NgR-Dependent plasticity in a dCA3^CaMKIIα → VTA^DA → NAc circuit underlies stress resilience and susceptibility

Abstract

The Nogo receptor (NgR), a core receptor for myelin-derived axon growth inhibitors, has been implicated not only in the regulation of synaptic plasticity but also in the pathogenesis of various neuropsychiatric disorders. However, its specific role and the underlying circuit mechanisms in the development of depression remain elusive. This study used a chronic unpredictable mild stress (CUMS) model in male mice and found that the expression of the NgR in the dorsal CA3 (dCA3) region of the hippocampus was significantly increased in depressive-like mice, and it was mainly enriched in glutamatergic neurons with reduced excitability. Downregulation of NgR in dCA3 or enhancement of glutamatergic neuronal excitability significantly alleviated depressive-like behaviors, while upregulation of NgR or inhibition of glutamatergic neuronal excitability promoted stress-induced depressive-like behaviors. Notably, dCA3 glutamatergic neurons specifically project to the dopaminergic neuronal pathway in the ventral tegmental area (VTA) and its downstream dopaminergic projections to the nucleus accumbens (NAc). The inhibition of D1/D2 receptors in the NAc is required for mediating depressive-like behaviors. Furthermore, downregulation of NgR in dCA3 exerts antidepressant-like effects by disinhibiting the neurons projecting to the VTA → NAc pathway. This study systematically reveals the role of NgR in dCA3 glutamatergic neurons and the dCA3^CaMKIIα → VTA^DA → NAc circuit in depressive-like behaviors, providing a new theoretical basis for developing potential therapeutic strategies targeting myelin-related signaling pathways for depression.

Introduction

Depression is a highly heterogeneous mental disorder that affects over 300 million people worldwide [1]. It is characterized by core symptoms including lack of motivation, anhedonia, and social withdrawal [2]. In the complex etiology of depression, chronic stress is recognized as a major risk factor [3, 4]. Evidence suggests that stress induces dysfunctions at cellular, molecular, and neural circuit levels, contributing to the pathophysiology of depression [5, 6]. Targeting stress-mediated mechanisms has thus emerged as a promising therapeutic strategy for the treatment of depressive disorders.

The Nogo receptor (NgR), a high-affinity common receptor for myelin-derived axonal growth inhibitors (MAIs), plays a pivotal role in regulating growth cone dynamics and suppressing neurite outgrowth [7, 8]. Over the past two decades, research in this field has primarily focused on the inhibitory function of the Nogo-A/NgR pathway in limiting axonal regeneration following central nervous system injury, along with related interventional strategies [9]. However, recent advances have unveiled non-canonical functions of NgR in neurodevelopment, synaptic plasticity, and the regulation of neuropsychiatric disorders, broadening our understanding of its multidimensional physiological and pathological roles [10–14]. Accumulating evidence suggests that NgR may serve as a genetic susceptibility factor for neurodevelopmental disorders such as schizophrenia, with relevant phenotypes observed in animal models [13]. In addiction research, the dynamic regulation of NgR expression by psychoactive substances offers a novel molecular perspective for understanding drug-induced neuroplasticity [14]. Furthermore, NgR signaling contributes to epileptogenesis by modulating neuronal excitability thresholds, underscoring its critical role in maintaining nervous system homeostasis [15]. Nevertheless, whether the NgR signaling pathway is involved in regulating depression-related behavioral phenotypes remains a critical and unresolved scientific question in the field of neuropsychiatric research.

In the mammalian central nervous system, the fine-tuned regulation of emotional stress responses relies on synergistic interactions among multiple brain regions, including the medial prefrontal cortex, amygdala, hippocampus, ventral tegmental area (VTA), nucleus accumbens (NAc), and striatum [16, 17]. Prolonged abnormal stress load may exacerbate atrophy of dorsal CA3 (dCA3) pyramidal neurons and hippocampal volume reduction, thereby playing a critical role in the mechanisms underlying stress-related cognitive impairment and emotional pathology [18, 19]. Resting-state functional magnetic resonance imaging (RS-fMRI) studies have revealed enhanced functional coupling within the VTA-hippocampus circuit in individuals with psychiatric disorders [20], which plays a pivotal role in integrating reward processing and stress responses [21–23]. Furthermore, the dopamine (DA) neurons’ projection from the VTA to γ-aminobutyric acid (GABA) neurons in the NAc has been established as a neural substrate for reward-mediated anxiolysis, underscoring its importance in emotional homeostasis [6, 24]. Based on this evidence, we hypothesize that NgR, specifically expressed in the dCA3 region, may function as a downstream target of myelin-associated genes and contribute to the pathogenesis of depression by regulating neural circuit function.

To test this hypothesis, we established CUMS model of depression and employed a multidisciplinary approach combining optogenetic and chemogenetic manipulations, viral-mediated conditional knockdown and overexpression of NgR, comprehensive behavioral analyses, electrophysiological recordings, and molecular assays. Our results demonstrate that NgR deletion in the dCA3 region significantly enhances behavioral adaptability to acute stress, reduces passive coping behaviors, and elicits sustained antidepressant-like effects. Mechanistically, these behavioral phenotypes depend on the regulatory function of the dCA3^CaMKIIα → VTA^DA → NAc neural circuit.

Materials and methods

Animals

Adult male C57BL/6J mice aged 7–8 weeks were maintained under standard housing conditions with controlled temperature (24–26 °C) and relative humidity (40–50%), on a 12/12-hour light-dark cycle. Animals were housed in groups of 4–5 per cage and provided with free access to rodent chow and drinking water.

Ethics statement

All animal studies and experimental procedures were approved by the Animal Care and Use Committee of Shandong Medical and Pharmaceutical University (2023-065). Written informed consent was not required, as no human participant was involved in this study. All methods were performed in accordance with the relevant guidelines and regulations.

CUMS and subthreshold unpredictable mild stress (SUMS)

Over 10-day (CUMS) or 3-day (SUMS) periods, mice received two randomly assigned mild stressors daily. The regimen comprised: physical confinement (2 h); tail suspension (15 min); continuous illumination (24 h); 45° cage tilt (24 h); damp bedding (24 h); elevated platform exposure (30 min); foot shock (15 min); and social isolation (for the entire 10-day or 3-day experimental period).

Behavioral testing

Social interaction in mice was assessed through the social interaction test (SIT). Depression was evaluated via the forced swim test (FST) and tail suspension test (TST). Anxiety-like behavior was assessed using the elevated plus maze (EPM) and open field test (OFT). Detailed methodologies are provided in the Supplementary Information.

Stereotaxic surgery and virus injection

Stereotaxic surgery for viral vector delivery was performed following an established protocol. Mice were anesthetized with sodium pentobarbital (50 mg kg⁻¹, i.p) and firmly position them in a stereotaxic frame (RWD, China). Burr holes were carefully drilled above the target regions using a fine cranial drill. Stereotaxic coordinates relative to bregma were determined as follows: dCA3 (anteroposterior [AP]: –1.94 mm; mediolateral [ML]: ±2.48 mm; dorsoventral [DV]: –2.24 mm), VTA (AP: –3.2 mm; ML: ±0.5 mm; DV: –4.3 mm), and NAc (AP: +1.2 mm; ML: ±0.7 mm; DV: –4.3 mm). The virus was injected at a rate of 100 nl min⁻¹ using a microinjection pump (WPI, USA). To minimize backflow and ensure sufficient viral diffusion, the needle was left in place for 10 min before being removed. Detailed information and principles regarding the viruses used in this study is provided in the Supplementary Information.

Cannula implantation and intracranial microinfusion

Anesthetize mice using sodium pentobarbital and firmly position them in a stereotaxic frame. Exposing the skull, and burr holes were drilled using a cranial drill. A cannula was inserted into the target brain region, and two anchor screws were inserted into the skull. Dental cement was used to fix the trocar and screws in place. To prevent occlusion, a catheter-compatible obturator was inserted into the cannula. The cannula was then connected to a syringe for drug delivery. For intra-NAc drug infusions, SCH-23390 (0.5 μg side⁻¹ in 0.25 μL, MCE) or Sulpiride (0.5 μg side⁻¹ in 0.25 μL, MCE) was administered into the NAc using a syringe pump at an infusion rate of 0.2 μL min⁻¹.

Optogenetic and chemogenetic

For neuronal activation, blue light (465 nm, 2 mW) was delivered as 15-ms pulses at 20 Hz; for neuronal inhibition, continuous yellow light (589 nm, 5 mW) was applied. All light delivery parameters were controlled via Transistor-Transistor Logic signals generated by the system’s integrated software. For chemogenetic activation or inhibition, clozapine N-oxide (CNO) was injected 30 min before behavioral testing. Detailed methodologies are provided in the Supplementary Information.

Golgi staining

Neuronal morphology and dendritic spine density in the dCA3 region were analyzed using the FD Rapid GolgiStain Kit (FD NeuroTechnologies), with sectioning performed on a Leica VT1000S vibrating microtome and imaging conducted using a Pannoramic MIDI Scanner (3DHistech). Detailed methodologies are provided in the Supplementary Information.

Fiber photometry recordings and analysis

Fiber photometry recordings were performed using a tricolor, single-channel fiber photometry system. Prior to formal experiments, optical fibers were photobleached by exposure to intense illumination for approximately 30 min to stabilize the signal baseline. During recording sessions, the 405 nm wavelength was used as an isosbestic control channel, while the 470 nm wavelength served as the excitation channel for GCaMP6s fluorescence detection. The output power was set to 2–3 μW for the 405 nm channel and 20 mW for the 470 nm channel. Calcium transients were recorded throughout the TST to capture neural dynamics during struggle episodes. Acquired signals were processed and analyzed in MATLAB (R2021a, MathWorks) and visualized as z-scored heatmaps to represent population-level activity.

Immunofluorescence

Brain sections (15 μm) were cut using a cryostat, air-dried for 1 h, and washed three times (10 min each) in PBST. After blocking with 5% normal goat serum in PBST for 90 min at room temperature, the sections were incubated overnight at 4 °C with primary antibodies diluted in PBST containing 2% normal goat serum. The following primary antibodies were used: rabbit anti-NgR (Merck, #AB15138; 1:100), mouse anti-GAD67 (Abcam, #ab26116; 1:100), mouse anti-CaMKIIα (CST, #50049; 1:200), mouse anti-TH (Merck, #MAB318; 1:200), mouse anti-GFAP (CST, #3670; 1:200), and mouse anti-IBA1 (CST, #58410; 1:200). The sections were then incubated with Alexa Fluor 488-, 594-, or 647-conjugated goat anti-rabbit or anti-mouse IgG (1:200; Invitrogen, Carlsbad, CA, USA) for 2 h at room temperature. Nuclei were counterstained with DAPI (1 μg mL⁻¹), and sections were mounted with antifade reagent (Solarbio).

In vitro electrophysiological recordings

Neuronal excitability was examined by recording miniature excitatory postsynaptic currents (mEPSCs) from VTA, dCA3 and dorsal CA1 (dCA1) neurons. Detailed methodologies are provided in the Supplementary Information.

Western-blot

Protein expression of NgR in the dCA3 region was analyzed by western blot using specific antibodies, with protein bands visualized by Tanon Image Analysis Software. Detailed methodologies are provided in the Supplementary Information.

Statistical analysis

Data with normal distribution and equal variance were presented as mean ± SEM. Graphical representations were prepared with GraphPad Prism 8.0, while statistical evaluations were carried out using SPSS 20.0. The comparison between two groups was performed using unpaired/paired Student’s t-tests. For comparisons among more than two groups, one-way analysis of variance was employed, followed by Tukey’s post hoc test for multiple comparisons where applicable. A probability threshold of p < 0.05 defined statistical significance.

Results

Chronic stress upregulates NgR expression in glutamatergic neurons of the dCA3

To investigate the effects of chronic stress on the ventral hippocampus, we adopted the classic CUMS paradigm. This model reliably induces depressive-like behaviors and associated neurobiological abnormalities, serving as a key experimental paradigm for elucidating the pathological mechanisms of depression [25]. Following 10 days of chronic stress, mice exhibited significant behavioral deficits, as evidenced by reduced interaction time in the SIT (Fig. 1A), decreased center time in the OFT (Fig. 1B, C), prolonged time in the closed arms of the EPM (Fig. 1D), increased immobility time in both the TST (Fig. 1E) and FST (Fig. 1F). Body weight measurements indicated a significant reduction in CUMS mice compared to the control group (Fig. 1G), further supporting the presence of stress-induced physiological and behavioral changes. Given that myelin-derived signaling mechanisms play a crucial regulatory role in the neurobiology of stress [26], we examined the expression of the NgR, a coreceptor for myelin proteins, following exposure to chronic stress. The results showed that NgR expression did not change significantly in brain regions such as the prefrontal cortex, VTA, and dCA1 following CUMS treatment (Supplementary Fig. 1). However, both immunofluorescence and Western Blot analyses revealed that stress significantly upregulated NgR protein levels in the mouse dCA3 region (Fig. 1H–J). Correlation analysis showed that NgR levels in the dCA3 were negatively correlated with body weight, and positively correlated with the degree of behavioral despair (immobility time in the FST/TST) (Fig. 1K–M).

Fig. 1. Chronic stress increases dCA3 glutamatergic neurons NgR expression.

A-G Animal behaviors in SIT, OFT, EPM, TST, FST, and weight (SIT, t = 2.800, p = 0.0135, n = 7 in the Control group, n = 10 in the CUMS group; OFT, t = 2.811, p = 0.0012, n = 7 in the Control group, n = 14 in the CUMS group; OFT total diatance, t = 1.745, p = 0.0980, n = 8 in the Control group, n = 12 in the CUMS group; EPM, t = 3.622, p = 0.0019, n = 7 in the Control group, n = 13 in the CUMS group; TST, t = 7.182, p < 0.001, n = 7 per group; FST, t = 2.915, p = 0.0113, n = 7 in the Control group, n = 9 in the CUMS group; weight, t = 2.179, p = 0.0379, n = 15 per group). H Mean fluorescence intensity in dCA3 of control and CUMS mice (t = 2.689, p = 0.0361, n = 4 per group). I Representative images of NgR in dCA3. Scale bars: 40 μm. J Representative Western blot micrographs and densitometry analysis of NgR (t = 5.664, p < 0.001, n = 6 per group). K Negative correlation of NgR level in dCA3 with weight (R² = 0.5960, p = 0.0054, n = 11). L-M Positive correlation of NgR level in dCA3 with immobility time in TST and FST (TST, R² = 0.4078, p = 0.0469, n = 10, FST, R² = 0.4907, p = 0.0241, n = 10). N Representative images showing NgR colocalization of CaMKIIα or GAD67. Scale bars: 40 μm. O Quantification of NgR cells immunoreactive to CaMKIIα or GAD67 (t = 37.19, p < 0.001, n = 4 per group). Data are expressed as mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, by a unpaired Student’s t-tests (A-H, J, O), a correlation analysis (K-M). White arrows indicate cells co-expressing NgR and CaMKIIα or GAD67.

To delineate NgR’s cellular specificity in dCA3, we analyzed hippocampal sections and found NgR co-localized with both glutamatergic and GABAergic neurons, but predominantly enriched in CaMKIIα-positive populations (64.19% ± 1.28%) versus minimal GABAergic presence (9.47% ± 0.73%). Additionally, we further examined NgR expression in glial cells and found that its colocalization with the astrocyte marker GFAP and the microglial marker IBA1 was relatively low (Supplementary Fig. 2). These results suggest that NgR is primarily expressed in glutamatergic neurons within the dCA3 region, while its expression in glial cells is relatively low. This further supports the critical role of NgR on glutamatergic neurons in the pathological mechanisms of depression (Fig. 1N–O).

dCA3 glutamatergic neurons are involved in the regulation of depressive-like behavior

Although extensive studies have documented the sensitivity of hippocampal neurogenesis to stress and emotional stimuli, neuronal excitability in this region has received considerably less attention. Therefore, we investigated the contribution of dCA3 neuronal excitability to depressive-like behaviors. We co-injected AAV9-DIO-GCaMP6s-mCherry and AAV9-CaMKIIα-Cre viruses into the dCA3 region of mice and implanted an optical fiber cannula at the same site (Fig. 2A). In the TST assay, the calcium activity of dCA3 neurons in CUMS-induced mice was significantly attenuated compared with that in the control group (Fig. 2B–D). Whole-cell recordings also showed a significant reduction in the amplitude of mEPSCs, while frequency remained unchanged, indicating a reduction in postsynaptic efficacy (Fig. 2E–G). These results indicate that the robust activation of dCA3 glutamatergic neurons under stress may contribute to neural regulation by modulating postsynaptic mechanisms, further suggesting a critical role of this neuronal population in the pathogenesis of depression-related states.

Fig. 2. dCA3 glutamatergic neurons are involved in the modulation of depressive-like behavior.

A Representative image showing GCaMP6s expression in dCA3 glutamatergic neurons. Scale bars: 200 μm. B Heatmaps demonstrating Ca²⁺ response during TST, time point zero is defined as the resting state. C Peri-event plot of Ca²⁺ response. D Quantification of area under the curve (AUC, 0–4 s) of fluorescence changes after the resting state during TST (t = 3.115, p = 0.0264, n = 4 in Control, n = 3 in CUMS). E Representative traces of mEPSCs in glutamatergic neurons. F Cumulative probability plots for the amplitude in glutamatergic neurons (t = 2.682, p = 0.0315, n = 4 in the Control group, n = 5 in the CUMS group). G Cumulative probability plots for the frequency in glutamatergic neurons (t = 0.5598, p = 0.5931, n = 4 in the Control group, n = 5 in the CUMS group). H Left: schematic of viral injection. Right: expression of hM4Di in dCA3 glutamatergic neurons. Scale bars: 200 μm. I Top: schematic of behavioral tests. Bottom: quantification of membrane action potentials before and after perfusion of CNO (120 pA, t = 6.708, p < 0.001, 140 pA, t = 2.445, p = 0.0292; 160 pA, t = 2.739, p = 0.0204; 180 pA, t = 2.150, p = 0.0421, n = 6 per group). J-O Behavioral results from SIT, OFT, EPM, TST, and FST following neuronal inhibition (SIT, t = 5.745, p < 0.001, n = 10 per group; OFT, t = 1.775, p = 0.0464, n = 10 per group; OFT total distance, t = 0.7385, p = 0.4697, n = 10 per group; EPM, t = 3.873, p = 0.0017, n = 10 per group; TST, t = 1.962, p = 0.0327, n = 10 per group; FST, t = 3.271, p = 0.0056, n = 8 per group). P Left: Schematic of viral injection. Right: representative image showing hM3Dq expression in dCA3 glutamatergic neurons. Scale bars: 200 μm. Q Top: schematic diagram of depression-and anxiety-related behavioral tests. Bottom: quantification of membrane action potentials before and after perfusion of CNO (120 pA, t = 2.303, p = 0.0274, 140 pA, t = 2.683, p = 0.0314; 160 pA, t = 2.205, p = 0.0316; 180 pA, t = 2.091, p = 0.0407, n = 6 per group). R-W Animal behaviors in SIT, OFT, EPM, TST, and FST (SIT, t = 3.269, p = 0.0067, n = 7 per group; OFT, t = 2.495, p = 0.0282, n = 7 per group; OFT total diatance, t = 1.958, p = 0.0739, n = 7 per group; EPM, t = 0.4480, p = 0.6621, n = 7 per group; TST,t = 2.892, p = 0.0135, n = 7 per group; FST, t = 3.799, p = 0.0025, n = 7 in per group). Data are expressed as mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, by a paired Student’s t-tests (I, Q), an unpaired Student’s t-tests (D, F, G, J-O, R-W).

To determine whether dCA3 glutamatergic neurons regulate depression-like behaviors, we microinjected a virus encoding (a designed Gi-coupled human muscarinic receptor) along with CaMKIIα-Cre into the dCA3 region of mice to achieve chemogenetic inhibition of this neuronal population (Fig. 2H, I). The SUMS model, which typically does not induce the pronounced anhedonia or classic behavioral despair seen in the conventional CUMS model, is particularly suited for probing the mechanisms underlying the development of disease susceptibility [27]. In the SUMS model, following suppression of dCA3 glutamatergic neuronal activity by CNO, mice exhibited reduced social interaction time, decreased time spent in the center zone of the OFT, diminished open arm exploration in the EPM, and significantly prolonged immobility in both the TST and FST (Fig. 2J–O). To elucidate the regulatory role of these neurons, we further employed chemogenetic approaches to activate dCA3 glutamatergic neurons and examined their impact on depression-like behaviors (Fig. 2P, Q). The experiments confirmed that activation of dCA3 glutamatergic neurons in the CUMS model significantly increased social interaction time, prolonged duration in the central zone, and decreased immobility time in both the TST and FST, while no statistically significant difference was observed in the EPM (Fig. 2R–W). Electrophysiological and calcium imaging evidence indicates reduced excitability of CA3 neurons in CUMS mice, while behavioral findings further demonstrate that chronic stress-induced alterations in dCA3 glutamatergic neuronal activity directly regulate depressive-like behaviors. Notably, there was no statistically significant difference in the total distance traveled in the OFT among the groups (Fig. 2L, T), which rules out the confounding effect of locomotor activity changes on the assessment of depressive-like behaviors and confirms that dCA3 neuronal excitability specifically influences depressive phenotypes.

Modulation of depressive-like behavior by dCA3 glutamatergic NgR

To investigate the role of NgR in glutamatergic neurons in regulating depressive-like behaviors, AAV9-CaMKIIα-shNgR-GFP or AAV9-CaMKIIα-scramble-GFP was bilaterally injected into the dCA3 region of mice to specifically suppress NgR expression in dCA3 glutamatergic neurons (Fig. 3A). GFP fluorescence confirmed accurate targeting of the injection site (Fig. 3A). Western blot demonstrated efficient knockdown of NgR in dCA3 (Fig. 3B). Whole-cell recordings revealed that genetic knockdown of NgR in dCA3 glutamatergic neurons reversed the reduction in mEPSCs amplitude, while leaving the frequency unchanged, suggesting that NgR knockdown rescues CUMS-induced decrease in neuronal excitability (Fig. 3C–E). Behavioral tests demonstrated that knockdown of the NgR gene in dCA3 glutamatergic neurons ameliorated depression-like behaviors in CUMS-exposed mice (Fig. 3F–J). These findings indicate that downregulation of NgR in dCA3 glutamatergic neurons alleviates stress-induced depression-like phenotypes, which may be associated with enhanced neuronal excitability.

Fig. 3. Knockdown or overexpression of NgR in dCA3 glutamatergic neurons modulates depressive-like behavior.

A Top: Behavioral testing scheme. Bottom: Targeted expression of NgR-shRNA in the dCA3. Scale bars: 200 μm. B Western blot and quantification of NgR protein levels (t = 5.095, p = 0.0022, n = 4 per group). C Example mEPSCs traces from glutamatergic neurons. D Cumulative probability plots for the amplitude in glutamatergic neurons (F(2, 9) = 26.97, p < 0.001, n = 4 cells from 4 mice per group). E Cumulative probability plots for the frequency in glutamatergic neurons (F(2, 11) = 0.9709, p = 0.4090, n = 4 cells from 4 mice in Control-scramble and CUMS-shNgR, n = 6 cells from 4 mice in CUMS-scramble). F-J Behavioral profiles in SIT, OFT, EPM, TST, and FST (SIT, F(2, 18) = 9.459, p = 0.0016, n = 7 per group; OFT, F(2, 18) = 23.97, p < 0.001,n = 7 per group; EPM, F(2, 21) = 9.753, p = 0.0010, n = 7–9 per group; TST, F(2, 21) = 10.45, p < 0.001, n = 7–9 per group; FST, F(2, 19) = 6.252, p = 0.0082, n = 7–9 per group). K Top: behavioral paradigm. Bottom: Targeted expression of OE-NgR in the dCA3. Scale bars: 200 μm. L Representative Western blot micrographs and densitometry analysis of protein levels of NgR (t = 7.372, p < 0.001, n = 5 in SUMS-vector, n = 4 in SUMS-OE NgR). M Sample mEPSCs recordings. N Cumulative probability plots for the amplitude in glutamatergic neurons (t = 5.245, p = 0.0012, n = 4 cells from 4 mice in SUMS-vector, n = 5 cells from 4 mice in SUMS-OE NgR). O Cumulative probability plots for the frequency in glutamatergic neurons (t = 2.074, p = 0.0767, n = 4 cells from 4 mice in SUMS-vector, n = 5 cells from 4 mice in SUMS-OE NgR). P-T Animal behaviors in SIT, OFT, EPM, TST, and FST (SIT, t = 2.323, p = 0.0337, n = 9 per group; OFT, t = 0.04752, p = 0.9626, n = 9 in SUMS-vector, n = 12 in SUMS-OE NgR; EPM, t = 0.5683, p = 0.5765, n = 9 in SUMS-vector, n = 12 in SUMS-OE NgR; TST, t = 2.973, p = 0.0116, n = 7 per group; FST, t = 2.552, p = 0.0230, n = 7 in SUMS-vector, n = 9 in SUMS-OE NgR). Data are expressed as mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, by an unpaired Student’s t-tests (B, L-T), a one-way ANOVA with Tukey post-hoc test (D, E, F-J).

To further verify whether upregulation of NgR in dCA3 glutamatergic neurons is sufficient to induce depression-like behaviors, we combined the SUMS paradigm with specific overexpression of NgR in this neuronal population (Fig. 3K). Western blot analysis confirmed robust upregulation of NgR expression in dCA3 neurons (Fig. 3L). Whole-cell patch-clamp recordings revealed that NgR overexpression significantly reduced neuronal excitability in glutamatergic neurons of SUMS mice (Fig. 3M–O). Behavioral analyses demonstrated that this manipulation significantly facilitated the emergence of depression-like phenotypes in SUMS mice, without affecting anxiety-related behaviors (Fig. 3P–T). Integrated findings establish NgR as a pivotal bidirectional modulator of glutamatergic circuit homeostasis, whose overexpression triggers synaptic dysfunction and network-level hypoexcitability to drive depression-like pathogenesis.

Optogenetic control reveals chronic stress suppresses dCA3^CaMKIIα → VTA^DA circuit to promote depression

The hippocampal dopaminergic pathway, which originates from the VTA, has been established as a critical neural circuit of depression-like behaviors comorbid with chronic pain [23]. To investigate the potential role of the dCA3 → VTA circuit in depression-like behaviors, we stereotaxically injected the retrograde tracing virus AAV-retro-hsyn-RFP into the VTA. Subsequent immunofluorescence examination of the dCA3 region identified labeled neuronal somata predominantly localized to glutamatergic populations, thereby providing initial anatomical evidence for a direct synaptic projection from the dCA3 to VTA (Fig. 4A).

Fig. 4. Optogenetic activation or inhibition of dCA3^CaMKIIα → VTA^DA circuit bidirectionally regulates depressive-like behavior in mice.

A Left: viral injection strategy. Right: representative images showing retrogradely labeled dCA3 neurons (red) co-localized with CaMKIIα (green). Scale bars: 40 μm. B Left: schematic of viral injection. Right: AAV-retro-DIO-eNPHR3.0-EYFP-labeled axons surround TH-positive neurons in the VTA. Scale bars: 200 μm. C Top: Behavioral testing timeline for depression-related assessments. Bottom: action potentials in response to yellow light stimulation in eNPHR-expressing VTA DA neurons and quantitative analysis of AP frequency (t = 4.881, p = 0.0164, n = 4 cells from 4 mice per group). D Sample mEPSCs recordings from dopaminergic neurons. E Cumulative probability plots for the amplitude in DA neurons (t = 0.1880, p = 0.8533, n = 9 cells from 6 mice per group). F Cumulative probability plots for the frequency in DA neurons (t = 2.308, p = 0.0347, n = 9 cells from 6 mice per group). G Representative action potential traces from dopaminergic cells. H Firing frequency analysis (t = 2.588, p = 0.0322, n = 5 cells from 5 mice per group). I-L Behavioral performance in SIT, OFT, EPM, and TST (SIT, t = 3.225, p = 0.0073, n = 7 per group; OFT, t = 1.217, p = 0.2471, n = 7 per group; EPM, t = 0.7669, p = 0.4580, n = 7 per group; TST, t = 3.483, p = 0.0045, n = 7 per group). M Left: viral injection schematic. Right: AAV-retro-DIO-hChR2-EYFP-labeled axons surround TH-positive neurons in the VTA. Scale bars: 200 μm. N Top: schematic diagram of depression-related behavioral tests. Bottom: action potentials in response to blue light stimulation in hChR2-expressing VTA DA neurons and quantitative analysis of AP frequency (t = 23.04, p < 0.001, n = 4 cells from 4 mice per group). O Example mEPSCs traces from dopaminergic neurons. P Cumulative probability plots for the amplitude in DA neurons (t = 0.8591, p = 0.4153, n = 5 cells from 5 mice per group). Q Cumulative probability plots for the frequency in DA neurons (t = 2.001, p = 0.0402, n = 5 cells from 5 mice per group). R Action potential traces from dopaminergic neurons. S Firing frequency quantification. (t = 4.355, p < 0.001, n = 11 cells from 6 mice per group). T-W Behavioral results from SIT, OFT, EPM, and TST (SIT, t = 2.543, p = 0.0245, n = 7 in EYFP, n = 8 in ChR2; OFT, t = 2.740, p = 0.0140, n = 9 in EYFP, n = 10 in ChR2; EPM, t = 3.379, p = 0.0025, n = 8 in EYFP, n = 7 in ChR2; TST, t = 2.971, p = 0.0090, n = 9 per group). Data are expressed as mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, by a paired Student’s t-tests (C, N), an unpaired Student’s t-tests (E-L, P, Q, S-W).

To functionally interrogate the dCA3 → VTA neural circuit, we injected AAV9-CaMKIIα-Cre into the dCA3 subregion and implanted Cre-dependent AAV-retro-DIO-eNPHR3.0-EYFP into the VTA of adult mice. Immunofluorescence analysis revealed that these fibers projected predominantly to TH-positive dopaminergic neurons in the VTA (Fig. 4B), with minimal targeting of GABAergic or glutamatergic neurons (Supplementary Fig. 4). These findings suggest that dCA3 glutamatergic neurons exhibit projection specificity toward VTA dopaminergic neurons. Based on this specific projection, we proceeded to functionally modulate the dCA3 → VTA DA pathway. The mice were subjected to a SUMS paradigm, followed by optical fiber implantation above the VTA subregion (Fig. 4C). Following the SUMS sessions, yellow light (589 nm) was applied to optogenetically inhibit the glutamatergic projection from the dCA3 to the VTA, enabling inhibition of this neural circuit (Fig. 4C). Whole-cell patch-clamp recordings revealed that optogenetic inhibition of axon terminals from eNPHR-expressing CaMKIIα-positive neurons in the dCA3 subregion significantly reduced the frequency of light-evoked mEPSCs in VTA dopaminergic neurons, without altering their amplitude (Fig. 4D–F). These findings suggest a presynaptic reduction in glutamate release probability, while postsynaptic receptor functionality remained intact. Furthermore, in eNPHR-expressing mice, spontaneous firing activity of VTA dopaminergic neurons was also markedly suppressed (Fig. 4G, H), further supporting the role of the dCA3 → VTA pathway in driving dopaminergic excitability. Collectively, CaMKIIα-positive neurons in the dCA3 subregion modulate the excitatory output of VTA dopaminergic neurons by forming excitatory synaptic connections with them. Notably, we further examined the effect of altered dCA3 neuronal activity on the downstream dCA1 region. The results showed that inhibition of dCA3 glutamatergic neurons in the SUMS model significantly reduced the frequency of mEPSCs in dCA1 neurons, while activation of these neurons in the CUMS model significantly increased mEPSCs frequency, with no changes in amplitude (Supplementary Fig. 3). These findings indicate that regulation of dCA3 neuronal activity can be transmitted along neural pathways to the downstream dCA1 region, suggesting that dCA3 may participate in the regulation of stress-related behaviors through multiple output pathways.

To investigate the regulatory role of the dCA3 → VTA pathway in depression-like behaviors, we conducted a series of behavioral assays in mice subjected to optogenetic inhibition of this circuit. Results demonstrated that yellow light (589 nm) suppression of the dCA3 → VTA dopaminergic pathway in eNPHR-expressing mice significantly reduced social interaction and increased immobility time in the TST (Fig. 4I, L), whilst having no effect on time spent in the central area during the OFT and time spent in the open arms of the EPM (Fig. 4J, K). These behavioral alterations indicate that optogenetic inhibition of the dCA3^CaMKIIα → VTA^DA circuit can elicit depressive-like phenotypes. To investigate whether activation of the dCA3^CaMKIIα → VTA^DA pathway produces antidepressant-like effects, hChR2–EYFP-expressing mice were subjected to CUMS (Fig. 4M, N). Whole-cell patch-clamp recordings confirmed that optogenetic stimulation of hChR2 effectively activated glutamatergic projections from the dCA3 to the VTA (Fig. 4N). This activation significantly increased the frequency—but not the amplitude—of light-evoked mEPSCs in VTA dopaminergic neurons and enhanced their spontaneous firing activity (Fig. 4O–S). Following 10 days of chronic stress, optogenetic activation of the dCA3CaMKIIα → VTADA circuit robustly ameliorated depression-like behaviors (Fig. 4T–W).

Chemogenetic inhibition of dCA3 → VTA dopamine (DA) pathway mimics chronic stress-induced depression

Given that acute optogenetic modulation of the dCA3^CaMKIIα → VTA^DA circuit was sufficient to bidirectionally regulate depressive-like phenotypes, we investigated whether sustained chemogenetic inhibition of this pathway would produce analogous and stable behavioral effects. Specifically, the Cre-dependent recombinant viral vectors AAV-retro-hsyn-DIO-hM4Di -mCherry (for inhibition) and AAV-retro-hsyn-DIO-mCherry (as control) were injected into the VTA, while AAV1-CaMKIIα-Cre under the control of the CaMKIIα promoter was delivered into the dCA3 to achieve targeting of the glutamatergic projections from dCA3 to VTA (Fig. 5A). Immunofluorescence analysis revealed that axon terminals of dCA3 glutamatergic neurons exhibited mCherry-positive signals and were densely distributed around the somata of dopaminergic neurons in the VTA (Fig. 5A). Electrophysiological recordings confirmed that perfusion of CNO onto brain slices effectively suppressed the excitability of VTA dopamine neurons (Fig. 5B). Following 19 days of viral expression, mice underwent SUMS and received intraperitoneal CNO administration 30 min before behavioral tests to selectively inhibit the dCA3 → VTA glutamatergic pathway (Fig. 5C). Whole-cell patch-clamp recordings in VTA DA neurons from hM4Di-mCherry expressing mice revealed that chemogenetic inhibition significantly decreased the frequency of mEPSCs, without affecting amplitude (Fig. 5D–F), consistent with a presynaptic mechanism of modulation and aligning with our optogenetic observations. Behaviorally, chemogenetic inhibition of the dCA3 → VTA pathway via CNO administration in hM4Di -mCherry–expressing mice elicited significant behavioral deficits, including increased social avoidance (Fig. 5G), heightened anxiety-like behaviors (reduced center time in the OFT and decreased open arm time in the EPM) (Fig. 5H, I), and increased despair-like behaviors, prolonged immobility time in both the TST and FST (Fig. 5J, K). These behavioral impairments consistently mirrored the phenotypes induced by optogenetic inhibition of the same pathway, demonstrating that both acute and sustained suppression of the dCA3 → VTA circuit reliably provoke depression-like behaviors.

Fig. 5. Chemogenetic activation or inhibition of dCA3^CaMKIIα → VTA^DA circuit bidirectionally modulates depressive-like behavior in mice.

A Left: schematic of viral injection. Right: representative image showing axons of hM4Di-expressing neurons surround TH-positive neurons in the VTA. Scale bars: 100 μm. B Left: action potentials in response to CNO in hM4Di-expressing VTA DA neurons. Right: quantitative analysis of AP frequency (t = 5.563, p = 0.0115, n = 4 cells from 4 mice per group). C Schematic diagram of depression-related behavioral tests. D Sample mEPSCs recordings from dopaminergic neurons. E Cumulative probability plots for the amplitude in DA neurons (t = 0.5267, p = 0.6099, n = 6 cells from 6 mice per group). F Cumulative probability plots for the frequency in DA neurons (t = 2.079, p = 0.0321, n = 6 cells from 6 mice per group). G-K Animal behaviors in SIT, OFT, EPM, TST, and FST (SIT, t = 4.110, p = 0.0011, n = 8 per group; OFT, t = 2.506, p = 0.0276, n = 7 per group; EPM, t = 2.672, p = 0.0182, n = 8 per group; TST, t = 2.554, p = 0.0230, n = 8 per group; FST, t = 4.060, p = 0.0016, n = 7 per group). L Left: schematic of viral injection. Right: representative image showing axons of hM3Dq-expressing neurons surround TH-positive neurons in the VTA. Scale bars: 100 μm. M Enhancement of AP frequency by CNO (t = 4.770, p = 0.0175, n = 4 cells from 4 mice per group). N Behavioral test timeline. O Sample mEPSCs recordings from dopaminergic neurons. P, Q Cumulative probability plots of mEPSCs amplitude and frequency in DA neurons (amplitude, t = 1.243, p = 0.2819, n = 5 cells from 5 mice per group; frequency, t = 2.006, p = 0.0399, n = 5 cells from 5 mice per group). R-V Behavioral results from SIT, OFT, EPM, TST, and FST (SIT, t = 2.969, p = 0.0101, n = 8 per group; OFT, t = 2.866, p = 0.0142, n = 7 per group; EPM, t = 2.673, p = 0.0203, n = 7 per group; TST, t = 3.184, p = 0.0079, n = 7 per group; FST, t = 4.583, p < 0.001, n = 7 per group). Data are expressed as mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, and ***p < 0.001, by a paired Student’s t-tests (B, M), an unpaired Student’s t-tests (E-K, P-V).

These results establish that suppression of the dCA3 → VTA pathway induces depressive-like behavior, prompting us to investigate pharmacological activation of this circuit could rescue depression-like phenotypes under CUMS using chemogenetic excitation. Following CUMS, CNO was administered to activate the circuit (Fig. 5L–N). This intervention significantly increased the frequency—but not amplitude—of mEPSCs in VTA dopaminergic neurons (Fig. 5O–Q). Behaviorally, chemogenetic activation robustly prevented stress-induced behavioral deficits, manifested as a significant attenuation of social avoidance, anxiety-like behaviors, and despair-like responses (Fig. 5R–V). Collectively, these data demonstrate that targeted excitation of the dCA3 → VTA pathway rescues depressive-like behavior changes under CUMS. These chemogenetic findings highlight the pathway’s robust regulatory capacity and translational potential for developing novel antidepressant strategies.

VTA → NAc dopaminergic inputs regulate depressive-like behaviors through D1/D2 receptor signaling

The mesolimbic dopamine system (VTA → NAc) constitutes a core pathway for motivation and reward processing [24, 28]. To delineate its circuit mechanisms underlying depression-like behaviors, we employed retrograde tracing from NAc and identified GRB-labeled somata in VTA dopaminergic (TH⁺) neurons, confirming a specific anatomical VTA→NAc connection (Fig. 6A, B). To functionally interrogate this pathway, we employed optogenetic approaches for precise circuit manipulation. DAT-Cre mice received intra-NAc injections of Cre-dependent inhibitory optogenetic virus AAV-retro-DIO-eNPHR3.0-EYFP, followed by optical fiber implantation to enable targeted modulation of the VTA→NAc pathway (Fig. 6C). Yellow light-induced inhibition of VTA → NAc dopaminergic inputs significantly reduced social interaction time and increased immobility time in the TST in SUMS mice, while having no significant effect on anxiety-like behaviors (Fig. 6D), indicating that suppression of this pathway rapidly induces depression-like behaviors. Conversely, optogenetic activation of the same pathway produced significant antidepressant effects, as evidenced by the amelioration of anxiety- and depression-like behaviors in CUMS mice (Fig. 6E, F). To further evaluate the sustained regulatory role of this circuit, we implemented chemogenetic inhibition for long-term modulation. hM4Di - mCherry virus was injected into the VTA, with pharmacological intervention achieved through local CNO administration in the brain region (Fig. 6G). Chronic chemogenetic inhibition of the VTA → NAc pathway elicited significant depression-like phenotypes, as evidenced by enhanced social avoidance, increased anxiety-like behaviors (reduced center time in the OFT and increased close arm time in the EPM), and exacerbated despair-like behaviors (prolonged immobility in the TST and FST) in SUMS mice, demonstrating that functional suppression of this circuit is sufficient to promote depression-like behaviors (Fig. 6H).

Fig. 6. NgR in dCA3 glutamatergic neurons drives depression-like behaviors by suppressing the VTA-NAc dopaminergic pathway.

A Retrobeads injection strategy. B Representative images demonstrate GRB⁺ neurons and TH-positive immunoreactivity in the VTA. Scale bars: 40 μm and 200 μm. C Left: schematic for optogenetic inhibition. Right: AAV-retro-DIO-eNPHR3.0-EYFP-labelled axon terminals expressed in the NAc. Scale bars: 200 μm. D Behavioral performance in SIT, OFT, EPM, and TST (SIT, t = 2.689, p = 0.0197, n = 7 per group; OFT, t = 1.041, p = 0.3185, n = 7 per group; EPM, t = 1.233, p = 0.2458, n = 6 per group; TST, t = 4.921, p < 0.001, n = 7 per group). E Left: schematic for optogenetic activation. Right: AAV-retro-DIO-hChR2-EYFP-labelled axon terminals expressed in the NAc. Scale bars: 200 μm. F Behavioral profiling in SIT, OFT, EPM, and TST (SIT, t = 3.609, p = 0.0028, n = 8 per group; OFT, t = 3.961, p = 0.0011, n = 9 per group; EPM, t = 2.644, p = 0.0206 n = 7 per group; TST, t = 2.390, p = 0.0341, n = 7 per group). G Left: schematic of viral injection and cannula implantation. Right: hM4Di-labelled axon terminals expressed in the NAc. Scale bars: 200 μm. H Behavioral analysis across five tests: SIT, OFT, EPM, TST, and FST (SIT, t = 2.558, p = 0.0126, n = 7 per group; OFT, t = 3.685, p = 0.0016, n = 7 per group; EPM, t = 2.859, p = 0.0144, n = 6 per group; TST, t = 2.696, p = 0.0194, n = 7 per group; FST, t = 2.525, p = 0.0266, n = 7 per group). I Left: schematic of viral injection, fiber, and cannula implantation. Right: AAV-retro-DIO-hChR2-EYFP-labelled axon terminals expressed in the NAc. Scale bars: 200 μm. J Behavioral results from SIT, OFT, EPM and TST (SIT, t = 3.002, p = 0.0110, n = 7 in EYFP and ChR2, t = 7.741, p < 0.001, n = 7 in ChR2 and n = 6 in ChR2 + SCH-23390, t = 2.994, p = 0.0112, n = 7 in ChR2 and ChR2 + Sulpiride; OFT, t = 4.414, p < 0.001, n = 7 in EYFP and ChR2, t = 4.382, p < 0.001, n = 7 in ChR2 and ChR2 + SCH-23390, t = 6.482, p < 0.001, n = 7 in ChR2 and ChR2 + Sulpiride; EPM, t = 3.885, p = 0.002 n = 7 in EYFP and ChR2, t = 5.088, p < 0.001, n = 7 in ChR2 and ChR2 + SCH-23390; t = 4.577, p < 0.001, n = 7 in ChR2 and ChR2 + Sulpiride; TST, t = 2.677, p = 0.0202, n = 7 in EYFP and ChR2, t = 3.948, p = 0.0019, n = 7 in ChR2 and ChR2 + SCH-23390; t = 2.551, p = 0.0254, n = 7 in ChR2 and ChR2 + Sulpiride). K Viral injection strategy. L Representative images of TH⁺, RFP⁺, EYFP⁺, and DAPI co-staining in the VTA. Scale bars: 40 μm. M Left: schematic of viral injection, fiber, and cannula implantation. Middle: NgR-shRNA expression in dCA3 glutamatergic neurons. Right: hM4Di expression in VTA dopaminergic neurons. Scale bars: 200 μm. N Timeline of behavioral tests for depression-like phenotypes. O-R Animal behaviors in SIT, OFT, EPM, and TST (SIT, t = 2.394, p = 0.0312, n = 8 in CUMS-scramble and CUMS-scramble + CNO, t = 2.974, p = 0.0136, n = 8 in CUMS-scramble and n = 9 in CUMS-shNgR, t = 2.842, p = 0.0139, n = 9 in CUMS-shNgR and n = 6 in CUMS-shNgR+CNO; OFT, t = 1.006, p = 0.3314, n = 8 in CUMS-scramble and CUMS-scramble + CNO, t = 0.1432, p = 0.8882, n = 8 in CUMS-scramble and in CUMS-shNgR, t = 2.123, p = 0.0520, n = 8 in CUMS-shNgR and CUMS-shNgR + CNO; EPM, t = 5.063, p < 0.001, n = 7 in CUMS-scramble and CUMS-scramble + CNO, t = 2.554, p = 0.0253, n = 7 in CUMS-scramble and CUMS-shNgR, t = 2.836, p = 0.0150, n = 7 in CUMS-shNgR and CUMS-shNgR + CNO; TST, t = 2.275, p = 0.0405, n = 8 in CUMS-scramble and n = 7 CUMS-scramble + CNO, t = 2.350, p = 0.0319, n = 8 in CUMS-scramble and n = 10 CUMS-shNgR, t = 4.751, p < 0.001, n = 10 in CUMS-shNgR and n = 7 CUMS-shNgR + CNO). Data are expressed as mean ± SEM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, by an unpaired Student’s t-tests (D, F, H, J, O-R).

The neurons of the NAc are primarily divided into two subtypes: D1-type neurons expressing dopamine D1 receptors and D2-type neurons expressing dopamine D2 receptors [29]. To investigate whether dopaminergic inputs from the VTA to the NAc modulate stress-induced depression-like phenotypes via D1 or D2 receptor signaling, we injected AAV-retro-DIO-hChR2-EYFP or AAV-retro-DIO-EYFP into the NAc of DAT-Cre mice (Fig. 6I). Optical fibers were implanted above the VTA, and guide cannulas were positioned above the NAc for local pharmacological interventions. Optogenetic activation of VTA dopaminergic terminals robustly reversed depression-like behaviors. However, this antidepressant-like effect was significantly attenuated by intra-NAc infusion of either the D1 receptor antagonist SCH-23390 or the D2 receptor antagonist Sulpiride (Fig. 6J). These findings demonstrate that the behavioral benefits of VTA → NAc dopaminergic activation require functional engagement of either D1 or D2 receptor subtypes, highlighting their role in mediating the antidepressant effects of mesolimbic dopamine signaling.

To further validate the regulatory role of DA receptors under physiological conditions, we performed local microinfusions of receptor antagonists into the NAc of control mice. Intra-NAc administration of either SCH-23390 or Sulpiride induced profound depression-like phenotypes (Supplementary Fig. 5, 6). These results indicate that pharmacological blockade of either D1 or D2 receptor signaling in the NAc is sufficient to elicit depression-like behaviors even in the absence of stress, underscoring the essential role of dopaminergic signaling in the maintenance of emotional homeostasis.

Reduction of NgR alleviates depressive-like behavior via dCA3^CaMKIIα → VTA^DA → NAc circuit

To structurally validate the hypothesized dCA3 → VTA → NAc circuit connectivity, we employed a viral strategy combining retrograde trans-synaptic tracing and Cre-dependent anterograde labeling. AAV9-CaMKIIα-Cre was injected into the dCA3 region, while the VTA received the Cre-dependent reporter virus AAV-retro-DIO-EYFP, and the retrograde tracer AAV-retro-hsyn-RFP was delivered into the NAc, enabling specific labeling of VTA neurons that receive synaptic input from dCA3 and project to the NAc (Fig. 6K). Three weeks post-injection, we identified a distinct neuronal population characterised by RFP expression within the neurons themselves, whilst the perinuclear regions exhibited expression of the EYFP within the VTA. Immunofluorescence analysis confirmed that these dual-labeled neurons were co-localized with tyrosine hydroxylase (TH⁺), indicating their dopaminergic identity (Fig. 6L). These results provide anatomical evidence that CaMKIIα-expressing neurons in the dCA3 form synaptic excitatory connections with dopaminergic neurons in the VTA that subsequently project to the NAc, thereby establishing the structural basis of the dCA3^CaMKIIα → VTA^DA → NAc functional circuit.

Our results demonstrate that CUMS upregulates dCA3 NgR expression and significantly suppresses the activity of the dCA3 → VTA → NAc neural circuit. To investigate whether NgR modulates depression-like behaviors through this pathway, we combined targeted gene knockdown and chemogenetic intervention for functional validation. Specifically, AAV-CaMKIIα-shNgR was injected into the dCA3 region to selectively knock down NgR expression in glutamatergic neurons, while AAV9-DIO-hM4Di -mCherry and AAV9-TH-Cre were co-injected into the VTA to suppress the excitability of dopaminergic neurons. Subsequent local administration of CNO into the NAc was performed to chemogenetically suppress VTA → NAc dopaminergic transmission. Immunofluorescence analysis confirmed the expression efficiency of both viruses, as expected fluorescent signals were observed in the dCA3 and VTA regions (Fig. 6M). Behavioral assays revealed that the antidepressant-like effects induced by NgR knockdown in the dCA3 were abolished upon chemogenetic inhibition of the VTA → NAc pathway, indicating that the behavioral benefits require intact dopaminergic output from the VTA to the NAc (Fig. 6N–R).

Building on evidence that D1/D2 receptor signaling in the NAc underlies the antidepressant effects of mesolimbic dopaminergic activation [30], we investigated whether the VTA → NAc pathway mediates the behavioral consequences of dCA3 NgR knockdown. Bilateral microinfusion of AAV9-CaMKIIα-shNgR into the dCA3, combined with intra-NAc administration of either the D1 antagonist SCH-23390 or D2 antagonist Sulpiride (Supplementary Fig. 7), revealed that blockade of either dopamine receptor subtype completely abolished the antidepressant-like behaviors induced by NgR knockdown (Supplementary Fig. 8, 9). These results indicate that the behavioral effects of NgR knockdown in the dCA3 are dependent on the activation of either D1 or D2 receptors in the NAc, establishing coordinated terminal dopamine receptor signaling as a necessary component of the dCA3 → VTA → NAc circuit’s functional output.

Having established the role of NgR in regulating depression-like behaviors through the dCA3 → VTA → NAc circuit, we next investigated whether structural plasticity in the NAc—the terminal region of this pathway—is involved in this process. We found that CUMS-exposed mice exhibited significantly reduced neuronal complexity in the NAc (Supplementary Fig. 6A, B). Dendritic spine analysis revealed a marked increase in the density of immature spines (filopodia types), whereas the density of mature spines (stubby types) is reduced (Supplementary Fig. 6C), indicating impaired spine maturation. Notably, hippocampal-specific overexpression of NgR fully recapitulated these structural deficits in the NAc (Supplementary Fig. 6D–F), demonstrating that NgR signaling from the hippocampus is sufficient to drive structural remodeling in a distal brain region. These findings collectively delineate a coherent pathological cascade: the core pathology originates from upregulation of NgR in the dCA3, which subsequently impairs the functional integrity of the dCA3→VTA→NAc circuit. This dysfunction ultimately leads to structural deficits in dendritic spines, thereby driving the expression of depression-like and anxiety-like behaviors.

Discussion

In this study, we revealed the roles of NgR and dCA3^CaMKIIa → VTA^DA → NAc circuit in regulating depressive-like behavior in male mice. Several key findings were demonstrated: (1) chronic stress increased NgR in dCA3 glutamatergic neurons; (2) dCA3 glutamatergic neurons’ NgR bidirectionally regulated depressive-like behavior; (3) NgR regulates depressive-like behavior through excitability regulation in dCA3 glutamatergic neurons; (4) Knockdown NgR alleviated depressive-like behavior by activating the dCA3^CaMKIIa → VTA^DA → NAc circuit.

NgR, a key component of the myelin-derived inhibitory signaling pathway, functions not only in the classical inhibition of axonal regeneration but has also recently been implicated in the regulation of neuronal excitability and neuropsychiatric disorders [13–15]. NgR is expressed in neurons, astrocytes, and microglia [31–33] and has been shown to modulate neuronal excitability and synaptic transmission [34, 35], its precise role in regulating excitatory homeostasis within specific neural circuits, as well as its specific mechanisms in the pathophysiology of psychiatric disorders, remains to be fully elucidated. To identify the brain region where NgR exerts its role in the pathogenesis of depression, we examined NgR expression changes in the cortex, VTA, dCA1, and dCA3 following CUMS. The results showed a significant upregulation of NgR specifically in the dCA3 subregion, leading us to focus our subsequent investigations on this area. Previous studies have established that chronic stress impairs dCA3 structure and function, compromising its role in stress-information processing and promoting depression pathogenesis [36, 37]. Building on this, our study further revealed a key role for NgR in the stress response of the dCA3 region: in a chronic stress model, upregulation of NgR in dCA3 glutamatergic neurons was associated with reduced neuronal excitability and the emergence of depressive-like behaviors. Notably, regarding the site of action of NgR, our electrophysiological data revealed that chronic stress induced changes in dCA3 neuronal mEPSCs characterized by a decrease in amplitude without a change in frequency. In synaptic physiology, this pattern classically suggests the involvement of postsynaptic mechanisms, although it does not completely rule out the possibility that enhanced NgR expression also occurs at presynaptic terminals. Future studies employing immunoelectron microscopy and other techniques are warranted to further delineate the precise localization of NgR at pre- and postsynaptic compartments and their respective functional contributions. Building upon these findings, our functional experiments further validated the core role of NgR in stress-induced pathophysiology: knockdown of NgR under these conditions effectively rescued the deficits in mEPSCs amplitude and alleviated depressive-like phenotypes. Conversely, NgR overexpression in the dCA3 of SUMS mice was sufficient to recapitulate both the neuronal excitability deficits and depressive-like behaviors. These results establish a close association between NgR-mediated synaptic dysfunction and stress-induced depressive phenotypes. It is worth further exploration that, although this study focused on the effects of stress on CA3 glutamatergic neurons, our immunofluorescence results also revealed low levels of NgR expression in GABAergic neurons. This observation suggests that NgR may also indirectly influence glutamatergic neuron output by modulating local inhibitory microcircuits.

Based on the above findings, we further investigated the downstream neural circuit through which NgR regulates depressive-like behavior. Given that dCA3 is an important subregion of the hippocampus, its neural projection to the VTA has been reported to participate in reward and motivation processing [38], however, its role in stress-induced depressive behaviors remains poorly understood. Our results demonstrate that VTA dopaminergic neurons serve as a critical downstream mediator for the transmission of stress-related information from dCA3 glutamatergic neurons, and that activation of this pathway is sufficient to induce depressive-like behaviors. Strikingly, the dopaminergic projection from the VTA to the NAc plays a central role in reward and motivation processing, and dysfunction of this circuit has been extensively implicated in depression-related anhedonia and behavioral despair [24, 39, 40]. Using chemogenetic approaches, we further verified that neurotransmitter release from VTA dopaminergic neurons in the NAc directly contributes to the formation of depressive-like behaviors, thereby providing direct evidence at the circuit level for the proposed mechanism. DA neurons in VTA send dense projections to the NAc, where neuronal subtypes expressing either D1 or D2 receptors constitute the direct and indirect pathways, respectively [41–43]. In our study, the projection from VTA dopaminergic neurons to the NAc played a critical role in regulating stress-induced depressive-like behaviors through the activation of either D1 or D2 receptors. This study integrated the polysynaptic dCA3^CaMKIIα → VTA^DA → NAc neural pathway and systematically elucidated the functional connectivity among these three brain regions within a stress-induced depression model, revealing myelin-associated signaling molecules as key regulators of neural circuit plasticity—thereby achieving cross-level integration from molecular.

While this study provides compelling evidence for NgR-mediated regulation of depression-like behaviors through the dCA3^CaMKIIα → VTA^DA → NAc circuit, several directions remain open for further investigation. Future studies would benefit from including female subjects to systematically evaluate potential sex-dependent differences in NgR function and circuit regulation. In addition, although our findings indicate that the behavioral effects mediated by this circuit require the involvement of either D1 or D2 receptors, and structural alterations in dendritic spines were observed in the NAc, the precise link between dopaminergic receptor signaling and spine plasticity within this pathway has yet to be fully defined. In future experiments, we will further integrate optogenetic manipulation with two-photon in vivo imaging, which is expected to clarify how dynamic D1/D2 receptor activation regulates the morphology, density, and functional plasticity of dendritic spines in NAc neurons, thereby offering a more complete synaptic-level understanding of how receptor signaling translates into structural remodeling in depression-related circuits.

Author contributions

J.Z., H.L., and B.W. contributed equally to this work. R.J. conceived the idea and designed the study. L.S., H.S., and R.J. acquired the funding and supervised the study. J.Z., H.L., B.W., and D.Z. performed the experiments. S.G., G.M., and Z.C. performed statistical analysis. R.J., J.Z., and H.S. wrote and revised the manuscript. All authors read and approved the manuscript before submission.

Funding

R.J. discloses support for the research of this work from the National Natural Science Foundation of China (grant No. 82301726), the Natural Science Foundation of Shandong Province (grant No. ZR2022QH087), and the Shandong Provincial Medical and Health Science and Technology Project. H.L. discloses support for the research of this work from the National Natural Science Foundation of China (grant No. 82470587). B.W. discloses support for the research of this work from the National Natural Science Foundation of China (grant No. 82301700). H.S. discloses support for the research of this work from the National Natural Science Foundation of China (grant No. 81971281). J.Z., D.Z., S.G., Z.C., G.M., and L.S. declare no relevant funding.

Data availability

All data needed for the conclusion in the present study were presented in the manuscript and Supplementary Information. Additional data to this paper may be requested from the corresponding authors.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41380-026-03670-9.