the lateral septum (LS) and ventromedial hypothalamus.26, 41 This indicates a differential modulatory capacity of these regions to acquire, express, and extinguish conditioned social fear compared to contextual or cued fear across age. Thus, the observed social avoidance behavior after SFC in EA/29d and EA/57d male mice might be based on the age‐specific development of underlying neurocircuits within the limbic system, known to be significantly remodeled in adolescence.42
SFC induces differential HPA axis stress responses in EA and AD mice
4.2
In our study, the inclusion of SFC^−^ control mice allowed the analysis of HPA axis responses to exposure to either a non‐social (EPF) or social stressor across development, that is, in EA and AD mice. Both EA/29d and AD mice showed lower plasma CORT levels after social stressor compared to EPF exposure, which might be explained by the comparatively low severity of the social stressor. In our experimental design, the stimulus mouse was restrained within a small wire‐mesh cage and presented within the home cage of the experimental mouse. However, after 4 weeks of social isolation, that is, in EA/57d mice, the differences between CORT responses to the social compared to EPF exposure disappeared. Post‐weaning social isolation of male mice was shown to persistently increase not only anxiety‐related behavior, but also neuroendocrine functions, especially the corticotropin‐releasing factor activity, which is a major regulator of the HPA axis.43 In contrast, exposure to the non‐social stressor (EPF) led to increased plasma CORT levels in all age and SFC groups, which is in line with numerous previous studies.44, 45
The adolescent rodent is suggested to be more susceptible to stressors aresulting ito high glucocorticoid levels compared to the adult brain, which ultimately leads to short‐ and long‐term behavioral alterations, such as anxiety‐related behavior, exploratory behavior, and avoidance learning.46 Moreover, in adolescent humans, acute stress results in a CORT hyper‐response and poor recovery compared to basal levels.47 As most studies are based on chronic stress in adolescent rodents or humans, knowledge on the HPA axis regulation after acute adolescent stress remains sparse.
In EA, the low plasma CORT response to a social challenge found in SFC^−^EA/29d mice, turned hyper‐responsive after social trauma exposure, i.e, in SFC^+^ EA/29d mice (Figure 3C). The high CORT response in SFC^+^EA mice was found despite the fact that SFC^+^EA mice did not or only very little investigate the social stimulus. Thus, a negative correlation between social investigation time and plasma CORT levels in EA/29d mice was revealed indicating that the pure presence of the social stimulus in their home cage is highly stressful 1 day after social fear conditioning (Figure 3D).
However, the high responsiveness of the HPA axis to the social stimulus was absent in SFC^+^EA/57d and SFC^+^AD mice. The observed SFC‐induced dysregulation of the HPA axis, reflected by reduced plasma CORT, is found in adolescents in response to social interaction48, 49 and is suggested to increase the vulnerability for SAD by altering a limbic 5‐HT receptor in humans.50
Effects of SFC on plasma OXT responses to social vs. non‐social stressors in EA and AD
4.3
To the best of our knowledge, this is the first study showing an age‐dependent release pattern of OXT after a singular social encounter. Thus, exposure to a social stimulus only increased plasma OXT in all adult control mice, that is, in SFC^−^EA/57d and SFC^−^AD mice, but not in adolescent mice (SFC^−^EA/29d; Figure 4A). The non‐responsiveness of the OXT system to a social stimulus in EA/29d further supports a dynamic developmental profile of this system, as age‐dependent distribution patterns of the OXT receptor in the brain have been described, which may at least partly underly the age‐specific socio‐emotional behavior.19, 22 Interestingly, irrespective of the conditioning status, the percentage of OXT^+^ cells within the PVN, but not SON, decreased subtly with age (Figure 4C,D). Thus, our results indicate that the increased plasma OXT release in SFC^−^EA/57d in contrast to their SFC^+^ counterparts, and the absence of this conditioning effect in EA/29d mice is not reflected by differences in OXT cell number.
OXT is capable of non‐specifically binding to arginine vasopressin (AVP) receptors, which are also dynamically expressed through development.19 Both neuropeptides, OXT and AVP, are expressed in the developing brain. However, in rodents, AVP emerges prenatally, whereas OXT is only detectable postnatally.19 The high level of social contact observed in SFC^−^EA/29d mice might be at least partly mediated by AVP or other factors, such as testosterone. As one of the major sex hormones, testosterone significantly rises during male puberty, thus affecting socio‐emotional behaviors in rodents and humans.51 The OXT and testosterone systems exert partly opposite effects on multiple social behaviors in adulthood.52, 53 However, we did not analyze plasma testosterone levels in this study.
Although we have only monitored peripheral OXT release, brain OXT plays an important role in regulating social behavior, especially the reversal of social fear and the promotion of social interactions. In general, exposure to socio‐emotional stress activates both intracerebral and peripheral OXT release, although in a stressor and region‐dependent manner.16, 54, 55 For example, in male rats, exposure to social defeat results in OXT release within the hypothalamic PVN, accompanied by OXT secretion into blood.55 Here, we show for the first time that also a mild social stimulus, that is, 5‐min exposure to a conspecific, results in elevated plasma OXT concentrations in adult mice. Similarly, social investigation during repeated exposure to social stimuli stimulates the release of OXT within the mouse LS as revealed by microdialysis,39 suggesting a simultaneous peripheral and central OXT release in response to social interactions. OXT signaling in the LS has been repeatedly shown to be essential for social approach and the reversal of social fear.26, 28, 39 However, OXT is also released within the LS in adult female rats when encountering an intruder and during the display of aggression.56 Thus, it remains to be shown, to which extend the exposure to a social stimulus might be perceived as a threat or whether social motivation leads to the observed rise in plasma OXT in adult SFC^−^ controls in our experiment.
Interestingly, social trauma either in EA or AD completely prevented the rise in plasma OXT in response to social exposure in adult mice. This is in line with a previous study revealing an abolished OXT release within the LS in adult SFC^+^ mice in response to repeated social contact during social fear extinction training.39 This indicates that SFC is effective in altering the functionality of the OXT system, thus affecting both peripheral secretion as well as release within distinct brain regions. However, plasma OXT levels positively correlated with the investigation time in EA/57d mice during exposure to a social stimulus (Figure 4B), which may indicate that the lack of social interaction and social contact is causal for the lack of OXT secretion in adult SFC^+^ mice. Interestingly, in humans, the experience of social trauma during childhood has been correlated with lower basal plasma OXT levels in adulthood,57 which we did not confirm in our study.
In conclusion, the SFC paradigm is a suitable animal model to generate acute and persistent social avoidance in EA mice, hence increasing its translational value with regard to the development of novel treatment options in adolescent patients. Moreover, we revealed age‐dependent responses of the HPA axis activity and the OXT system to a social, but not non‐social stressor. Experience of a social trauma in EA significantly interfered with the neuroendocrine development suggesting plasma CORT and/or OXT levels as biomarkers of trauma in puberty.
AUTHOR CONTRIBUTIONS
Conceptualization: Anna Bludau and Inga D. Neumann. Methodology: Anna Bludau, Melanie Kabas, Rohit Menon, and Inga D. Neumann. Formal analysis: Anna Bludau and Melanie Kabas. Software: Anna Bludau. Investigation: Anna Bludau and Melanie Kabas. Writing—original draft: Anna Bludau, Rohit Menon, and Inga D. Neumann. Writing—review and editing: Anna Bludau, Melanie Kabas, Rohit Menon, and Inga D. Neumann. Visualization: Anna Bludau. Funding acquisition: Inga D. Neumann. Supervision: Inga D. Neumann.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Supporting information
Figure S1. Relative spleen and adrenal weight of adolescent (EA) and adult (AD) social fear‐conditioned (SFC^+^) and unconditioned (SFC^−^) mice after 5‐min stressor exposure. Male mice underwent SFC in EA or AD and were either exposed to a 5‐min heterotypic (elevated platform; EPF) or homotypic (conspecific; SOC) stressor or remained undisturbed (BAS) on the subsequent day (EA/29d and AD) or 4 weeks later (EA/57d). (A) Relative spleen weight (in mg/g), and (B) relative adrenal weight (in mg/g) under BAS conditions or 10 min after exposure to a EPF or SOC stressor in SFC^+/−^EA/29d, SFC^+/−^EA/57d, and SFC^+/‐^AD mice. Data represents mean + SEM; n = 4–8/group.
Figure S2. Plasma hormone levels, social investigation time, as well as relative organ weights of adolescent (EA) and adult (AD) social fear‐conditioned (SFC^+^) and unconditioned (SFC^−^) mice after 5‐min stressor exposure. (A) Male mice underwent SFC in EA or AD and were either exposed to a 5‐min heterotypic (elevated platform; EPF) or homotypic (conspecific; SOC) stressor or remained undisturbed (BAS) on the subsequent day (EA/29d and AD) or 4 weeks later (EA/57d). Hormone plasma levels were obtained 30 min post stressor exposure. The given days represent the animal's age. (B) Plasma corticosterone (CORT, in ng/ml), (C) plasma oxytocin (OXT, in pg/ml), (D) social investigation time, (E) correlation of plasma OXT and social investigation time (in %) during SOC exposure, (F) relative thymus weight (in mg/g), (G) relative adrenal weights (in mg/g), (H) relative spleen weights (in mg/g), and (I) relative splenocyte number (in cells/mg) under BAS conditions or 30 min after exposure to the EPF or SOC either in SFC^+/−^EA/29d, SFC^+/−^EA/57d, and SFC^+/‐^AD mice. *p < .05, **p < .01 SFC^+^ vs. SFC^−^, ^#^ p < .05, ^##^ p < .01 stress vs. respective basal; ^+^ p < .05 vs. respective EPF. Data represents mean + SEM; n = 6–8/group.
Table S1. Statistical details corresponding to all figures.
Table S2. Group sizes corresponding to all figures.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Forbes EE , Dahl RE . Pubertal development and behavior: hormonal activation of social and motivational tendencies. Brain Cogn. 2010;72:66‐72. doi:10.1016/j.bandc.2009.10.007 19942334 PMC 3955709 · doi ↗ · pubmed ↗
- 2Steinberg L . A social neuroscience perspective on adolescent risk‐taking. Dev Rev. 2008;28:78‐106. doi:10.1016/j.dr.2007.08.002 18509515 PMC 2396566 · doi ↗ · pubmed ↗
- 3Adriani W , Laviola G . Elevated levels of impulsivity and reduced place conditioning with d‐amphetamine: two behavioral features of adolescence in mice. Behav Neurosci. 2003;117:695‐703. doi:10.1037/0735-7044.117.4.695 12931955 · doi ↗ · pubmed ↗
- 4Kessler RC , Mc Laughlin KA , Green JG , et al. Childhood adversities and adult psychopathology in the WHO world mental health surveys. Br J Psychiatry. 2010;197:378‐385. doi:10.1192/bjp.bp.110.080499 21037215 PMC 2966503 · doi ↗ · pubmed ↗
- 5Koyuncu A , İnce E , Ertekin E , Tükel R . Comorbidity in social anxiety disorder: diagnostic and therapeutic challenges. Drugs Context. 2019;8:212573. doi:10.7573/dic.212573 30988687 PMC 6448478 · doi ↗ · pubmed ↗
- 6Haller J , Harold G , Sandi C , Neumann ID . Effects of adverse early‐life events on aggression and anti‐social Behaviours in animals and humans. J Neuroendocrinol. 2014;26:724‐738. doi:10.1111/jne.12182 25059307 · doi ↗ · pubmed ↗
- 7Borsini A , Giacobbe J , Mandal G , Boldrini M . Acute and long‐term effects of adolescence stress exposure on rodent adult hippocampal neurogenesis, cognition, and behaviour. Mol Psychiatry. 2023;28:4124‐4137. doi:10.1038/s 41380-023-02229-2 37612364 PMC 10827658 · doi ↗ · pubmed ↗
- 8Lui P , Padow VA , Franco D , et al. Divergent stress‐induced neuroendocrine and behavioral responses prior to puberty. Physiol Behav. 2012;107:104‐111. doi:10.1016/j.physbeh.2012.06.011 22728428 PMC 3409319 · doi ↗ · pubmed ↗