Hippocampal astrocytes modulate anxiety-like behavior

The study investigates whether hippocampal astrocytes contribute to the regulation of anxiety-like behavior. While astrocytes are known to modulate synaptic transmission within tripartite synapses and have been implicated in psychiatric disorders, their direct role in affective behaviors, particularly anxiety, has not been clearly defined. The hippocampus is a key structure for anxiety regulation, with DG granule cell activation being anxiolytic and ventral CA1 neurons linked to anxiety responses. The authors hypothesize that hippocampal astrocyte calcium activity reflects anxiety state and that activating these astrocytes can modulate anxiety-like behavior by altering synaptic transmission in DG granule cells via gliotransmitter release.

Prior work shows astrocytes regulate synaptic transmission and behavior via gliotransmitters. Postmortem studies in mood disorders report altered astrocyte morphology and volume; animal models of PTSD show reduced hippocampal GFAP levels and anxiety-like behaviors. Astrocyte activation in the central amygdala can reduce anxiety-like behaviors. In the hippocampus, astrocyte vesicular release affects recognition memory, and chemogenetic activation enhances contextual fear memory, indicating astrocytes modulate hippocampal-dependent functions. Neural circuit studies implicate hippocampus–hypothalamic and hippocampus–prefrontal pathways in anxiety. Collectively, these findings suggest astrocyte involvement in affective disorders but the mechanisms linking hippocampal astrocytes to anxiety remained unresolved prior to this work.

• Transgenic mice: Generated astrocyte-specific reporter and effector lines by crossing hGFAP-CreERT2 with floxed-stop-GCaMP6s (hGFAP-GCaMP6s) for Ca2+ imaging and floxed-stop-ChR2-EYFP (hGFAP-ChR2) for optogenetic activation. Tamoxifen induction ensured astrocyte-specific expression (validated by GFAP/S100B colocalization; absence in NeuN/Iba1).
• In vivo two-photon Ca2+ imaging: Implanted hippocampal cranial windows in head-fixed, awake hGFAP-GCaMP6s mice. Recorded astrocyte Ca2+ signals (ΔF/F) in CA1 stratum radiatum and stratum lacunosum-moleculare during virtual reality (VR) navigation emulating elevated plus maze (closed arms vs open, bright center). Motion correction (NoRMCorre) and ROI extraction (CalmAn); event detection based on ΔF/F thresholds and position alignment; classification of pre- vs post-responsive cells relative to entering the center.
• Optogenetics: Implanted optic fibers bilaterally/unilaterally over dorsal or ventral hippocampus. Delivered continuous 473 nm light (1–3 mW, 5 min) to hGFAP-ChR2 astrocytes. Verified Ca2+ elevation upon stimulation in primary astrocyte cultures (Rhod-2AM imaging) and acute slices using AAV-GFAP-jRGECO1a imaging.
• Behavioral assays: Elevated plus maze (EPM) with light-off/on/off epochs (5/5/5 min), Open field test (OFT) with similar epochs and extended 1 h sessions to assess duration of effects, T-maze for spatial working memory, and real-time place preference (RTPP) to assess place preference.
• Neuronal activation mapping: c-Fos immunohistochemistry 1 h after astrocyte stimulation to assess activation in dorsal/ventral hippocampus (DG, CA1, CA3).
• Electrophysiology: Whole-cell voltage-clamp recordings of sEPSCs from DG granule cells (and CA1 neurons in supplementary experiments) in acute slices during optogenetic astrocyte stimulation (10 min protocol: 3 min pre, 5 min light, 2 min post). Examined effects along dorsoventral axis, repeated stimulation, and assessed intrinsic excitability in current-clamp. Tested contribution of extracellular K+ by elevating [K+]o (7 or 9 mM) and monitored inward currents and sEPSCs.
• Gliotransmitter assays: Measured ATP release using bioluminescence from primary astrocytes and hippocampal slices during light stimulation; tested for glutamate and D-serine release (colorimetric/fluorometric kits). Pharmacology: Applied ATP (100 µM) during sEPSC recordings; used PPADS (100 µM), a purinergic receptor antagonist, in slices and via hippocampal cannula in vivo during behaviors to probe ATP signaling.
• Statistics: One-way and two-way repeated measures ANOVA with post hoc Bonferroni or LSD tests; Mann–Whitney U, independent and paired t-tests; Kolmogorov–Smirnov tests for cumulative distributions. Significance at p<0.05.

• Hippocampal astrocytes exhibit increased Ca2+ activity in anxiogenic contexts: During VR-EPM, astrocytes showed significantly more Ca2+ peaks in the open, bright center compared to corner or closed areas (ANOVA p=0.000; LSD p=0.000 center vs corner; p=0.000 center vs closed; n=87 cells). Upon entering the center, 75.2% of hippocampal astrocytes responded versus 23.3% when entering the corner (n=129 events). Among center-responsive cells, 19.6% were pre-responsive (before entry) and 80.4% post-responsive (after entry).
• Optogenetic activation of hippocampal astrocytes induces anxiolytic behavior: In EPM, hGFAP-ChR2 mice spent more time in open arms during light-on than controls (two-way RM ANOVA p=0.000; U-test p=0.001 light-on; effect partially persisted into the subsequent light-off epoch, p=0.003). Time in center also increased (ANOVA p=0.002; U-test p=0.046 light-on; p=0.001 last light-off). Travel speed increased during and after stimulation (ANOVA p=0.002; U-test p=0.010 light-on; p=0.001 last light-off). Similar effects occurred with ventral hippocampal stimulation.
• Increased exploratory drive in OFT: Total distance traveled increased during light-on (ANOVA p=0.007; U-test p=0.018) and remained elevated after light-off (p=0.006). The increase persisted for ~10 min after light cessation in 1 h sessions (p=0.018 at 5–10 min; p=0.006 at 10–15 min). Center distance changes were not significant.
• Neuronal activation and synaptic effects: c-Fos expression increased in DG after astrocyte stimulation (Mann–Whitney U p=0.008), with broader dorsal–ventral effects. Whole-cell recordings showed increased sEPSC frequency (but not amplitude) in DG granule cells during light in hGFAP-ChR2 slices (paired t-test p=0.000; amplitude p=0.566), with rapid onset (peak within 20–50 s) and return toward baseline during the 5-min stimulation. Similar sEPSC frequency increases occurred in both dorsal and ventral hippocampus and in CA1 pyramidal neurons. Repeated stimulation produced comparable effects, indicating no light-induced neuronal damage.
• Elevated extracellular K+ alone does not account for synaptic changes: While optogenetic stimulation induced inward holding currents consistent with [K+]o elevation, experimentally raising [K+]o to 7 mM increased inward currents without changing sEPSC frequency or amplitude; at 9 mM, both frequency and amplitude increased. Thus, [K+]o changes do not fully explain selective frequency increases observed.
• ATP mediates astrocyte–neuron effects: Light stimulation increased ATP release from ChR2-expressing primary astrocytes (p=0.001) and hippocampal slices (p=0.044); no significant glutamate or D-serine release was detected. Exogenous ATP (100 µM) increased sEPSC frequency (one-way ANOVA p=0.004; Pre vs ATP p=0.007) without affecting amplitude. PPADS blocked the light-induced sEPSC frequency increase (light-on p=0.021 vs vehicle; amplitudes n.s.). In vivo, PPADS largely abolished the astrocyte activation-induced increase in time spent in open arms in EPM, while effects on center time and speed induction were less affected; PPADS attenuated maintenance of increased center time and speed during the second light-off epoch. OFT distance/speed increases were not blocked by PPADS.
• Specificity: Optogenetic astrocyte activation did not alter T-maze performance (spatial working memory) or induce real-time place preference, suggesting specificity for anxiety-related behavior rather than general memory or reward effects.

Findings demonstrate that hippocampal astrocytes detect and react to anxiogenic environments with Ca2+ elevations and, when activated, causally reduce anxiety-like behavior. The behavioral anxiolysis aligns with increased excitatory synaptic transmission in DG granule cells and elevated neuronal activation (c-Fos). Mechanistically, astrocyte-derived ATP acting on purinergic receptors increases sEPSC frequency, mediating at least part of the anxiolytic effect—particularly open-arm exploration in EPM and the persistence of effects after stimulation. Elevated extracellular K+ from optogenetic activation contributes to inward currents but does not explain selective increases in sEPSC frequency, supporting a gliotransmitter mechanism. Effects were observed with both dorsal and ventral astrocyte stimulation, suggesting astrocyte modulation of affective processing across the hippocampal axis and potential network-level propagation (e.g., via astrocytic gap junctions or recurrent DG circuitry). The results provide a framework linking astrocytic activity and gliotransmission to affective behavior and suggest astrocytes as targets for anxiolytic interventions.

The study reveals that hippocampal astrocytes respond to anxiogenic visual contexts with intracellular Ca2+ increases and that their optogenetic activation produces anxiolytic behavioral effects. These effects are mediated by astrocyte-derived ATP, which enhances excitatory synaptic transmission in DG granule cells. The work establishes a mechanistic link between hippocampal astrocyte activity, ATP signaling, synaptic homeostasis, and anxiety-like behavior, highlighting astrocytes as potential therapeutic targets for anxiety disorders. Future studies should define upstream triggers of astrocyte Ca2+ signals in anxiogenic states, delineate purinergic receptor subtypes and downstream circuits, resolve cellular and spatial specificity of ATP release, and investigate long-lasting behavioral effects and potential impacts on related behaviors such as social interaction and memory under varying contexts.

• The VR-based anxiogenic paradigm, while validated behaviorally, may not fully replicate real-world EPM contexts.
• Cellular and spatial resolution of ATP measurements was limited; precise identification of releasing astrocyte subpopulations and target synapses remains unresolved.
• PPADS did not block all behavioral consequences (e.g., OFT exploration increases, induction of center time/speed), suggesting incomplete blockade in vivo, involvement of additional gliotransmitters, or PPADS-insensitive purinergic/adenosine receptors.
• The mechanism underlying persistence of anxiolytic effects after light offset is unknown; potential synaptic plasticity was not directly tested.
• Contributions of elevated extracellular K+ and possible off-target effects of prolonged light were explored but not exhaustively quantified across cell types and regions.
• Sex- and age-specific effects were not assessed (male mice only), limiting generalizability.