Prefrontal oxytocin receptor positive cells mediate stress-induced anxiety in tuberous sclerosis complex

Abstract

Stress is a major risk factor for maladaptive processes such as pathological anxiety, which is highly prevalent in Tuberous Sclerosis Complex (TSC), a neurodevelopmental disorder caused by Tsc1/Tsc2 mutations. To investigate underlying mechanisms, we modeled Tsc2 haploinsufficiency in oxytocin receptor-expressing cells (OTRCs). Conditional deletion of Tsc2 in OTRCs induced hyperactivation of mTORC1 and PERK-mediated integrated stress response (ISR), impairing protein synthesis and suppressing medial prefrontal cortex (mPFC) circuits. Under chronic social isolation stress, male mutants exhibited anxiety-like behaviors, reduced motivation, and prefrontal hypoactivity, whereas females showed resilience to motivational deficits but diminished social preference. Pharmacological PERK inhibition, and OTRC-specific Rheb manipulation in mPFC, restored normative behavior and mPFC excitability, implicating the TSC–Rheb–PERK axis as a regulator of sex-specific stress vulnerability. These findings highlight integrated stress response modulation in OTRCs as a potential therapeutic strategy for anxiety linked to prefrontal dysfunction.

Mechanistic studies in a mouse model of tuberous sclerosis complex reveal that heightened integrated stress response in oxytocin receptor positive neurons disrupts prefrontal cortex function and produces sex-specific anxiety behaviors under social isolation.

Introduction

Social affiliation constitutes an essential biological drive in social species, including both rodents and humans, encompassing behaviors such as affiliative grooming, social play, vocal communication, and reciprocal social interactions. Disruption of these interactions acts as a potent psychosocial stressor, eliciting diverse behavioral outcomes contingent upon the duration and context of deprivation, consistent with the Yerkes-Dodson law’s inverted U-shaped relationship between arousal and performance¹. Acute social isolation often induces a compensatory increase in social motivation, enhancing affiliative behaviors to restore social homeostasis^2–5. In contrast, chronic social isolation imposes a sustained allostatic load, contributing to maladaptive affective states and heightened social withdrawal, particularly in individuals with underlying genetic susceptibility^6,7. In Tuberous Sclerosis Complex (TSC), a neurodevelopmental disorder with a high prevalence of pathological anxiety, this may reflect dysregulated affective reactivity to sustained environmental stress. While numerous studies using TSC rodent models have elucidated impairments in cognitive domains—including learning and memory, cognitive flexibility, and social recognition^8–12 —there remains a significant gap in our understanding of the neurobiological substrates underlying TSC-associated anxiety. Specifically, the multi-level neural correlates spanning molecular signaling pathways, neuromodulatory dynamics, and circuit-level dysfunctions remain poorly characterized. Notably, TSC-associated neuropsychiatric disorders (TAND)^13–15, of which anxiety disorders represent a major component, account for the greatest clinical burden in this complex multisystem condition^14,16,17.

The hypothalamic neuropeptide oxytocin (OT) plays a critical role in modulating stress responsivity^18–21 and enhancing the salience of social stimuli²². OT exerts its central effects via its membrane-bound G-protein coupled receptor (OTR), which is expressed in a subset of brain cells—oxytocin receptor-expressing cells (OTRCs)—in a manner that is both experience- and sex-dependent. At the systems level, OT modulates neural circuits by enhancing the salience of neuronal activity, primarily through GABAergic mechanisms that increase the signal-to-noise ratio; however, recent studies have also implicated OT in modulating glutamatergic transmission and astrocytic function^22–24. During stress exposure, OT is released from the paraventricular nucleus (PVN) and acts to attenuate activation of the hypothalamic-pituitary-adrenal (HPA) axis^25–27. OTRCs in the midbrain contribute to social motivation following acute social isolation⁵, whereas chronic isolation leads to downregulation of Oxtr gene expression in the central amygdala, contributing to anxiety-like phenotypes²⁸. The medial prefrontal cortex (mPFC), which integrates inputs from sensory, motor, and limbic circuits, plays a key role in top-down regulation of emotion during both acute and chronic stress. Although direct evidence linking prefrontal OTRCs to the effects of social isolation is lacking, optogenetic inhibition of mPFC OTRCs in adult mice induces sex-dependent anxiety-like behaviors, mainly via interactions with corticotropin-releasing factor (CRF)-mediated stress signaling pathways^29,30. Developmental timing also influences susceptibility to stress: chronic social isolation during adolescence reduces the excitability of deep-layer mPFC neurons that project to subcortical regions such as the paraventricular thalamus³¹ and nucleus accumbens³², impairing social recognition. On the other hand, similar isolation in male adult, but not juvenile mice, induces anxiety- and depression-like behaviors³³, suggesting an age-dependent and domain-specific vulnerability to prolonged social isolation³⁴. Furthermore, the mPFC itself is a target of maladaptive stress responses, undergoing structural and functional remodeling—including dendritic retraction and hypomyelination—after chronic emotional stress exposure^35–38. The OT system also exhibits plasticity under chronic stress conditions by dynamically regulating Oxtr expression, modifying the intrinsic excitability of OT neurons, and adapting afferent signaling to OTRCs, all of which serve to recalibrate physiological set points in response to environmental demands^22,39,40. Importantly, pharmacological or genetic enhancement of OT-OTR signaling has demonstrated efficacy in ameliorating socioemotional dysfunction across various neuropathological conditions^41,42.

To investigate the molecular and cellular mechanisms underlying TSC-associated pathological anxiety, we selectively inactivated Tsc2 in the OT system. Tsc2 encodes a GTPase-activating protein (GAP) for the small G-protein Rheb, which activates both the mammalian target of rapamycin complex 1 (mTORC1)⁴³ and PKR-like endoplasmic reticulum kinase (PERK)⁴⁴. Activation of PERK initiates the integrated stress response (ISR), a protective signaling cascade that reduces global protein synthesis while selectively enhancing the translation of stress-adaptive mRNAs to support cell recovery and survival^45,46. However, chronic ISR activation impairs neuronal function and disrupts plasticity mechanisms critical for resilience to psychological stress^47,48. We hypothesized that cell type-specific disruption of Tsc2 in OTRCs would increase vulnerability to social isolation and trigger maladaptive anxiety-like behaviors by hyper-activating pathways normally restrained by the TSC complex. To test this, we performed heterozygous deletion of Tsc2 in OTRCs and examined the behavioral impact of prolonged social isolation in both male and female mice. Through pharmacological manipulation, we identified PERK-ISR signaling as the key molecular driver of heightened emotional susceptibility in males. Further, we identified OTRCs in mPFC as the cellular substrates that mediate the behavioral and electrophysiological signatures of ISR-induced network suppression and stress-induced anxiety.

Results

Cell type-specific deletion of Tsc2 in OTRCs alters energy balance during stress

We initially used OTR-Cre × GFP-L10 double transgenic mice to map the distribution of OTR-expressing cells (OTRCs)⁴⁹. We quantified OTR+ neurons in the medial prefrontal cortex (mPFC) in male and female mice, with males expressing significantly higher OTR+ neurons than females in the prelimbic subregion (Fig. 1a). To selectively manipulate OTRCs and explore their role in TSC, we crossed OTR-Cre mice with Tsc2 floxed mice⁵⁰, generating a line with cell type-specific heterozygous deletion of the Tsc2 gene in OTRCs (Fig. 1b). To assess TSC2 protein levels in mPFC OTR+ neurons, we labeled these cells using a virally delivered double-floxed mCherry construct in adult wild-type (WT) and conditional heterozygous Tsc2 (cHET) mice (Fig. 1c). The heterozygous deletion of Tsc2 led to a notable 32% reduction in TSC2 protein in OTR+ neurons (Fig. 1d).

Fig. 1. Cell type-specific heterozygous deletion of Tsc2 in OTRCs.

a Cre recombinase activity in the medial prefrontal cortex (mPFC) of the OTR-Cre driver line (#ON66), validated by crossing with a floxed GFP-tagged ribosomal protein L10 reporter line. Quantification of OTR+ cells (right panel) in the prelimbic region (PL) of mPFC indicate higher expression in males compared to females. b Breeding strategy for generating mice with conditional heterozygous deletion of Tsc2 in oxytocin receptor-expressing cells (OTRCs): OTR-Cre^⁺/⁻; Tsc2 ^f/+ (cHET) and Cre-positive controls with intact Tsc2 alleles: OTR-Cre⁺/⁻; Tsc2^+/+ (WT). c AAV-mediated expression of a Cre-dependent (floxed) mCherry reporter selectively targets OTRCs in the mPFC. d Representative immunofluorescence images of mPFC from WT and cHET mice show reduced TSC2 expression in OTRCs in cHETs (left panel). Scale bar = 50 μm. Quantification of TSC2 immunoreactivity (right panel) in prefrontal OTRCs indicates effective knockdown of TSC2 in cHET mice. e Schematic of behavioral timeline used to assess the effects of prolonged social isolation stress (SIS) during adult life. f Violin plots show that WT males display normative weight gain over the 2-month SIS period, an effect absent in cHET males (left panel). Food intake remains unchanged between genotypes (right panel), suggesting that impaired weight gain in cHETs is not attributable to caloric intake. g Violin plots show SIS-induced suppression of weight gain in WT females, while cHET females maintain normal weight gain, indicating potential stress resilience (left panel). Food consumption remains comparable across genotypes (right panel). h Heatmaps of open field activity in SIS-exposed WT and cHET male mice (left). SIS reduces exploratory behavior similarly in both genotypes (middle-left); however, cHET males exhibit a pronounced increase in thigmotaxis, indicating elevated anxiety-like behavior (middle-right and right panels). RM Two-way ANOVA for Open field activity–Interaction: Time × Genotype + Housing, F (6, 96) = 1.448; ns; Time: F (1.665, 79.94) = 40.33; ****p < 0.0001; Genotype+Housing: F (3, 48) = 2.161; ns. RM Two-way ANOVA for Open field thigmotaxis–Interaction: Time ×  Genotype+Housing, F (6, 94) = 0.6238; ns; Time: F (1.837, 86.33) = 8.092; **p < 0.01; Genotype+Housing: F (3, 47) = 17.39; ***p < 0.0001. i Representative images from the marble burying assay (left). SIS-exposed cHET males bury significantly fewer marbles than WT counterparts, indicative of diminished motivation to perform an ethologically relevant task (middle and right panels). RM Two-way ANOVA for marbles buried over time–Interaction: Time × Genotype + Housing, F (18, 348) = 1.830; *p < 0.05; Time: F (2.270, 131.7) = 101.9; ****p < 0.0001; Genotype + Housing: F (3, 58) = 3.243; *p < 0.05. j Heatmaps of open field activity in SIS-exposed WT and cHET female mice (left). Both locomotor activity and thigmotaxis remain comparable between genotypes (middle and right). RM Two-way ANOVA for Open field activity–Interaction: Time × Genotype + Housing, F (6, 110) = 0.4198; ns; Time: F (1.732, 95.28) = 88.94; ****p < 0.0001; Genotype+Housing: F (3, 55) = 8.151; ****p < 0.0001. k Representative images from the marble burying test in SIS-exposed female mice (left). WT females bury fewer marbles compared to cHETs, suggesting enhanced behavioral resilience in cHET females under chronic stress conditions. Statistical tests: d, f (right), g (right) - Student’s unpaired t test; a, f (left), g (left), h–k (right panels)—Two-way ANOVA with Bonferroni post-hoc test; h–k (left panels) RM Two-way ANOVA with Bonferroni post-hoc test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns not significant. Sample size: d n = 189 WT cells from 4 mice, and 89 cHET cells from 3 mice; f–k n = 8–18 mice/group.

We next investigated how Tsc2 haploinsufficiency in OTR-expressing cells (OTRCs) influences responses to prolonged social isolation. Beginning at 3 months of age, mice were singly housed while receiving identical environmental enrichment—such as bedding and nesting materials—and ad libitum access to food, consistent with group-housed conditions (Fig. 1e). Body weight was recorded at 3 months (baseline) and again at 5 months following the social isolation stress (SIS) period. Wild-type (WT) males showed a normal increase in body weight over the two-month isolation period. In contrast, male conditional heterozygous (cHET) mice failed to show similar weight gain (Fig. 1f). To determine whether this was due to altered feeding behavior, socially isolated males were food-restricted for 24 h, followed by a 10-min measurement of chow consumption in their home cages. No differences in food intake were observed, suggesting that the lack of weight gain in cHET males may stem from increased energy expenditure as a result of sustained stress. In female mice, the pattern differed: cHET females gained weight during the isolation period, whereas WT females did not. Similar to males, food consumption after isolation was comparable between genotypes (Fig. 1g). These findings suggest that Tsc2 heterozygous deletion in OTRCs leads to sex-specific alterations in energy balance regulation under chronic stress, enhancing stress susceptibility in males while promoting stress resilience in females.

Sex differences in stress susceptibility in anxiety-like behaviors

To examine sex-specific behavioral consequences of Tsc2 haploinsufficiency under social isolation stress (SIS), we conducted a longitudinal behavioral analysis of adult WT and cHET mice using a battery of anxiety- and motivation-related assays under both group-housed and post-isolation (10-day SIS) conditions. In male mice, Tsc2 genotype did not influence baseline anxiety-like behavior under group housing (Fig. 1h). However, SIS triggered robust anxiogenic responses in the open field test, with increased thigmotaxis and reduced center zone occupancy. This effect was significantly exacerbated in cHET males, who failed to exhibit the exploratory adaptation for the center zone seen in WT counterparts. In the elevated plus maze, SIS significantly reduced the percentage of time spent in open arms in cHET males (Supplementary Fig. 1a). Furthermore, SIS selectively diminished marble-burying behavior in cHET males but not in WT mice (Fig. 1i), pointing to a deficit in motivated, goal-directed behavior. This reduction occurred independently of self-grooming behavior, which remained unaltered across conditions (Supplementary Fig. 1b), suggesting a specific impairment in motivation rather than a generalized effect on repetitive behaviors.

In contrast, female mice exhibited a distinct behavioral profile. Under group-housed conditions, cHET females spent significantly more time in the center zone during open field exploration, consistent with reduced anxiety-like behavior, whereas SIS led to a pronounced increase in locomotor activity across both genotypes (Fig. 1j), suggestive of elevated arousal or exploratory drive. In the elevated plus maze, SIS unexpectedly increased time spent in the open arms in cHET females compared to group-housed controls; however, SIS did not yield differences between WT and cHET females (Supplementary Fig. 1c). In the marble burying assay, SIS significantly reduced burying behavior in WT females; however, this effect was absent in cHET females, indicating a protective effect of Tsc2 haploinsufficiency against motivational disruption (Fig. 1k). Self-grooming remained stable across conditions (Supplementary Fig. 1d), ruling out stereotypic compensation. Together, these results reveal a striking sex-dependent divergence in behavioral responses to social isolation stress mediated by Tsc2 haploinsufficiency in oxytocin receptor-expressing cells (OTRCs): while cHET males show heightened vulnerability, cHET females display resilience against motivational consequences of SIS.

To discern whether the elevated anxiety-like behavior observed in socially isolated male cHET mice is accompanied by aberrant sociability, we employed the three-chamber social interaction (3CSI) assay. This paradigm assesses social motivation by offering subjects a choice between a novel, sex-matched conspecific and an inanimate object. Under group-housed conditions, both male and female cHET mice exhibited robust social preference, spending significantly more time with the novel conspecific (Fig. 2a, b). Following a 10-day period of social isolation, male cHETs retained a high social preference index comparable to WT controls, indicating preserved social motivation. To further interrogate the nature of this behavior, socially isolated male cHETs were evaluated in a home-cage dyadic interaction assay. These mice demonstrated a significant increase in affiliative social behaviors, including social sniffing and allogrooming, without concomitant increases in aggressive interactions (Fig. 2c, Supplementary Fig. 2a). Importantly, this heightened social engagement did not extend to reproductive behaviors, as socially isolated male cHETs exhibited normal courtship interactions with estrous females (Fig. 2d, Supplementary Fig. 2b). These findings suggest that male cHETs seek social interaction as a coping mechanism to buffer stress induced by anxiogenic contexts, rather than for reproductive purposes. Conversely, socially isolated female cHETs—despite exhibiting resilience to anxiety in open field and elevated plus maze assays—showed a marked reduction in social preference in the 3CSI test, failing to differentiate between a novel female conspecific and an inanimate object (Fig. 2e, f). Given that social interaction is considered an evolutionarily conserved, homeostatically regulated behavior, this lack of social approach indicates a potential shift in behavioral set-point toward an asocial phenotype. This may reflect a long-term adaptive mechanism whereby stress-exposed female cHETs downregulate social motivation as part of a broader strategy to cope with persistent social deprivation. While an alternative explanation could involve emergence of social anxiety, this is less likely in light of their preserved behavioral performance in non-social anxiogenic tasks. Together, these results suggest that Tsc2 haploinsufficiency in OTRCs produces sex-divergent adaptations to chronic social isolation: stress-induced hyper-sociality in males and a blunted social drive in females, dissociable from general anxiety phenotypes.

Fig. 2. Sex differences in social interaction of cHET mice.

a Schematic of the three-chamber social interaction (3CSI) test and representative heatmap illustrating zone preference (left). Both group-housed and socially isolated male mice spent significantly more time in the interaction zone with a novel male conspecific compared to an inanimate object (right), indicating preserved social preference. b All male groups, including WT and cHET mice, demonstrated robust sociability in the 3CSI assay. Sociability index ≥ 0.5 signifies social preference. c In the dyadic social interaction test, socially isolated cHET males exhibited increased affiliative behaviors—spending more time investigating (sniffing and grooming) the unfamiliar male—compared to WT controls (left). No differences were observed in the number of aggressive bouts between genotypes (right). d Socially isolated cHET males exhibited normal mating behavior, with comparable investigation times and mounting durations toward sexually receptive female conspecifics relative to WT controls. e Schematic of the 3CSI test for female cohorts and representative heatmaps of interaction patterns (left). While group-housed female cHET mice displayed normal social preference, socially isolated female cHETs showed reduced interaction with the unfamiliar female, favoring the object zone (right). f A significant reduction in social preference is observed in socially isolated cHET females, indicative of stress-induced social deficits. Sociability index ≥ 0.5 signifies social preference. Statistical tests: a, b, e, f Two-way ANOVA with Bonferroni post-hoc test; c, d Mann-Whitney test. *p < 0.05, **p < 0.01, ****p < 0.001, ns not significant. Sample size: n = 7–15 mice/ group.

Selective anxiolytic effects of pharmacological inhibition of downstream TSC effectors

Having identified that social isolation selectively induces a persistent anxiety-like state in male cHET mice, we next sought to investigate the underlying molecular mechanisms and identify potential targets for behavioral rescue in anxiogenic contexts. Specifically, we analyzed the dysregulation of canonical TSC2-controlled signaling pathways within OTRCs in the mPFC. TSC2 encodes a GTPase-activating protein (GAP) that negatively regulates the small GTPase Rheb (Ras homolog enriched in brain). In its active, GTP-bound form, Rheb stimulates two downstream effectors: mammalian target of rapamycin complex 1 (mTORC1) and PKR-like endoplasmic reticulum kinase (PERK), initiating parallel but functionally divergent signaling cascades^44,51. Rheb is perinuclearly enriched in the cytoplasm and transient interactions with lysosome-localized mTORC1 are sufficient to activate the complex⁵². This activation facilitates substrate binding, including 4E-binding proteins (4E-BPs) and S6 kinase 1 (S6K1), thereby promoting cap-dependent translation, ribosomal biogenesis, and cellular growth.

Concurrently, Rheb can also directly engage PERK—an ER-resident transmembrane kinase that initiates the integrated stress response (ISR) via phosphorylation of eukaryotic initiation factor 2α (eIF2α) at Ser51. Activation of the ISR leads to global translational suppression, while allowing selective translation of stress-responsive mRNAs. Thus, through its interaction with Rheb, the TSC complex exerts dual regulatory control over protein synthesis, simultaneously gating both anabolic (via mTORC1) and stress-adaptive (via PERK) translational responses. To assess the impact of Tsc2 deletion on these pathways within OTRCs, we stereotactically injected an AAV vector expressing a double-floxed mCherry reporter into the mPFC of wild-type and cHET mice to label OTRCs. Following 10 days of social isolation, brain sections were processed for immunohistochemistry and probed for phospho-S6 ribosomal protein (Ser235/236) and phospho-eIF2α (Ser51) as readouts of mTORC1 and PERK activity, respectively. We observed a significant upregulation of both phospho-S6RP ( ↑ 70%) and phospho-eIF2α (↑67%) in cHET OTRCs relative to controls (Fig. 3a, b), indicating concomitant activation of both pathways. These findings demonstrate that the TSC-Rheb axis does not operate as a binary molecular switch but instead exerts coordinated, parallel regulation of mTORC1- and PERK-dependent translational programs. This dual activation state in Tsc2-deficient OTRCs likely contributes to the maladaptive behavioral phenotype observed in male cHETs under stress. Accordingly, these data provide a rationale for testing pharmacological interventions targeting mTORC1 and/or PERK effectors as potential therapeutic strategies for restoring normative affective behaviors.

Fig. 3. Systemic inhibition of mTORC1 and PERK pathways.

a Schematic illustrating bilateral intra-mPFC injections of AAV9.hSyn.DIO.mCherry into WT and cHET mice (left). Representative immunohistochemical images show labeling for phosphorylated S6 ribosomal protein (p-S6RP) and phosphorylated eIF2α (p-eIF2α), markers of mTORC1 and PERK pathway activation, respectively (right). b Quantification of immunofluorescence reveals significantly increased p-S6RP levels in OTRCs within the cHET mPFC, consistent with hyperactivation of mTORC1 signaling (left). Similarly, p-eIF2α levels intensity is significantly elevated in cHET mPFC, indicating enhanced activation of the PERK arm of the ISR (middle). Schematic diagram depicts the downstream consequences of TSC2 disruption in OTRCs, impacting both mTORC1 and PERK signaling pathways (right). c XY trajectory plots showing locomotor activity and center-to-total distance ratio in SIS-exposed male mice following subchronic systemic administration of rapamycin, a selective mTORC1 inhibitor. RM Two-way ANOVA for Open field activity–Interaction: Time ×  Genotype + Drug, F (6, 110) = 1.404; ns; Time: F (1.704, 93.73) = 73.80; ****p < 0.0001; Genotype + Drug, F (3, 55) = 6.247; **p < 0.001. RM Two-way ANOVA for Open field thigmotaxis–Interaction: Time × Genotype + Drug, F (6, 110) = 2.315; *p < 0.05; Time: F (1.906, 104.8) = 8.676; ***p < 0.001; Genotype + Drug: F (3, 55) = 1.472; ns. d Rapamycin treatment reduces spontaneous locomotor activity in both WT and cHET males (left) but has no significant effect on thigmotaxis behavior (right). e Rapamycin restores motivational behavior in socially isolated cHET males, as evidenced by increased marble burying in the marble burying assay. RM Two-way ANOVA for marbles buried over time–Interaction: Time × Genotype + Drug, F (18, 300) = 2.465; ***p < 0.001; Time: F (2.532, 126.6) = 75.46; ****p < 0.0001; Genotype + Drug: F (3, 50) = 3.120; *p < 0.05. f XY trajectory plots illustrating exploratory activity and center preference in SIS-exposed male mice following systemic administration of GSK2606414 (GSK), a selective PERK inhibitor. RM Two-way ANOVA for Open field activity–Interaction: Time × Genotype + Drug, F (6, 76) = 4.620; ***p < 0.001; Time: F (1.858, 70.60) = 88.71; ****p < 0.0001; Genotype + Drug: F (3, 38) = 2.160; ns. RM Two-way ANOVA for Open field thigmotaxis–Interaction: Time × Genotype+Drug, F (6, 78) = 0.9493; ns; Time: F (1.877, 73.21) = 8.316; ***p < 0.001; Genotype + Drug, F (3, 39) = 4.265; *p < 0.05. g PERK inhibition does not significantly alter general locomotor activity in cHET males but rescues anxiety-life behavior by increasing time spent in the center zone of the open field arena. h PERK inhibition with GSK also rescues goal-directed behavior in cHET males, increasing marble burying activity following SIS. RM Two-way ANOVA for marbles buried over time–Interaction: Time × Genotype + Drug, F (18, 288) = 2.695; ***p < 0.001; Time: F (2.285, 109.7) = 93.15; ****p < 0.0001; Genotype + Drug, F (3, 48) = 5.437; **p < 0.01. i PERK inhibition, but not rapamycin, restores social preference in socially isolated cHET female mice, measured by increased time spent interacting with a novel conspecific versus an inanimate object. Conversely, PERK inhibition in WT females abolishes normal social preference, suggesting a critical role of PERK signaling in modulating sociability in a context- and genotype-dependent manner. j GSK, but not rapamycin, effectively rescues social preference deficits in socially isolated cHET female mice. Statistical tests: b Unpaired t test; c, e, f, h—RM Two-way ANOVA with Bonferroni post-hoc test; d, g Two-way ANOVA with Bonferroni post-hoc test. *p < 0.05, **p < 0.01. ***p < 0.001. b n = 54 cells from WT mice, and 21 cells from cHET mice (3 mice/group). c–j n = 6-14 mice/group.

Previous studies have demonstrated that pharmacological inhibition of mTORC1 signaling can ameliorate cognitive and social deficits in adult Tsc2 heterozygous mice^8,53. Although not previously evaluated in TSC models, acute inhibition of PERK has been shown to enhance cognitive performance by promoting translational output^54,55. However, prolonged inhibition of PERK has been associated with impairments in behavioral flexibility and working memory^56,57. Based on these findings, we hypothesized that subchronic inhibition of either mTORC1 or PERK may alleviate social isolation-induced anxiety in male Tsc2 cHET mice and social avoidance in female cHET mice. To test this, we administered either the mTORC1 inhibitor rapamycin (5 mg/kg, intraperitoneally) or vehicle once daily for 3 days to socially isolated WT and cHET mice. Behaviorally, rapamycin attenuated SIS-induced hyperlocomotion in the open field test but did not rescue center avoidance (thigmotaxis) in male cHETs (Fig. 3c, d). Notably, this suppression of hyperactivity was also observed in WT males, suggesting a genotype-independent effect. In contrast, rapamycin restored marble burying behavior in cHET males, indicative of a rescue in motivational drive, but did not significantly impact WT males (Fig. 3e).

Subchronic PERK inhibition with 3 days of oral administration of GSK2606414 had no effect on open field locomotor activity, yet it normalized center exploration in cHET males, indicating reduced anxiety-like behavior (Fig. 3f, g). In WT males, PERK inhibition paradoxically increased locomotion without altering thigmotaxis. Moreover, PERK inhibition rescued marble burying behavior in cHET males, while further elevating marble-burying and stereotypy in WT males (Fig. 3h). Thus, mTORC1 and PERK inhibition exhibited distinct yet complementary effects on exploratory behavior, with both treatments converging on restoration of ethologically relevant behaviors in cHET mutants (Supplementary Fig. 3a-f). In female cHET mice, PERK inhibition—but not rapamycin—rescued SIS-induced social avoidance, reinstating normal social preference (Fig. 3i). Importantly, rapamycin had no adverse effect on social behavior in WT females, whereas PERK inhibition abolished social preference in these animals (Fig. 3j), indicating a narrow optimal window of PERK activity required for intact social interaction. Together, these findings demonstrate that while both mTORC1 and PERK pathways are hyperactivated downstream of Rheb in Tsc2-deficient OTRCs, the PERK-mediated integrated stress response (ISR) plays a primary and context-specific role in the manifestation of affective and social behavioral abnormalities in cHET males and females.

Cell type-specific knockdown of Rheb in mPFC OTRCs rescues stress-induced behavioral avoidance

To localize the neural substrates mediating stress-induced behavioral avoidance in TSC, we selectively manipulated Rheb expression in prefrontal OTRCs of Tsc2 cHET and WT mice. A dual viral strategy was employed involving bilateral stereotaxic injection of a lentiviral construct expressing a Cre-dependent shRNA targeting Rheb mRNA (shRheb) alongside an adeno-associated virus (AAV) encoding a Cre-dependent mCherry reporter, enabling both gene knockdown and identification of OTRCs in the mPFC (Fig. 4a). Quantitative analysis revealed that targeted Rheb knockdown in mPFC OTRCs reduced Rheb protein levels by 23.7% in WT mice and by 46.6% in cHET mice relative to mCherry-expressing controls (Fig. 4b). Although the greater knockdown efficiency in cHET mice was unexpected, these measurements likely reflect combined effects of translational output and degradation dynamics. Loss of a Tsc2 allele expands the pool of active, GTP-bound Rheb, which can be rapidly cycled back to the GDP-bound state to maintain homeostasis. TSC, together with the E3 ubiquitin ligase RNF152, is proposed to ubiquitinate GDP-bound Rheb, targeting it for degradation⁵⁸. These results raise key questions about the regulation of Rheb degradation under Tsc2 haploinsufficiency and point to important directions for future investigation.

Fig. 4. Cell type-specific knockdown of Rheb in prefrontal OTRCs.

a Schematic illustrating viral delivery of mCherry and pSICO-shRheb constructs into mPFC oxytocin receptor-expressing cells (OTRCs) of WT and cHET mice (left). b Representative immunohistochemical images show co-labeling of mCherry and RHEB, confirming effective RHEB knockdown in OTRCs following shRheb expression (left). Quantification of RHEB immunofluorescence intensity reveals significant reduction in RHEB levels in both WT and cHET mice upon shRheb expression (middle). Diagram depicting the proposed mechanism by which Rheb knockdown counteracts the downstream consequences of Tsc2 disruption in OTRCs (right). c XY trajectory plots showing locomotor activity and thigmotactic behavior in SIS-exposed male WT and cHET mice following OTRC-specific Rheb knockdown in the mPFC. RM Two-way ANOVA for Open field activity–Interaction: Time × Genotype + RhebKD, F (6, 88) = 1.088; ns; Time: F (1.745, 76.78) = 68.12; ****p < 0.0001; Genotype+RhebKD: F (3, 44) = 4.395; **p < 0.01. RM Two-way ANOVA for Open field thigmotaxis–Interaction: Time × Genotype + RhebKD, F (6, 88) = 0.5973; ns; Time: F (1.982, 87.22) = 13.30; ****p < 0.0001; Genotype+RhebKD: F (3, 44) = 4.585; **p < 0.01. d OTRC-specific Rheb knockdown enhances spontaneous locomotor activity in SIS-exposed WT males but has no significant effect on overall locomotion in cHET males (left). Targeted Rheb knockdown in OTRCs increases center zone exploration in the open field test, indicative of reduced anxiety-like behavior (right). e Targeted knockdown of Rheb in OTRCs of cHET female mice results in increased time spent interacting with a female conspecific relative to an inanimate object. f Rheb knockdown in OTRCs restores social preference in socially isolated cHET female mice, without affecting social approach behavior in WT females. Statistical tests: a, d, f Two-way ANOVA with Bonferroni post-hoc test; c RM Two-way ANOVA with Bonferroni post-hoc test. *p < 0.05, **p < 0.01. ***p < 0.001, ****p < 0.0001. Sample size: a n = 33-99 cells/group from 3 mice/ group; c–f) n = 6–10 mice/ group.

We next evaluated whether this manipulation could reverse stress-induced behavioral alterations in cHET mice following social isolation. In male cHETs, Rheb knockdown in mPFC OTRCs resulted in increased locomotor activity and enhanced exploration of the center zone in the open field test, resembling the effects of systemic PERK inhibition (Fig. 4c, d). In WT males, Rheb knockdown increased general locomotion without affecting thigmotaxis. On the other hand, in female cHETs, Rheb knockdown in mPFC OTRCs restored normal social preference following SIS exposure, rescuing the social approach deficit observed under control conditions (Fig. 4e, f). No effect on social behavior was observed in WT females, indicating specificity of the manipulation to the mutant phenotype. Collectively, these results implicate the TSC-Rheb signaling axis in mPFC OTRCs as a key node regulating stress-induced behavioral avoidance in both sexes. However, the failure to rescue marble burying behavior in male cHETs (Supplementary Fig. 4a) suggests that motivational deficits are mediated by OTRCs outside the mPFC—potentially within striatal circuits previously shown to require TSC2 signaling for normal execution of repetitive behaviors⁵⁹.

Dysfunctional OTRCs in mPFC mediate network suppression

Given that activation of the PERK-mediated integrated stress response (ISR) in oxytocin receptor-expressing cells (OTRCs) is a primary driver of stress-induced behavioral phenotypes in Tsc2 cHET mice, we hypothesized that protein homeostasis is disrupted in these cells under conditions of chronic social isolation. To test this, we utilized viral-mediated labeling of OTRCs in the medial prefrontal cortex (mPFC) via Cre-dependent mCherry expression, followed by exposure to 10 days of social isolation. Nascent protein synthesis in mPFC slices was then assessed ex vivo using fluorescent non-canonical amino acid tagging (FUNCAT) for a 2-h incubation period (Fig. 5a). Click chemistry-based visualization of FUNCAT-labeled peptides combined with mCherry immunofluorescence revealed a significant reduction in de novo protein synthesis in cHET OTRCs, with an average 28% decrease relative to WT controls (Fig. 5b). These findings suggest impaired translational capacity in mPFC OTRCs, likely compromising their ability to modulate local network activity, particularly in circuits supporting synaptic plasticity and stress adaptation.

Fig. 5. Dysfunctional OTRCs in mPFC mediate network suppression.

a Schematic depicting viral-mediated expression of mCherry in OTRCs within the mPFC (left), along with ex vivo fluorescent non-canonical amino acid tagging (FUNCAT) to assess de novo protein synthesis in acute mPFC slices from WT and cHET mice (right). b Representative immunofluorescence images showing co-localization of anti-mCherry signal and FUNCAT labeling via click chemistry, marking nascent protein synthesis in virally labeled OTRCs (left). Quantification of FUNCAT fluorescence intensity reveals a significant reduction in nascent protein synthesis in OTRCs from cHET mice compared to WT controls. c Schematic illustrating electrophysiological recording of evoked field potentials in layer 5 (L5) of the mPFC following electrical stimulation of layer 2/3 (L2/3). Representative image shows a coronal mPFC slice during ex vivo field potential recording. d TGOT superfusion in male brain slices leads to a sharp decrease in output spiking activity in pyramidal layer of cHET mPFC while it leads to a delayed excitation in WT mPFC (left). Bar plots showing TGOT-elicited decline in network excitation in cHET SIS mPFC 20 min post TGOT treatment (right). e Both WT and cHET female mice exhibit comparable modulation of evoked field potentials in the inner layers of the mPFC following TGOT application (left). No significant genotype-dependent change in field potential amplitude in response to TGOT is observed (right), indicating preserved oxytocin receptor-mediated synaptic responsiveness. f Oral administration of PERK inhibitor, GSK2606414, rescues TGOT-induced network suppression in the mPFC of male cHET mice (left). Bar graphs showing the recovery of network activity 35 min following TGOT application, indicating reinstatement of oxytocin receptor-mediated synaptic modulation (right). g PERK inhibition does not impact TGOT-induced modulation of evoked network activity in the mPFC of WT male mice (left). Both vehicle- and GSK2606414-treated WT animals exhibit comparable network excitation in response to TGOT, indicating that PERK inhibition does not alter oxytocin receptor-mediated synaptic dynamics in the WT mPFC. h Pre-incubation of mPFC slices from cHET mice with Agatoxin IVA (AgTx-IVA), a P/Q-type calcium channel blocker, restores OT-evoked network suppression and rescues the characteristic delayed excitation in the deep pyramidal layer (left). Bar graphs showing a significant increase in network activity following TGOT application in AgTx-IVA pre-treated cHET slices, indicating functional recovery of oxytocin-mediated synaptic modulation (right). i Model of cortical network disruption in the conditional OTRC:Tsc2 cHET mouse model. Statistical tests: b Mann-Whitney test, d–h (left) - RM Two-way ANOVA with Bonferroni post-hoc test. d–h (right)–Two-way ANOVA with Bonferroni post-hoc test. Sample size: b n = 259 WT cells and 274 cHET cells from 4 mice/group. d–h n = 6–14 slices from 3–5 mice/group.

To assess whether this translational dysregulation impacts synaptic function, we performed extracellular field recordings of evoked orthodromic synaptic potentials in acute mPFC slices from socially isolated WT and cHET mice (Fig. 5c). Electrical stimulation was applied to superficial cortical layers (L2/3), and field excitatory postsynaptic potentials (fEPSPs) were recorded from deep layer 5 (L5) in the mPFC. Following a stable 20-min baseline period, slices were perfused with the selective OTR agonist [Thr4, Gly7]-oxytocin (TGOT) for 10 min, and fEPSPs were recorded for an additional 30 min. In WT male mPFC, TGOT application induced a significant increase in fEPSP amplitude, indicative of enhanced network excitability. This effect was absent in cHET males, suggesting a functional deficit in OTRC-mediated excitation of cortical circuits (Fig. 5d). Although the increased excitatory output from mPFC in male WT may appear paradoxical given that majority of OTR-expressing cells (OTRCs) in this region are inhibitory interneurons, primarily somatostatin-positive (SST+), the functional circuitry provides a plausible explanation. SST interneurons are known to innervate not only excitatory pyramidal neurons (PNs) but also parvalbumin-expressing (PV+) fast-spiking interneurons, thereby mediating disinhibition of pyramidal output through PV suppression^60,61. Interestingly, TGOT failed to elicit a similar excitatory response in the female WT mPFC, and no significant difference in evoked activity was observed between WT and cHET females following TGOT stimulation (Fig. 5e). These observations are consistent with prior findings by Li et al.²⁹, who reported enhanced inhibitory postsynaptic currents (IPSCs) from OTR-expressing interneurons onto L5 pyramidal neurons in females, potentially counteracting excitatory network recruitment upon OTR activation. Collectively, these data implicate a cell-type-specific deficit in protein synthesis and circuit-level dysfunction in prefrontal OTRCs of male Tsc2 cHET mice, underlying their impaired ability to mount adaptive responses to social stress.

Given the upregulation of the PERK-mediated integrated stress response (ISR) observed in the medial prefrontal cortex (mPFC) of Tsc2 cHET males, we next evaluated whether pharmacological inhibition of PERK could restore prefrontal network activity. Field excitatory postsynaptic potentials (fEPSPs) were recorded from mPFC slices of socially isolated male mice following subchronic oral administration of the selective PERK inhibitor GSK2606414. As anticipated, PERK inhibition rescued the diminished evoked synaptic responses in cHET males (Fig. 5f), whereas it had no significant effect on network excitability in WT controls (Fig. 5g). Although its effects were not restricted to the mPFC, we observed a rescue of mPFC excitability, suggesting that PERK inhibition either directly restored excitability within the mPFC or did so indirectly via effects on connected circuits.

To test whether the impaired network activity in cHET males results from a failure of OTRCs to inhibit PV interneurons, we pre-treated acute mPFC slices with the selective P/Q-type calcium channel blocker ω-Agatoxin IVA (AgTx-IVA), which preferentially suppresses PV interneuron output due to their high dependence on P/Q-type Ca²⁺ channels^62,63. In cHET mPFC slices, this intervention rescued TGOT-induced network suppression and significantly enhanced evoked responses, indicating that excessive PV interneurons-mediated inhibition underlies the observed network hypoactivity (Fig. 5h). These findings support a cortical microcircuit model in which OTRCs modulate mPFC output through disinhibition of L5 pyramidal neurons by sending strong inhibitory input to PV interneurons (Fig. 5i). Disruption of this inhibitory control in Tsc2-deficient OTRCs leads to hyperactivity of PV cells and consequent suppression of pyramidal output. This pathological shift is exacerbated under stress conditions—such as social isolation or pharmacological OTR stimulation with TGOT—due to heightened oxytocin signaling. Since oxytocin negatively regulates TSC complex activity, its engagement in a Tsc2 haploinsufficient background further amplifies PERK activation and ISR output. Thus, at the network level, Tsc2 loss in OTRCs leads to dysregulation of excitation-inhibition balance in the mPFC, culminating in reduced cortical output under stress, and representing a key mechanism underlying affective dysfunction in Tsc2 mutant males.

Discussion

In this study, we examined the interaction between Tsc2 haploinsufficiency and chronic social isolation stress (SIS) in shaping affective behaviors and underlying cortical circuit function. We identified a pronounced suppression of mPFC network activity in Tsc2 cHET males following SIS, which was associated with heightened behavioral avoidance and increased motivation for social interaction, likely reflecting an attempt to buffer against stress. These results align with prior literature demonstrating that chronic social isolation reduces the excitability of deep-layer pyramidal neurons in the mPFC, contributing to hypofrontality and associated behavioral inflexibility. For example, Park et al.⁶⁴ reported that sustained anxiety states attenuate spontaneous firing in prefrontal neuronal subpopulations, impairing top-down regulation required for adaptive decision-making. While previous studies have reported that haploinsufficiency of Tsc2 in males leads to reduced social interaction^53,65,66, these findings were based on global Tsc2 heterozygous mice, in contrast to the OTRC-specific deletion examined in our study.

Despite similar molecular activation of the ISR in the mPFC across both sexes, only male cHET mice exhibited pronounced anxiety-like behaviors and mPFC hypoactivity, whereas female cHETs maintained behavioral resilience under stress. This sexual dimorphism suggests sex-specific circuit-level adaptations to Tsc2 loss and chronic stress. Previous work by Li et al.²⁹ identified that OTR-expressing interneurons in the mPFC uniquely express corticotropin-releasing hormone-binding protein (CRHBP), which serves to antagonize the effects of the stress hormone CRH. While CRHBP is expressed in OTRCs of both sexes, Li et al. demonstrated that CRH-evoked excitation of L2/3 pyramidal neurons—via CRHR1—is significantly more robust in males than females, suggesting a sex-biased vulnerability to CRH-mediated excitation. It is plausible that during chronic stress and sustained ISR, male OTR interneurons in cHETs undergo downregulation of CRHBP, thereby enhancing CRH-driven excitatory input onto L2/3 pyramidal neurons. This dysregulation could impair the modulation of L2/3 → L5 synaptic communication, disrupting prefrontal output and contributing to the observed hypofrontality and maladaptive behavioral responses. Future studies are warranted to examine the dynamic regulation of CRHBP in OTRCs during chronic stress, and to test its causal role in sex-specific vulnerability to prefrontal dysfunction and stress-induced anxiety.

Our results demonstrate that selective Tsc2 haploinsufficiency in OTRCs potentiates the unfolded protein response (UPR) via the PERK signaling axis, leading to maladaptive behavioral outcomes in male mice subjected to subthreshold chronic social isolation stress (10 days). Specifically, PERK-dependent activation of the ISR was associated with elevated anxiety-like avoidance and reduced ethologically relevant motivated behaviors. These findings align with prior evidence linking chronic social isolation to UPR activation in other model organisms; for instance, prolonged social isolation in Drosophila has been shown to fragment sleep architecture and upregulate all three major UPR branches—including BiP elevation, Xbp1 splicing, and phosphorylation of eIF2α at Ser51 via PERK⁶⁷. Although mTORC1 signaling is known to contribute to ER stress through hyperactivation of cap-dependent translation and increased proteotoxic load, our findings indicate that in Tsc2-deficient OTRCs, the ISR likely does not originate from mTORC1-driven proteotoxic burden. While both mTORC1 and PERK pathways were upregulated in cHET OTRCs, we observed a net reduction in global protein synthesis, suggesting that ISR-mediated translational repression overrides any anabolic input from mTORC1. This interpretation is further supported by our behavioral pharmacology experiments: rapamycin, a selective mTORC1 inhibitor, was effective in restoring motivational drive (e.g., marble burying behavior) in male cHETs but failed to rescue center-avoidant behavior in the open field, indicating that mTORC1 and PERK are activated in parallel, yet functionally dissociable, signaling pathways. Overall, our findings demonstrate that cell-type-specific Tsc2 haploinsufficiency in OTRCs modifies the susceptibility to social isolation. Given the high prevalence of anxiety disorders, it is imperative to investigate the subtypes of pathological anxiety that are driven by distinct circuit-level and molecular mechanisms.

Methods

Animals

Mice were provided with food and water ad libitum and were maintained in a 12 h/12 h light/dark cycle at New York University or Stony Brook University at stable temperature (78^oF) and humidity (40–50%). All mice were backcrossed to C57Bl/6J strain for at least 5 generations. Both male and female mice, aged 3-6 months, were used in all experiments. Floxed Tsc2 mice were kindly provided by Dr. Michael Gambello (Emory University). OTR Cre BAC transgenic mice (Founder line #ON66) were generated by GENSAT and kindly provided by Dr. Nathaniel Heintz (The Rockefeller University). Wildtype C57Bl/6 J mice (stock #000664) were purchased from Jackson labs. All procedures involving the use of animals were performed in accordance with the guidelines of the National Institutes of Health and were approved by the University Animal Welfare Committee of New York University and Stony Brook University.

Drugs and chemicals

Rapamycin stock (50 mg/mL) was dissolved to 1 mg/mL in 5% Tween-80, 15% polyethene glycol 400 (PEG-400), and 0.9% saline. Rapamycin at 5 mg/kg (LC Laboratories #R-5000) was injected intraperitoneally for 3 days. PERK inhibitor GSK2606414 (MedChem Express #HY-18072) was first dissolved in DMSO to prepare the stock concentration of 100 mg/ml. The GSK2606414 stock was diluted to 10 mg/ml with 0.5% HPMC, 0.1% Tween-80 and ddH₂O, and administered to mice at 50 mg/kg by oral gavage, once a day for 3 days. Thr⁴, Gly⁷-Oxytocin (TGOT) (Bachem, # 4013837) was dissolved in ddH₂O at a concentration of 0.4 mM for stock solution and was further diluted in sterile saline to a final concentration of 16 μM for slice electrophysiology. ω-Agatoxin IVA was dissolved in ddH₂O to prepare 5 µM stock, which was further delivered in the bath for a final concentration of 0.5 µM. Azidohomoalanine (AHA) (Click Chemistry Tools, #1066-100) was dissolved in ddH₂O at a stock concentration of 100 mM and diluted to 1 mM in sterile saline. Stock solution of aqueous 32% paraformaldehyde (EMS, #15714) was freshly diluted to 4% in 0.1 M PBS for transcardial perfusions and post-fixation of brain slices.

Stereotaxic surgeries

Mice were anesthetized with the mixture of Ketamine (100 mg/kg) and Xylazine (10 mg/kg) in sterile saline (i.p. injection). Adequate sedation was determined by a lack of gentle toe pinch withdrawal reflex. A lubricant eye ointment (Genteal Tears; Alcon) was applied on the eyes to prevent ocular dehydration. Visual monitoring of respiration occurred periodically throughout surgery to confirm survival. Stereotaxic surgeries were carried out inside a Class II A2 biosafety cabinet (LabRepCo, # NU-677-500) on the Kopf stereotaxic instrument (Model #942), which was equipped with a Nanojector UMP3TA (WdI). Viral vectors were injected intracranially using 2.0 μl Neuros syringe (Hamilton, #65459-02) at 1 nl/s. Before removing the needle, an interval of 10 min was allotted for viral diffusion. Upon completion of intracranial injections, the incision of the scalp was closed using a tissue adhesive (3 M Vetbond). Normothermia during surgery was maintained at 36.6 °C by resting the mouse on a covered heating pad (7 cm × 7 cm) connected to a rodent warmer console (Stoelting, # 53800 M). To prevent somatic dehydration, each mouse was subcutaneously injected with 500 μl of sterile saline before surgery, and further provided with hydrogel (Clear H2O, # 70-01-5022) in the recovery cage. General monitoring and analgesia using 3 mg/kg subcutaneous Ketoprofen (Zoetis) were provided up to three days post-surgery. To knock down Rheb in OTRCs, 200 nl of lenti pSICO.DIO.shRheb (1.00 × 1013 GC/ml, packaged by Vigene, Addgene #81087) was bilaterally injected into the medial prefrontal cortex [−2.00 mm anterior posterior (AP), ±0.25 mm mediolateral (ML) and −1.60 mm dorsoventral (DV)] of either OTR Cre^+/- wildtype od OTR Cre^+/-.Tsc2^f/+ (cHET) mice. The same mice also went through a prior surgery three weeks earlier with AAV9.hSyn.DIO.mCherry.WPRE (1.00 × 1013 GC/ml; Addgene #50459) to label OTRCs in mPFC, since the pSICO vector has no fluorophore expressed in Cre+ cells that are targeted for Rheb knockdown.

Behavior

All behavior sessions were conducted during the light cycle (7:00 AM–7:00 PM EST). Both male and female mice were included in behavior experiments. Tsc2 cHET and control mice were randomly assigned to experimental conditions including drug and vehicle administration, and for the order of testing in any given paradigm. Behavior experiment data were collected by experimenters that were blind to the experimental conditions and genotype. Following weaning, mice were group-housed with up to three sex-matched littermates in standard cages (8 in × 13 in × 5 in) fitted with corn cob bedding and Enviro-dri enrichment. Acclimation of the mice in the behavior room lasted at least 30 min prior to testing. A white noise machine (Yogasleep Dohm) was used to blunt any environmental noise that may induce confounding anxiety. All items in the behavior equipment were cleaned with 30% ethanol unless otherwise stated. Behavior testing occurred in the order of least to most demanding to the mouse subjects.

Social Isolation stress

Adult mice (males and females, ~3 months old) were removed from their group-housed cage and singly housed in a new standard cage for the remainder of their life-span. They were provided with food and water ad libitum and maintained in a 12 h/12 h light dark cycle at a stable temperature (78°F) and humidity (40–50%). The cages were provided with Enviro-dri enrichment identical to group-housed mice. Following 10 days of social isolation stress, mice were subjected to a series of behavioral experiments or euthanized to perform histology and slice electrophysiology.

Feeding assay

Singly housed mice were food-deprived for 24 h, following which they were provided with a single chow pellet in the center of their home cage. The amount of chow consumed by the animal over 10 min was determined by weighing the remaining chow pellet and subtracting from the initial weight and recorded by an experimenter blind to the genotype.

Open field activity

Mice were placed in the center of an open field (27.31 cm × 27.31 cm × 20.32 cm) for 15 min during which a computer-operated optical system (EthoVision XT15 software, Noldus) monitored the spontaneous movement of the mice as they explored the arena. The arena was divided into center zone (inner square: 13.67 cm × 13.67 cm) and total zone encompassing the whole arena of open field. The parameters measured were distance traveled, and the ratio of distance traveled in the center to that in the total zone. Further, the data was analyzed as distance traveled across three time-bins, 5 min each.

Elevated plus maze

Mice were placed in an elevated plus maze (38.5 cm above the floor) and allowed to explore freely for 5 min. Each open and closed arm measured 5 cm × 30.5 cm, with a central zone measuring 5 cm × 5 cm. Behavioral activity was recorded and analyzed using EthoVision XT15 video tracking software, which quantified the duration and frequency of entries into the open and closed arms.

Marble-burying test

Mice were placed into a standard polycarbonate rat cage (28 cm × 19 cm × 15 cm) equipped with 2 inches of fresh bedding (ALPHA-dri) and 20 black glass-marbles (14 mm diameter) arranged as a matrix with 4 rows and 5 columns. The video of the trials was captured using a computer-operated video-tracking software (Ethovision XT15). Each trial was carried out for 30 min. The number of marbles buried every 5 min and the latency to bury the first marble were manually scored.

Self-grooming test

Self-grooming behavior was assessed in 18 cm × 18 cm × 30 cm cuboid chambers (Coulbourn Instruments) housed within sound-attenuated cubicles. Illumination was provided by a white houselight, and the chamber floor was covered with a white Plexiglas platform. To minimize exploratory activity and promote grooming behavior, no environmental enrichment was provided. Each session lasted 30 min. Grooming behavior was recorded using USB video cameras (AM2-CA01, Actimetrics), and videos were manually scored for latency to initiate grooming and total grooming duration during the final 10 min of the session.

Three-chamber social interaction test

Each mouse was introduced into a 3-chamber social box (Harvard Apparatus). The arena was divided into three chambers—two 42.5 × 19.5 ×23 cm side chambers (left and right) and one 42.5 × 18 × 23 cm center chamber with two 10 (W) × 23 (H) cm removable doors that separate the center chamber from the side chambers. Each side chamber had a cylindrical wire enclosure placed in the outer corner. In the first trial, the enclosures were empty, and the subject mouse was allowed to freely explore the arena spanning three chambers. In the second trial, an enclosure in one of the lateral chambers had a novel object whereas the other enclosure in the contralateral chamber had a sex-matched stranger mouse. The locations of the novel object and mouse were counterbalanced across trials. The subject mouse was allowed to freely explore the social and non-social targets. Each trial ran for 5 min. A video-tracking software was used to track the movements of the mice (Ethovision XT15). The duration and frequency of the subject mouse in the zone of interaction (1 inch radius outside of the footprint of enclosure) were assessed. Sociability index was calculated as follows:

$$\mathrm{Sociability\; index}=\frac{S1\left(ZI\right)}{O\left(ZI\right)+S1\left(ZI\right)}$$

where S1(ZI) is the duration of the subject mouse in stranger mice zone of interaction, and O(ZI) is the duration of the subject mouse in the object zone of interaction.

Resident-intruder test

For the resident-intruder test, an unfamiliar male BALB/cJ mouse of similar body weight and age was introduced into the home cage of the resident animal for 10 min. The enrichment, as well as food and water bottle were removed from the cage during the experiment. The dyadic interaction between the resident mouse and intruder in the rearing cage were video-recorded using GigE USB camera and Ethovision XT13 software. Following parameters were manually scored—the amount of time that the resident animal spends investigating the intruder (following, sniffing, grooming) and the amount of time spent on physical attacks (biting, clawing, wrestling, and chasing).

Mating test

For the mating test, a sexually receptive C57Bl/6J female in estrus phase was selected as the mating partner and introduced into the home cage of the male subject for 10 min. The enrichment, as well as food and water bottle were removed from the cage during the experiment. The dyadic interaction between the mice were video-recorded using GigE USB camera and Ethovision XT13 software. An experimenter blind to the genotype manually scored the time the male subject spends investigating the female mouse (following, sniffing, grooming) and the amount of time spent on mounting and consummatory behaviors.

Slice Electrophysiology

Fresh brain slices from conditional heterozygote mice (OTR Cre^+/-.Tsc2^f/+cHET) and their wildtype littermates were used for ex vivo slice electrophysiology experiments. All mice were singly housed for at least 10 days to induce social isolation stress before the recordings. Mice were deeply anesthetized with halothane and perfused with ice-cold slicing solution (85 mM NaCl, 2.5 mM KCl, 1.2 mM NaH₂PO₄, 25 mM NaHCO₃, 0.5 mM CaCl₂, 7 mM MgSO₄, 25 mM glucose, 75 mM sucrose, and 0.5 mM ascorbic acid). The brain was then decapitated and extracted into a beaker filled with ice-cold slicing solution equilibrated with 95% O₂/5% CO₂. The chilled brains were blocked at the level of the optic chiasm and sectioned coronally from rostral to caudal into 300 μm slices using a vibratome. A selected coronal PFC slice was maintained on an interface chamber at 31°C for at least 1 h with artificial cerebrospinal fluid (ACSF) flow rate of 1.5–2.5 ml/min for recording. The ACSF contained 125 mM NaCl, 3.3 mM KCl, 1.2 mM NaH₂PO₄, 25 mM NaHCO₃, 2 mM CaCl₂, 1.2 mM MgSO₄, 15 mM glucose, and 0.5 mM ascorbic acid. A glass recording electrode (filled with ACSF, 5 MΩ resistance) was guided to the inner layers (L5) of the medial prefrontal cortex, and a bipolar stainless steel stimulating electrode (75  Ω) was positioned to target layer 2/3 of mPFC. Orthodromic synaptic potentials were evoked via an isolated current generator (100 μs pulses of 0.3–0.7 mA; Digitimer). Evoked field potentials were recorded with an Axoclamp 2B amplifier and Axon WCP software (Molecular Devices). Field EPSP (fEPSP) was measured as a change in evoked field potential amplitude. Baseline responses were recorded for 20 min at 0.05 Hz with a stimulus intensity of 40–50% of maximum fEPSP. When stated, slices were pre-incubated with 0.5 µM ω-Agatoxin IVA. At the 20 min mark of the baseline recording, a selective oxytocin receptor agonist, [Thr4,Gly7]- oxytocin (TGOT) (2 µM) (Bachem, #4013837), was superfused for 10 min with the same intensity and pulse duration as the baseline stimuli. fEPSP was recorded continuously for 30 min thereafter. Data were analyzed offline using WCP PeakFit (Molecular Devices). Peak amplitude was determined by measuring the maximum deviation from the baseline potential of the evoked fEPSP waveform. The n refers to the slice number. All fEPSP data are presented as mean ± standard error of the mean (SEM). Average traces were calculated over 1 min (n = 6, consecutive sweeps) of stimulation and plotted. These averaged traces were normalized to the mean peak amplitude that was recorded over the 20 min baseline. Experimental groups were compared using a two-tailed t-test.

Fluorescent non-canonical amino acid tagging (FUNCAT)

300 μm-thick brain slices containing medial prefrontal cortex (mPFC) [Bregma +2.34 mm to +1.94 mm] were prepared in cold (4^oC) carboxygenated (95% O₂, 5% CO₂) cutting solution (110 mM sucrose, 60 mM NaCl, 3 mM KCl, 1.25 mM NaH₂PO₄, 28 mM NaHCO₃, 5 mM Glucose, 0.6 mM Ascorbate, 7 mM MgCl₂ and 0.5 mM CaCl₂) using a VT1200S vibratome (Leica). Slices were recovered in oxygenated ACSF solution (125 mM NaCl, 2.5 mM KCl, 1.25 mM NaH₂PO₄, 25 mM NaHCO₃, 25 mM glucose, 1 mM MgCl₂ and 2 mM CaCl₂) at 32 °C for 2 h. 1 mM AHA was then added to the ACSF solution for an additional 2.5 h to label nascent protein synthesis. Following incubation, the mPFC was dissected out and submerged in 4% paraformaldehyde (EMS) at 4 °C overnight. The fixed micro-slices were embedded in 3% agarose, re-cut into 40 μm sections using a vibratome, and stored at −20 °C in the cryosolution overnight. The mPFC sections were blocked and permeabilized in 5% IgG protease-free BSA, 3% normal goat serum, 0.3% Triton X-100 in 1X PBS at RT for 90 min with agitation. The brain sections were then subjected to click chemistry using Cell Reaction kit (Thermo Fisher, #C10269) with 25 μM Alkyne Alexa fluor 405 (Click Chemistry tools, #1309-1), CuSO₄ and additive overnight at 4 °C. The following day, the sections were washed 3 times with 1X PBS for 10 min per wash. Sections were blocked with 1% normal goat serum (NGS) in 1X PBS for 1 h followed by incubation with primary antibodies [rat anti-mCherry (Thermo Fisher, #M11217)] overnight at 4 °C for standard immunohistochemistry.

Immunohistochemistry

Mice were deeply anesthetized with a mixture of ketamine (150 mg/kg) and xylazine (15 mg/kg), and transcardially perfused with 0.1 M PBS followed by 4% paraformaldehyde (EMS) in PBS. Brains were removed and postfixed in 4% PFA for 24 h. The PFA is then replaced with 30% sucrose solution for another 24 h. 40 μm free-floating coronal brain sections containing PFC were collected using Leica vibratome (VT1000s) and stored in 1X PBS containing 0.05% sodium azide at 4 °C. After blocking in 5% normal goat serum in 0.1 M PBS with 0.1% Triton X-100, brain sections were probed overnight with primary antibodies [chicken anti-EGFP (Abcam #ab13970), rabbit anti-EGFP (Thermo Fisher #G10362), rabbit anti-p-eIF2α S51 (Cell Signaling #9721), mouse anti-NeuN (Millipore Sigma #MAB377), rat anti-mCherry (Thermo Fisher #M11217), rabbit anti-p-S6 (240/244) (Cell Signaling #5364S), mouse anti-p-S6 (S235/6) (Cell Signaling # 62016S), rabbit anti-RHEB (Fisher #PIPA520129) and rabbit anti-TSC2 (Sigma #HPA030409). After washing three times in 0.1 M PBS, brain sections were incubated with Alexa Fluor conjugated secondary antibodies 1:200 (Thermo Fisher #A-11001, #A32931, #A32733, #A48255, #A48254, #A32740) in blocking buffer for 1.5 h at RT and mounted using Prolong Gold antifade mountant with or without DAPI (Fisher # P36931, # P36930).

Image acquisition and analysis

Imaging data from immunohistochemistry experiments were acquired using a laser scanning microscopy (LSM) confocal microscope (Zeiss) with 20× objective lens (with 1× or 2× zoom) and z-stacks (approximately 6 optical sections with 0.563 μm step size) for three coronal sections per mouse from AP Bregma +1.98 mm to + 1.78 mm (n = 3 mice) were collected. Imaging data was analyzed with ImageJ using the Bio-Formats importer plugin. Maximum projection of the z-stacks was generated followed by manual outline of individual cells and mean fluorescence intensity measurements using the drawing and measure tools. Mean fluorescence intensity values for all cell measurements were normalized to the mean fluorescence intensity for controls. In the case of verifying Rapamycin’s effect on mTORC1 pathway in mPFC, field fluorescence of the entire image area was measured, instead of outlining each cell. This was done to prevent underrepresentation of images with fewer labeled cells. Each image was opened with ImageJ, and the 6 z-stacks of the images were subjected to maximum projection with maximum intensity. A rectangle was used to outline the entire perimeter of the image, and the fluorescence was measured. The area, and mean intensity of the image were measured.

Statistics

Statistical analyses were performed using GraphPad Prism 8 (GraphPad software). Data are expressed as mean ± SEM. Data from two groups were compared using two-tailed Unpaired Student’s t test. Multiple group comparisons were conducted using one-way ANOVA, two-way ANOVA, or RM Two-way ANOVA with post-hoc tests as described in the appropriate figure legend. Statistical analysis was performed with an α level of 0.05. P values < 0.05 were considered significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Conceptualization: P.S. and EK Methodology: P.S., O.T., S.L., R.D.R.T., M.H., A.F., A.G., K.S.A.R., M.M., M.D., and M.M.O. Investigation: P.S., O.T., S.L., R.D.R.T., M.H., A.F., A.G., K.S.A.R., M.M., M.D., and M.M.O. Funding acquisition: P.S. and E.K. Supervision: P.S. and E.K. Writing – original draft: P.S. Writing – review & editing: E.K., O.T., S.L., R.D.R.T., M.H., A.F., A.G., K.S.A.R., M.M., M.D., and M.M.O.

Peer review

Peer review information

Communications Biology thanks Molly Huntsman, Zhanyan Fu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editor: Benjamin Bessieres. A peer review file is available.

Data availability

All source data supporting the findings are available within the article and Supplementary data. No custom scripts were generated for data analysis. Custom pipelines for data analysis are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-025-09193-3.