A discrete serotonergic circuit regulates vulnerability to social stress

Stress promotes adaptation but prolonged stress contributes to psychiatric disorders such as anxiety, PTSD, and depression. Individuals differ in stress responses, with resilient versus susceptible phenotypes, consistent with diathesis-stress frameworks proposing biological predispositions to stress effects. The serotonergic system regulates diverse behaviors and is implicated in neuropsychiatric disorders; deficits in serotonin and SERT gene variants increase depression risk following stress, and SSRIs target this system. Most serotonergic neurons reside in the dorsal raphe (DR), whose activity influences reward and emotions; low serotonin increases susceptibility to psychosocial stress. Although DR GABAergic modulation can affect acquisition of social avoidance, the direct role of DR serotonergic neurons in stress vulnerability remains unclear. DR serotonergic neurons receive broad limbic inputs and project to the ventral tegmental area (VTA), a hub for motivation and reward that exhibits dopamine neuron adaptations after chronic social defeat stress (CSDS). DR→VTA serotonergic projections are abundant; global serotonergic inhibition elevates VTA dopamine firing, and DR→VTA projections mediate reward, but their role in stress is unknown. The study tests whether DR→VTA serotonergic neurons mediate effects of social stress and how they adapt electrophysiologically, hypothesizing that this circuit regulates vulnerability and resilience to social defeat.

Prior work shows: (1) Serotonergic dysregulation is linked to anxiety and depression; subsets of depressed patients show serotonin deficiency and SERT gene variants interact with stress to elevate depression risk; SSRIs are frontline treatments. (2) DR contains many serotonergic neurons influencing reward and emotion; decreasing serotonin increases stress susceptibility; disinhibiting DR 5-HT neurons via GABAergic silencing prevents social avoidance acquisition. (3) VTA dopamine neurons undergo maladaptive changes after CSDS and contribute to depression-like behaviors; inputs to VTA shape these responses. (4) DR→VTA projections are dense; pharmacological serotonin depletion elevates VTA dopamine firing; DR→VTA stimulation mediates reward. (5) Heterogeneity exists within DR 5-HT neurons, including co-expression of VGluT3 enabling glutamate co-release, and pathway-specific effects to NAc and mPFC in social reward and emotion. Gaps: direct contribution of DR→VTA serotonergic neurons to social stress vulnerability and their activity-dependent adaptations were unclear.

Animals: Male Sert-Cre mice (8–12 weeks) and male C57BL/6J mice used under approved institutional protocols, 12-h light/dark cycle, food and water ad libitum. Viral tools: Cre-dependent AAVs for expression of eArch-GFP, ChR2-mCherry, eNpHR3.0-mCherry, eYFP, and intersectional Flp/fDIO vectors sourced from BrainVTA, Sunbio, and Taitool. AAVretro used for projection-specific labeling/manipulation. Stereotaxic injections: Mice anesthetized with pentobarbital and injected at DR (AP −5.2 mm, ML 0, DV −2.6 mm, 15° angle), VTA (AP −3.1 mm, ML ±0.5 mm, DV −4.6 mm), NAc (AP +1.6, ML ±0.7, DV −4.2), or mPFC (AP +1.8, ML ±0.3, DV −2.3). Injection volumes 300–500 nL at 100 nL/min. Optical ferrules implanted 0.2 mm above DR for somatic stimulation. Expression allowed 4–8 weeks prior to experiments. Circuit mapping: Dual-site retrograde labeling with AAVretro-DIO-eYFP/mCherry into VTA with mPFC or NAc in Sert-Cre mice to identify collateralization. Intersectional strategy: AAVretro-DIO-Flp in VTA plus AAV-fDIO-eYFP in DR to selectively label DR 5-HT neurons projecting to VTA and trace their axonal arbors. Histology: Perfusion, fixation, sectioning (40 μm), immunostaining for Tph2 and GFP, confocal imaging (Nikon A1), quantification of labeled cells/fibers. Electrophysiology: Acute brain slices (300 μm) containing DR or VTA prepared in sucrose ACSF; recordings at 32–34 °C with MultiClamp 700B. Cell-attached recordings for spontaneous firing of DR 5-HTDR-VTA neurons identified via retrograde labeling. Whole-cell current-clamp to assess intrinsic excitability (current steps −40 to 340 pA; threshold and rheobase measured). Synaptic recordings: Voltage-clamp in VTA neurons with Cs-methanesulfonate internal; optogenetic activation of DR terminals expressing ChR2 with blue light (5 ms pulses) to evoke EPSCs (−70 mV) and IPSCs (0 mV). Pharmacology: CNQX (10 μM) for AMPAR, ketanserin (10 μM) for 5-HT2A/2C-mediated currents; TTX (1 μM) and 4-AP (200 μM) to confirm monosynaptic inputs; kynurenic acid and bicuculline to validate sEPSCs/sIPSCs. eNpHR validation with 593 nm light. Behavior: Chronic social defeat stress (CSDS) for 10 days with screened aggressive CD1 males; classification into susceptible (SI<1) and resilient (SI>1) using two-stage social interaction (SI) test. Subthreshold social defeat stress (two 3-min physical defeats with sensory exposure) used to probe vulnerability without persistent phenotype unless combined with optogenetic manipulation. Other assays: Open field test (OFT) for locomotion/anxiety, elevated plus maze (EPM), tail suspension test (TST), sucrose preference test (SPT). Optogenetics: DR 5-HT neuron inhibition (eArch/eNpHR) or activation (ChR2) with defined protocols—typically 20 Hz, 5 ms blue light for activation; continuous 593 nm yellow light for inhibition. Acute stimulations during SI test; chronic paradigms (20 min/day for 1 week inhibition; 2×/day, 10 min per session for 1 week activation). Statistics: Two-sided tests; t-tests and ANOVAs with Fisher’s LSD post hoc where appropriate; data as mean ± SEM; analyses via SPSS and GraphPad Prism. Sample size not predetermined by power calculation; experiments repeated independently at least three times; exclusion criteria for electrophysiology included Rs>30 MΩ or instability >20%.

- Distinct DR 5-HT subpopulation projects to VTA (5-HTDR-VTA): Dual retrograde labeling revealed limited collateralization; ~22.57–31.62% were 5-HTDR-VTA, ~4.16% DR-mPFC only, ~3.42% DR→VTA/mPFC, ~12.03% DR-NAc only, ~2.33% DR-VTA/NAc; 54.02–69.85% AAV-negative Tph2+ neurons. 5-HTDR-VTA neurons were ventrally distributed relative to DR-mPFC/DR-NAc populations and showed primary projection to VTA using intersectional AAVretro-DIO-Flp/fDIO-eYFP strategy, with sparse axons in mPFC/NAc. - Neurotransmitter identity: Optogenetic activation of DR 5-HT terminals in VTA evoked fast EPSCs abolished by CNQX (AMPA/kainate antagonist) (n=8 cells; paired t-test t=4.172, P=0.004) and slow outward currents reduced by ketanserin (5-HT2A/2C antagonist) (n=15; paired t-test, P=9.2e-5), indicating co-release of glutamate and serotonin. Response distribution in VTA neurons: 35.8% received both EPSC and IPSC components (34/95), 27.4% IPSC only, 8.4% EPSC only. TTX blocked light-evoked EPSCs, rescued by 4-AP; short latencies and low jitter supported monosynaptic DR→VTA inputs. - CSDS alters 5-HTDR-VTA physiology: After 10-day CSDS, mice classified as susceptible showed social avoidance (reduced SI ratio) versus resilient and controls. In retrogradely labeled 5-HTDR-VTA neurons, susceptible mice had decreased spontaneous firing (control mean ~2.68±0.34 Hz; one-way ANOVA with LSD post hoc: control vs susceptible P=0.046; susceptible vs resilient P=0.003). Interspike intervals increased in susceptible mice. Intrinsic excitability reduced in susceptible mice: decreased depolarization-evoked firing (two-way ANOVA post hoc, control vs susceptible P=0.044; susceptible vs resilient P=0.024); increased action potential threshold and rheobase. Synaptic inputs: resilient mice had increased sEPSC frequency onto 5-HTDR-VTA neurons; sIPSC frequency and amplitudes unchanged across groups. Blocking fast synaptic transmission did not normalize reduced excitability, implicating intrinsic changes. - Inhibiting DR 5-HT neurons promotes susceptibility after subthreshold stress: Somatic eArch inhibition of DR 5-HT neurons had no baseline effect on OFT/EPM/SI, but during subthreshold defeat produced robust social avoidance and decreased SI ratio (one-way ANOVA interaction F1,34=20.751, P<0.001; SI ratio F1,34=6.037, P=0.019), whereas subthreshold defeat alone did not. - Projection-specific inhibition of 5-HTDR-VTA neurons promotes susceptibility: eNpHR expression in DR→VTA neurons validated by light-evoked hyperpolarization (n=5). Acute inhibition during SI did not alter behavior (F1,22≈0.18–0.22, ns), but inhibition following subthreshold defeat induced social aversion with reduced SI ratio (interaction F1,33=18.287, P<0.001; SI ratio F1,33=28.155, P<0.001). No baseline changes in locomotion, anxiety, TST, or SPT in stress-naive mice. - Bidirectional control of susceptibility/resilience: In resilient mice post-CSDS, acute inhibition of 5-HTDR-VTA during SI had no effect (interaction F1,14=0.95, P=0.346), but chronic inhibition (20 min/day, 1 week) decreased social interaction and converted to susceptible-like phenotype (interaction F1,13=0.613, P=0.006; SI ratio F1,13=5.546, P=0.035). In susceptible mice post-CSDS, acute ChR2 activation (20 Hz) increased time in interaction zone and SI ratio (interaction F1,14=9.497, P=0.008; SI ratio F1,14=7.375, P=0.017), and chronic activation for 1 week similarly promoted resilience (interaction F1,12=4.773, P=0.049; SI ratio F1,12=5.989, P=0.031). Control animals showed no behavioral changes with stimulation. - Overall, decreased activity of 5-HTDR-VTA neurons characterizes susceptibility; activation confers resilience; inhibition (especially chronic) promotes susceptibility.

The study identifies a discrete DR serotonergic subpopulation projecting to the VTA that bidirectionally regulates vulnerability to social stress. Susceptible mice exhibit reduced spontaneous firing and intrinsic excitability of 5-HTDR-VTA neurons, independent of immediate synaptic input changes, suggesting intrinsic cellular adaptations. Conversely, resilient mice show enhanced excitatory synaptic input onto these neurons. DR 5-HT terminals in VTA co-release glutamate and 5-HT, providing both fast excitation (AMPAR-mediated) and slower inhibition (5-HT2A/2C-mediated), positioning this pathway to finely modulate VTA circuits implicated in stress-related behaviors. The behavioral causality is demonstrated: inhibiting 5-HTDR-VTA neurons after subthreshold defeat induces social avoidance, while activating this pathway reverses susceptibility to resilience. Chronic, but not acute, inhibition can convert resilient to susceptible phenotypes, implying plasticity-dependent processes and potentially stronger or selectively adapted connections in resilient states. These findings integrate with prior evidence that serotonergic tone modulates VTA dopamine neuron activity and stress susceptibility, suggesting that reduced DR→VTA serotonergic drive may disinhibit or enhance VTA dopamine output characteristic of susceptible animals. The data highlight functional heterogeneity within DR 5-HT neurons and emphasize pathway-specific roles in stress, reward, and emotional regulation.

This work delineates a discrete DR→VTA serotonergic circuit as a key determinant of individual vulnerability to social stress. It shows that 5-HTDR-VTA neuron activity decreases with susceptibility, while enhancing their activity promotes resilience; bidirectional optogenetic manipulation correspondingly modulates social interaction behavior after stress. The circuit co-releases serotonin and glutamate onto VTA neurons, indicating multiple modes of control over VTA function in stress. These findings suggest that targeted activation of the DR→VTA serotonergic pathway could be a therapeutic strategy for stress-related disorders. Future research should map upstream inputs that remodel 5-HTDR-VTA neurons in resilience (e.g., cortical projections), define downstream VTA cell-type-specific targets and synaptic mechanisms, dissect contributions of SERT-only versus SERT–VGluT3 co-expressing DR neurons, and test generalizability across diverse stress models.

- Initial dual-site retrograde tracing may under-detect collateralization because only one or two targets can be sampled simultaneously; intersectional strategies partly address this. - Sample sizes were not predetermined by power analysis; some animals were excluded due to failed injections or lack of dual labeling, which may introduce bias. - Findings are in male mice with optogenetic manipulations; generalizability to females, other species, and different stress paradigms remains to be established. - Acute versus chronic manipulation produced different outcomes (notably, chronic inhibition required to convert resilience), indicating complex plasticity that was not fully elucidated. - Upstream sources of increased excitatory input in resilient mice and the specific downstream VTA cell types mediating behavioral effects were not identified; co-release contributions of serotonin versus glutamate were not separately quantified behaviorally.