Skip to main content
NCBI home page
Molecules and Cells logoLink to Molecules and Cells
. 2026 Jan 21;49(3):100320. doi: 10.1016/j.mocell.2026.100320

Amygdala-hippocampus circuit regulates stress coping via mGluR5-dependent BDNF signaling

Tae-Eun Kim 2,3, Tae-Yong Choi 2,4, Ja Wook Koo 2,3,⁎, Jeongseop Kim 1,⁎
PMCID: PMC12907709  PMID: 41577137

Abstract

Stress-related psychiatric disorders are underpinned by dysfunction in the prefrontal cortex and hippocampus; however, the underlying circuit-specific mechanisms remain ill-defined. Here, we identified the basolateral amygdala (BLA)-to-ventral hippocampus (vHPC) circuit as a critical regulator of stress-coping behaviors. Although chronic social defeat stress reduced the mGluR5 expression in both the vHPC and medial prefrontal cortex (mPFC), our circuit-specific behavioral analysis revealed that the activation of the BLA-vHPC circuit produced a significantly greater improvement in coping behavior compared with the activation of the BLA-mPFC circuit. Subsequently, we mechanistically demonstrated that reduced mGluR5 in the vHPC directly impairs CREB-mediated brain-derived neurotrophic factor (BDNF) transcription, a molecular cascade tightly linked to passive coping. These findings reveal a novel circuit-specific molecular mechanism governing stress recovery, positioning the mGluR5-BDNF pathway as a highly specific and promising therapeutic target for future gene therapy interventions.

Keywords: Basolateral amygdala, Brain-derived neurotrophic factor, Metabotropic glutamate receptor 5, Stress-coping behavior, Ventral hippocampus

Graphical Abstract

ga1

Open in a new tab

INTRODUCTION

Stress is best understood as a dynamic transaction between an individual and their environment, where a stimulus is cognitively appraised as either a challenge or a threat (Lazarus, 1966, Lazarus, 2006). When the demands exceed the available resources, individuals employ coping strategies, defined as continuous efforts to manage the stressor (Lazarus and Folkman, 1984). Although active, problem-focused coping can be adaptive, a hallmark of mood disorders such as major depressive disorder (MDD) is reliance on passive coping (Aldao et al., 2010), which involves avoidance or suppression (Carver et al., 1989). Mounting evidence links these passive strategies to the core symptoms of depression, including helplessness and diminished motivation, ultimately worsening psychological well-being (Aldao et al., 2010).

To investigate the mechanisms of MDD, animal models have been widely developed (Nestler and Hyman, 2010). A cornerstone of this research is the forced swim test (FST), pioneered in the 1970s (Bogdanova et al., 2013, Porsolt et al., 1978a, Porsolt et al., 1978b). In this paradigm, rodents placed in an inescapable water cylinder eventually exhibit immobility, a passive floating posture with minimal movement (Porsolt et al., 1977, Slattery and Cryan, 2012). This was originally interpreted as behavioral despair since antidepressants significantly reduce immobility (Bogdanova et al., 2013, Porsolt et al., 1978a, Porsolt et al., 1978b). However, this interpretation has been increasingly challenged due to its low construct validity; empirical evidence suggests that immobility is a memory-dependent learned response rather than a pathological state (Commons et al., 2017, De Pablo et al., 1989, Jefferys and Funder, 1994, West, 1990). A more widely accepted view now considers immobility not as despair but as an adaptive, energy-conserving response (a learned strategy to maximize survival in a futile situation) (Commons et al., 2017). This reinterpretation is particularly compelling because it suggests that immobility in the FST is, in essence, a behavioral manifestation of passive coping strategies observed in human depression (de Kloet and Molendijk, 2016). Accordingly, in this study, immobility in the FST is conceptualized as a passive coping strategy, rather than as an index of behavioral despair.

This conceptual convergence between passive coping in humans and behavioral immobility in animal models, particularly as measured by the forced swim test (FST), presents a compelling opportunity for integrated research. Previous studies have shown that metabotropic glutamate receptor 5 (mGluR5) signaling contributes to depressive-like phenotypes in both humans (Deschwanden et al., 2011, Quiroz et al., 2016) and rodents (Hughes et al., 2013, Kim et al., 2022, Kim et al., 2025, Li et al., 2006). Although pharmacological inhibition or genetic deletion of mGluR5 has been shown to shift responses from passive to active coping under inescapable stress (Hughes et al., 2013, Kim et al., 2025, Li et al., 2006), its precise circuit-specific role in passive coping remains unclear. This represents a critical gap, especially since passive coping and other depressive symptoms, such as anhedonia, reflect distinct facets of the disorder (Kim et al., 2024). Therefore, a more comprehensive understanding of mGluR5 function requires examining its influence on both dimensions. Furthermore, although prior studies have established that mGluR5 signaling in the basolateral amygdala (BLA) projection to the medial prefrontal cortex (mPFC) and ventral hippocampus (vHPC) contributes to stress resilience (Kim et al., 2022), their specific roles in the mediation of passive coping behaviors have yet to be investigated.

In the present study, we address this gap by directly comparing the contributions of the BLA-mPFC and BLA-vHPC circuits in mediating passive coping following chronic social defeat stress (CSDS). Our findings reveal that the BLA-vHPC circuit exerts a dominant influence on passive coping responses, whereas the BLA-mPFC circuit contributes more selectively to social interaction. Mechanistically, we demonstrate that stress susceptibility is associated with enhanced mGluR5-dependent signaling in the vHPC. This increased signaling promotes CREB phosphorylation and drives brain-derived neurotrophic factor (BDNF) transcription via promoter IV. In contrast, these molecular changes are not observed in the mPFC. By uncovering the circuit- and pathway-specific mechanisms underlying passive coping, this study establishes a framework for linking animal models to human symptom clusters. These insights not only refine our understanding of depression pathophysiology but also highlight mGluR5-CREB-BDNF signaling in the BLA-vHPC circuit as a tractable target for symptom-specific, circuit-guided interventions in MDD.

MATERIALS AND METHODS

Animal Subjects

Male C57BL/6N mice (8 weeks old, Orient Bio) and male CD1 aggressors (3-5 months old, Orient Bio) were used for the study. All mice were provided with food and water ad libitum and were kept on a 12-hour light/dark cycle. Temperature and humidity were strictly controlled. CD1 mice were housed individually, whereas C57BL/6N mice were group-housed before the experiments and singly housed afterward. All animal procedures were approved by the Korea Brain Research Institute's Institutional Animal Care and Use Committee (IACUC-24-00008), with an emphasis on minimizing animal suffering and the number of animals used.

Surgical Procedures

Stereotaxic Surgeries

To achieve optical activation or inactivation of the axon terminals from BLA neurons projecting to either the mPFC or vHPC, adeno-associated viruses (AAVs) were bilaterally infused into the BLA. All viral injections were performed using a 5-µL Hamilton syringe (7641-01) fitted with a 33-gauge needle (7762-06). The stereotaxic coordinates for viral injection from bregma were as follows: BLA (AP, −1.6 mm; ML, ±3.4 mm; DV, −4.9 mm; angle, 7°), as previously described (Kim et al., 2022).

We used AAV5 vectors expressing the excitatory opsin hChR2-EYFP, the inhibitory opsin eNpHR3.0-EYFP, or a control fluorophore (AAV5-hSyn-EGFP). Notably, 1 week after viral infusion, the mice were bilaterally implanted with optic fibers targeting either the mPFC or vHPC. The stereotaxic coordinates for optic fiber implantation from bregma were as follows: mPFC (AP, +1.7 mm; ML, ±0.75 mm; DV, −2.2 mm; angle, 15°); vHPC (AP, −3.7 mm; ML, ±3.0 mm; DV, −4.5 mm; angle, 0°).

For the localized overexpression or knockdown of mGluR5, viral vectors were bilaterally infused into either the mPFC or vHPC. The stereotaxic coordinates for these injections from bregma were as follows: mPFC (AP, +1.7 mm; ML, ±0.75 mm; DV, −2.5 mm; angle, 15°); vHPC (AP, −3.7 mm; ML, ±3.0 mm; DV, −4.8 mm; angle, 0°). A lentiviral vector was used to overexpress mGluR5, whereas an AAV5 vector encoding an shRNA sequence was used for gene knockdown. All viral infusions were delivered at a constant rate of 0.1 µL/min.

For experiments combining optogenetic manipulation and mGluR5 knockdown within the BLA-mPFC or BLA-vHPC pathways, all AAVs were injected into the respective regions on the same day. Notably, 1 week after viral infusion, the mice were bilaterally implanted with optic fibers targeting either the mPFC or vHPC.

Cannula Implantation and Infusions

For our drug administration protocols, we performed bilateral intra-mPFC or vHPC infusions approximately 15 minutes before the subthreshold social defeat stress (SSDS) tests. We used a 33-gauge internal cannula to deliver a PI3K/AKT inhibitor (LY294002) and a MAPK/ERK inhibitor (U0126) at behaviorally active doses. Specifically, LY294002 (1 µg/0.25 µL/side) and U0126 (1 µg/0.25 µL/side) were dissolved in a vehicle solution of 10% DMSO in saline. The infusions were administered at a continuous rate of 0.25 µL/min via a micro-infusion pump (KD Scientific, Legato 200). After the infusion, the injector needles were left in place for an additional 5 minutes to allow for drug diffusion before being withdrawn. Control animals received the vehicle solution only. After the cannulas were removed, the mice were left undisturbed for 5 minutes before the SSDS tests began.

Behavioral Protocols

Chronic Social Defeat Stress

The CSDS protocol involved exposing the C57BL/6N mice to 10 minutes of physical aggression from a CD1 mouse for 10 consecutive days. After each session, the mice were housed in the same cage as the aggressor but separated by a divider. The control mice were placed in similar cages with nonaggressive littermates. After the 10-day period, the mice were classified as susceptible (social interaction ratio < 1.0) or resilient (social interaction ratio > 1.0) based on their social interaction test (SIT) results (Kim et al., 2022). Animals with a social interaction ratio equal to 1.0 were excluded from further analyses.

Subthreshold Social Defeat Stress

The SSDS protocol involved 3 social defeat sessions, each consisting of 5 minutes of physical defeat followed by 15 minutes without physical contact. After the sessions, the mice were returned to their home cages (Kim et al., 2022).

Social Interaction Test

The SIT, performed 24 hours after the last defeat session, consisted of two 2.5-minute sessions in a square arena. In the first session, the mice explored the arena with an empty wire mesh enclosure. In the second session, a novel CD1 mouse was placed inside the enclosure. The time spent in the designated social interaction zone was measured. The social interaction ratio was calculated to classify the mice as susceptible or resilient (Kim et al., 2022).

Forced Swim Test

The FST was performed following the established protocols (Kim et al., 2024). The mice were individually placed in a 2,000-mL Pyrex glass beaker filled with 1,200 mL of water for 6 minutes, with the water replaced between subjects. The test sessions were recorded by a video camera positioned at the side of the beaker, and immobility during the last 4 minutes was scored by an observer blind to the genotype.

Molecular and Histological Analysis

Optogenetic Manipulation

To manipulate the BLA to mPFC or BLA to vHPC pathways, we used in vivo optogenetics for 5 minutes immediately following each daily defeat session during both the 10-day CSDS and 1-day SSDS protocols. For stimulation, a 473-nm blue laser was used to deliver light pulses at a frequency of 20 Hz. These pulses were delivered in bursts of 5, each lasting 40 ms, with a power output of 15-20 mW, and repeated every 10 seconds. For photoinhibition, a 594-nm laser was used to deliver a constant yellow light at 5 mW. Custom patch cords were used to connect the lasers to implanted optic fibers for precise light delivery (Kim et al., 2022).

Brain Slice Imaging

The mice were anesthetized with CO2 gas and transcardially perfused with 0.1 M phosphate-buffered saline, followed by 4% (wt/vol) paraformaldehyde in phosphate-buffered saline, 24 hours after the behavioral test. Brains were postfixed in 4% paraformaldehyde for several hours and cryoprotected in 30% sucrose at 4 °C overnight. Coronal brain sections were prepared using a cryostat (Leica) and stored at −20 °C in cryoprotectant solution until use.

After fixation was carried out, the brain samples were imaged to verify viral fluorescence expression and drug injection sites. The brain sections were either counterstained and mounted with HardSet Antifade Mounting Medium containing 4',6-diamidino-2-phenylindole (DAPI; Vector Laboratories, H-1500) or immunostained with a mouse anti-NeuN primary antibody (1:100, Millipore, MAB377), followed by an Alexa Fluor 594--conjugated donkey anti-mouse IgG secondary antibody (1:500, Invitrogen) to visualize the neuronal populations. Confocal images were acquired using an A1R-MP microscope (Nikon), and panoramic scans were obtained with a Panoramic Slide Scanner (3DHistech) under identical gain and exposure settings.

Quantitative Chromatin Immunoprecipitation

Chromatin immunoprecipitation was performed with modifications based on a previously described protocol (Koo et al., 2015). In brief, the tissues were cross-linked with 1% formaldehyde and quenched with glycine (final concentration: 125 mM) for 5 minutes at room temperature. The fixed tissues were washed twice with ice-cold phosphate-buffered saline containing a protease inhibitor cocktail (cOmplete Mini, Roche, 11836153001) and sonicated using a Bioruptor (Diagenode) to shear the chromatin into fragments of approximately 500 bp at 4 °C.

The supernatant containing transcription factor-bound DNA was incubated overnight at 4 °C with antibodies against total CREB (Abcam, ab32515) or phosphorylated CREB (Abcam, ab32096). The immune complexes were captured with agarose beads (Santa Cruz Biotechnology, sc-2003) and sequentially washed with low-salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.1, and 150 mM NaCl), high-salt buffer (same composition except 500 mM NaCl), LiCl buffer (250 mM LiCl, 1% IGEPAL CA630, 1% sodium deoxycholate, and 1 mM EDTA), and TE buffer (10 mM Tris-HCl, pH 8.1).

Post washing, the immune complexes were eluted from the beads, and cross-links were reversed by incubation at 65 °C overnight. The eluate was treated sequentially with RNase-free DNase (Qiagen, 79254) for 1 hour at 37 °C and with Proteinase-K for 2 hours at 37 °C. DNA was purified by extraction with phenol:chloroform:isoamyl alcohol (Sigma-Aldrich, 77617), followed by cleanup using a polymerase chain reaction (PCR) purification kit (Qiagen, 28104). The PCR primers (Supplementary Table 1) were designed according to the sequences described in previous reports (Koo et al., 2015).

Western Blot

Proteins were extracted from the mPFC and vHPC by homogenization in SynPER Synaptic Protein Extraction Reagent (Thermo Scientific, 87793) supplemented with a protease inhibitor cocktail (Roche, 11697498001). The supernatant was collected by centrifugation, and the protein concentration was determined using a bicinchoninic acid (BCA) protein assay kit (BCA1-1KT, Thermo Scientific).

Equal amounts of protein (20 µg per lane) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions and transferred onto polyvinylidene difluoride membranes. The membranes were blocked and incubated overnight at 4 °C with primary antibodies against mGluR5 (1:2,000, Abcam, ab76316), total CREB (1:1,000, Abcam, ab32515), phosphorylated CREB (1:1,000, Abcam, ab32096), and β-actin (1:6,000, Santa Cruz Biotechnology, sc-47778). The following day, membranes were rinsed 3 times with Tris-buffered saline containing 0.05% Tween-20 (TBS-T, prepared from 10X TBS-T buffer, LPS solution, CBT007) and then incubated for 1 hour at room temperature with horseradish peroxidase-conjugated goat anti-mouse or goat anti-rabbit IgG secondary antibodies (1:5,000, Bethyl Laboratories, A90-116P or A120-101P). After 3 additional washes with TBS-T, immunoreactive bands were visualized using a chemiluminescent horseradish peroxidase substrate (Millipore, P9072). Band intensities were quantified using ImageJ software (NIH). The relative protein expression levels were normalized to β-actin or total protein, as the total protein levels remained unchanged across experimental groups.

Statistical Analysis

All the behavioral data were recorded and analyzed using the automated SMART3.0 video tracking system (Panlab). All values are expressed as the mean ± standard error of the mean (SEM). We performed statistical analyses using GraphPad Prism 10. Prior to analysis, the normality and homoscedasticity of the data were assessed using the Shapiro-Wilk test and Bartlett’s test (or F-test), respectively. Outliers were identified and removed with Grubb’s test. Specific statistical tests and their results are detailed in the figure legends.

RESULTS

mGluR5 in BLA-Recipient Circuits Regulates Stress-Induced Social Avoidance and Coping Strategies

Recent studies have shown that mGluR5 in the mPFC and vHPC, both receiving inputs from the BLA, can act as a molecular switch between susceptibility and resilience to CSDS (Kim et al., 2022). Moreover, mGluR5 has been reported to engage distinct downstream signaling pathways in the mPFC and vHPC to regulate CSDS-induced depressive-like behaviors (Kim et al., 2022). To validate these findings, we combined circuit-specific optogenetic stimulation with an immunoblot analysis. The optogenetic activation of BLA-mPFC or BLA-vHPC projections (20 Hz, 5 pulses) significantly ameliorated social avoidance behavior and increased mGluR5 protein expression in both the mPFC (Supplementary Fig. 1) and vHPC (Supplementary Fig. 2). However, the contribution of mGluR5 within these neural circuits to other behavioral dimensions, particularly passive coping as assessed by the FST, remains largely unknown.

To address this, we manipulated mGluR5 expression in the mPFC and vHPC following CSDS (Fig. 1A and B). The overexpression of mGluR5 using an LV-EF1a-mGluR5-IRES-ZsGreen1 virus in either the mPFC or vHPC significantly rescued CSDS-induced social avoidance (Fig. 1C and E), which was pronounced in CSDS+ZsGreen1 mice. In the subsequent FST, CSDS+ZsGreen1 mice exhibited increased immobility, reflecting a shift toward passive coping. Notably, mGluR5 overexpression in both the mPFC and vHPC not only prevented social avoidance but also promoted active coping by attenuating immobility (Fig. 1D and F).

Fig. 1.

Fig. 1

Open in a new tab

mGluR5-dependent signaling in BLA-recipient circuits regulates stress-induced social avoidance and coping.(A) Schematic of the stress paradigm and viral strategy used for mGluR5 overexpression. (B) Representative confocal images of LV-EF1α-mGluR5-IRES-ZsGreen1 injection sites in the mPFC and vHPC. Scale bars, 500 µm and 250 µm. (C) mGluR5 overexpression in the mPFC prevents CSDS-induced social avoidance (Control+ZsGreen1, n = 8; CSDS+ZsGreen1, n = 7; CSDS+mGluR5 OE, n = 8 mice; 1-way ANOVA with Fisher’s LSD, F2, 20 = 4.427, P = .0256). (D) mGluR5 overexpression in the mPFC prevents CSDS-induced passive coping (n = 8 mice per group; 1-way ANOVA with Tukey, F2, 21 = 10.82, P < .001). (E) mGluR5 overexpression in the vHPC prevents CSDS-induced social avoidance (Control+ZsGreen1, n = 7; CSDS+ZsGreen1, n = 8; CSDS+mGluR5 OE, n = 8 mice; 1-way ANOVA with Fisher’s LSD, F2, 20 = 5.306, P = .0142). (F) mGluR5 overexpression in the mPFC prevents CSDS-induced passive coping (Control+ZsGreen1, n = 8; CSDS+ZsGreen1, n = 7; CSDS+mGluR5 OE, n = 8 mice; 1-way ANOVA with Tukey, F2, 20 = 7.751, P = .0032). (G) Schematic of the stress paradigm and viral strategy used for mGluR5 knockdown. (H) Representative confocal images of the AAV5-hSyn1-mGluR5-shRNA injection sites in the mPFC and vHPC. Scale bars, 500 and 250 µm. (I) mGluR5 knockdown in the mPFC induced social avoidance (Control+GFP, n = 6; SSDS+GFP, n = 7; SSDS+mGluR5 KD, n = 8 mice; 1-way ANOVA with Kruskal-Wallis, H = 7.739, P = .0146). (J) mGluR5 knockdown in the mPFC modestly increased passive coping, but not beyond SSDS levels (Control+GFP, n = 6; SSDS+GFP, n = 7; SSDS+mGluR5 KD, n = 8 mice; 1-way ANOVA with Tukey, F2, 18 = 5.563, P = .0132). (K) mGluR5 knockdown in the vHPC induced social avoidance (Control+GFP, n = 6; SSDS+GFP, n = 7; SSDS+mGluR5 KD, n = 8 mice; 1-way ANOVA with Fisher’s LSD, F2, 18 = 4.781, P = .0216). (L) vHPC mGluR5 knockdown increased the immobility, indicating passive coping after SSDS (Control+GFP, n = 6; SSDS+GFP, n = 7; SSDS+mGluR5 KD, n = 7 mice; 1-way ANOVA with Fisher’s LSD, F2, 17 = 3.728, P = .0454). ∗P < .05, ∗∗P < .01, ∗∗∗P < .001. Data are presented as mean ± SEM.

To further test the role of endogenous mGluR5 in stress susceptibility, we used an AAV5-hSyn1-mGluR5-shRNA-GFP virus to knock down mGluR5 expression in the mPFC and vHPC under SSDS conditions (Fig. 1G and H). As anticipated, SSDS alone did not induce significant behavioral deficits. However, mGluR5 knockdown unmasked region-specific vulnerabilities. In the SIT, mGluR5 knockdown in either the mPFC or vHPC was sufficient to induce social avoidance (Fig. 1I and K). In the FST, mGluR5 knockdown in the mPFC produced only a modest, nonsignificant increase in the immobility compared with SSDS+GFP (Fig. 1J), whereas mGluR5 knockdown in the vHPC resulted in a robust increase (Fig. 1L). To directly test the causal role of mGluR5 in mediating the effects of BLA inputs, we combined an AAV-based mGluR5 knockdown in the mPFC or vHPC (Supplementary Figs. 3A and 4A) with photostimulation of the BLA projections (BLA-mPFC: Supplementary Fig. 3B; BLA-vHPC: Supplementary Fig. 4B). This setup allowed us to investigate whether mGluR5 directly mediates the modulatory effect of BLA projection activity on stress-related social avoidance behavior. Consistent with previous reports, CSDS induced social avoidance, and mGluR5 knockdown suppressed recovery (Supplementary Figs. 3C and 4C). Importantly, even with robust photostimulation of the BLA-mPFC (Supplementary Fig. 3C) or BLA-vHPC circuits (Supplementary Fig. 4G), social avoidance persisted when mGluR5 expression was inhibited in the mPFC (Supplementary Fig. 3D and E) or vHPC (Supplementary Fig. 4H and I). These results demonstrate that mGluR5 is essential for the BLA inputs to modulate social behavior, confirming a direct causal link between mGluR5 and the glutamatergic neural circuit's effect on stress resilience. Together, our results show that although mGluR5 in both the mPFC and vHPC is critical for mitigating social avoidance, mGluR5 in the vHPC plays a disproportionately greater role in suppressing passive coping strategies under stress.

BLA-vHPC Circuit Exerts a Stronger Influence on Stress Coping Than the BLA-mPFC Circuit

To investigate the causal contribution of BLA projections to the mPFC and vHPC in stress-induced behaviors, we injected an anterograde AAV carrying ChR2 or NpHR into the BLA and implanted optic fibers above either the mPFC and/or vHPC to selectively manipulate these projections (Fig. 2A and B). The optogenetic activation of BLA axon terminals in the mPFC or vHPC with ChR2 (20 Hz, 5 pulses for 5 minutes, delivered immediately after the 10 minutes of physical defeat and followed by 24 hours of sensory stress each day; Fig. 2C) effectively rescued CSDS-induced social avoidance compared with CSDS+EGFP mice (Fig. 2D). These results indicate that BLA-recipient circuits are sufficient to abolish stress-induced social avoidance, consistent with previous findings.

Fig. 2.

Fig. 2

Open in a new tab

The BLA-vHPC circuit predominates over the BLA-mPFC circuit in controlling stress coping.(A) Schematics of the CSDS paradigm and optogenetic stimulation of the BLA-mPFC and BLA-vHPC pathways. (B) Representative confocal images showing rhodopsin virus injection in the BLA with expression in the mPFC and vHPC. Scale bar, 500 µm. (C) Optogenetic stimulation of the BLA projections to the mPFC and vHPC. (D) Optogenetic activation of BLA projections to the mPFC and vHPC rescued social avoidance after CSDS [Control+EGFP, n = 8; CSDS+EGFP, n = 10; CSDS+ChR2 (mPFC), n = 6; CSDS+ChR2 (vHPC), n = 6 mice; 1-way ANOVA with Fisher’s LSD, F3, 26 = 13.94, P < .0001]. (E) Optogenetic activation of BLA projections during CSDS reduced the duration of immobility in both the mPFC and vHPC, with a stronger effect in the BLA-vHPC pathway [Control+EGFP, n = 10; CSDS+EGFP, n = 10; CSDS+ChR2 (mPFC), n = 7; CSDS+ChR2 (vHPC), n = 7 mice; 1-way ANOVA with Fisher’s LSD, F3, 30 = 13.28, P < .0001]. (F) Schematic of simultaneous optogenetic inhibition of BLA-mPFC projections and activation of BLA-vHPC projections. (G) BLA-mPFC inhibition induced social avoidance after CSDS, whereas BLA-vHPC activation rescued it. When both circuits were activated, the vHPC activity predominated, preventing social avoidance [Control+EGFP, n = 8; CSDS+EGFP, n = 10; CSDS+NpHR (mPFC), n = 6; CSDS+ChR2 (vHPC), n = 7; CSDS+ChR2 (vHPC)+NpHR (mPFC), n = 6 mice; 1-way ANOVA with Fisher’s LSD, F4, 32 = 10.39, P < .0001]. (H) BLA-mPFC inhibition promoted passive coping after CSDS, whereas BLA-vHPC activation shifted the behavior to active coping. With simultaneous activation, the vHPC activity predominated, maintaining active coping [Control+EGFP, n = 8; CSDS+EGFP, n = 10; CSDS+NpHR (mPFC), n = 5; CSDS+ChR2 (vHPC), n = 7; CSDS+ChR2 (vHPC)+NpHR (mPFC), n = 6 mice; 1-way ANOVA with Fisher’s LSD, F4, 31 = 7.787, P < .001]. (I) Schematic of simultaneous optogenetic activation of BLA-mPFC projections and inhibition of BLA-vHPC projections. (J) BLA-mPFC activation suppressed social avoidance after CSDS, whereas BLA-vHPC inhibition induced it. With simultaneous activation, the mPFC activity predominated, preventing social avoidance [Control+EGFP, n = 10; CSDS+EGFP, n = 8; CSDS+ChR2 (mPFC), n = 7; CSDS+NpHR (vHPC), n = 6; CSDS+ChR2 (mPFC)+NpHR (vHPC), n = 7 mice; 1-way ANOVA with Fisher’s LSD, F4, 33 = 14.87, P < .0001]. (K) BLA-mPFC activation promoted active coping after CSDS, whereas BLA-vHPC inhibition induced passive coping. With simultaneous activation, the vHPC activity predominated, maintaining passive coping [Control+EGFP, n = 10; CSDS+EGFP, n = 8; CSDS+ChR2 (mPFC), n = 7; CSDS+NpHR (vHPC), n = 6; CSDS+ChR2 (mPFC)+NpHR (vHPC), n = 6 mice; 1-way ANOVA with Fisher’s LSD, F4, 32 = 6.489, P < .001]. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. Data are presented as mean ± SEM.

When the coping behavior was assessed in the FST, photoactivation of both BLA-recipient circuits significantly reduced immobility compared with that in CSDS+EGFP mice, thereby enhancing active coping (Fig. 2E). Notably, activation of the BLA-vHPC circuit produced a greater improvement in the coping behavior than activation of the BLA-mPFC circuit, indicating that the BLA-vHPC circuit is more sensitive to the modulation of stress-coping strategies.

To identify which circuit is more critical for passive coping, we simultaneously manipulated the BLA-mPFC and BLA-vHPC circuits (Fig. 2F). The optogenetic inhibition of BLA axon terminals in the mPFC using NpHR (constant yellow light, 5 mW) exacerbated CSDS-induced social behavioral deficits (Fig. 2G). Conversely, the optogenetic activation of the BLA axon terminals in the mPFC restored CSDS-induced social behavior. When both manipulations were applied concurrently, namely the inhibition of BLA-mPFC with activation of BLA-vHPC, there were no social behavioral deficits.

Subsequently, we evaluated the coping behavior after 10 days of CSDS and optic manipulation. The CSDS+EGFP group, as well as the group with BLA-mPFC inhibition, displayed increased immobility compared with other groups (Fig. 2H). However, activation of the BLA-vHPC circuit reversed passive coping behavior. Importantly, when both circuits were manipulated simultaneously, activation of the BLA-vHPC circuit still reduced immobility despite concurrent inhibition of the BLA-mPFC circuit.

To further clarify the relative contributions of these circuits, we directly compared the effects of concurrent manipulation of the BLA-mPFC and BLA-vHPC circuits (Fig. 2I). Our findings reveal a clear functional distinction, as the 2 circuits exhibited a compensatory relationship in social behavior, with activation of the BLA-mPFC circuit enhancing social interaction and inhibition of the BLA-vHPC pathway inducing social avoidance (Fig. 2J). When both manipulations were combined (BLA-mPFC activated, BLA-vHPC inhibited), the social behavior was maintained at control levels, suggesting that these circuits exert redundant or complementary influences on social interaction.

By contrast, the coping behavior revealed a hierarchical relationship dominated by the BLA-vHPC circuit. Inhibition of the BLA-vHPC circuit promoted robust passive coping, and this effect persisted even when the BLA-mPFC circuit was simultaneously activated, which typically reduces immobility (Fig. 2K). Thus, the influence of the BLA-vHPC circuit on coping behavior was prepotent, overriding the opposing effect of the BLA-mPFC circuit. These results indicate that, unlike social behavior, passive coping is primarily governed by the BLA-vHPC circuit, highlighting a distinct functional hierarchy between these 2 circuits.

mPFC and vHPC Engage Distinct Molecular Mechanisms in Stress-Induced Behaviors

Our prior findings revealed a striking divergence in how mGluR5 mediates the stress-induced behavioral changes in the mPFC and vHPC, despite both regions receiving glutamatergic inputs from the BLA (Kim et al., 2022). However, the underlying circuit-specific molecular mechanisms for these behavioral changes, particularly those related to passive coping, remain unknown.

To address this gap, we sought to identify a signaling pathway that regulates passive coping behavior. Previous studies have established that mGluR5 activation leads to ERK1/2 phosphorylation and the upregulation of transcription factors such as CREB (Jong et al., 2009, Mao et al., 2005, Wang et al., 2007). Furthermore, our previous work demonstrated a functional divergence between the BLA-mPFC and BLA-vHPC pathways, with the mPFC circuit mediating depressive-like behavior through an mGluR5-Homer1b/c-AKT signaling pathway, whereas the vHPC circuit engaging an ERK-dependent mechanism (Kim et al., 2022). Building on this evidence, we hypothesized that CREB, a key downstream effector of mGluR5, could explain differential contributions of the mPFC and vHPC. Given the established link between mGluR5-ERK signaling and CREB phosphorylation, we specifically hypothesized that CREB plays a more significant role in the BLA-vHPC pathway.

To test this hypothesis, we used SSDS to investigate whether the AKT and ERK pathways influence circuit-specific passive coping behavior. Following bilateral cannula implantation in the mPFC or vHPC, mice received infusions of either the PI3K-Akt inhibitor LY294002 or the MAPK/ERK inhibitor (U0126 Fig. 3A). Notably, 24 hours post SSDS, we measured the social interaction and coping behaviors in mice infused bilaterally with pathway-specific inhibitors (Fig. 3B and E). Our results revealed a clear dissociation between the 2 regions. The infusion of LY294002 into the mPFC significantly reduced social interaction (Fig. 3C) but had no effect on the coping behavior (Fig. 3D). Conversely, the infusion of U0126 into the vHPC not only decreased the social interaction (Fig. 3F), but also significantly increased passive coping, as reflected by increased immobility in the FST (Fig. 3G).

Fig. 3.

Fig. 3

Open in a new tab

Distinct molecular pathways in the mPFC and vHPC mediate stress-induced behaviors.(A) Schematics illustrating the experimental design for pharmacological inactivation during subthreshold social defeat stress. Representative images of the bilateral cannula implantation sites in the (B) mPFC and (E) vHPC. Scale bar, 2 mm. (C) Infusion of the AKT antagonist LY294002 into the mPFC significantly reduced social interaction after subthreshold social defeat stress, whereas the MAPK/ERK antagonist U0126 had no effect (Control+Veh, n = 8; SSDS+Veh, n = 8; SSDS+LY294002, n = 7; SSDS+U0126, n = 8 mice; 1-way ANOVA with Fisher’s LSD, F3, 27 = 4.616, P = .0099). (D) Infusion of LY294002 or U0126 into the mPFC had no effect on FST immobility after subthreshold social defeat stress (Control+Veh, n = 8; SSDS+Veh, n = 8; SSDS+LY294002, n = 8; SSDS+U0126, n = 7 mice). (F) Infusion of U0126 into the vHPC significantly reduced social interaction after subthreshold social defeat stress, whereas LY294002 had no effect (Control+Veh, n = 7; SSDS+Veh, n = 8; SSDS+LY294002, n = 8; SSDS+U0126, n = 7 mice; 1-way ANOVA with Fisher’s LSD, F3, 26 = 3.659, P = 0.0253). (G) Infusion of U0126 into the vHPC markedly increased the duration of immobility even under subthreshold stress, an effect not observed in other groups, highlighting the importance of the ERK pathway downstream of mGluR5 in the vHPC (Control+Veh, n = 5; SSDS+Veh, n = 8; SSDS+LY294002, n = 7; SSDS+U0126, n = 8 mice; 1-way ANOVA with Fisher’s LSD, F3, 24 = 4.323, P = .0143). ∗P < .05, ∗∗P < .01. Data are presented as mean ± SEM.

Together, these results suggest that the PI3K-Akt pathway in the mPFC and the MAPK/ERK pathway in the vHPC play distinct roles in stress-induced behaviors. Although both regions are critical for behavioral regulation under stress, their divergent downstream signaling cascades govern different components of the response, with the vHPC circuit specifically mediating passive coping strategies.

Chronic Social Defeat Stress Does Not Alter Bdnf Transcription in the mPFC

Given the established link between mGluR5-ERK signaling and CREB phosphorylation, together with our finding that selective ERK inhibition affected passive coping only in the vHPC, we subsequently investigated the underlying molecular mechanism in these regions. We hypothesized that CREB plays a significant role in the BLA-vHPC circuit but not the BLA-mPFC circuit. To test this, we performed a western blot analysis to examine the CREB phosphorylation in the mPFC (Fig. 4A) and vHPC (Fig. 5A) of stress-susceptible and resilient mice. As anticipated, no group differences were detected in either the total CREB (tCREB) or phospho-CREB (p-CREB) levels in the mPFC (Fig. 4B-D).

Fig. 4.

Fig. 4

Open in a new tab

CSDS does not alter Bdnf transcription in the mPFC.(A) Representative western blots and quantification of mGluR5, p-CREB/CREB, proBDNF, and mature BDNF expression in the mPFC. (B) Quantification of the p-CREB/β-actin levels in the mPFC; no group differences observed. (C) Quantification of the tCREB/β-actin levels in the mPFC; no group differences observed. (D) Ratio of p-CREB/tCREB in the mPFC; no group differences observed. (E) proBDNF/β-actin expression in the mPFC; no group differences observed. (F) Susceptible mice showed reduced mature BDNF compared with the controls, but not compared with resilient mice (1-way ANOVA with Tukey, F2, 6 = 8.070, P = .0199). (G) The ratio of mature BDNF/proBDNF was decreased in susceptible mice relative to other groups (1-way ANOVA with Tukey, F2, 6 = 17.50, P = .0031). (H) Schematic showing the relative positions of amplicons (green bars) generated by the primers used for ChIP-qPCR. Exons are depicted as boxes, introns as connecting lines, and exon numbers in Roman numerals. The CREB-binding sites (red circles) are indicated at Bdnf promoter regions relative to the transcription start site: exon I (−96 to −84 bp/−90 to −78 bp), exon II (−317 to −307 bp), exon IV (−42 to −33 bp/−36 to −26 bp), and exon VI (−84 to −73 bp). (I) tCREB binding to the BDNF promoters I, II, IV, and VI in the mPFC; no group differences observed. (J) p-CREB binding to Bdnf promoters I, II, IV, and VI in the mPFC; no group differences observed. n = 3 mice per group for all panels. ∗P < .05, ∗∗P < .01. Data are presented as mean ± SEM.

Fig. 5.

Fig. 5

Open in a new tab

CSDS reduces CREB-dependent Bdnf transcription in the vHPC.(A) Representative western blots and quantification of mGluR5, p-CREB/CREB, proBDNF, and mature BDNF expression in the vHPC. (B) Susceptible mice showed reduced p-CREB/β-actin expression compared with that in the control and resilient groups (1-way ANOVA with Tukey, F2, 6 = 106.3, P < .0001). (C) Quantification of the tCREB/β-actin levels in the vHPC; no group differences observed. (D) In the vHPC, susceptible mice showed a significantly reduced p-CREB/tCREB ratio compared with other groups (1-way ANOVA with Tukey, F2, 6 = 26.99, P = .0010). (E) proBDNF/β-actin expression in the vHPC; no group differences observed. (F) Susceptible mice showed reduced mature BDNF compared with that in other groups (1-way ANOVA with Tukey, F2, 6 = 36.46, P < .001). (G) The ratio of mature BDNF/proBDNF was decreased in susceptible mice relative to other groups (1-way ANOVA with Tukey, F2, 6 = 30.88, P < .001). (H) Schematic of Bdnf promoter regions analyzed by ChIP-qPCR, showing exon structure and CREB-binding sites. (I) tCREB binding to Bdnf promoters I, II, IV, and VI in the vHPC; no group differences observed. (J) In the vHPC, susceptible mice showed reduced p-CREB binding at Bdnf promoter IV, leading to decreased Bdnf transcription, whereas binding at promoters I, II, and VI was unchanged (2-way ANOVA with Fisher’s LSD, F6, 18 = 3.341, P = .0216). n = 3 mice per group for all panels. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, ∗∗∗∗P < .0001. Data are presented as mean ± SEM.

Because CREB is a central regulator of BDNF-induced gene expression (Finkbeiner et al., 1997), we next assessed the BDNF protein levels. Although the precursor BDNF (proBDNF/β-actin) did not differ among the groups (Fig. 4E), we identified a critical defect in post-translational processing. Specifically, the ratio of mature BDNF monomer to β-actin was significantly reduced in susceptible mice compared with that in controls, though not relative to resilient mice (Fig. 4F). This suggests that although both susceptible and resilient mice ultimately achieve comparable levels of mature BDNF, the efficiency of processing is markedly reduced in the susceptible group (Fig. 4G). Given that mature BDNF promotes neuronal survival and plasticity, whereas proBDNF can induce cell death and synaptic pruning, a reduced conversion efficiency likely shifts the functional balance toward a proapoptotic state, providing a compelling molecular explanation for the behavioral vulnerability of the susceptible phenotype. However, because both susceptible and resilient mice were exposed to the same stress and ultimately displayed similar levels of mature BDNF, this post-translational defect alone is unlikely to fully explain their divergent behavioral outcomes.

To determine whether these effects extend to transcriptional regulation, we focused on the key Bdnf promoters identified in prior research. Previous studies have demonstrated that chronic stress paradigms such as social defeat decrease Bdnf mRNA expression via decreased activity at promoters III and IV (Tsankova et al., 2006), whereas chronic antidepressant treatment with imipramine increases the expression through the same promoters (Tsankova et al., 2006). Furthermore, we included promoters I, II, IV, and VI, which are known to be activity-regulated and subject to transcriptional suppression through RNA polymerase II stalling and reduced CREB binding following chronic morphine exposure (Koo et al., 2015). By focusing on these stress- and activity-responsive promoters, our goal was to determine whether our model exhibited a similar pattern of Bdnf regulation.

Our subsequent quantitative chromatin immunoprecipitation analysis, using the antibodies against total and phospho-CREB, confirmed that the observed changes in BDNF expression were independent of CREB. As anticipated, we found no significant CREB binding at Bdnf promoters I, II, IV, or VI in the mPFC (Fig. 4H-J). Collectively, these data point to an alternative, CREB-independent transcriptional mechanism regulating Bdnf expression in this region.

Chronic Social Defeat Stress Alters Bdnf Transcription in the vHPC

In marked contrast to our findings in the mPFC, the western blot analysis revealed a significant reduction in p-CREB levels in the vHPC of susceptible mice, whereas the tCREB levels remained unchanged (Fig. 5B-D). This selective decrease in CREB phosphorylation supports our hypothesis that the BLA-vHPC circuit, unlike the BLA-mPFC circuit, critically depends on CREB signaling for stress-induced behavioral regulation.

Next, we examined the BDNF protein expression in the vHPC. Consistent with the reduction in p-CREB, susceptible mice showed a pronounced decrease in the mature BDNF monomer levels relative to both resilient and control groups (Fig. 5E-G). In contrast, precursor BDNF (proBDNF/β-actin) did not differ across groups, suggesting that the deficit arises not from impaired synthesis of BDNF protein but from impaired processing of proBDNF into its mature form. Because mature BDNF promotes synaptic potentiation and neuronal survival, whereas proBDNF facilitates apoptotic and synaptic pruning, this imbalance strongly implicates a shift in the molecular milieu of the vHPC toward a proapoptotic, stress-vulnerable state.

To test whether this deficit was transcriptionally mediated, we performed quantitative chromatin immunoprecipitation assays targeting activity- and stress-sensitive Bdnf promoters (I, II, IV, and VI), previously implicated in stress-induced transcriptional repression (Fig. 5H). Notably, we observed a selective reduction in p-CREB binding at promoter IV in the vHPC of susceptible mice, whereas this effect was absent at promoters I, II, and VI, with the total CREB binding remaining unchanged across all the promoters (Fig. 5I and J). Thus, promoter IV emerges as a critical locus where the stress-induced reductions in activated CREB selectively impair Bdnf regulation under chronic stress conditions. Taken together, these findings delineate a molecular cascade in which CSDS reduces the p-CREB levels in the vHPC, diminishes the CREB occupancy at promoter IV, and consequently downregulates mature BDNF expression. This CREB-dependent repression of Bdnf in the vHPC stands in sharp contrast to the mPFC, where neither the CREB phosphorylation nor BDNF transcription was affected. Therefore, the vHPC functions as a stress-sensitive hub within the BLA-vHPC circuit. In susceptible mice, impaired CREB signaling directly impairs the BDNF availability through the mGluR5-CREB-BDNF pathway, a molecular mechanism that drives the behavioral phenotype of high passive coping in the FST (Fig. 6). These results provide a clear link between a specific neural circuit, its underlying molecular defect, and behavioral vulnerability to CSDS.

Fig. 6.

Fig. 6

Open in a new tab

Proposed model illustrating the mechanism underlying CSDS-induced passive coping behavior. Schematic showing the proposed mechanism mediated by mGluR5 in CSDS-induced passive coping behavior. Both BLA-projecting circuits contribute to behavioral change in social interaction; however, the BLA-vHPC circuit plays a more dominant role than the BLA-mPFC in regulating passive coping behavior.

DISCUSSION

In this study, we systematically investigated the distinct contributions of the BLA-mPFC and BLA-vHPC circuits to stress-induced behavioral and molecular changes, offering a refined neurobiological framework for understanding the symptom-specific vulnerabilities in MDD. Our findings reveal a functional hierarchy and molecular dissociation between these 2 circuits, challenging the notion of a monolithic depression circuit and providing a circuit-based model of psychopathology (Lv et al., 2022).

For decades, the FST has been central to antidepressant drug screening, with immobility traditionally interpreted as behavioral despair (Nestler and Hyman, 2010, Yankelevitch-Yahav et al., 2015). However, the FST has been criticized for its low face and construct validity as a model of depression (Commons et al., 2017, de Kloet and Molendijk, 2016). More recent perspectives emphasize that immobility in the FST reflects an adaptive, energy-conserving passive coping strategy rather than a simple lack of hope (Yankelevitch-Yahav et al., 2015). This reinterpretation aligns with our findings that mGluR5 modulation selectively shapes the coping style, paralleling the passivity and reduced motivational engagement observed in clinical depression (de Kloet and Molendijk, 2016).

Our optogenetic and pharmacological manipulations revealed a functional segregation between the BLA-mPFC and BLA-vHPC circuits. For social avoidance, a core symptom of MDD (Lv et al., 2022), the circuits displayed compensatory or redundant roles: the activation of either circuit rescued stress-induced deficits (Anderson et al., 2020). This suggests that social behavior, a highly conserved process (Kim et al., 2022), is maintained by a resilient, multicircuit system that can withstand partial impairment (Zhang et al., 2019).

In contrast, the coping behavior is governed by a prepotent hierarchy dominated by the BLA-vHPC circuit. The inhibition of this pathway robustly promotes passive coping even when the BLA-mPFC circuit is simultaneously activated (Felix-Ortiz et al., 2013), positioning BLA-vHPC as a dominant arbiter of survival strategies under inescapable stress (Felix-Ortiz et al., 2013). Crucially, we argue that passive coping is not simple motor quiescence for energy saving but an integrated conservation---a withdrawal state that reflects an allostatic recalibration in which resources are reallocated when escape is judged improbable. Within this framework, the amygdala provides salience- and valence-weighted threat signals, and the vHPC encodes contextual uncertainty and anxiety. Their convergence biases downstream nodes (eg, BNST, hypothalamus, and periaqueductal gray) toward reduced exploration, parasympathetic predominance, altered HPA dynamics, and a metabolic downshift that minimizes detection risk while preserving energy for future adaptive responses (Anastasiades and Carter, 2021). This hierarchical model offers a mechanistic account for patients who can maintain normal social function yet remain profoundly passive and unmotivated, representing a debilitating MDD symptom cluster.

At the molecular level, we uncover a dissociation in downstream mGluR5 signaling. In the mPFC, the social behavior aligns with PI3K-Akt signaling, consistent with prior work that mGluR5-Homer interactions can selectively engage this pathway. In the vHPC, by contrast, the coping behavior critically depends on the MAPK/ERK-CREB-BDNF cascade (Wang and Mao, 2019), as evidenced by ERK inhibition reducing social interaction and increasing passive coping (Wagner et al., 2013). This aligns with studies showing that stress-driven changes in Homer proteins disrupt the hippocampal synaptic function (Reshetnikov and Bondar, 2021). Thus, the same receptor (mGluR5) maps onto circuit-specific signaling architectures that differentially gate behavior.

Crucially, the 2 regions converge on mature BDNF deficiency via distinct mechanisms. In the vHPC of susceptible mice, we observed reduced p-CREB and diminished p-CREB binding at BDNF promoter IV, linking transcriptional failure directly to high passive coping (Alboni et al., 2011). In the mPFC, CREB phosphorylation and BDNF transcription were unaltered; instead, the deficit lay in the post-translational processing of proBDNF into mature BDNF (Fan et al., 2025). These distinctions reinforce the view that MDD reflects heterogeneous, circuit-specific vulnerabilities rather than uniform breakdown (Miao et al., 2020). Conceptually, our data position the mGluR5-CREB-BDNF axis as a biochemical substrate through which the BLA-vHPC hierarchy enforces a system-wide low-expenditure policy, whereas the PI3K-Akt perturbations associated with the mPFC selectively impair social function.

Our findings suggest a new avenue for targeted therapeutics for MDD. The identification of the BLA-vHPC circuit as a prepotent regulator of coping makes it an attractive treatment target. Pharmacological strategies could include positive allosteric modulators (PAM) of mGluR5 to enhance synaptic plasticity and resilience (Cleva and Olive, 2011, Kim et al., 2025). This is supported by the evidence showing that ketamine and its metabolites directly bind TrkB, potentiating BDNF signaling (Casarotto et al., 2022, Duman et al., 2021, Kim et al., 2024). Such mechanisms point to a paradigm shift in antidepressant development, beyond monoaminergic modulation toward interventions that promote activity-dependent neuroplasticity. In parallel, circuit-specific neuromodulation, such as repetitive transcranial magnetic stimulation protocols (Fitzsimmons et al., 2024) targeting the BLA-vHPC circuit, could yield symptom-specific benefits for passivity and anhedonia.

This study has a few limitations that warrant consideration. The CSDS/SSDS models, despite ethological validity, can yield dynamic trajectories with phenotype switching over time (Wendelmuth et al., 2020), and the historical emphasis on males limits the generalizability, underscoring the need for robust female defeat paradigms. Methodologically, optogenetics requires invasive implants and offers limited tissue penetration (Lehtinen et al., 2021, Malyshev et al., 2015); viral gene transfer can elicit immune responses that compromise expression. Future work should address these constraints with multimodal approaches of cell-type-resolved transcriptomics to pinpoint vulnerable ensembles and human neuroimaging to map circuit biomarkers and validate translational relevance.

In conclusion, our findings delineate how distinct BLA-recipient circuits govern specific depressive symptoms through divergent molecular pathways. By moving beyond a monolithic view of depression, these findings provide a refined neurobiological framework for understanding symptom-specific vulnerabilities and highlight circuit- and pathway-level heterogeneity as a core feature of MDD. Beyond this mechanistic insight, our work also provides a roadmap for precision interventions, where therapeutic strategies can be tailored to specific circuits and signaling pathways. By linking animal models to molecular and circuit biomarkers and by integrating future single-cell transcriptomics and human neuroimaging, these findings set the stage for translational advances by which symptom-specific interventions can reshape antidepressant development and clinical practice.

Funding and Support

This research was supported by the grants of the National Research Foundation (2021R1A2C2094627 and RS-2024-000460958) and Korea Brain Research Institute research program (26-BR-05-02), and received APC from Daegu Catholic University School of Medicine. We thank Prof. Shinwoo Kang at Soonchunhyang University for technical support.

CRediT Authorship Contribution Statement

Tae-Eun Kim: Writing---review and editing, writing---original draft, visualization, resources, methodology, investigation, formal analysis, and data curation. Tae-Yong Choi: Writing---review and editing, writing---original draft, validation. Ja Wook Koo: Writing---review and editing, writing---original draft, validation, supervision, investigation, and funding acquisition. Jeongseop Kim: Writing---review and editing, writing--- original draft, visualization, validation, supervision, resources, project administration, methodology, investigation, funding acquisition, formal analysis, data curation, and conceptualization.

Declaration of Competing Interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Footnotes

Appendix A

Supplemental material associated with this article can be found online at doi:10.1016/j.mocell.2026.100320.

Contributor Information

Ja Wook Koo, Email: jawook.koo@kbri.re.kr.

Jeongseop Kim, Email: jeongseop@cu.ac.kr.

Appendix A. Supplementary material

Supplementary material

mmc1.pptx (3.2MB, pptx)

.

References

  1. Alboni S., Tascedda F., Corsini D., Benatti C., Caggia F., Capone G., Barden N., Blom J.M., Brunello N. Stress induces altered CRE/CREB pathway activity and BDNF expression in the hippocampus of glucocorticoid receptor-impaired mice. Neuropharmacology. 2011;60:1337–1346. doi: 10.1016/j.neuropharm.2011.01.050. [DOI] [PubMed] [Google Scholar]
  2. Aldao A., Nolen-Hoeksema S., Schweizer S. Emotion-regulation strategies across psychopathology: a meta-analytic review. Clin. Psychol. Rev. 2010;30:217–237. doi: 10.1016/j.cpr.2009.11.004. [DOI] [PubMed] [Google Scholar]
  3. Anastasiades P.G., Carter A.G. Circuit organization of the rodent medial prefrontal cortex. Trends Neurosci. 2021;44:550–563. doi: 10.1016/j.tins.2021.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Anderson K.M., Collins M.A., Kong R., Fang K., Li J., He T., Chekroud A.M., Yeo B.T.T., Holmes A.J. Convergent molecular, cellular, and cortical neuroimaging signatures of major depressive disorder. Proc. Natl. Acad. Sci. U. S. A. 2020;117:25138–25149. doi: 10.1073/pnas.2008004117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bogdanova O.V., Kanekar S., D'Anci K.E., Renshaw P.F. Factors influencing behavior in the forced swim test. Physiol. Behav. 2013;118:227–239. doi: 10.1016/j.physbeh.2013.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Carver C.S., Scheier M.F., Weintraub J.K. Assessing coping strategies: a theoretically based approach. J. Pers. Soc. Psychol. 1989;56:267–283. doi: 10.1037//0022-3514.56.2.267. [DOI] [PubMed] [Google Scholar]
  7. Casarotto P., Umemori J., Castren E. BDNF receptor TrkB as the mediator of the antidepressant drug action. Front. Mol. Neurosci. 2022;15 doi: 10.3389/fnmol.2022.1032224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cleva R.M., Olive M.F. Positive allosteric modulators of type 5 metabotropic glutamate receptors (mGluR5) and their therapeutic potential for the treatment of CNS disorders. Molecules. 2011;16:2097–2106. doi: 10.3390/molecules16032097. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Commons K.G., Cholanians A.B., Babb J.A., Ehlinger D.G. The Rodent forced swim test measures stress-coping strategy, not depression-like behavior. ACS Chem. Neurosci. 2017;8:955–960. doi: 10.1021/acschemneuro.7b00042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. de Kloet E.R., Molendijk M.L. Coping with the forced swim stressor: towards understanding an adaptive mechanism. Neural Plast. 2016;2016 doi: 10.1155/2016/6503162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. De Pablo J.M., Parra A., Segovia S., Guillamon A. Learned immobility explains the behavior of rats in the forced swimming test. Physiol. Behav. 1989;46:229–237. doi: 10.1016/0031-9384(89)90261-8. [DOI] [PubMed] [Google Scholar]
  12. Deschwanden A., Karolewicz B., Feyissa A.M., Treyer V., Ametamey S.M., Johayem A., Burger C., Auberson Y.P., Sovago J., Stockmeier C.A., et al. Reduced metabotropic glutamate receptor 5 density in major depression determined by [(11)C]ABP688 PET and postmortem study. Am. J. Psychiatry. 2011;168:727–734. doi: 10.1176/appi.ajp.2011.09111607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Duman R.S., Deyama S., Fogaca M.V. Role of BDNF in the pathophysiology and treatment of depression: activity-dependent effects distinguish rapid-acting antidepressants. Eur. J. Neurosci. 2021;53:126–139. doi: 10.1111/ejn.14630. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fan X., Wang H., Lv X., Wang Q., Yu B., Li X., Li L., Zhang Y., Ma N., Lu Q., et al. The pCREB/BDNF pathway in the hippocampus is involved in the therapeutic effect of selective 5-HT reuptake inhibitors in adult male rats exposed to blast traumatic brain injury. Brain Sci. 2025;15:236. doi: 10.3390/brainsci15030236. PMID: 40149758; PMCID: PMC11940387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Felix-Ortiz A.C., Beyeler A., Seo C., Leppla C.A., Wildes C.P., Tye K.M. BLA to vHPC inputs modulate anxiety-related behaviors. Neuron. 2013;79:658–664. doi: 10.1016/j.neuron.2013.06.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Finkbeiner S., Tavazoie S.F., Maloratsky A., Jacobs K.M., Harris K.M., Greenberg M.E. CREB: a major mediator of neuronal neurotrophin responses. Neuron. 1997;19:1031–1047. doi: 10.1016/s0896-6273(00)80395-5. [DOI] [PubMed] [Google Scholar]
  17. Fitzsimmons S., Oostra E., Postma T.S., van der Werf Y.D., van den Heuvel O.A. Repetitive transcranial magnetic stimulation-induced neuroplasticity and the treatment of psychiatric disorders: state of the evidence and future opportunities. Biol. Psychiatry. 2024;95:592–600. doi: 10.1016/j.biopsych.2023.11.016. [DOI] [PubMed] [Google Scholar]
  18. Hughes Z.A., Neal S.J., Smith D.L., Sukoff Rizzo S.J., Pulicicchio C.M., Lotarski S., Lu S., Dwyer J.M., Brennan J., Olsen M., et al. Negative allosteric modulation of metabotropic glutamate receptor 5 results in broad spectrum activity relevant to treatment resistant depression. Neuropharmacology. 2013;66:202–214. doi: 10.1016/j.neuropharm.2012.04.007. [DOI] [PubMed] [Google Scholar]
  19. Jefferys D., Funder J. The effect of water temperature on immobility in the forced swimming test in rats. Eur. J. Pharmacol. 1994;253:91–94. doi: 10.1016/0014-2999(94)90761-7. [DOI] [PubMed] [Google Scholar]
  20. Jong Y.J., Kumar V., O'Malley K.L. Intracellular metabotropic glutamate receptor 5 (mGluR5) activates signaling cascades distinct from cell surface counterparts. J. Biol. Chem. 2009;284:35827–35838. doi: 10.1074/jbc.M109.046276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kim J., Kang S., Choi T.Y., Chang K.A., Koo J.W. Metabotropic glutamate receptor 5 in amygdala target neurons regulates susceptibility to chronic social stress. Biol. Psychiatry. 2022;92:104–115. doi: 10.1016/j.biopsych.2022.01.006. [DOI] [PubMed] [Google Scholar]
  22. Kim J., Lee J., Song H., Koo J.W., Kang S. mGluR5 as a potential orchestrator of astrocyte interactions in neurological disorders. Neural Plast. 2025;2025 doi: 10.1155/np/7259018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kim J., Seol S., Kim T.E., Lee J., Koo J.W., Kang H.J. Synaptotagmin-4 induces anhedonic responses to chronic stress via BDNF signaling in the medial prefrontal cortex. Exp. Mol. Med. 2024;56:329–343. doi: 10.1038/s12276-024-01156-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Koo J.W., Mazei-Robison M.S., LaPlant Q., Egervari G., Braunscheidel K.M., Adank D.N., Ferguson D., Feng J., Sun H., Scobie K.N., et al. Epigenetic basis of opiate suppression of Bdnf gene expression in the ventral tegmental area. Nat. Neurosci. 2015;18:415–422. doi: 10.1038/nn.3932. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lazarus R.S. Psychological stress and the coping process. McGraw-Hill; New York, NY: 1966. [Google Scholar]
  26. Lazarus R.S. Stress and emotion: a new synthesis. Springer publishing company; Princeton, NJ: 2006. [Google Scholar]
  27. Lazarus R.S., Folkman S. Stress, appraisal, and coping. Springer publishing company; Princeton, NJ: 1984. [Google Scholar]
  28. Lehtinen K., Nokia M.S., Takala H. Red light optogenetics in neuroscience. Front. Cell. Neurosci. 2021;15 doi: 10.3389/fncel.2021.778900. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Li X., Need A.B., Baez M., Witkin J.M. Metabotropic glutamate 5 receptor antagonism is associated with antidepressant-like effects in mice. J. Pharmacol. Exp. Ther. 2006;319:254–259. doi: 10.1124/jpet.106.103143. [DOI] [PubMed] [Google Scholar]
  30. Lv Q.Y., Chen M.M., Li Y., Yu Y., Liao H. Brain circuit dysfunction in specific symptoms of depression. Eur. J. Neurosci. 2022;55:2393–2403. doi: 10.1111/ejn.15221. [DOI] [PubMed] [Google Scholar]
  31. Malyshev A., Goz R., LoTurco J.J., Volgushev M. Advantages and limitations of the use of optogenetic approach in studying fast-scale spike encoding. PLoS One. 2015;10 doi: 10.1371/journal.pone.0122286. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mao L., Yang L., Tang Q., Samdani S., Zhang G., Wang J.Q. The scaffold protein Homer1b/c links metabotropic glutamate receptor 5 to extracellular signal-regulated protein kinase cascades in neurons. J. Neurosci. 2005;25:2741–2752. doi: 10.1523/JNEUROSCI.4360-04.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Miao Z., Wang Y., Sun Z. The relationships between stress, mental disorders, and epigenetic regulation of BDNF. Int. J. Mol. Sci. 2020;21:1375. doi: 10.3390/ijms21041375. PMID: 32085670; PMCID: PMC7073021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nestler E.J., Hyman S.E. Animal models of neuropsychiatric disorders. Nat. Neurosci. 2010;13:1161–1169. doi: 10.1038/nn.2647. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Porsolt R.D., Bertin A., Jalfre M. Behavioral despair in mice: a primary screening test for antidepressants. Arch. Int. Pharmacodyn. Ther. 1977;229:327–336. [PubMed] [Google Scholar]
  36. Porsolt R.D., Anton G., Blavet N., Jalfre M. Behavioural despair in rats: a new model sensitive to antidepressant treatments. Eur. J. Pharmacol. 1978;47:379–391. doi: 10.1016/0014-2999(78)90118-8. [DOI] [PubMed] [Google Scholar]
  37. Porsolt R.D., Bertin A., Jalfre M. "Behavioural despair" in rats and mice: strain differences and the effects of imipramine. Eur. J. Pharmacol. 1978;51:291–294. doi: 10.1016/0014-2999(78)90414-4. [DOI] [PubMed] [Google Scholar]
  38. Quiroz J.A., Tamburri P., Deptula D., Banken L., Beyer U., Rabbia M., Parkar N., Fontoura P., Santarelli L. Efficacy and safety of basimglurant as adjunctive therapy for major depression: a randomized clinical trial. JAMA Psychiatry. 2016;73:675–684. doi: 10.1001/jamapsychiatry.2016.0838. [DOI] [PubMed] [Google Scholar]
  39. Reshetnikov V.V., Bondar N.P. The role of stress-induced changes of Homer1 expression in stress susceptibility. Biochemistry. 2021;86:613–626. doi: 10.1134/S0006297921060018. [DOI] [PubMed] [Google Scholar]
  40. Slattery D.A., Cryan J.F. Using the rat forced swim test to assess antidepressant-like activity in rodents. Nat. Protoc. 2012;7:1009–1014. doi: 10.1038/nprot.2012.044. [DOI] [PubMed] [Google Scholar]
  41. Tsankova N.M., Berton O., Renthal W., Kumar A., Neve R.L., Nestler E.J. Sustained hippocampal chromatin regulation in a mouse model of depression and antidepressant action. Nat. Neurosci. 2006;9:519–525. doi: 10.1038/nn1659. [DOI] [PubMed] [Google Scholar]
  42. Wagner K.V., Hartmann J., Mangold K., Wang X.D., Labermaier C., Liebl C., Wolf M., Gassen N.C., Holsboer F., Rein T., et al. Homer1 mediates acute stress-induced cognitive deficits in the dorsal hippocampus. J. Neurosci. 2013;33:3857–3864. doi: 10.1523/JNEUROSCI.4333-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Wang J.Q., Fibuch E.E., Mao L. Regulation of mitogen-activated protein kinases by glutamate receptors. J. Neurochem. 2007;100:1–11. doi: 10.1111/j.1471-4159.2006.04208.x. [DOI] [PubMed] [Google Scholar]
  44. Wang J.Q., Mao L. The ERK pathway: molecular mechanisms and treatment of depression. Mol. Neurobiol. 2019;56:6197–6205. doi: 10.1007/s12035-019-1524-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wendelmuth M., Willam M., Todorov H., Radyushkin K., Gerber S., Schweiger S. Dynamic longitudinal behavior in animals exposed to chronic social defeat stress. PLoS One. 2020;15 doi: 10.1371/journal.pone.0235268. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. West A.P. Neurobehavioral studies of forced swimming: the role of learning and memory in the forced swim test. Prog Neuropsychopharmacol. Biol. Psychiatry. 1990;14:863–877. doi: 10.1016/0278-5846(90)90073-p. [DOI] [PubMed] [Google Scholar]
  47. Yankelevitch-Yahav R., Franko M., Huly A., Doron R. The forced swim test as a model of depressive-like behavior. J. Vis. Exp. 2015;2:52587. doi: 10.3791/52587. PMID: 25867960; PMCID: PMC4401172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zhang J.Y., Liu T.H., He Y., Pan H.Q., Zhang W.H., Yin X.P., Tian X.L., Li B.M., Wang X.D., Holmes A., et al. Chronic stress remodels synapses in an amygdala circuit-specific manner. Biol. Psychiatry. 2019;85:189–201. doi: 10.1016/j.biopsych.2018.06.019. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material

mmc1.pptx (3.2MB, pptx)

Articles from Molecules and Cells are provided here courtesy of Korean Society for Molecular and Cellular Biology

RESOURCES