Prefrontal cortex astrocytes modulate distinct neuronal populations to control anxiety-like behavior

Kim, Eunyoung; Du, Hairuo; Tan, Yanqi; Brown, Brandon L.; Chang, Yaowen; Díaz-Castro, Blanca; Sweedler, Jonathan V.; Yu, Xinzhu

doi:10.1038/s41467-025-63131-9

Download PDF

Article
Open access
Published: 21 August 2025

Prefrontal cortex astrocytes modulate distinct neuronal populations to control anxiety-like behavior

Nature Communications volume 16, Article number: 7819 (2025) Cite this article

6638 Accesses
6 Altmetric
Metrics details

Subjects

Abstract

Accumulating evidence has supported diverse regulatory functions of astrocytes in different neural circuits as well as various aspects of complex behaviors. However, little is known about how astrocytes regulate different neuronal subpopulations that are linked to specific behavioral aspects within a single brain region. Here, we show that astrocytes in the medial prefrontal cortex (mPFC) encode anxiogenic environmental cues in freely behaving mice. Silencing mPFC astrocyte Ca²⁺ signaling heightens anxiety-like behavior and triggers opposing functional responses in excitatory and inhibitory neurons. Moreover, neuronal subpopulations tuned to anxiety-like behavior are differentially modulated by mPFC astrocytes at single cell and network levels. Using cell type-specific proximity biotinylation approaches, we identified significant intracellular and intercellular proteomic alterations in mPFC astrocytes and at the astrocyte-neuron interface associated with anxiety. Collectively, our findings uncover mechanisms underpinning the heterogenous astrocyte-neuron interaction that is behaviorally relevant and offer critical insights into the pathophysiology of emotional disorders.

Neonatal cortical astrocytes possess intrinsic potential in neuronal conversion in defined media

Article Open access 05 February 2021

Multicore fiber optic imaging reveals that astrocyte calcium activity in the mouse cerebral cortex is modulated by internal motivational state

Article Open access 08 April 2024

Three-photon in vivo imaging of neurons and glia in the medial prefrontal cortex with sub-cellular resolution

Article Open access 23 May 2025

Introduction

The mPFC plays a crucial role in cognitive flexibility, decision-making, and socioemotional regulation, all fundamental components of anxiety^1,2,3,4. Both clinical and basic research studies have reported that defective prefrontal function results in aberrant processing of potentially aversive stimuli, presumably contributing to the core symptoms of pathological anxiety^5,6. It is generally believed that an imbalance between excitatory and inhibitory systems in the mPFC underlies anxiety behavior⁷, but information from the published literature is contradictory and inconclusive. For instance, it was found that neuronal activities shifted towards excitation in the mPFC of mice exposed to anxiogenic stimuli^8,9, whereas evidence of mPFC hypoactivity arising from increased GABAergic neuronal activity was found to be accompanied by stress-induced anxiety^10,11. These contradictory findings imply a complex interplay among different populations of prefrontal neurons that may be further modulated by additional neural circuit components such as glial cells.

Astrocytes are an abundant glial cell type in the central nervous system, where they intimately associate with neurons to monitor, integrate, and maintain neuronal activity^12,13. A key mechanism of such interactions is astrocyte Ca²⁺ signaling^14,15, which has both intracellular and intercellular functions. Astrocyte Ca²⁺ signaling is evoked by neurotransmitters, and is altered in distinct neurological and psychiatric disorders^16,17. It has been shown that genetic deletion of a transcription factor in astrocytes resulted in abnormal synaptic transmission of mPFC neurons and increased anxiety-like behavior¹⁸. However, whether and how mPFC astrocyte Ca²⁺ signaling modulates different neuronal populations in aiding or inhibiting anxiety-like behavior remains largely unknown.

In this study, we found a critical role for mPFC astrocyte Ca²⁺ signaling in modulating anxiety-like behavior in mice and further identified underlying cellular and molecular mechanisms. We first revealed a physiological relationship between mPFC astrocyte Ca²⁺ signaling and anxiety-like behavior in freely behaving mice. Using a genetic approach, we found that silencing astrocyte Ca²⁺ signaling in the mPFC promoted anxiety-like behavior, presumably via increasing and decreasing the activities of excitatory and inhibitory neurons, respectively. Interestingly, neuronal ensembles associated with physiological correlates of anxiety-like behavior were altered by silencing mPFC astrocyte Ca²⁺ signaling in a task- and cell-specific manner. Finally, we provide a comprehensive and unbiased characterization of molecular alterations within mPFC astrocytes as well as at the astrocyte-neuron interface with proximity labeling proteomic approaches. Together, our data indicate that astrocytes are essential regulators of anxiety-relevant neural circuits.

Results

mPFC astrocyte Ca²⁺ responses to anxiety

To determine whether and how mPFC astrocytes respond to naturally occurring anxiety-like behavior, we monitored Ca²⁺ signals from a group of mPFC astrocytes in freely behaving mice with fiber photometry (Fig. 1a). An optical fiber was implanted in the mPFC of wild-type mice (C57BL/6J) at 4–6 weeks of age following microinjection of an adeno-associated virus (AAV) with an astrocyte-specific promoter GfaABC₁D to express a genetically encoded green Ca²⁺ indicator GCaMP6f in mPFC astrocytes (Fig. 1b). Ca²⁺ signals were recorded while mice explored different compartments (closed arms, open arms, and a center platform) of an elevated plus maze (EPM) over 10 min, which is a well-established method to assess anxiety-like behavior in rodents¹⁹ (Fig. 1c). The average numbers of entry were 6.1 ± 0.8 for the open arm and 17.8 ± 2.1 for the closed arms. Strikingly, mPFC astrocytes exhibited significantly elevated Ca²⁺ signals during exploration in the anxiogenic open arms, but not in the closed arms (Fig. 1d–f). To further explore this observation, we employed an elevated zero maze (EZM), which eliminated the center platform and was composed of adjacent open and closed quadrants (Fig. 1g). Consistently, Ca²⁺ signals in the mPFC astrocytes were considerably greater during exploration in the open area than the closed area (Fig. 1h, i). The increase in astrocyte Ca²⁺ signals appeared to precede open arm/area entry, suggesting the presence of additional neuromodulatory inputs triggered by anxiogenic cues, such as norepinephrine^20,21,22,23, that may subsequently activate mPFC astrocytes during behavior. Together, these data indicate a physiological correlation between prefrontal astrocyte Ca²⁺ signaling and distinct anxiety states.

**Fig. 1: mPFC astrocyte Ca²⁺ signaling is associated with anxiety-like behavior in vivo.**

To investigate whether astrocyte Ca²⁺ responses to open area exploration in anxiety-like behavioral tests were simply mirroring local population-level neuronal activities, we performed in vivo Ca²⁺ imaging of mPFC neurons with fiber photometry (Supplementary Fig. 1a). Immunohistochemistry (IHC) analysis showed that more than 96.8% of NeuN-positive cells express a red Ca²⁺ indicator jRGECO1a (Supplementary Fig. 1b). Unlike astrocytes, neurons imaged in the mPFC as a group did not exhibit differential activity between open and closed arms of the EPM (Supplementary Fig. 1c–e). This phenotype was also observed in different neuronal populations expressing GCaMP6f with AAVs carrying hSyn, CaMKII, and mDlx promoters, respectively (Supplementary Fig. 1f–o). Collectively, these results suggest different Ca²⁺ activities during anxiety-like behavior between neurons and astrocytes on the population level that could arise from intrinsically low and asynchronous firing rates of prefrontal neurons^24,25 versus extensive astrocytic network that is tightly coupled by gap junction²⁶.

In addition to emotional regulation, the mPFC is engaged in a range of cognitive processes including working memory²⁷, which is sensitive to emotional states. To examine whether mPFC astrocytes were also involved in spatial working memory, we analyzed astrocyte Ca²⁺ signals in a T-maze spontaneous alternation test, which is based on the natural propensity of rodents to explore a novel arm over a familiar one and does not require reward or pre-training²⁸ (Fig. 1j). After 8 days of habituation, individual mice were placed at the start arm and allowed to explore either the left or right horizontal “goal” arms across multiple trials on a single day. Error alternation was defined as choosing the same arm as the previous trial, while correct alternation was defined as choosing the unexplored arm from the previous trial. Interestingly, mPFC astrocytes showed comparable Ca²⁺ responses between correct and error alternations (Fig. 1k, l), suggesting that strong astrocyte recruitment is not necessary during this task for working memory unlike the EPM and EZM tests for anxiety.

Circuit-specific silencing of mPFC astrocytes

The overall increased activity of prefrontal astrocyte Ca²⁺ signals during exploration of anxiety-evoking areas could be interpreted in two ways: (1) as a response to anxiogenic inputs; or (2) as an innate motivational drive to facilitate approach behavior. To discriminate between these two possibilities, we performed a loss-of-function manipulation of mPFC astrocyte Ca²⁺ signaling and investigated the effects on anxiety-like behavior (Fig. 2a).

**Fig. 2: Reducing Ca²⁺ signaling of mPFC astrocyte enhances anxiety-like behavior.**

To silence mPFC astrocyte activity, we employed a Cre-dependent transgenic mouse line (CalEx^flox mice)²⁹ that we recently developed and characterized to express an efficient Ca²⁺ pump, CalEx, on the plasma membrane to remove Ca²⁺ without causing astrogliosis^29,30. Either a control (tdTomato) or Cre AAV with the astrocyte-specific promoter (GfaABC₁D) was delivered bilaterally into the mPFC of wild-type or CalEx^flox mice (Fig. 2b). Subsequent imaging and behavioral experiments were conducted 2–3 weeks after the surgeries. IHC analysis confirmed the cell-type and anatomical specificity of CalEx expression (Fig. 2c). We refer to CalEx mice as those with CalEx expression in the mPFC astrocytes. To functionally validate CalEx expression, we co-expressed GCaMP6f in mPFC astrocytes (Fig. 2d). Ex vivo Ca²⁺ imaging with confocal microscopy revealed a robust reduction in spontaneous Ca²⁺ signals across all astrocyte compartments in CalEx mice compared to controls, demonstrating the efficacy of the CalEx pump in causing broad suppression of astrocyte Ca²⁺ activity (Fig. 2e, f, and Supplementary Fig. 2a–c and Supplementary Table 1). Consistently, an event-based analysis³¹ detected significantly fewer Ca²⁺ events in CalEx mPFC astrocytes (274.75 ± 52.15 events vs. 37.65 ± 6.48 events, P < 0.0001 with Mann-Whitney test, Fig. 2g and Supplementary Videos 1, 2). Moreover, the spatial and temporal density of co-occurring astrocyte Ca²⁺ events was greatly reduced in CalEx mice (Fig. 2h). To confirm whether astrocyte Ca²⁺ signals were attenuated by CalEx in vivo, we recorded from CalEx mice subjected to the EPM test with fiber photometry and did not detect any Ca²⁺ responses to open arm exploration in CalEx mice (Fig. 2i, j). Together, these results confirm that CalEx expression resulted in a broad reduction of spontaneous Ca²⁺ signaling specifically in mPFC astrocytes both ex vivo and in vivo.

While the basal Ca²⁺ activity was markedly reduced with CalEx, it may be possible that receptor-mediated events are preserved. To assess this, we applied phenylephrine (PE), a potent selective α1-adrenoceptor agonist, to specifically activate G protein-coupled receptor (GPCR) pathways that normally elevate intracellular Ca²⁺ levels³². We found that control astrocytes showed robust Ca²⁺ responses evoked by PE application (10 μM), whereas CalEx astrocytes displayed minimal Ca²⁺ changes that corresponded to ~90% reduction of the control PE responses (Supplementary Fig. 2d–f and Supplementary Table 1). Collectively, these data demonstrate that CalEx substantially attenuated both spontaneous and GPCR-mediated Ca²⁺ signals in the mPFC astrocytes.

Heightened anxiety-like behavior by silencing mPFC astrocyte Ca²⁺ signaling

To evaluate whether global mPFC astrocyte Ca²⁺ attenuation had any effects on anxiety-like behavior, we conducted a battery of behavioral tests. In an open field test (OFT), CalEx mice travelled similar distance as the control mice (Fig. 2k, l). However, they spent significantly less time in the center of the arena (Fig. 2m), which is an indicator of increased anxiety levels in rodents³³. In the EPM, CalEx mice showed decreased duration and number of entries to the open arms compared with the control mice (Fig. 2n, o). Consistently, in the EZM, CalEx mice spent significantly less time exploring the open areas than the control mice (Fig. 2p, q). Thus, silencing mPFC astrocyte Ca²⁺ signaling enhanced anxiety-like behavior without altering locomotion activities, indicating a necessary role of mPFC astrocytes in maintaining normal anxiety responses¹⁸.

Both anxiety and depression are associated with mPFC dysfunction and frequently co-morbid⁷. To assess whether attenuated Ca²⁺ signals in the mPFC astrocytes also induced depression-like phenotypes, we performed sucrose preference test (SPT), forced swim test (FST), and tail suspension test (TST). In the SPT, we found comparable preference ratio of water and sucrose consumption between control and CalEx mice (Supplementary Fig. 3a, b), suggesting that silencing Ca²⁺ signaling in mPFC astrocytes did not induce anhedonia-like behavior, which reflects the inability to experience pleasure and is a core symptom of major depressive disorder³⁴. Consistently, CalEx mice behaved similarly in the FST and TST as control mice (Supplementary Fig. 3c–f), demonstrating that attenuated Ca²⁺ signals in the mPFC astrocytes did not produce depressive-like behavior. Moreover, CalEx mice performed equivalently as control mice in the T-maze spontaneous alternation test (Supplementary Fig. 3g, h), suggesting that short-term spatial working memory was unaffected by silencing prefrontal astrocyte Ca²⁺ signaling, consistent with the absence of correlated astrocyte Ca²⁺ activity during spontaneous alternation (Fig. 1j–l).

Our previous studies showed that silencing astrocyte Ca²⁺ signaling with CalEx in the dorsolateral striatum triggered excessive self-grooming behavior^29,30. However, we did not find any changes in innate behaviors including self-grooming and rearing when attenuating astrocyte Ca²⁺ signaling specifically in the mPFC (Supplementary Fig. 3i–k), highlighting circuit-specific effects of astrocytes on behavior.

Collectively, these data establish a physiological relationship between astrocyte Ca²⁺ signaling and approach-avoidance behavior under anxiogenic conditions and further suggest that mPFC astrocytes represent an essential cellular component of anxiety regulation.

Behavioral changes by chemogenetic activation of mPFC astrocytes

To investigate the potential impact of activating mPFC astrocytes on anxiety-like behavior, we utilized Designer Receptors Exclusively Activated by Designer Drugs (DREADDs)³⁵. Either hM3Dq DREADD or control protein (tdTomato) was expressed in the mPFC astrocytes of wild-type mice with AAV microinjection 2–3 weeks before subsequent evaluation (Supplementary Fig. 4a). hM3Dq expression was astrocyte-specific and detected in ~99.2% of mPFC astrocytes (Supplementary Fig. 4b). Systemic administration of a synthetic ligand, clozapine N-oxide (CNO, 1 mg/kg, i.p. injection), upregulated c-Fos expression in mPFC astrocytes expressing hM3Dq 2 h but not 30 min after CNO injection (Supplementary Fig. 4c), consistent with previous findings from various brain regions^36,37,38,39. Moreover, ex vivo Ca²⁺ imaging showed that CNO application (1 µM) evoked robust Ca²⁺ responses in hM3Dq-expressing mPFC astrocytes but not in control mPFC astrocytes (Supplementary Fig. 4d–g), as an expected consequence of activating Gq-coupled GPCR pathways in astrocytes. Thus, hM3Dq DREADD expression allows for the selective activation of Ca²⁺ signaling in the mPFC astrocytes.

Next, we prepared mice with bilateral expression of hM3Dq or control protein in mPFC astrocytes and assessed anxiety-like behavior 30 min and 2 h after CNO administration (1 mg/kg, i.p. injection). In the OFT, hM3Dq + CNO mice showed increased ambulation compared with control + CNO mice 2 h but not 30 min after the CNO injection (Supplementary Fig. 4h, i, o, p), in line with different levels of cellular activation revealed by c-Fos expression (Supplementary Fig. 4c). In contrast, there was no difference in time spent in the center between hM3Dq + CNO and control + CNO groups at either time point (Supplementary Fig. 4j, q). Consistently, hM3Dq + CNO mice had similar performance as control + CNO mice in the EPM and EZM tests 30 min and 2 h after CNO injection (Supplementary Fig. 4k–n, r–u), indicative of normal anxiety-like behavior.

Together with the data showing heightened anxiety-like behavior by silencing mPFC astrocytes (Fig. 2), these results suggest a necessary role of mPFC astrocytes in maintaining normal anxiety states, but their activation alone is not sufficient to further reduce anxiety-like behavior beyond the baseline.

Opposing effects of mPFC astrocyte Ca²⁺ silencing on excitatory and inhibitory neurons

Astrocytic territories encompass synapses from both excitatory and inhibitory neurons. To explore how silencing astrocyte Ca²⁺ signaling affected dynamics of defined neuronal populations in the mPFC, we performed in vivo Ca²⁺ imaging of glutamatergic and GABAergic neurons in freely behaving mice that were subjected to the EZM with head-mounted microendoscopes (Fig. 3a). To image neural dynamics from the mPFC, we implanted a gradient index (GRIN) lens following the AAV microinjection to express the Ca²⁺ indicator GCaMP6f in either glutamatergic or GABAergic neurons of control and CalEx mice (Fig. 3b).

**Fig. 3: Opposite changes of excitatory and inhibitory neuronal activities by silencing mPFC astrocytes.**

First, we examined the overall activities of different neuronal populations based on the locations of mice on the EZM (Fig. 3c, g). We found that glutamatergic neurons in CalEx mice displayed significantly greater Ca²⁺ signals compared to control mice during the exploration of the closed area but not the open area (Fig. 3d), suggesting location-specific neuronal alterations by silencing astrocytes. Because neuronal Ca²⁺ activity is associated with bursts of action potentials⁴⁰, this result suggests hyperactivity of glutamatergic neurons in the mPFC caused by silencing astrocytes. In contrast, GABAergic neurons in CalEx mice showed significantly reduced Ca²⁺ activity during the exploration of both the closed and open areas (Fig. 3h), indicative of a hypoactivity of inhibitory neurons. Thus, silencing astrocyte Ca²⁺ signaling in the mPFC exerts opposing effects on excitatory and inhibitory neurons, leading to an overall prefrontal over-excitation.

Next, to investigate the potential impact of silencing mPFC astrocytes on synaptic connections, we analyzed both excitatory and inhibitory synapses with confocal microscopy. Presynaptic marker VGLUT1 and postsynaptic marker PSD95 were used to label glutamatergic synapses, while presynaptic marker VGAT and postsynaptic marker gephyrin (GPHN) were used to identify GABAergic synapses (Fig. 3e, i). We quantified puncta numbers of individual presynaptic, postsynaptic, and colocalized puncta within individual astrocytic territories in the mPFC from both control and CalEx mice. The density (per 10 μm³) of VGLUT1, PSD95, and colocalized (VGLUT1 + PSD95) puncta of glutamatergic synapses were significantly greater in CalEx mice compared to control mice (Fig. 3f). On the other hand, density of markers of GABAergic synapses (VGAT, GPHN, and VGAT + GPHN colocalization) was significantly reduced in CalEx mice (Fig. 3j). These data are consistent with neuronal activity changes in vivo revealed by the microendoscopic Ca²⁺ analysis.

Taken together, these results suggest that silencing mPFC astrocyte activity differentially regulates excitatory and inhibitory neurons at both the cellular and synaptic levels, leading to an imbalance of excitation and inhibition in the prefrontal neural circuit that potentially underpins the heightened anxiety state.

Changes of neuronal subpopulations encoding anxiety

The observation that mPFC neurons imaged as a whole with fiber photometry did not exhibit dynamic patterns associated with anxiety-like behavior (Supplementary Fig. 1) raised a possibility that the encoding of anxiety-like behavior is associated with unique subsets of mPFC neurons. To test this idea, we examined whether any individual neurons imaged with microendoscopes responded to open and closed area exploration in the EZM, using a receiver operating characteristic (ROC) analysis⁴¹. Of all the imaged glutamatergic neurons, 7.9% were tuned to open area exploration and 4.0% were tuned to closed area exploration in the control mice (Fig. 4a, b and Supplementary Fig. 5a–e). Cells that were active, but did not show selective tuning to either open or closed intervals were referred to as “other cells”. Thus, identification of behavior-related neuronal subsets was dependent on specific time-locked responses. Moreover, open area-excited glutamatergic neurons exhibited greater Ca²⁺ activities than closed area-excited neurons (Fig. 4c), suggesting distinct intrinsic properties of neuronal ensembles encoding different measures of anxiety-like behavior. We found that silencing mPFC astrocytes did not change the distribution or mean activities of glutamatergic neurons that were behaviorally tuned (Fig. 4b, c). Next, we examined whether the activity patterns of GABAergic neurons were affected by silencing mPFC astrocytes. In the control group, we identified 20.5% and 2.7% of GABAergic neurons tuned to open area and closed area intervals, respectively (Fig. 4i, j and Supplementary Fig. 5f–j). This distribution pattern was unaltered in CalEx mice. Interestingly, while open area-excited and closed area-excited GABAergic neurons exhibited similar activity levels in control mice, open area-excited GABAergic neurons displayed significantly greater activity in CalEx mice (Fig. 4k). These results suggest that silencing mPFC astrocytes differentially changes excitatory and inhibitory neurons that selectively encode anxiety-like behavior.

**Fig. 4: Distinct changes of excitatory and inhibitory neuronal subpopulations encoding anxiety-like behavior.**

It has been shown that cortical astrocytes are capable of network-level communication in response to neuronal inputs⁴², raising a possibility that silencing mPFC astrocytes may alter information representation by prefrontal neuronal networks on a population level. To address this, we constructed a binary Fisher’s Linear Discriminant (FLD) model to predict general measures of anxiety-like behavior (open vs. closed areas) based on population activities. Open area locations in the EZM were predicted by decoders trained with Ca²⁺ signals of glutamatergic neurons (Fig. 4d, e), with significantly greater accuracy than models constructed using data with randomized labels (Fig. 4f). Similarly, decoders trained with Ca²⁺ signals of GABAergic neurons achieved significantly higher performance than shuffled data (Fig. 4l–n). Interestingly, CalEx expression in mPFC astrocytes did not compromise the decoding accuracy (Fig. 4f, n), indicating that neural representations of general information of anxiety-like behavior are stable and resilient to attenuated astrocyte signaling.

In the EZM, we noticed that mice often wandered at the transition zones from closed areas to open areas instead of moving directly across. Based on this observation, we next examined whether neuronal activities and information encoded were distinct between the transition and the non-transition zones (Fig. 4g). We defined transition zones as within 0.5 s before entering open areas and non-transition zones as > 2 s after entering and before leaving the closed areas. For both glutamatergic and GABAergic neurons, open area-excited cells displayed significantly larger Ca²⁺ signals in the transition zone than non-transition zones (Fig. 4g, o). Interestingly, the opposite was observed for closed area-excited cells, i.e., smaller amplitudes of neuronal Ca²⁺ activity in the transition zones. While closed area-excited GABAergic neurons retained this pattern in CalEx mice (Fig. 4o), closed area-excited glutamatergic neurons did not discriminate between transition and non-transition zones anymore (Fig. 4g). Moreover, using a random forest model, decoders trained with GABAergic neuronal data significantly outperformed models constructed with shuffled data (Fig. 4p). In contrast, decoders using glutamatergic neuronal data failed to predict transition zones in CalEx group (Fig. 4h), suggesting a loss of information encoded in the glutamatergic neurons to represent defined locations associated with different anxiety levels.

Together, these data demonstrate that silencing mPFC astrocytes impairs network dynamics of excitatory neurons that selectively encode information of the transition zone, i.e., the location presumably associates with decision-making into the anxiogenic area. These findings also highlight that prefrontal astrocytes differentially modulate the activity of excitatory and inhibitory neurons at both the single cell and population levels.

Proteomic alterations of mPFC astrocytes

Our findings indicate an imperative role of astrocyte Ca²⁺ signaling in regulating anxiety-like behavior (Figs. 1 and 2). Given that many signaling pathways and pathological events are protein-driven, to comprehensively understand the molecular mechanisms underlying astrocytic regulation of anxiety, we performed cell-specific proteomics with a proximity labeling approach to determine protein changes induced by silencing mPFC astrocytes. To capture astrocyte protein identities in vivo, we generated an AAV to express a biotin ligase called TurboID⁴³ in the cytosol of mPFC astrocytes (Fig. 5a). Biotin was given 3 weeks after the AAV microinjection. In the presence of biotin, TurboID biotinylated adjacent proteins that were subsequently purified with streptavidin-conjugated beads (Fig. 5b).

**Fig. 5: Astrocyte-specific proteomes reveals cytosolic changes by silencing Ca²⁺ signaling.**

Using liquid chromatography tandem mass spectrometry (LC-MS/MS), we identified 955 proteins from control and CalEx mPFC astrocytes (n = 5 replicates, each replicate contained mPFC tissue from 3 mice). From these detected proteins, 358 proteins were commonly shared by control and CalEx mPFC astrocytes, and 247 and 350 proteins were specific to control and CalEx mPFC astrocytes, respectively (Fig. 5c). Based on the top 100 most abundant common proteins, we assembled a putative association network and highlighted proteins related to several core functions of astrocytes, including metabolism, synapse regulation, homeostasis maintenance, and cell morphogenesis (Fig. 5d). Within the association network, many proteins displayed relative changes between the control and CalEx astrocytes, coded by color. Gene ontology (GO) biological pathway analysis and the KEGG and reactome pathway analysis showed enrichments in pathways related to synaptic function and neurological disorders (Supplementary Fig. 6a), highlighting physiological importance of mPFC astrocytes in these processes.

By performing label free quantification (LFQ), we found that 71 out of the 358 common proteins were differentially regulated between control and CalEx mPFC astrocytes (P < 0.05), with 25 proteins upregulated and 46 proteins downregulated in CalEx mPFC astrocytes (Fig. 5e, and Supplementary Fig. 6b and Supplementary Data 1). The top differentially regulated proteins were assigned multiple biological functions (Fig. 5f). For instance, metabolic proteins (PCCA, ATPB and ENOA) were upregulated in CalEx mPFC astrocytes, while scaffold proteins (MAP4 and MAP6) and proteins related to neuroinflammation and neurological diseases (TRIM9, TAU and SYUA) were downregulated in CalEx mPFC astrocytes. Among identified unique proteins, cell adhesion molecules (NLGN2 and NLGN3) and GPCRs (GRM3 and GRM5) were found to be specific to CalEx mPFC astrocytes. In contrast, synapse modifying proteins (MEG10 and MERTK) were only detected in control mPFC astrocytes (Fig. 5g). Finally, the differences between control and CalEx mPFC astrocytes were noted in the top enriched biological pathways as well as the classification of protein classes (Fig. 5h and Supplementary Fig. 6c, d).

We also highlighted proteins related to the synthesis and transport of neurotransmitters including glutamate, GABA, and ATP, that were present in both control and CalEx mPFC astrocytes (Supplementary Fig. 7a). Most of these proteins were not differentially regulated by CalEx expression, such as glutamine synthetase (GLNA/GS), glial glutamate transporters (EAA1/GLAST and EAA2/GLT-1), and GABA transporter (S6A11/GAT3). Interestingly, ATP synthase subunit beta (ATPB) was significantly upregulated in CalEx mPFC astrocytes (Supplementary Fig. 7b), suggesting a potential change of purinergic pathways by silencing astrocyte Ca²⁺ signaling. Next, to explore potential mechanisms mediating neuroactive substance release from astrocytes, we analyzed proteins involved in vesicular transport, neurotransmitter homeostasis, as well as core physiological functions related to ion, glucose, and lipid homeostasis by astrocytes^38,44,45 (Supplementary Fig. 7c, d). We identified 34 proteins related to vesicular transport, most of which were comparable between control and CalEx astrocytes. Two proteins (Complexin 1/CPLX1 and Piccolo/PCLO) were differentially regulated between control and CalEx astrocytes. PCLO was found to negatively regulate synaptic vesicle exocytosis⁴⁶ and was detected in a small subpopulation of hippocampal astrocytes capable of releasing glutamate via exocytosis⁴⁵. CPLX1 is involved in vesicle fusion and neurotransmitter release at synapses⁴⁷, yet its role in astrocytes remains to be discovered. In addition, we identified 19 proteins involved in neurotransmitter transport and metabolism with 3 proteins (VGLU2, SC6A9, and GABT) uniquely expressed in CalEx astrocytes, suggesting potential changes in neurotransmitter balance caused by astrocyte Ca²⁺ silencing that may underlie neuronal and synaptic alterations. Lastly, we constructed protein networks based on GO enrichment analysis of proteins and pathways related to exocytosis, a key process that potentially regulates neuroactive substance release (Supplementary Fig. 7e). Further examinations will be needed to determine whether and how these proteins may mediate astrocyte-neuron crosstalk in the context of anxiety.

Moreover, several proteins involved in microtubule organization, including microtubule-associated proteins MAP4, MAP6, and TAU, were downregulated in CalEx astrocytes. In addition, ATAT (alpha-tubulin N-acetyltransferase 1) and MAP9 (microtubule-associated protein 9) were uniquely expressed in control astrocytes. These scaffolding protein changes may underlie the reduced territory size in CalEx-expressing mPFC astrocytes (Supplementary Fig. 8). In contrast, GFAP (glial fibrillary acidic protein), an intermediate filament protein that is predominantly expressed in the somata and major branches of astrocytes, was unaffected by CalEx expression. Consistently, CalEx astrocytes displayed normal cell body sizes and numbers of major branches (Supplementary Fig. 8).

The deficiency of NLGN2 in cortical astrocytes was found to suppress excitatory synaptic transmission while promoting inhibitory synaptic transmission⁴⁸. Consistently, our data show that NLGN2 was specifically enriched in CalEx mPFC astrocytes, accompanied by increased excitation and decreased inhibition of neuronal function (Fig. 3). Moreover, astrocytes were found to actively engulf synapses in an activity-dependent manner via the MEGF10 and MERTK phagocytic receptors⁴⁹, which both appeared to be uniquely expressed in control mPFC astrocytes but not CalEx mPFC astrocytes, suggesting a dysregulation of phagocytosis-mediated synapse elimination without proper astrocyte activity. Our results showing increased excitatory synaptic density in the mPFC of CalEx mice also supported previous findings that mice lacking MEGF10 in astrocytes had significantly more excitatory synapses in the hippocampus⁴⁹.

Taken together, these data demonstrate multifaceted changes of astrocyte proteomes on individual protein, signaling pathway, and protein network levels, and suggest that the physiological functions of mPFC astrocytes may be altered in multiple domains to accompany the synaptic and neuronal alterations underpinning the anxiety phenotype.

Proteomic alterations at astrocyte-neuron interface

The finding that mPFC astrocytes differentially modulated glutamatergic and GABAergic neurons associated with anxiety-like behavioral measures (Figs. 3 and 4) raises the question of how astrocytes differentially regulate excitatory and inhibitory neurons. To answer this, we adopted a split-TurboID approach⁵⁰ to characterize the composition of proteomes at the interface between astrocytes and neurons in the mPFC. We generated AAVs expressing N- and C-terminal TurboID fragments specifically in neurons and astrocytes, respectively (Fig. 6a). Expression of both N- and C-terminal fragments at the astrocyte-neuron interface reconstituted the enzymatic activity of TurboID, reflected by stronger streptavidin signals from the IHC analysis (Fig. 6b). This allows us to isolate proteins specifically from the interacting locations between astrocytes and neurons within the mPFC excluding distal projections.

**Fig. 6: Astrocyte-neuron interface proteomes identifies cell-specific changes.**

Using this split-TurboID approach followed by an LC-MS downstream analysis, we identified 112 common interface proteins, 108 interface proteins unique to the control group, and 72 interface proteins unique to the CalEx group (each replicate contained mPFC tissue from 4 mice; Fig. 6c). The most abundant common interface proteins were associated with key biological functions, such as ion transport, secretion, and cell development (Supplementary Fig. 9a). Moreover, 16 out of the 112 common interface proteins displayed differential regulation between control and CalEx groups (P < 0.05; Fig. 6d, Supplementary Fig. 9b and Supplementary Data 2). One of the most significantly enriched proteins in the CalEx group appeared to be AT2B2, which was the plasma membrane Ca²⁺ pump CalEx, proving the validity of our proteomics approach. In addition, lactate dehydrogenase (LDHA), a key metabolic enzyme in lactate synthesis was downregulated in the CalEx group, suggesting a disruption in energy supply between astrocytes and neurons. Interestingly, several synaptic proteins, including MPP2, SHAN3, STXB5 and VGLU2, were uniquely identified at the astrocyte-neuron interface in either control or CalEx groups (Fig. 6d and Supplementary Fig. 9c, d), suggesting relocations of synaptic proteins as a consequence of silencing astrocyte Ca²⁺ signaling. To further validate the proteomic analysis in the brain tissue, we performed IHC analysis based on antibody availability. A significant reduction of LDHA expression was detected in the mPFC of CalEx mice, whereas increased VGLU2 expression was found in the territories of CalEx mPFC astrocytes (Supplementary Fig. 9e), in accordance with the proteomic data.

Next, to investigate the cell-specific composition of proteins detected at the astrocyte-neuron interface, we performed cell-type classification of interface proteins based on our cytosolic astrocyte proteome analysis as well as published proteomic datasets^51,52,53,54. Out of all the identified interface proteins, 85 proteins were found to be expressed in astrocytes, 121 proteins expressed in excitatory neurons and 71 proteins expressed in inhibitory neurons (Fig. 6e). Protein association network analysis with differentially regulated interface proteins and unique interface proteins further revealed that proteins associated with biological processes, such as modulation of chemical synaptic transmission, were enriched across all cell types (Supplementary Fig. 9f). Additionally, cell-cell adhesion proteins were concentrated in astrocytes, proteins related to vesicle mediated transport were enriched in excitatory neurons, and proteins involved in regulation of postsynaptic organization were found in inhibitory neurons, indicative of cell type-specific physiological needs (Fig. 6f–h). Furthermore, cell-specific interface proteins associated with various functions displayed distinct changes by silencing mPFC astrocytes (Fig. 6i). Finally, to identify potential protein pairs mediating interactions between astrocytes and different neuronal populations, we referenced to a biological repository containing over 2 million known protein, genetic, and chemical interactions⁵⁵. Our interface proteomic data contained 126 pairs of protein interaction, including 88 pairs between astrocytes and excitatory neurons and 38 pairs between astrocytes and inhibitory neurons (Fig. 6j). Strikingly, more than 90% of the proteins constituting interaction pairs were uniquely found in either control or CalEx groups. For instance, DP13A (Adaptor protein, phosphotyrosine interacting with pH domain and leucine zipper 1, also called APPL1) was uniquely expressed at the astrocyte-neuron interface in CalEx group. It was found that DP13A/APPL1 was enriched at synapses and its overexpression increased synapse density and PSD95 clusters in hippocampal neurons⁵⁶, which is consistent with our results showing higher density of PSD95 puncta in the CalEx group. LIS1 (Lissencephaly-1, also called PAFAH1B1) was a unique protein identified at the astrocyte-neuron interface under the control condition. Mice with LIS1 deficiency displayed increased excitatory synaptic currents and decreased inhibitory synaptic currents in the dentate gyrus that resulted a shift of synaptic E/I balance^57,58. Thus, the loss of LIS1 at the astrocyte-neuron interface in the CalEx group could potentially contribute to the opposite activity changes of glutamatergic and GABAergic neurons during anxiety-like behavior. Collectively, these data indicate substantial cell-specific alterations at the astrocyte-neuron interface induced by silencing mPFC astrocytes that may underlie task-selective changes of neuronal activities associated with anxiety.

Links with human diseases

While our analysis identified changes of astrocyte proteins (Fig. 5) and astrocyte-neuron interface proteins triggered by silencing mPFC astrocytes (Fig. 6) in a mouse model, some significantly altered proteins had been identified as risk genes for neurological disorders in human, such as SYUA for Parkinson’s disease (PD) and TAU for Alzheimer’s disease (AD). To explore the association between astrocyte silencing-induced changes and mPFC-related neuropsychiatric disorders, we surveyed several databases (Supplementary Data 2) to identify disease-associated genes that were subsequently mapped with our proteomic data (Fig. 6k). Both astrocyte cytosolic proteins and astrocyte-neuron interface proteins were concentrated to multiple mPFC-related diseases, such as anxiety disorders, major depressive disorder (MDD), schizophrenia (SCZ), post-traumatic stress disorder (PTSD), and AD. For example, an extracellular matrix protein TNR (tenascin-R) was uniquely expressed at the astrocyte-neuron interface of control mice and identified as a risk factor for anxiety disorder, bipolar disorder, and SCZ. Consistently, the deficiency of TNR enhanced excitatory synaptic transmission and elevated anxiety-like behavior in mice^59,60. The water channel AQP4 (aquaporin 4), which is an essential regulator of water homeostasis, blood brain barrier, and neuronal excitability in the brain⁶¹, was downregulated in CalEx astrocytes and linked to multiple diseases such as MDD, SCZ, PTSD and dementia. Together, these results suggest that altered astrocyte Ca²⁺ signaling may be a common feature of diverse neurological and psychiatric disorders.

Discussion

Anxiety is an evolutionarily conserved and adaptive response to potential threats that is essential for survival. However, uncontrolled and persistently elevated anxiety leads to anxiety disorders, which are the most common and disabling mental health disorders worldwide with inadequate therapeutic options^62,63,64. Numerous studies indicate essential neuronal contributions to anxiety, highlighting an imbalance between excitatory and inhibitory systems in the mPFC⁷, yet the mechanisms of non-neuronal cells governing this emotional state remain poorly understood.

Combining in vivo Ca²⁺ imaging with genetic silencing of astrocytes, we identified an essential role of astrocyte Ca²⁺ signaling within the mPFC in anxiety regulation. mPFC astrocyte Ca²⁺ signals in freely behaving mice selectively responded to natural exploration of environments prone to trigger anxiety. On the population level, astrocyte Ca²⁺ changes did not merely mirror nearby neuronal responses and possibly convey information that is not present in population neuronal Ca²⁺ signaling. Moreover, silencing astrocyte Ca²⁺ signaling in the mPFC with a genetic loss-of-function approach enhanced anxiety-like behavior. Although continuous Ca²⁺ extrusion by CalEx could alter extracellular Ca²⁺ concentration, our previous study and others showed that CalEx expression in striatal and cortical astrocytes did not change fast synaptic transmission^30,65. This is presumably because the extracellular Ca²⁺ concentration is substantially larger ( ~ 10⁴ times) than the intracellular Ca²⁺ levels within astrocytes⁶⁶ and Ca²⁺ exported by CalEx is likely dissipated by diffusion into the extracellular space. Our finding also recalls a previous gain-of-function study showing that activation of hippocampal astrocytes with optogenetic stimulation increased open arm exploration in a VR-simulated EPM test⁶⁷. Neuronal projection from the ventral hippocampus to the mPFC has been implicated in anxiety-related firing properties and network synchrony^68,69. Whether and how astrocytes residing in distributed neural circuits of anxiety synergize to represent aversive information remains to be elucidated.

The role of mPFC astrocytes in anxiety-like behavior is not explained by a simple binary framework. Selective activation of Gq-GPCR pathways with hM3Dq DREADDs in mPFC astrocytes did not produce anxiolytic effects beyond the baseline level. These results corroborate previous findings that optogenetic and chemogenetic activation of mPFC astrocytes did not alter anxiety-like behavior^65,70. However, a previous study paradoxically reported that hM3Dq activation in mPFC astrocytes enhanced anxiety-like behavior⁷¹. This discrepancy may arise from differences in promoter specificity, CNO regimens, and stereotaxic coordinates. It is noteworthy that gain-of-function and loss-of-function manipulations of astrocytes do not necessarily yield opposing changes on the transcriptomic or behavioral levels as revealed by studies focusing on striatal astrocytes^38,72. In line with the striatal work, two recent studies reported that both activation of hippocampal astrocytes with channelrhodopsin-2 (ChR2) and attenuation of hippocampal astrocyte Ca²⁺ signaling using conditional Itpr2 knockout mice induced anxiolytic effects^67,73. Together, these data indicate that the influences of astrocytes on anxiety-like behavior is complex and multifaceted.

Our data show that suppression of mPFC astrocyte activity increased glutamatergic neuronal activity and decreased GABAergic neuronal activity, giving rise to an over-excitation in the mPFC that could underlie the heightened anxiety state. This suggests that mPFC astrocytes may engage in a homeostatic plasticity mechanism to maintain excitation and inhibition in balance. Conversely, neural networks with deficient astrocyte signaling may lose the ability to compensate and adjust neuronal activity changes triggered by environmental threats, socioemotional challenges, and pathological insults. This is of great importance because excitatory-inhibitory balance (E/I balance) of the prefrontal circuit are essential for various behaviors, including emotions, social behavior, and cognition. For instance, optogenetic elevation but not reduction of E/I balance in the mPFC of mice impaired sociability and conditioned fear learning⁷⁴. Chemogenetic inhibition of mediodorsal thalamic neurons led to elevated E/I balance in the mPFC by reducing inhibitory synaptic transmission and subsequently working memory deficits⁷⁵. Computational modeling revealed distinct patterns of decision-making dysfunction caused by increased and decreased E/I ratio⁷⁶. Furthermore, besides anxiety disorders, imbalance of excitation and inhibition of the prefrontal circuit has been implicated in many neurological and psychiatric disorders, such as autism spectrum disorders, MDD, SCZ, and AD^77,78,79. Further investigation into astrocyte-neuron crosstalk may provide critical insight into the neurobiological mechanisms underlying E/I balance regulation for essential physiological functions as well as novel strategies for restoring the E/I balance under disease conditions.

Silencing astrocyte Ca²⁺ signaling with CalEx in the striatum was demonstrated to reduce activity of GABAergic medium spiny neurons in freely behaving mice without altering fast synaptic transmission (EPSCs or IPSCs) or presynaptic neurotransmitter release probability in brain slices³⁰. Consistently, a recent study showed that CalEx expression in mPFC astrocytes did not change mEPSC or mIPSC of cortical neurons or extracellular glutamate and ATP levels under baseline condition⁶⁵. These electrophysiological and biochemical analyses in brain slices suggest minimal effects of silencing astrocytes on synaptic transmission or extracellular neurotransmitter levels. The discrepancy of neuronal changes between ex vivo and in vivo evaluations could be attributed to several factors. First, neuronal communication occurs in multiple ways that are independent from synaptic transmission, such as tonic modulation, extracellular vesicles, and membrane receptor trafficking⁸⁰. Second, neuronal activity in the intact brain relies on temporally coordinated inputs from multiple brain regions, as well as local neuromodulator and neurotransmitter levels that are influenced by environmental cues and behavioral contexts. These factors are largely absent or greatly suppressed in ex vivo brain slice preparations. As a result, ex vivo electrophysiological recordings may not fully capture the dynamic range of neuronal activity that occurs in a more integrated and complex in vivo environment.

Both astrocytes and neurons are remarkably diverse in their molecular composition within and across brain regions. Recently, interactions between neurons and molecularly defined astrocyte subpopulations have been found to drive animal behaviors. For instance, Crym-positive astrocytes enriched in the central striatum contributed to perseveration behavior via a selective presynaptic regulation of orbitofrontal cortex to striatum projections⁸¹. In the basolateral amygdala, a subset of neurons and their neighboring astrocytes coordinated transcriptomic changes to encode long-term fear memory⁸². In support of these findings, we found that silencing mPFC astrocytes differentially regulates glutamatergic and GABAergic neurons that were preferentially tuned to certain aspects of anxiety-like behavior. This cell type-specific modulation of mPFC astrocytes on neuronal function could be attributable to the intrinsic molecular heterogeneity of astrocytes and neurons, which collectively determine the diversity of signaling pathways mediating various responses to physiological alterations. It is also feasible that astrocytes may discriminate between synapses from distinct neuronal populations based on local signals, and therefore establish intercellular interactions in a synapse-specific manner. These models are not mutually exclusive and can be tested in future studies.

Lastly, using cell-specific proteomics, we identified protein expression profiles from astrocytes as well as at the astrocyte-neuron interface, providing insight on cell-autonomous and non-cell-autonomous mechanisms underlying anxiety phenotype. For instance, we identified astrocyte-specific and interface enriched proteins, such as NLGN2, MEGF10, DP13A and LIS1, that may account for the opposing effects of astrocyte silencing on excitatory and inhibitory neuronal activity. In addition, we also discovered cell adhesion proteins enriched at the astrocyte-neuron interface that could mediate astrocyte-neuron interactions. Importantly, cell adhesion proteins that were differentially regulated by silencing mPFC astrocytes, such as NLGN2 and NEGR1, have been implicated in multiple neuropsychiatric disorders^83,84. TAU and SYUA have been implicated in the most prevalent neurodegenerative disorders, AD and PD, but have not yet been examined in the context of anxiety. Surprisingly, both proteins appeared to be significantly downregulated by silencing mPFC astrocyte Ca²⁺ signaling. These data not only indicate essential physiological functions of astrocytes across various neurological and psychiatric diseases⁴⁴, but also suggest that astrocyte Ca²⁺ signaling pathway may represent potential therapeutic targets for diverse disorders dependent on the prefrontal function.

Methods

Mouse models

All animal experiments were approved by the Illinois Institutional Animal Care and Use Committee (IACUC) at the University of Illinois at Urbana-Champaign (UIUC) and executed under the National Institute of Health Guide for the Care and Use of Laboratory Animals. Both male and female mice were used in this study. Mice were managed by the Division of Animal Resources at UIUC with a 12 h light-dark cycle, 22 °C ambient temperature, 55% ambient humidity and ad libitum feeding. Wild-type C57BL/6 J mice and homozygous transgenic CalEx^flox mice were generated and maintained in in-house breeding colonies. PCR was used to determine the genotypes using genomic DNA extracted from ear biopsies. A forward primer 5′-CCTTTCTGGGAGTTCTCTGCTGC-3′ and a reverse primer 5′-CGTAAGTTATGTAACGCGGAACTCC-3′ (217 bp amplicon) were used to confirm the CalEx^flox allele.

Plasmid constructs

The cDNA of the TurboID enzyme gene was cloned into pZac2.1 containing the astrocyte specific promoter GfaABC₁D. V5-TurboID-NES-pCDNA3 was kindly shared by Dr. Alice Ting (Addgene Cat# 107169) and we cloned the cDNA of the TurboID enzyme gene into a pZac2.1 plasmid containing the astrocyte specific promoter GfaABC₁D. To do so, the sequence containing the V5 Tag-TurboID-HA Tag-nuclear export signal (NES) was amplified by PCR using the following primers: Forward: 5′ CCTCGAGCTCGGATCCATGGGCAAGCCCATCCCCAA 3′, Reverse: 5′ TAAGCGAATTGGATCCTTAGTCCAGGGTCAGGCGCTCCAGGGG 3′. The vector GfaABC₁D-BioID2-13xLinker-BioID2, kindly shared by Dr. Baljit S. Khakh, was cut with the restriction enzyme BamHI to remove the BioID2 gene. The V5-TurboID-NES-HA fragment was then inserted into the BamHI pZac2.1-GfaABC₁D vector using the In-fusion® HD Cloning kit (Takara Cat# 639650). Plasmids were confirmed by sequencing (Source Bioscience). pZac2.1-GfaABC₁D-TurboID plasmid was deposited to Addgene (Cat# 229038). pZac2.1-GfaABC₁D-TurboID(C)-HA-GPI and AAV-hSynI-V5-TurboID(N)-GPI was a gift from Scott Soderling (Addgene Cat# 163697 and Cat# 163696)⁵⁰.

Surgical procedures

All surgical procedures complied with the IACUC protocol. In brief, mice (4–6 weeks old) were anesthetized continuously with inhalation of isoflurane on a stereotaxic frame (David Kopf Instruments). Analgesia (Carprofen, 5 mg/kg) was administered subcutaneously before the surgery. Craniotomies with appropriate sizes were created using a high-speed dental drill (Henry Schein Inc.).

For viral microinjection, a pre-pulled glass pipette (1B100-4; World Precision Instruments) filled with AAVs was placed into the prelimbic (PL) subregion of the mPFC (AP + 2.1 from the bregma, ML ± 0.3 from the midline, DV −1.3 from the pial surface) and AAVs were microinjected at 200 nl/min rate using a microinjection syringe pump (UMP3; World Precision Instruments). The glass pipette was withdrawn after 10 min, and the incised skin was cleaned and sutured for healing.

For optic fiber implantation, a 200 μm fiber optic cannula (RWD, NA = 0.39) was implanted into the mPFC at the same coordinates immediately after the AAV microinjection and secured to the skull with tissue adhesive and dental cement.

Gradient refractive index (GRIN) lens implantation was performed 1–2 weeks after AAV microinjection. Dexamethasone (0.06 mg/kg) and Carprofen were administered subcutaneously before the surgery. A 1 mm diameter craniotomy was created in the mPFC and the cortical tissue underneath was aspirated up to 1 mm in depth using a blunt syringe needle connected to a vacuum pump. Cortex buffer (NaCl 135 mM, KCl 5 mM, HEPES 5 mM, CaCl₂ 2.38 mM, MgCl₂ 2.14 mM, pH 7.4) was continuously applied to the surgical site to prevent tissue from dehydration. A 1 mm GRIN lens (Inscopix) was implanted at the same coordinates as AAV microinjection and secured. 2–3 weeks after GRIN lens implantation, a baseplate was placed above the lens and fixed at an optimal focal distance of the miniature microendoscope. Subsequent behavioral and imaging experiments were conducted 2–3 weeks after the surgeries.

Both ex vivo and in vivo Ca²⁺ imaging was recorded from the PL subregion of the mPFC. It is noteworthy that CalEx expression was restricted within the dorsomedial PFC but not limited to the PL subregion due to the spread of virus.

Viruses used in this study include: AAV2/5-GfaABC₁D-PI-Cre (1.29 × 10¹³ GC/ml), AAV2/5-GfaABC₁D-tdTomato (2.1 × 10¹³ GC/ml), AAV2/5-GfaABC₁D-cyto-GCaMP6f (2.0 × 10¹³ GC/ml), AAV2/1-Syn-NES-jRGECO1a (1.9 × 10¹³ GC/ml), AAV2/9-hSyn-GCaMP6f (1.4 × 10¹³ GC/ml), AAV2/9-CaMKII-GCaMP6f (2.7 × 10¹³ GC/ml), AAV2/9-mDlx-GCaMP6f (2.3 × 10¹³ GC/ml), AAV2/5-GfaABC₁D-hM3Dq-mCherry (1.8 × 10¹³ GC/ml), AAV2/5-GfaABC₁D-cyto-TurboID (6.4 × 10¹² GC/ml), AAV2/5-GfaABC₁D-TurboID(C)-HA-GPI (7.3 × 10¹² GC/ml), AAV2/1-hSyn1-V5-TurboID(N)-GPI (8.7 × 10¹² GC/ml). Viruses were injected with a total volume of 200–550 nl per site. For morphological analysis, AAV2/5-GfaABC₁D-cyto-GCaMP6f or AAV2/5-GfaABC₁D-tdTomato viruses were diluted to a 1:10 ratio for sparse labeling.

Ex vivo astrocyte Ca²⁺ imaging with confocal microscopy and analysis

Mice were sacrificed between 8 AM and 10 AM during the light cycle. Brains were dissected out and sliced coronally with 300 μm thickness in ice-cold sucrose cutting solution (30 mM NaCl, 4.5 mM KCl, 1.2 mM NaH₂PO₄, 26 mM NaHCO₃, 10 mM D-glucose, 194 mM sucrose, and 1 mM MgCl₂) with a vibratome (MicroSlicer Zero 1 N; Ted Pella). Coronal slices containing the mPFC were incubated in artificial cerebrospinal fluid (ACSF) (124 mM NaCl, 4.5 mM KCl, 1.2 mM NaH₂PO₄, 26 mM NaHCO₃, 10 mM D-glucose, 1 mM MgCl_2, and 2 mM CaCl₂) for 30 min at 30–32 °C and then 1 h at room temperature before recording. All solutions were oxygenated with 95% O₂/5% CO₂. Brain slices for imaging were used within 7 h after sectioning.

Astrocyte Ca²⁺ imaging in acute mPFC slices was performed using a confocal microscope (Fluoview 3000; Olympus) equipped with a perfusion system to supply oxygenated ACSF. A laser with 488 nm wavelength was used to image GCaMP6f fluorescent signals and a laser with 594 nm wavelength was used to confirm mCherry tagged CalEx expression. A 40X water immersion objective lens (NA = 0.8, Olympus) was used for imaging and astrocytes were chosen at least 20 μm below the tissue surface. Images were acquired at 1 frame per second. To activate GPCR-mediated Ca²⁺ responses in astrocytes, 10 μM phenylephrine (PE, Tocris Bioscience) was bath-applied in the presence of 0.25 μM tetrodotoxin citrate (TTX, Cayman Chemical) to inhibit neuronal activities. To activate hM3Dq-mediated Ca²⁺ responses in astrocytes, 1 μM clozapine N-oxide (CNO, Tocris Bioscience) was bath-applied in the presence of 0.25 μM TTX.

A custom Plugin in ImageJ (NIH) was used to correct slow drifts in cell position. For region of interest (ROI)-based analysis, the somata, major branches and territories of the mPFC astrocytes were manually demarcated and astrocyte Ca²⁺ signals were measured by the fluorescence intensity of ROIs over time. Ca²⁺ signals were presented as the relative changes in fluorescence (ΔF/F). Peak amplitude, half-width, and frequency of Ca²⁺ signals for spontaneous activity and integrated areas for PE and CNO responses were analyzed in Origin 2021b. For event-based analysis, spontaneous Ca²⁺ signals were analyzed with AQuA with parameters optimized for ex vivo cytosolic GCaMP6f signals³¹.

Immunohistochemistry (IHC) analysis

Brains were extracted after transcardial perfusion with 1X phosphate-buffered saline (PBS) followed by 10% formalin. Perfused brains were post-fixed in 10% formalin overnight and then cryoprotected in 30% sucrose. Serial brain sections with 40 μm thickness were collected with a cryostat (CM3050 S; Leica). Brain sections were incubated with the following primary antibodies overnight at 4 °C: chicken anti-GFP (1:1000, ab13970; Abcam), rabbit anti-c-Fos (1:1000, ABE457; Millipore), rabbit anti-RFP (1:1000, 600-401-379; Rockland), rabbit anti-S100β (1:1000, PA5-78161; Invitrogen), mouse anti-S100β (1:1000, S2532; Sigma-Aldrich), rabbit anti-LDHA (1:1000, AB52488; Millipore), mouse anti-NeuN (1:1000, MAB377; Millipore), mouse anti-GPHN (1:250, 147011; Synaptic systems), mouse anti-PSD95 (1:250, MA1-046; Thermo Fisher Scientific), mouse anti-ATPB (1:1000, AB14730; Abcam), guinea pig anti-EAA2/GLT-1 (1:1000, AB1783; Millipore), guinea pig anti-VGAT (1:500, 131004; Synaptic systems), guinea pig anti-VGLUT1 (1:500, AB5905; Millipore) and guinea pig anti-VGLUT2 (1:1000, AB2251-I; Millipore). After washing 3 times in 1X PBS for 10 min each, sections were incubated with the following secondary antibodies (1:1000) at room temperature for 2 h: Alexa Fluor 488 goat anti-chicken (A11039; Thermo Fisher Scientific), Alexa Fluor 488 goat anti-mouse (A10680; Thermo Fisher Scientific), Alexa Fluor 594 goat anti-rabbit (A11037; Thermo Fisher Scientific), Alexa Fluor 647 goat anti-mouse (A21235; Thermo Fisher Scientific), Alexa Fluor 647 goat anti-guinea pig (A21450; Thermo Fisher Scientific). For biotinylated protein staining, Alexa Fluor 647 streptavidin (1:200, S21374; Thermo Fisher Scientific) was added together with other Alexa-conjugated secondary antibodies during the 2 h incubation. Confocal images were acquired using FV3000 (Olympus) with a 10X or 40X oil immersion objective lens (NA = 0.4 and 1.4, respectively). ImageJ and Imaris (Oxford Instruments) were used to process the images.

To quantify synapse puncta numbers, confocal images were acquired with a 40X oil immersion objective lens (Olympus, NA = 1.4). Channels containing presynaptic and postsynaptic puncta were split and a colocalization assay was performed using the DiAna plugin⁸⁵. The numbers of individual excitatory (VGLUT1 and PSD95) and inhibitory (VGAT and gephyrin) synaptic puncta, as well as colocalized puncta, were measured. Imaris software was utilized to reconstruct three-dimensional synaptic puncta with an automatic spot-detection function. The diameter thresholds for puncta detection were set to 0.5 μm for the presynaptic markers (VGLUT1 and VGAT) and 0.25 μm for postsynaptic markers (PSD95 and gephyrin). To quantify protein expression, individual astrocytes were outlined based on S100β signals, and the intensity of protein of interest was measured within the astrocyte territories in ImageJ.

To analyze astrocyte morphologies, the confocal images were acquired with a 40X oil immersion objective lens. Individual astrocytes expressing control or CalEx were outlined based on the GCaMP6f fluorescence, and the soma size and the number of major branches were quantified in ImageJ (NIH). For territory size estimation, z-projection of individual astrocytes were thresholded to remove background signals and territory sizes were determined by measuring the area of a region of interest (ROI) surrounding the thresholded fluorescence profile of astrocytes.

Behavioral tests

Behavioral tests were performed during the light cycle with dim light illumination. Both male and female mice (7–14 weeks old) were used in behavioral tests, and all mice were transferred to the behavior testing room at least 30 min before the experiments to acclimatize to the environment.

Open-field test (OFT)

Mice were not habituated in the OFT apparatus prior to the test. Each mouse was placed in the center of an open field chamber (38 cm × 38 cm) enclosed by walls (20 cm) to explore freely for 20 min. Travel distances, time spent in the center, rearing, and self-grooming were analyzed. TopScan tracking software (Clever Sys) was used to record and analyze the travel distances and time spent in the center. The number and duration of the bouts of innate behaviors, including rearing and self-grooming, were counted manually. Rearing behavior was defined as either the mouse standing on both hind limbs in a vertical upright position against the walls of the chamber or without contacting the walls. Self-grooming behavior included licking paws, stroking the nose, mouth, and face, moving paws over the head and behind the ears, licking the body fur, licking the tail, and cleaning the genitals. Grooming episodes were classified as separated events if the interval between them was longer than 5 s or if behaviors other than self-grooming were observed.

Elevated plus maze (EPM) test

Mice were not habituated in the EPM prior to the test. The cross-shaped elevated plus maze apparatus consisted of two open arms (30 cm × 30 cm), two perpendicular closed arms (30 cm × 30 cm) with walls (19 cm) to enclose the arms, and a center area (6 cm × 6 cm). The apparatus was raised 60 cm from the floor. The experimental mouse was placed in the center area with its head directed to the same closed arm and allowed to explore freely for 10 min. Anxiety-like behavior was manually evaluated with the entry numbers and time spent in each compartment.

Elevated zero maze (EZM) test

Mice were not habituated in the EZM prior to the test. The EZM test used an apparatus that was elevated 50 cm from the floor and consisted of two open areas and two closed areas with walls (19.5 cm) that formed a circle. The experimental mouse was placed in the same closed arm and allowed to explore freely for 10 min. Similar to the EPM test, anxiety-like behavior was analyzed manually by counting the entry numbers and time spent in each compartment.

T-maze spontaneous alternation test

The T-maze apparatus consisted of a start arm (36 cm × 7.5 cm), left and right (28 cm × 7.5 cm) arms with vertical sliding doors. Prior to the testing, mice were handled 5 min per day for 8 days to habituate for the experimenter and handling process. To assess the working memory, a sample trial and 10 test trials were conducted with an intertrial interval of 10 s. The cut-off time for each trial was 120 s. Correct and error alternations were determined for the test trials.

Sucrose preference test (SPT)

The SPT was conducted for 3 consecutive days in the home cage and therefore did not require habituation. On day 1, mice were given free access to two bottles of water for 24 h. On day 2, one bottle was replaced with 1% (wt/vol) sucrose solution (S0389; Millipore). On day 3, the 1% sucrose water bottle was replaced by a water bottle. The bottles were weighed every day to determine the amount of liquid consumption. The ratio of water to water and water to sucrose consumption were used to calculate the preference ratio.

Tail suspension test (TST)

Mice were not habituated prior to the test. The TST was conducted in a chamber (64.5 cm × 43 cm) with separated compartments (20.5 cm × 43 cm) with walls to prevent animals from observing or interacting with each other. Mice were suspended by their tails using a paper tape, and a plastic cylinder (1 cm in diameter) was attached to the mouse’s tail to prevent it from climbing upward. The test duration was 6 min, and immobility time was scored by manually.

Forced swim test (FST)

Mice were not habituated prior to the test. An inescapable transparent cylindrical tank (20 cm diameter × 38.5 cm height) was filled with room temperature water with a depth of 15 cm. The water was changed between trials. Individual mouse was placed in the center of the water tank and recorded for 6 min. Immobility time was scored manually for the last 4 min of the test.

Fiber photometry recording and analysis

Mice were handled and habituated for at least 3 days before the recording. The implanted fiber optic cannula was connected to an optical fiber (Doric, NA = 0.37) to record the fluorescent signals with a Tricolor multichannel fiber photometry system (RWD R820) during behavioral tests. The excitation wavelength for GCaMP6f was 470 nm, the wavelength for jRGECO1a was 560 nm, and the isosbestic signal was 410 nm. Fluorescent signals were processed with custom Python codes.

All the fluorescent signals were normalized as ΔF/F using the following formula, where the baseline F(0) was defined as the lowest fluorescence value averaged over 10 s.

$${\Delta }F/F=\frac{F\left(t\right)-F(0)}{F(0)}$$

The Z-score was calculated using the average and standard deviation of the ΔF/F.

$$Z{\mbox{-}}{score}=\frac{\frac{\Delta F}{F}-{mean}\frac{\Delta F}{F}}{{SD}\frac{\Delta F}{F}}$$

Miniature microendoscopic recording and analysis

Miniature microendoscopes (Inscopix nVista or UCLA MINIscope v4) were used to record neuronal Ca²⁺ signals in freely behaving mice. Mice were handled and habituated for 3–5 days before the experiments. On the day of imaging, a microendoscope was fitted onto the baseplate and secured. GCaMP6f fluorescent signals and behaviors in the EZM test were recorded simultaneously for 10 min.

Fluorescent signals were processed using Inscopix Data Processing Software with spatial and temporal down-sampling by a factor of two. Motion correction and ΔF/F processing were executed followed by ROI identification to segment neurons. Further analyses were performed using custom Python codes. The receiver operating characteristic (ROC) analysis was used to quantify single neuronal responses during behavior events (entry of either open area or closed area). Behavior event identification based on neuronal activity was assessed using the true positive rate (TPR) and the false positive rate (FPR) over all the time points after applying a binary threshold to the ΔF/F signal and comparing it to a binary event vector that indicated behavior bouts. Plotting the TPR versus the FPR across a range of binary thresholds spanning the neuronal signal’s minimum and maximum values provided a ROC curve that explained how effectively the neuronal signals detected behavior events at various thresholds. The area under the ROC curve (auROC) was employed as a metric to determine how strongly neurons were influenced by each behavior. When the auROC equals 0.5, the neurons were identified as “not correlated with the behavior”, and when the AUC is more than 0.5 (with P < 0.05), the neurons were considered to be associated with either open area or closed area.

Statistical models were applied to predict behavioral events based on neuronal activities. A binary Fisher’s Linear Discriminant (FLD) model was used to differentiate between behaviors (open area vs. closed area), and a random forest model was employed to predict the location in the transition vs. non-transition zones within the closed area. Training sets were created using population vectors during behavior bouts, with negative training data sampled from the baseline. For each cross-validation fold, 90% of the data was used for training and 10% for testing. Model performance was evaluated using the area under the auROC. Null models were created using training data with shuffled class labels.

Protein purification with TurboID

Proximity-dependent biotinylation labeling experiments with TurboID were performed as previously described⁵⁴ with some modifications. Three weeks after AAV microinjection, biotin (Sigma; B4501) was subcutaneously injected at 24 mg/kg for 3 times over 24 h. For astrocyte protein purification, brain tissue containing the mPFC from 3 mice was pooled as one replicate (n = 5 replicates for both control and CalEx groups). For interface protein purification, mPFC brain tissue from 4 mice was pooled as one replicate (n = 5 replicates for the control group and n = 4 replicates for the CalEx group). Tissue was homogenized with 500 µl of lysis buffer-A (50 mM HEPES pH 7.5, 150 mM NaCl, 1 mM EDTA, 1X Halt Protease Inhibitor (Thermo; 78425)), an equal volume of lysis buffer-B (50 mM HEPES pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.4% SDS, 2% TritonX-100, 2% deoxycholate) was added. The homogenate was sonicated for 20 s and centrifuged at 15,000 g for 15 min at 4 °C. Supernatant was further ultracentrifuged at 100,000 g for 30 min at 4 °C (Beckman TLA-100 ultracentrifuge, TLA-100.3 rotor). SDS was added to the cleared supernatant to a final concentration of 1% and the samples were heated at 95 °C for 5 min. Samples were then cooled down on ice and incubated with 10 µl of Pierce Streptavidin magnetic beads (Thermo Fisher; 88817) at 4 °C overnight. Beads were washed twice with 2% SDS, twice with 1% TritonX-100, 1% deoxycholate, 25 mM LiCl, twice with 1 M NaCl, 5 times with 50 mM ammonium bicarbonate, and 3 times with 20 mM HEPES pH 7.5, 150 mM NaCl. Magnetic beads were then centrifuged at 2000 g for 2 min and trypsinized and subjected to the proteomics sample preparation steps.

Protein extraction and trypsin digestion

The proteomics workflow was adapted from previously published work⁸⁶. Protein digestion was carried out directly after protein purification. Dithiothreitol (DTT) was added to the beads to a final concentration of 10 mM. The samples were then incubated at 60 °C for 30 min, and iodoacetamide was added to the sample to a final concentration of 15 mM. Next, samples were incubated again for 30 min at room temperature in the dark. DTT was added to the samples to a final concentration of 20 mM. The samples were incubated with the DTT-containing buffer for 30 min at room temperature in the dark to quench the alkylation reaction. Sequencing grade modified trypsin (Promega, Madison, WI) was spiked into each sample at an enzyme/protein ratio (w/w) of 1:20. The samples were digested for 16 h in a 37 °C water bath. After the incubation, the enzymatic reaction was quenched by adding 1 μl of formic acid, leading to a pH adjustment to 3 − 4. The samples were dried in the vacuum concentrator (Eppendorf). Then, the dried protein digests were reconstituted in 200 μl of 0.1% formic acid in water, then the samples were desalted using Pierce peptide desalting spin columns (ThermoFisher Scientific), according to the manufacturer’s instructions. The tryptic peptides were eluted from the spin columns with 600 μl of 50/50/0.1 acetonitrile/water/FA mixture. The eluate was dried completely in the vacuum concentrator.

Quantitative LC–MS/MS analysis

Liquid chromatography separation of the tryptic peptides was performed on a nanoElute LC system (Bruker Daltonics). Mobile phases consisted of LC-MS grade water containing 0.1% formic acid (solvent A) and LC-MS grade acetonitrile containing 0.1% formic acid (solvent B). For each sample analysis, the peptide digest was reconstituted in 6 μl solvent A then was loaded on a PepMap trap column (Thermo Fisher, 300 μm × 5 mm, C18, 5 μm particles, 100 Å pore size). The peptides were then separated on a nanoElute FIFTEEN column (Bruker Daltonics, 75 μm × 150 mm, C18, 1.9 μm particles, 120 Å pore size). The separation was carried out at 40 °C with a flow rate of 300 nl/min using the following gradient: 2−10% B from 0 to 3 min, 10 − 40% B from 3–45 min, followed by wash step at 90% B. The LC was coupled to a timsTOF Pro 2 MS (Bruker Daltonics) equipped with a CaptiveSpray ion source. The MS was operated in parallel accumulation-serial fragmentation (PASEF) mode with dynamic exclusion, 10 PASEF MS/MS scans per 1.1 s cycle. Active exclusion for precursor ions was selected to release after 0.40 min, and precursors were reconsidered for analysis if the current intensity/previous intensity was greater than 4.0. The m/z range was set between 100 and 1700, and the ion mobility range was set between 0.60 and 1.60 V·s/cm² with a ramp time of 100.0 ms. The intensity threshold was set to 5000 units, and the target intensity was set to 20,000 units. The collision energy was ramped between 20.00–59.00 eV as a function of ion mobility.

Database search and label-free quantitation

The acquired raw data from each LC-MS analysis were imported onto PEAKS Online (Bioinformatics Solutions Inc., Waterloo, ON) for protein identifications⁸⁷. The selected protein database for Mus musculus was downloaded from UniProt. The following parameters were employed for de novo peptide sequencing and protein identification via database search: precursor mass error tolerance = 20 ppm, fragment mass error tolerance = 0.05 Da, enzyme = trypsin, digest mode = specific, peptide length was set between 6 and 45 residues, and the maximum number of missed cleavages was set to 3. Carbamidomethylation of cysteine residues was set as a fixed modification and oxidation of methionine residues was set a variable modification. Search result filters were set to a peptide discovery rate (FDR) of 1%, and a protein filter was set to −${\log }_{10}$(P value) > 20.

For label-free quantitation, the PEAKS LFQ module was employed to extract the peak areas of the peptides in each sample⁸⁸. Parameters for the PEAKS Q quantitation were as follows: mass error tolerance = 20 ppm, retention time shift tolerance was set to “auto detect”, collisional cross section (CCS) error tolerance = 0.05, and feature intensity ≥300. The following peptide filters were selected: average area ≥ 300, peptide ID count ≥ 2, charge between 1 and 10, have at least 2 confident samples per group, and use in group coefficient of variation filter. For protein abundance calculations, the top 3 highest peptide peak areas were averaged for each unique protein. The protein peak areas obtained from PEAKS LFQ were imported onto the Normalyzer (version 1.14.0) online tool for normalization, whereby peak areas were ${\log }_{2}$ transformed and then normalized by variance stabilization normalization (VSN). This normalization method has been shown to be well-suited for label free quantitation proteomics for reducing the intragroup variation for accurate quantitation⁸⁹.

The analysis for protein classes was performed using PANTHER Classification System (https://www.pantherdb.org/) with all Mus musculus genes used as reference. The KEGG and reactome pathway analysis was performed using STRING with FDR < 0.05 and with all Mus musculus genes used as reference. Gene ontology (GO) pathway analysis was performed using GO enrichment analysis (https://geneontology.org), with all Mus musculus genes used as reference.

Protein networks were created using Cytoscape (v3.10.1)⁹⁰, with nodes corresponding to the gene names for proteins identified in the analysis. A list of protein-protein associations and putative interactions from published datasets was assembled using STRING. In all networks, the node size was proportional to the absolute value of fold change between CalEx and control groups. Networks representing biological processes were created by selecting top terms from STRING GO analysis for “biological process” related to the functions listed. Summary of identified mPFC astrocyte proteins related to the synthesis and transport of glutamate, GABA, and ATPwas created in BioRender (https://BioRender.com/z9fpxtg).

Proteins related to GPCRs, neurotransmitter and ion transport, synapse regulation, Ca²⁺-dependent vesicle release, cell adhesion and lipid metabolism were identified based on previously published studies^38,51,72. Interface proteins were filtered with a confidence score ≥ 4 for extracellular or plasma membrane localization (https://www.genecards.org). Cell-type classification of identified interface proteins was based on our analysis with astrocyte cytosolic proteomes and published proteomic datasets^51,52,53,54. Interaction pairs of cell type-specific interface proteins were identified based on the Biological General Repository for Interaction Datasets⁵⁵ (BioGRID^4.4, https://thebiogrid.org/). Genes associated with mPFC related human diseases were obtained from Phenopedia (https://phgkb.cdc.gov; accessed February 2024), NCBI and previously published research^91,92,93.

Quantification and statistical analysis

All statistical tests were performed in Origin 2021b. The graphs were plotted in Origin 2021b or Python and assembled in CorelDraw 2021. All data were represented as mean ± SEM or otherwise stated in the figures. To compare data from different groups, a normality test was performed using Origin 2021b to determine whether the data were normally distributed or not. Depending on the distribution of data, two-sample Student’s t tests or non-parametric tests Mann-Whiteney tests were performed. When the comparisons were made among more than two conditions, one-way and two-way repeated ANOVA tests were executed and followed by Bonferroni post hoc tests. The significant difference was declared at *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. When the P value was greater than 0.05, it was stated as non-significant (n.s.).

Reporting summary

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

Data availability

All data are accessible upon request from the corresponding author. The astrocyte cytosolic and interface proteomic data generated in this study have been deposited in the ProteomeXchange database under accession code PXD055874 (https://www.proteomexchange.org/). The animal behavior data, imaging data and processed proteomic data generated in this study are provided in the Supplementary Information/Source Data file. Source data are provided with this paper.

Code availability

Code for imaging data analysis is available at Github (https://github.com/YuGliaLab/astrocyteanxiety).

References

Logue, S. F. & Gould, T. J. The neural and genetic basis of executive function: attention, cognitive flexibility, and response inhibition. Pharm. Biochem Behav. 123, 45–54 (2014).
Article CAS Google Scholar
Stuss, D. T. & Alexander, M. P. Executive functions and the frontal lobes: a conceptual view. Psychol. Res 63, 289–298 (2000).
Article CAS PubMed Google Scholar
Tovote, P., Fadok, J. P. & Luthi, A. Neuronal circuits for fear and anxiety. Nat. Rev. Neurosci. 16, 317–331 (2015).
Article CAS PubMed Google Scholar
Calhoon, G. G. & Tye, K. M. Resolving the neural circuits of anxiety. Nat. Neurosci. 18, 1394–1404 (2015).
Article CAS PubMed PubMed Central Google Scholar
Myers-Schulz, B. & Koenigs, M. Functional anatomy of ventromedial prefrontal cortex: implications for mood and anxiety disorders. Mol. Psychiatry 17, 132–141 (2012).
Article CAS PubMed Google Scholar
Robinson, O. J. et al. Towards a mechanistic understanding of pathological anxiety: the dorsal medial prefrontal-amygdala ‘aversive amplification’ circuit in unmedicated generalized and social anxiety disorders. Lancet Psychiatry 1, 294–302 (2014).
Article PubMed PubMed Central Google Scholar
Hare, B. D. & Duman, R. S. Prefrontal cortex circuits in depression and anxiety: contribution of discrete neuronal populations and target regions. Mol. Psychiatry 25, 2742–2758 (2020).
Article PubMed PubMed Central Google Scholar
Berg, L., Eckardt, J. & Masseck, O. A. Enhanced activity of pyramidal neurons in the infralimbic cortex drives anxiety behavior. PLoS One 14, e0210949 (2019).
Article CAS PubMed PubMed Central Google Scholar
Liu, W. Z. et al. Identification of a prefrontal cortex-to-amygdala pathway for chronic stress-induced anxiety. Nat. Commun. 11, 2221 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Shepard, R. & Coutellier, L. Changes in the prefrontal glutamatergic and parvalbumin systems of mice exposed to unpredictable chronic stress. Mol. Neurobiol. 55, 2591–2602 (2018).
Article CAS PubMed Google Scholar
Page, C. E., Shepard, R., Heslin, K. & Coutellier, L. Prefrontal parvalbumin cells are sensitive to stress and mediate anxiety-related behaviors in female mice. Sci. Rep. 9, 19772 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Siracusa, R., Fusco, R. & Cuzzocrea, S. Astrocytes: role and functions in brain pathologies. Front Pharm. 10, 1114 (2019).
Article CAS Google Scholar
Yu, X. & Khakh, B. S. SnapShot: astrocyte interactions. Cell 185, 220–220.e221 (2022).
Article CAS PubMed Google Scholar
Shigetomi, E., Patel, S. & Khakh, B. S. Probing the complexities of astrocyte calcium signaling. Trends Cell Biol. 26, 300–312 (2016).
Article CAS PubMed PubMed Central Google Scholar
Bazargani, N. & Attwell, D. Astrocyte calcium signalling: the third wave. Nat. Neurosci. 19, 182–189 (2016).
Article CAS PubMed Google Scholar
Khakh, B. S. & Sofroniew, M. V. Diversity of astrocyte functions and phenotypes in neural circuits. Nat. Neurosci. 18, 942–952 (2015).
Article CAS PubMed PubMed Central Google Scholar
Khakh, B. S. Astrocyte-neuron interactions in the striatum: insights on identity, form, and function. Trends Neurosci. 42, 617–630 (2019).
Article CAS PubMed PubMed Central Google Scholar
Li, X. et al. Loss of liver X receptor beta in astrocytes leads to anxiety-like behaviors via regulating synaptic transmission in the medial prefrontal cortex in mice. Mol. Psychiatry 26, 6380–6393 (2021).
Article CAS PubMed Google Scholar
Kraeuter, A. K., Guest, P. C. & Sarnyai, Z. The elevated plus maze test for measuring anxiety-like behavior in rodents. Methods Mol. Biol. 1916, 69–74 (2019).
Article CAS PubMed Google Scholar
Lefton, K. B. et al. Norepinephrine signals through astrocytes to modulate synapses. Science 388, 776–783 (2025).
Article CAS PubMed Google Scholar
Guttenplan, K. A. et al. GPCR signaling gates astrocyte responsiveness to neurotransmitters and control of neuronal activity. Science 388, 763–768 (2025).
Article CAS PubMed Google Scholar
Chen, A. B. et al. Norepinephrine changes behavioral state through astroglial purinergic signaling. Science 388, 769–775 (2025).
Article CAS PubMed PubMed Central Google Scholar
Xin, Q. et al. Neuron-astrocyte coupling in lateral habenula mediates depressive-like behaviors. Cell 188, 3291–3309, (2025).
Liu, L. et al. Cell type-differential modulation of prefrontal cortical GABAergic interneurons on low gamma rhythm and social interaction. Sci. Adv. 6, eaay4073 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Nougaret, S. & Genovesio, A. Learning the meaning of new stimuli increases the cross-correlated activity of prefrontal neurons. Sci. Rep. 8, 11680 (2018).
Article ADS PubMed PubMed Central Google Scholar
Szabo, Z. et al. Extensive astrocyte synchronization advances neuronal coupling in slow wave activity in vivo. Sci. Rep. 7, 6018 (2017).
Article ADS PubMed PubMed Central Google Scholar
Baddeley, A. Working memory. C. R. Acad. Sci. III 321, 167–173 (1998).
Article CAS PubMed Google Scholar
d’Isa, R., Comi, G. & Leocani, L. Apparatus design and behavioural testing protocol for the evaluation of spatial working memory in mice through the spontaneous alternation T-maze. Sci. Rep. 11, 21177 (2021).
Article ADS PubMed PubMed Central Google Scholar
Yu, X., Moye, S. L. & Khakh, B. S. Local and CNS-Wide Astrocyte Intracellular Calcium Signaling Attenuation In Vivo with CalEx(flox) Mice. J. Neurosci. 41, 4556–4574 (2021).
Article CAS PubMed PubMed Central Google Scholar
Yu, X. et al. Reducing astrocyte calcium signaling in vivo alters striatal microcircuits and causes repetitive behavior. Neuron 99, 1170–1187 (2018).
Article CAS PubMed PubMed Central Google Scholar
Wang, Y. et al. Accurate quantification of astrocyte and neurotransmitter fluorescence dynamics for single-cell and population-level physiology. Nat. Neurosci. 22, 1936–1944 (2019).
Article CAS PubMed PubMed Central Google Scholar
Porter, J. T. & McCarthy, K. D. Astrocytic neurotransmitter receptors in situ and in vivo. Prog. Neurobiol. 51, 439–455 (1997).
Article CAS PubMed Google Scholar
La-Vu, M., Tobias, B. C., Schuette, P. J. & Adhikari, A. To approach or avoid: an introductory overview of the study of anxiety using rodent assays. Front Behav. Neurosci. 14, 145 (2020).
Article PubMed PubMed Central Google Scholar
Gencturk, S. & Unal, G. Rodent tests of depression and anxiety: construct validity and translational relevance. Cogn. Affect Behav. Neurosci. 24, 191–224 (2024).
Article PubMed PubMed Central Google Scholar
Armbruster, B. N., Li, X., Pausch, M. H., Herlitze, S. & Roth, B. L. Evolving the lock to fit the key to create a family of G protein-coupled receptors potently activated by an inert ligand. Proc. Natl. Acad. Sci. USA 104, 5163–5168 (2007).
Article ADS PubMed PubMed Central Google Scholar
Kurogi, Y., Sanagi, T., Ono, D. & Tsunematsu, T. Chemogenetic activation of astrocytes modulates sleep-wakefulness states in a brain region-dependent manner. Sleep. Adv. 5, zpae091 (2024).
Article PubMed PubMed Central Google Scholar
Adamsky, A. et al. Astrocytic activation generates de novo neuronal potentiation and memory enhancement. Cell 174, 59–71 (2018).
Article CAS PubMed Google Scholar
Yu, X. et al. Context-specific striatal astrocyte molecular responses are phenotypically exploitable. Neuron 108, 1146–1162 e1110 (2020).
Article CAS PubMed PubMed Central Google Scholar
Erickson, E. K. et al. Cortical astrocytes regulate ethanol consumption and intoxication in mice. Neuropsychopharmacology 46, 500–508 (2021).
Article CAS PubMed Google Scholar
Yuste, R., MacLean, J., Vogelstein, J. & Paninski, L. Imaging action potentials with calcium indicators. Cold Spring Harb. Protoc. 2011, 985–989 (2011).
Article PubMed Google Scholar
Zweig, M. H. & Campbell, G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin. Chem. 39, 561–577 (1993).
Article CAS PubMed Google Scholar
Cahill, M. K. et al. Network-level encoding of local neurotransmitters in cortical astrocytes. Nature 629, 146–153 (2024).
Article ADS CAS PubMed PubMed Central Google Scholar
Branon, T. C. et al. Efficient proximity labeling in living cells and organisms with TurboID. Nat. Biotechnol. 36, 880–887 (2018).
Article CAS PubMed PubMed Central Google Scholar
Endo, F. et al. Molecular basis of astrocyte diversity and morphology across the CNS in health and disease. Science 378, eadc9020 (2022).
Article CAS PubMed PubMed Central Google Scholar
de Ceglia, R. et al. Specialized astrocytes mediate glutamatergic gliotransmission in the CNS. Nature 622, 120–129 (2023).
Article ADS PubMed PubMed Central Google Scholar
Leal-Ortiz, S. et al. Piccolo modulation of Synapsin1a dynamics regulates synaptic vesicle exocytosis. J. Cell Biol. 181, 831–846 (2008).
Article CAS PubMed PubMed Central Google Scholar
Li, Y. Z. et al. Complexin regulation of synaptic vesicle release: mechanisms in the central nervous system and specialized retinal ribbon synapses. Cell Commun. Signal 22, 581 (2024).
Article PubMed PubMed Central Google Scholar
Stogsdill, J. A. et al. Astrocytic neuroligins control astrocyte morphogenesis and synaptogenesis. Nature 551, 192–197 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Chung, W. S. et al. Astrocytes mediate synapse elimination through MEGF10 and MERTK pathways. Nature 504, 394–400 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Takano, T. et al. Chemico-genetic discovery of astrocytic control of inhibition in vivo. Nature 588, 296–302 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Soto, J. S. et al. Astrocyte-neuron subproteomes and obsessive-compulsive disorder mechanisms. Nature 616, 764–773 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
van Oostrum, M. et al. The proteomic landscape of synaptic diversity across brain regions and cell types. Cell 186, 5411–5427 e5423 (2023).
Article PubMed PubMed Central Google Scholar
Wang, L. et al. A cross-species proteomic map reveals neoteny of human synapse development. Nature 622, 112–119 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Uezu, A. et al. Identification of an elaborate complex mediating postsynaptic inhibition. Science 353, 1123–1129 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Oughtred, R. et al. The BioGRID database: A comprehensive biomedical resource of curated protein, genetic, and chemical interactions. Protein Sci. 30, 187–200 (2021).
Article CAS PubMed Google Scholar
Majumdar, D., Nebhan, C. A., Hu, L., Anderson, B. & Webb, D. J. An APPL1/Akt signaling complex regulates dendritic spine and synapse formation in hippocampal neurons. Mol. Cell Neurosci. 46, 633–644 (2011).
Article CAS PubMed PubMed Central Google Scholar
Dinday, M. T., Girskis, K. M., Lee, S., Baraban, S. C. & Hunt, R. F. PAFAH1B1 haploinsufficiency disrupts GABA neurons and synaptic E/I balance in the dentate gyrus. Sci. Rep. 7, 8269 (2017).
Article ADS PubMed PubMed Central Google Scholar
Hunt, R. F., Dinday, M. T., Hindle-Katel, W. & Baraban, S. C. LIS1 deficiency promotes dysfunctional synaptic integration of granule cells generated in the developing and adult dentate gyrus. J. Neurosci. 32, 12862–12875 (2012).
Article CAS PubMed PubMed Central Google Scholar
Freitag, S., Schachner, M. & Morellini, F. Behavioral alterations in mice deficient for the extracellular matrix glycoprotein tenascin-R. Behav. Brain Res 145, 189–207 (2003).
Article CAS PubMed Google Scholar
Gottschling, C., Wegrzyn, D., Denecke, B. & Faissner, A. Elimination of the four extracellular matrix molecules tenascin-C, tenascin-R, brevican and neurocan alters the ratio of excitatory and inhibitory synapses. Sci. Rep. 9, 13939 (2019).
Article ADS PubMed PubMed Central Google Scholar
Papadopoulos, M. C. & Verkman, A. S. Aquaporin water channels in the nervous system. Nat. Rev. Neurosci. 14, 265–277 (2013).
Article CAS PubMed PubMed Central Google Scholar
Craske, M. G. et al. Anxiety disorders. Nat. Rev. Dis. Prim. 3, 17024 (2017).
Article PubMed Google Scholar
Penninx, B. W., Pine, D. S., Holmes, E. A. & Reif, A. Anxiety disorders. Lancet 397, 914–927 (2021).
Article PubMed PubMed Central Google Scholar
Ansara, E. D. Management of treatment-resistant generalized anxiety disorder. Ment. Health Clin. 10, 326–334 (2020).
Article PubMed PubMed Central Google Scholar
Noh, K. et al. Cortical astrocytes modulate dominance behavior in male mice by regulating synaptic excitatory and inhibitory balance. Nat. Neurosci. 26, 1541–1554 (2023).
Article CAS PubMed Google Scholar
Breitwieser, G. E. Extracellular calcium as an integrator of tissue function. Int J. Biochem Cell Biol. 40, 1467–1480 (2008).
Article CAS PubMed PubMed Central Google Scholar
Cho, W. H. et al. Hippocampal astrocytes modulate anxiety-like behavior. Nat. Commun. 13, 6536 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Padilla-Coreano, N. et al. Direct ventral hippocampal-prefrontal input is required for anxiety-related neural activity and behavior. Neuron 89, 857–866 (2016).
Article CAS PubMed PubMed Central Google Scholar
Adhikari, A., Topiwala, M. A. & Gordon, J. A. Single units in the medial prefrontal cortex with anxiety-related firing patterns are preferentially influenced by ventral hippocampal activity. Neuron 71, 898–910 (2011).
Article CAS PubMed PubMed Central Google Scholar
Hao, T., Du, X., Yang, S., Zhang, Y. & Liang, F. Astrocytes-induced neuronal inhibition contributes to depressive-like behaviors during chronic stress. Life Sci. 258, 118099 (2020).
Article CAS PubMed Google Scholar
Gonzalez-Arias, C. et al. Dysfunctional serotonergic neuron-astrocyte signaling in depressive-like states. Mol. Psychiatry 28, 3856–3873 (2023).
Article CAS PubMed PubMed Central Google Scholar
Nagai, J. et al. Specific and behaviorally consequential astrocyte Gq GPCR signaling attenuation in vivo with ibetaARK. Neuron 109, 2256–2274 e2259 (2021).
Article CAS PubMed PubMed Central Google Scholar
Li, J. T. et al. Astrocytes in the ventral hippocampus bidirectionally regulate innate and stress-induced anxiety-like behaviors in male mice. Adv. Sci. (Weinh.) 11, e2400354 (2024).
PubMed Google Scholar
Yizhar, O. et al. Neocortical excitation/inhibition balance in information processing and social dysfunction. Nature 477, 171–178 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Ferguson, B. R. & Gao, W. J. Thalamic control of cognition and social behavior via regulation of gamma-aminobutyric acidergic signaling and excitation/inhibition balance in the medial prefrontal cortex. Biol. Psychiatry 83, 657–669 (2018).
Article CAS PubMed Google Scholar
Lam, N. H. et al. Effects of altered excitation-inhibition balance on decision making in a cortical circuit model. J. Neurosci. 42, 1035–1053 (2022).
Article CAS PubMed PubMed Central Google Scholar
Nelson, S. B. & Valakh, V. Excitatory/inhibitory balance and circuit homeostasis in autism spectrum disorders. Neuron 87, 684–698 (2015).
Article CAS PubMed PubMed Central Google Scholar
Sohal, V. S. & Rubenstein, J. L. R. Excitation-inhibition balance as a framework for investigating mechanisms in neuropsychiatric disorders. Mol. Psychiatry 24, 1248–1257 (2019).
Article PubMed PubMed Central Google Scholar
Ren, S. Q. et al. Amyloid beta causes excitation/inhibition imbalance through dopamine receptor 1-dependent disruption of fast-spiking GABAergic input in anterior cingulate cortex. Sci. Rep. 8, 302 (2018).
Article ADS PubMed PubMed Central Google Scholar
Vizi, E. S., Fekete, A., Karoly, R. & Mike, A. Non-synaptic receptors and transporters involved in brain functions and targets of drug treatment. Br. J. Pharm. 160, 785–809 (2010).
Article CAS Google Scholar
Ollivier, M. et al. Crym-positive striatal astrocytes gate perseverative behaviour. Nature 627, 358–366 (2024).
Article ADS CAS PubMed PubMed Central Google Scholar
Sun, W. et al. Spatial transcriptomics reveal neuron-astrocyte synergy in long-term memory. Nature 627, 374–381 (2024).
Article ADS CAS PubMed PubMed Central Google Scholar
Singh, K. et al. Neural cell adhesion molecule Negr1 deficiency in mouse results in structural brain endophenotypes and behavioral deviations related to psychiatric disorders. Sci. Rep. 9, 5457 (2019).
Article ADS PubMed PubMed Central Google Scholar
Ali, H., Marth, L. & Krueger-Burg, D. Neuroligin-2 as a central organizer of inhibitory synapses in health and disease. Sci Signal 13, eabd8379 (2020).
Ghosh, M. K., Kildsig, D. O. & Mitra, A. K. Preparation and characterization of methotrexate-immunoglobulin conjugates. Drug Des. Deliv. 4, 13–25 (1989).
CAS PubMed Google Scholar
Wu, Y. F. et al. Development of a cell-free strategy to recover aged skeletal muscle after disuse. J. Physiol. 601, 5011–5031 (2023).
Article CAS PubMed Google Scholar
Zhang, J. et al. PEAKS DB: de novo sequencing assisted database search for sensitive and accurate peptide identification. Mol. Cell Proteom. 11, M111 010587 (2012).
Article Google Scholar
Anapindi, K. D. B. et al. PACAP and other neuropeptide targets link chronic migraine and opioid-induced hyperalgesia in mouse models. Mol. Cell Proteom. 18, 2447–2458 (2019).
Article Google Scholar
Valikangas, T., Suomi, T. & Elo, L. L. A systematic evaluation of normalization methods in quantitative label-free proteomics. Brief. Bioinform 19, 1–11 (2018).
CAS PubMed Google Scholar
Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13, 2498–2504 (2003).
Article CAS PubMed PubMed Central Google Scholar
Karunakaran, K. B. & Amemori, K. I. Spatiotemporal expression patterns of anxiety disorder-associated genes. Transl. Psychiatry 13, 385 (2023).
Article PubMed PubMed Central Google Scholar
Ruzicka, W. B. et al. Single-cell multi-cohort dissection of the schizophrenia transcriptome. Science 384, eadg5136 (2024).
Article ADS CAS PubMed Google Scholar
Daskalakis, N. P. et al. Systems biology dissection of PTSD and MDD across brain regions, cell types, and blood. Science 384, eadh3707 (2024).
Article ADS CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank Justin Rhodes, Kai Zhang, Uwe Rudolph, and Patrick Sweeney for providing equipment access, Steve Hill for optimizing the proteomic protocol, Carlos Parra Pérez for generating the cytosolic TurboID plasmid, Mingyi Ma for help annotating behavior, and Baljit Khakh for insightful comments. This work was supported by an NIH NIMH R01 grant (R01 MH132556), an NIH Director’s New Innovator Award (DP2 NS136871D), a Whitehall Foundation research grant, a Brain and Behavior Research Foundation NARSAD Young Investigator grant, and a Brain Research Foundation grant (to X.Y.). The proteomics efforts were supported by NIH (P30DA018310 to J.V.S.).

Author information

Xinzhu Yu
Present address: Center for Neuroimmunology and Glial Biology, The Brown Foundation Institute of Molecular Medicine for the Prevention of Human Diseases, The University of Texas Health Science Center at Houston, Houston, TX, USA
These authors contributed equally: Eunyoung Kim, Hairuo Du.

Authors and Affiliations

Neuroscience Program, University of Illinois at Urbana-Champaign, Urbana, IL, USA
Eunyoung Kim, Hairuo Du & Xinzhu Yu
Department of Chemistry, Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, Urbana, IL, USA
Yanqi Tan & Jonathan V. Sweedler
Carle Illinois College of Medicine, University of Illinois at Urbana-Champaign, Urbana, IL, USA
Brandon L. Brown
Department of Computer Science, University of Illinois at Urbana-Champaign, Urbana, IL, USA
Yaowen Chang
UK Dementia Research Institute and Centre for Discovery Brain Sciences, University of Edinburgh, Edinburgh, UK
Blanca Díaz-Castro
Department of Molecular and Integrative Physiology, University of Illinois at Urbana-Champaign, Urbana, IL, USA
Xinzhu Yu

Authors

Eunyoung Kim
View author publications
Search author on:PubMed Google Scholar
Hairuo Du
View author publications
Search author on:PubMed Google Scholar
Yanqi Tan
View author publications
Search author on:PubMed Google Scholar
Brandon L. Brown
View author publications
Search author on:PubMed Google Scholar
Yaowen Chang
View author publications
Search author on:PubMed Google Scholar
Blanca Díaz-Castro
View author publications
Search author on:PubMed Google Scholar
Jonathan V. Sweedler
View author publications
Search author on:PubMed Google Scholar
Xinzhu Yu
View author publications
Search author on:PubMed Google Scholar

Contributions

E.K. performed stereotaxic surgeries, in vivo and ex vivo Ca²⁺ imaging, histological analysis, behavioral tests, data analysis, and made figures. H.D. performed stereotaxic surgeries, in vivo and ex vivo Ca²⁺ imaging, behavioral tests, protein extraction, sample preparation, data analysis, and made figures for proteomic experiments. Y.T. performed the LC-MS/MS experiments and analysis. B.L.B. and Y.C. analyzed in vivo Ca²⁺ imaging data. B.D.-C. provided technical support and reagents for proteomic experiments. J.V.S. supervised proteomic experiments. X.Y. conceived and supervised the study. X.Y. wrote the manuscript with inputs from all the authors.

Corresponding author

Correspondence to Xinzhu Yu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Sung Joong Lee, who co-reviewed with Kyungchul Noh, Gertrudis Perea and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Movie 1

Supplementary Movie 2

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Kim, E., Du, H., Tan, Y. et al. Prefrontal cortex astrocytes modulate distinct neuronal populations to control anxiety-like behavior. Nat Commun 16, 7819 (2025). https://doi.org/10.1038/s41467-025-63131-9

Download citation

Received: 05 September 2024
Accepted: 11 August 2025
Published: 21 August 2025
DOI: https://doi.org/10.1038/s41467-025-63131-9