GHS-R1a signaling drives anxiety-related behavior by shaping excitability of ventromedial hypothalamic neurons

Zhang, Meng; Yang, Liu; Mi, Xue; Hu, Gonghui; Lu, Yingchang; Wang, Chen; Yang, Jie; Sun, Xiaomin; Niu, Minglu; Li, Xianchao; Wang, Sihan; Zhang, Jingsai; Yu, Hanbing; Wang, Yuyang; Yu, Ming; Li, Nan; Zhou, Yu

doi:10.1038/s41467-025-59116-3

Download PDF

Article
Open access
Published: 24 April 2025

GHS-R1a signaling drives anxiety-related behavior by shaping excitability of ventromedial hypothalamic neurons

Meng Zhang ORCID: orcid.org/0009-0004-0684-472X^1,2,3^nAff8,
Liu Yang^2,3,
Xue Mi^2,3,
Gonghui Hu^2,3,
Yingchang Lu^2,3,
Chen Wang^2,3,
Jie Yang^2,3^nAff5,
Xiaomin Sun^2,3^nAff6,
Minglu Niu^2,3^nAff7,
Xianchao Li^2,3,
Sihan Wang^2,3,
Jingsai Zhang^2,3,
Hanbing Yu^2,3,
Yuyang Wang⁴,
Ming Yu¹,
Nan Li ORCID: orcid.org/0000-0002-5828-6975¹ &
…
Yu Zhou ORCID: orcid.org/0000-0002-8193-7561^1,2,3,4

Nature Communications volume 16, Article number: 3858 (2025) Cite this article

6310 Accesses
1 Citations
2 Altmetric
Metrics details

Subjects

Abstract

The neural substrates of anxiety are poorly understood, which hinders treatment of anxiety disorders. Here we found, αCaMKII⁺ neurons in the ventral medial hypothalamic nucleus (VMH) responded to stressors with increased activity in male mice, both under physiological conditions and after repeated restraint stress. Activation of VMH αCaMKII⁺ neurons were necessary and sufficient to ameliorate stress-induced anxiety. The peripheral metabolic hormone ghrelin and receptor GHS-R1a play a complex role in emotion regulation; however, the mechanism is uncertain. A delayed increase in GHS-R1a expression in VMH αCaMKII⁺ neurons coincided with the development of stress-induced enhancement of anxiety-related behavior. GHS-R1a expression in VMH αCaMKII⁺ neurons promoted anxiety-related behavior, whereas GHS-R1a knockdown had the opposite effect. GHS-R1a upregulation inhibited the excitability of VMH αCaMKII⁺ neurons. We conclude that GHSR1a signaling drives stress-induced anxiety by shaping the activity of VMH αCaMKII⁺ neurons. GHS-R1a may be a therapeutic target for treating anxiety disorders such as post-traumatic stress disorder.

Stress-induced GHS-R1a expression in medial prefrontal cortical neurons promotes vulnerability to anxiety in mice

Article Open access 13 March 2025

Serotonergic psychedelic 5-MeO-DMT alters plasticity-related gene expression and generates anxiolytic effects in stressed mice

Article 05 July 2024

Ventral tegmental area glutamate neurons mediate nonassociative consequences of stress

Article 28 November 2022

Introduction

Anxiety is a physiological adaptive response to external stressful stimuli (stressor), which is of great significance to the survival of the individual. However, inappropriate or abnormal anxiety is a core symptom of emotional disorders characterized by excessive worry or fear, leading to disrupted daily activities¹. Stress exposure, especially chronic and repeated stress, is the most common risk factor for pathological anxiety and contributes to both the onset and progression of anxiety disorders^2,3,4,5,6. Noticeably, the global incidence of anxiety disorders had risen by 25.6% since the beginning of the COVID-19 pandemic, causing a huge burden for post-pandemic society worldwide. The neural substrates of anxiety and related disorders remain largely unknown^7,8.

Ghrelin is a peripheral hormone that controls feeding and energy balance and may play a key role especially in stress-induced homeostatic processes⁹. Acyl-ghrelin (the active form) is elevated in mice exposed to chronic social defeat stress¹⁰ and chronic unpredictable mild stress¹¹. Importantly, the prolonged increase in circulating acyl-ghrelin occurs in vulnerable adolescent exposed to a chronic severe stressor¹², indicating that ghrelin is not only a “hunger” hormone but also a persistent biomarker for stress. However, how ghrelin participates in the stress response and anxiety regulation is not clear.

Growth hormone secretagogue receptor 1a (GHS-R1a), the only identified ghrelin receptor, is widely distributed in the brain, as well as various peripheral tissues such as cardiovascular system, immune system, bone, and gonads^13,14. In the brain, GHS-R1a is enriched in key nodes of neural circuits of emotion, cognition and reward, such as the hippocampus^15,16,17, and hypothalamus^18,19,20. It is generally believed that acyl-ghrelin exerts its function by binding to GHS-R1a and activating multiple downstream signaling cascades^21,22. Importantly, GHS-R1a is a member of the class A type G protein-coupled receptors, which exhibit the highest known constitutive activities, and thus its inverse agonists may be useful therapeutic agents²³.

The VMH is an evolutionarily conserved deep subcortical brain region that plays a crucial role in various physiological functions, including feeding control²⁴, metabolic regulation²⁵, and energy and skeletal homeostasis²⁶. The VMH is also a key hypothalamic node that participates in processing and expressing several innate behaviors. In particular, the ventrolateral part of the VMH is specifically involved in social behaviors^27,28,29, whereas the dorsomedial and central parts of the VMH (VMHdm/c) are closely associated with defensive behaviors and stress responses^30,31. Recent studies have demonstrated that direct activation of neurons expressing the nuclear receptor protein NR5A1 (also known as steroidogenic factor SF1) in the VMHdm/c induces defensive behaviors^32,33. In addition, activating VMH neurons expressing alpha-Ca²⁺/calmodulin-dependent protein kinase II (αCaMKII⁺) has been reported to induce anxiety-like behavior³⁴. These findings further highlighting an important role of the VMH in the emotional response or state. However, the molecular and cellular mechanisms remains elusive.

The ghrelin/GHS-R1a system is involved in regulating multiple brain functions, including stress response and anxiety¹⁸, by mechanisms that are not well characterized. In situ hybridization assays have revealed strong expression of Ghsr1a mRNA in the hypothalamic nuclei, including the VMH, in rodents^13,14,35. Moreover, GHS-R1a deficiency alleviates anxiety- and despair-like behaviors in mice exposed to chronic social defeat stress³⁶. In previous studies, we demonstrated GHS-R1a expression in hippocampal αCaMKII⁺ neurons and highlighted its critical role in memory regulation^17,37. αCaMKII, the most abundant protein in excitatory synapses within the CNS, is essential for detecting subtle intracellular calcium variations and is pivotal in regulating receptor trafficking, excitatory neurotransmission, synaptic plasticity, and neuronal signaling^38,39. In the VMH, particularly the VMHdm, aCaMKII⁺ neurons constitute a substantial population of excitatory neurons⁴⁰, and have been implicated in regulating wakefulness and anxiety³⁴. Here, we propose that ghrelin/GHS-R1a signaling in the VMH aCaMKII⁺ neurons regulates the stress response and anxiety-related behaviors. To test this hypothesis, we manipulated GHS-R1a expression in aCaMKII⁺ neurons in the VMH, specifically in the VMHdm/c subdivision. We combined optogenetic, chemogenetic, and pharmacological techniques; fiber photometry; electrophysiological recordings; and behavioral assays to elucidate how ghrelin/GHS-R1a signaling in the VMH αCaMKII⁺ neurons regulates anxiety-related behavior. Our findings show that GHS-R1a is an important molecular switch that promotes anxiety-related behavior by shaping the activity of αCaMKII⁺ neurons in the VMH, providing insights for the treatment of stress-induced anxiety disorders.

Results

VMHdm/c αCaMKII⁺ neurons undergo activity changes during EPM exploration

We first used in vivo fiber photometry to investigate the dynamic activity of αCaMKII⁺ neuron in the VMH, specifically in the VMHdm/c, as mice were exploring the elevated plus maze (EPM) (Fig. 1a), a well-accepted paradigm for evaluating anxiety-like behavior⁴¹. The population Ca²⁺ signals (averaged dF/F), an indicator of neuronal activity, increased whenever mice entered the open arms of the maze (Fig. 1b), and they decreased whenever mice entered the closed arms (Fig. 1c). In addition, VMH αCaMKII⁺ neurons had significantly higher overall Ca²⁺ signals (average dF/F) when the mice explored the open arms as compared with the closed arms (Fig. 1d, e), indicating increased activity of this neuronal population in response to an unsafe or stressful environment under physiological conditions. In neural circuits of anxiety, the hypothalamus is considered to participate mainly in evaluating stressful events and initiating a response⁴²; therefore, we proposed that αCaMKII⁺ neurons in the VMH work as a stress sensor that may be associated with anxiety-related behavior.

Fig. 1: Fiber photometry showing activity of αCaMKII+ neurons in mice exploring the EPM. — **Fig. 1: Fiber photometry showing activity of αCaMKII⁺ neurons in mice exploring the EPM.**

We confirmed robust αCaMKII expression in VMHdm/c neurons using Camk2a-Cre; Ai9-tdTomato mice, C57BL6 mice transfected with AAV-Camk2a-eGFP virus in the VMH, and αCaMKII immunostaining (Supplementary Fig. 1a-d). SF1 is a marker protein of those VMHdm/c neurons that are necessary for both defensive responses and a defensive emotional state^30,33. Therefore, we checked the cellular co-expression of Camk2a and Nr5a1 in VMH neurons by fluorescence in situ hybridization (FISH) (Fig. 1f). We found that over 82% of VMHdm/c neurons co-express Camk2a and Nr5a1 (Fig. 1g).

Chemogenetic manipulation of the activity of αCaMKII⁺ neuron in the VMHdm/c alters anxiety-related behavior

Next, we explored whether manipulating the activity of VMH αCaMKII⁺ neurons can affect anxiety-related behavior. AAV virus expressing Camk2a promoter-driven excitatory (hM3Dq-) or inhibitory (hM4Di-) DREADDs (designer receptors exclusively activated by designer drugs)⁴³, or control AAV-Camk2α-mCherry virus was delivered into the VMHdm/c of C57BL/6 mice 4 weeks before EPM and tail suspension (TS) tests (Fig. 2a). DREADD-mediated regulation of neuronal activity in VMHdm/c αCaMKII⁺ neurons were confirmed by ex vivo cell-attached voltage-clamp recordings (Fig. 2b-e). CNO administration increased the firing frequency of VMH αCaMKII⁺ neurons expressing hM3Dq-DREADD (Fig. 2c), whereas it inhibited firing of the same population of neurons expressing hM4Di-DREADD (Fig. 2d). In contrast, CNO had no effect on the firing frequency of VMH αCaMKII⁺ neurons that expressed only mCherry (Fig. 2e). During the EPM test, activation of VMH αCaMKII⁺ neurons by CNO (delivered intraperitoneally) increased the open-arm exploration time of hM3Dq-DREADD mice (Fig. 2f, g). Consistently, inactivating the same group of neurons by CNO reduced the open-arm exploration time of hM4Di-DREADD mice (Fig. 2f, g). Importantly, CNO administration had no effect on control mCherry-expressing mice (Fig. 2f, g), and neither group exhibited significant changes in total travel distance in the EPM following CNO administration. Thus, selectively activating VMH αCaMKII⁺ neurons attenuated anxiety-related behavior, whereas inactivating VMH αCaMKII⁺ neurons increased this behavior. In addition, CNO administration during TS tests reduced the immobility time of hM3Dq-expressing mice (Fig. 2h), whereas it had no effect on hM4Di-expressing mice (Fig. 2h). As expected, CNO injection did not alter the immobility time of control mice that expressed only mCherry without DREADD (Fig. 2h). Altogether, our findings indicated that activation of αCaMKII⁺ neurons in the VMH is both sufficient and necessary to suppress anxiety-like behavior in mice under the baseline condition (i.e., in the absence of pre-exposure to a stressor).

Fig. 2: Chemogenetic manipulation of αCaMKII+ neurons activity in the VMHdm/c. — **Fig. 2: Chemogenetic manipulation of αCaMKII⁺ neurons activity in the VMHdm/c.**

Optogenetic activation of VMH αCaMKII⁺ neurons suppresses anxiety-related behavior

As optogenetics has finer spatiotemporal resolution than chemogenetics with respect to regulating neuronal activity⁴⁴, we also investigated the effect of optogenetic activation of VMHdm/c αCaMKII⁺ neurons on anxiety-related behavior at the baseline state (Fig. 3a). Based on pilot study results (Supplementary Fig. 2a), we selected a photostimulation intensity of 1.2 mW/mm² to activate VMH αCaMKII⁺ neurons. This relatively low intensity did not induce noticeable freezing and/or activity bursts during EPM tests. First, we confirmed ex vivo with whole-cell current recordings that 473 nm photostimulation (1.2 mW/mm², 20 Hz, 20 ms pulse width) dramatically increased the firing of VMH αCaMKII⁺ neurons expressing virus-mediated ChR2, whereas it had no effect on the firing of control neurons expressing virus-mediated eYFP only (Fig. 3b). We then used the photostimulation approach to activate the VMH αCaMKII⁺ neurons in behaving mice. During the first phase of the EPM test, mice were allowed to explore the maze for 5 min in the absence of photostimulation; during the second phase, which also lasted for 5 min, optogenetic stimuli were manually delivered whenever mice left the closed arms and switched off immediately upon re-entering the closed arms (Fig. 3c). ChR2-expressing mice spent more time exploring the open arms during the second phase of the test than during the first phase (Fig. 3d, e). In contrast, the control eYFP-expressing mice had a similar open-arm exploration time during the two phases of the test (Fig. 3d, e). Photostimulation did not affect the total travel distance of either ChR2-expressing mice or control eYFP-expressing mice (Fig. 3e). We also applied photostimulation to separate groups of mice either during exploration of the closed arms or across all zones during the second phase of the EPM test. Consistently, photostimulation specifically increased open-arm exploration time in ChR2-expressing mice, not control eYFP-expressing mice (Supplementary Fig. 2b, c). Photostimulation did not significantly alter total travel distance in either group during the EPM test (Supplementary Fig. 2b, c). In addition, our TS tests showed that the same optogenetic stimuli reduced the immobility time of ChR2-expressing mice (Fig. 3f, g), whereas it had no effect on control eYFP-expressing mice (Fig. 3f, g). Therefore, our findings consistently demonstrate that optogenetic activation of VMH αCaMKII⁺ neurons suppresses anxiety-related behavior.

**Fig. 3: Optogenetic activation of VMH αCaMKII+ neurons.**

Previous studies have reported that optogenetic activation of SF1⁺ neurons in the VMHdm/c reduces the time mice spend in the photostimulating chamber during the real-time place preference (RTPP) test³². In our RTPP test, we found that activation of αCaMKII⁺ neurons in the VMHdm/c also induced acute aversion in ChR2-expressing mice under physiological conditions (Fig. 3h, i). In conclusion, our findings indicate that activation of αCaMKII⁺ neurons in the VMHdm/c alleviates anxiety-like behavior while inducing acute aversion.

RRS exposure induces activation of VMH αCaMKII⁺ neurons

To further explore the role of VMH αCaMKII⁺ neurons in regulating anxiety-related behavior, we subjected mice to repeated restraint stress (RRS) for 1 h per day over a 4-d period. Anxiety-related behavior was assessed using the EPM test at 90 min, 24 h, 3 d, or 6 d following RRS exposure. We observed a significant reduction in open-arm exploration time at 6 d post-stress exposure compared to earlier time points, suggesting a delayed onset of anxiety-related phenotypes (Fig. 4a). Furthermore, mice exposed to RRS exhibited enhanced anxiety-related behaviors 6 d after RRS exposure, including reduced open-arm exploration in the EPM test, decreased sociability in the social interaction (SI) test, and increased immobility time in the force swimming (FS) and TS test (Fig. 4b, c). To examine neuronal activation, we qualified c-Fos expression in VMH αCaMKII⁺ neurons by immunostaining at 90 min and 6 d following RRS exposure, immediately after an EPM test (Fig. 4d). Although anxiety-like behaviors were fully developed at 6 d (not 90 min) after RRS exposure (Fig. 4d), the number of c-Fos⁺ neurons in the VMHdm/c was similarly increased at both time points compared to unstressed controls (Fig. 4e, f). These findings indicate that αCaMKII⁺ neurons in the VMHdm/c are persistently activated after RRS exposure.

Fig. 4: RRS exposure activates VMH αCaMKII+ neurons. — **Fig. 4: RRS exposure activates VMH αCaMKII⁺ neurons.**

To further investigate RRS-induced dynamic activation of VMH αCaMKII⁺ neurons, we carried out ex vivo whole-cell current-clamp recordings selectively on stress-driven αCaMKII⁺ neurons identified with a combination of the Robust Activity Marking (RAM) reporter system (activity-dependent mKate2 expression)⁴⁵ and Camk2a promoter-driven eGFP expression (Fig. 4g). The firing frequency of Fos⁺αCaMKII⁺ neurons (identified as mKate2⁺GFP⁺) in the VMHdm/c was reduced at 6 d post-RRS exposure compared to 90 min post-RRS exposure (Fig. 4h-k), suggesting that reduced activity of stress-driven αCaMKII⁺ neurons in the VMHdm/c may correlate with enhanced anxiety-related behavior. Notably, previous studies have reported that chronic stress increases the proportion of burst-firing SF1⁺ neurons in the VMHdm⁴⁶. In our recordings, Fos⁺αCaMKII⁺ neurons in the VMHdm/c did exhibit three different firing patterns—tonic, burst, and silent—with tonic firing being the most common (Supplementary Fig. 3a). However, we did not observe any significant shifts in the firing pattern of Fos⁺αCaMKII⁺ neurons in the VMHdm/c from 90 min to 6 d after RRS exposure (Supplementary Fig. 3b).

Additionally, we examined the Ca²⁺ activity of VMHdm/c αCaMKII⁺ neuron using in vivo fiber photometry during open- and closed-arm exploration in the EPM at 6 d after RRS exposure (Fig. 5a). While the averaged Ca²⁺ activity of VMHdm/c αCaMKII⁺ neurons increased when stressed mice entered the open arms (Fig. 5b) and decreased when they entered the closed arms (Fig. 5c), the overall Ca²⁺ activity in the open arms was no longer significantly higher than in the closed arms (Fig. 5d, e). Furthermore, during open-arm exploration, VMHdm/c αCaMKII⁺ neurons in stressed mice exhibited reduced population Ca²⁺ activity compared to these in unstressed mice, whereas no significant differences were observed during closed-arm exploration (Fig. 5f). These results indicate that while VMH αCaMKII⁺ neurons still respond to acute environmental stress, their activity is diminished with the development of RRS-induced anxiety.

Fig. 5: Fiber photometry showing activity of VMHdm/c αCaMKII+ neurons during stressed mice exploring the EPM at 6 d after RRS exposure. — **Fig. 5: Fiber photometry showing activity of VMHdm/c αCaMKII⁺ neurons during stressed mice exploring the EPM at 6 d after RRS exposure.**

Activating VMH αCaMKII⁺ neurons ameliorates RRS-induced enhancement of anxiety-related behavior

Next, we tested whether manipulating the activity of VMH αCaMKII⁺ neurons affected RRS-induced abnormal anxiety. AAV virus expressing Camk2a promoter-driven hM3Dq- or hM4Di-DREADD or only mCherry was delivered into the VMH 4 weeks before RRS exposure. EPM and TS tests were done 6-10 d after RRS insult (Fig. 6a). Similar to what we found for the baseline state, activating VMH αCaMKII⁺ neurons by CNO injection increased the time the stressed mice spent in the open arms during the EPM test (Fig. 6b, c). Consistently, inactivating the same group of neurons decreased the open-arm exploration time for the stressed mice (Fig. 6b, c). CNO administration did not affect the open-arm exploration time of the stressed mice expressing only mCherry in VMH αCaMKII⁺ neurons (Fig. 6b, c). Moreover, activating VMH αCaMKII⁺ neurons by CNO injection reduced the immobility time of the stressed mice during the TS test (Fig. 6d). In contrast, inactivating VMH αCaMKII⁺ neurons by CNO injection did not affect the immobility time of the stressed mice during this test (Fig. 6d). As expected, CNO injection did not affect the immobility time of the stressed mice expressing only mCherry (Fig. 6d).

Fig. 6: Chemogenetic and optogenetic activation of VMHdm/c αCaMKII+ neurons in stressed mice. — **Fig. 6: Chemogenetic and optogenetic activation of VMHdm/c αCaMKII⁺ neurons in stressed mice.**

We also tested whether optogenetic activation of VMH αCaMKII⁺ neurons affected RRS-induced anxiety (Fig. 6e). Using ex vivo whole-cell current recordings, we confirmed that photostimulation significantly increased the firing of ChR2-expressing αCaMKII⁺ neurons, whereas it had no effect on the firing of control αCaMKII⁺ neurons expressing only eYFP in the VMHdm/c (Fig. 6f). Activating VMH αCaMKII⁺ neurons with 473-nm blue light increased open-arm exploration time in stressed mice without altering their total travel distance in the EPM test (Fig. 6g, h). To further explore the effect of photostimulation, we applied light to separate groups of stressed mice either while they explored the closed arms or across all zones during the second phase of the EPM test (Supplementary Fig. 4a, b). Consistent with findings in unstressed mice, photostimulation specifically increased open-arm exploration time in stressed mice expressing ChR2, but not in stressed mice expressing only eYFP (Fig. 6h and Supplementary Fig. 4a, b). In addition, photostimulating the VMHdm/c αCaMKII⁺ neurons as described above reduced the immobility time of the stressed mice during the TS test (Fig. 6i). As expected, photostimulation did not alter the immobility time of the stressed control mice expressing only eYFP in αCaMKII⁺ neurons (Fig. 6i). Additionally, optogenetic activation of αCaMKII⁺ neurons in VMHdm/c also triggered acute aversion in stressed mice expressing ChR2, but not in stressed mice expressing only eYFP (Fig. 6j). Altogether, our findings demonstrate that activating αCaMKII⁺ neurons in the VMHdm/c suppresses anxiety-like behavior both under baseline conditions and after RRS exposure. Therefore, we conclude that activation of VMH αCaMKII⁺ neurons may be essential for coping with stress-induced anxiety.

RRS induces prolonged upregulation of GHS-R1a expression in VMH αCaMKII⁺ neurons

Both mice and adolescent individuals that are exposed to chronic stress show a prolonged increase in circulating ghrelin^10,12. To investigate whether RRS alters GHS-R1a expression in the VMH, we examined its mRNA levels using RT-qPCR (Fig. 7a). Our results showed that GHS-R1a transcripts in the VMHdm/c were significantly increased only at 6 d after RRS insult, not at earlier time points (Fig. 7b). GHS-R1a transcripts in the ventral hippocampus (vHPC) were also increased at 6 d but not at 90 min following RRS exposure (Fig. 7c). In contrast, RRS exposure did not significantly affect GHS-R1a expression in the nucleus accumbens (NAc), a region that is also involved in anxiety regulation^47,48, at either time point (Fig. 7c). To confirm these dynamic changes, we further checked GHS-R1a expression in VMH αCaMKII⁺ neurons by fluorescence in situ hybridization (FISH) (Fig. 7d). Consistent with the RT-qPCR results, Ghsr1a expression in VMHdm/c αCaMKII⁺ neurons was elevated at 6 d but not at 90 min after RRS exposure (Fig. 7d, e). Specifically, both the proportion of Ghsr1a⁺Camk2a⁺ neurons among Camk2a⁺ neurons and the fluorescence intensity density (IntDensity) of Ghsr1a in individual Ghsr1a⁺Camk2a⁺ neurons were increased at 6 d but not 90 min after RRS exposure (Fig. 7e). Furthermore, Ghsr1a expression in VMHvl αCaMKII⁺ neurons was also elevated at 6 d compared to 90 min following RRS exposure, while the fluorescence IntDensity of individual neurons did not show significant changes (Fig. 7f). Additionally, we analyzed the proportions of Ghsr1a⁺Camk2a^- and Ghsr1a⁺Camk2a⁺ neurons within the Ghsr1a⁺ neuronal population in the VMHdm/c. Our results indicated that approximately 7% of Ghsr1a⁺ neurons were Camk2a^-, while 93% were Camk2a⁺ in unstressed mice. These proportions were not significantly altered following RRS exposure (Supplementary Fig. 5). Interestingly, a recent study using GHS-R1a-eGFP mice reported that 93% of GHS-R1a⁺CB1R⁺ neurons in the VMH were glutamatergic, while the remaining 7% were other types, with 3% being GABAergic⁴⁹. Altogether, our findings indicate that RRS exposure induces a delayed upregulation of GHS-R1a expression in VMH αCaMKII⁺ neurons, which occurs concurrently with the development of RRS-induced anxiety-like behavior.

Fig. 7: Changes of GHS-R1a expression in VMH αCaMKII+ neurons after RRS. — **Fig. 7: Changes of GHS-R1a expression in VMH αCaMKII⁺ neurons after RRS.**

Manipulating GHS-R1a expression in VMH αCaMKII⁺ neurons alters anxiety-related behavior

Next, we checked whether directly manipulating GHS-R1a expression in VMH αCaMKII⁺ neurons affected anxiety-related behavior. To disrupt endogenous GHS-R1a expression in αCaMKII⁺ neurons, we delivered either an AAV virus expressing Camk2a-driven Ghsr1a-specific shRNA to the VMHdm/c of C57BL6/J mice (Fig. 8a) or an AAV-Camk2a-Cre virus to the VMHdm/c of Ghsr1a^fl/fl mice (Fig. 8b). Both RNA interference and Cre-dependent knockdown of GHS-R1a expression in VMHdm/c αCaMKII⁺ neurons significantly increased the open-arm exploration time of unstressed mice under baseline conditions (Fig. 8a, b). In addition, to increase GHS-R1a expression in VMH αCaMKII⁺ neurons, we delivered either an AAV-DIO-hGhsr1a virus to the VMHdm/c of Camk2a-Cre mice (Fig. 8c) or an AAV-Camk2a-hGhsr1a virus to the VMHdm/c of C57BL6/J mice (Fig. 8d). Both Cre-dependent and Cre-independent GHS-R1a overexpression in VMH αCaMKII⁺ neurons reduced the open-arm exploration under baseline conditions (Fig. 8c, d). Cre-dependent GHS-R1a knockdown or Cre-independent GHS-R1a overexpression in VMHdm/c αCaMKII⁺ neurons were confirmed by FISH (Fig. 8e). Therefore, our findings indicated a causal relationship between GHS-R1a expressions in VMHdm/c αCaMKII⁺ neurons and anxiety-like behavior under baseline conditions.

**Fig. 8: Manipulation of GHS-R1a expression in VMH αCaMKII+ neurons.**

We then checked whether manipulating GHS-R1a expression in VMHdm/c αCaMKII⁺ neurons alters anxiety-related behavior in stressed mice. Disrupting GHS-R1a expression in VMH αCaMKII⁺ neurons by either RNA interference (Fig. 8f) or Cre-dependent knockdown (Fig. 8g) alleviated anxiety-like behavior in stressed mice when tested at 6 d after RRS exposure (Fig. 8f, g). In contrast, increasing GHS-R1a expression in VMHdm/c αCaMKII⁺ neurons facilitated RRS-induced anxiety-related behavior (Fig. 8h). As a control experiment, we also delivered AAV-DIO-Ghsr1a virus or control AAV-DIO-GFP virus to the VMHdm/c of Dlx5/6-Cre mice to increase GHS-R1a expression selectively in VMH inhibitory neurons, and found no behavioral difference between groups either under baseline conditions or after RRS exposure (Supplementary Fig. 6a-c). As endogenous GHS-R1a expression in the vHPC was also elevated at 6 d after RRS exposure, we also investigated whether virus-mediated upregulation of GHS-R1a expression in vHPC αCaMKII⁺ neurons affected anxiety-related behavior. An increase in GHS-R1a expression in the vHPC αCaMKII⁺ neurons did not affect anxiety-related behavior, either at baseline or after RRS (Supplementary Fig. 7a-g). Altogether, our findings clearly demonstrate a causal relationship between GHS-R1a expression in VMHdm/c αCaMKII⁺ neurons and anxiety-like behavior in both unstressed and stressed mice. We thus conclude that RRS-induced a delayed elevation of GHS-R1a expression in VMH αCaMKII⁺ neurons contributes to the development of enhanced anxiety-like behavior.

Increasing GHS-R1a expression inhibits excitability of VMH αCaMKII⁺ neurons

Ghrelin/GHS-R1a signaling in the hippocampus suppresses the intrinsic excitability of pyramidal neurons, and the GHS-R1a antagonist YIL781 abolishes this effect⁵⁰. Here we demonstrated that virus-mediated GHS-R1a expression reduced spontaneous firing of VMHdm/c αCaMKII⁺ neurons (Fig. 9a-c). In addition, local administration of YIL781 in the VMH eliminated the anxiogenic effect of GHS-R1a overexpression (Fig. 9d, e), whereas it had no effect on control mice that expressed only GFP (Fig. 9f).

Fig. 9: The effect of increased GHS-R1a expression on the excitability of VMH αCaMKII+ neurons. — **Fig. 9: The effect of increased GHS-R1a expression on the excitability of VMH αCaMKII⁺ neurons.**

We further examined whether chemogenetic activation of VMH αCaMKII⁺ neurons with increased GHS-R1a expression could attenuated anxiety-like behavior. Our ex vivo whole-cell current recordings confirmed that CNO administration increased the firing of VMHdm/c αCaMKII⁺ neurons that express both GHS-R1a and hM3Dq-DREADD (Fig. 9g-i). CNO administration also increased the open-arm exploration time both under baseline conditions and 6 d after RRS exposure (Fig. 9j). Meanwhile, we also tested whether chemogenetic inactivation of VMH αCaMKII⁺ neurons with increased GHS-R1a expression exacerbated anxiety-like behavior. As expectedly, CNO administration inhibited firing of VMHdm/c αCaMKII⁺ neurons expressing both GHS-R1a and hM4Di-DREADD (Fig. 9h, i). Interestingly, CNO administration significantly reduced the open-arm exploration time in the unstressed mice but not in the stressed mice (Fig. 9k). Our findings thus indicate that GHS-R1a signaling in the VMHdm/c promotes anxiety-like behavior by regulating the activity of αCaMKII⁺ neurons.

Discussion

Current findings regarding the role of ghrelin and GHS-R1a in aversive processing and emotional behaviors are contradictory. Ghrelin inhibits fear memory consolidation in unstressed rats⁵¹ and reduces anxiety after acute stress⁵². However, chronic stress increases circulating acyl-ghrelin and promotes central ghrelin resistance in the amygdala, enhancing fear memory formation^12,51,53,54. While GHS-R1a knockout alleviates anxiety- and depression-related behaviors following chronic social defeat stress³⁶, ghrelin may protect against depressive symptoms induced by chronic stress¹⁰. It seems that, although chronic stress triggers persistent elevation in circulating ghrelin, ghrelin/GHS-R1a signaling appears to play distinct roles, either maladaptive or adaptive, in anxiety and fear processing. How GHS-R1a sensors altered ghrelin level in the circulation and participates in emotion processing or expression is still a puzzle.

Our study demonstrates that RRS induces persistent anxiety-like behavior that fully develops by 6 d post-exposure. Unlike chronic stress models that primarily trigger depression-like phenotypes, such as chronic social defeat stress¹⁰ or chronic unpredictable mild stress¹¹, the RRS protocol in this study predominantly induces anxiety-related behavior. This 1 h/d for 4 d protocol is also more time-efficient than the 4 h/d for 14 d model used in post-traumatic stress disorder (PTSD) studies⁵⁴. More interestingly, we found that delayed upregulation of GHS-R1a in VMH αCaMKII⁺ neurons correlated with RRS exposure-induced anxiety. GHS-R1a upregulation may be linked to energy homeostasis in response to stress, given the VMH’s role in regulating both metabolic and defensive behaviors^9,27,32. Manipulating GHS-R1a in VMHdm/c αCaMKII⁺ neurons demonstrated its necessity and sufficiency in driving anxiety-like behavior in both unstressed and stressed mice. Our study provides the evidence that GHS-R1a signaling promotes anxiety-related behavior by modulating excitability of VMH αCaMKII⁺ neurons, offering insight into the molecular and cellular mechanisms of stress-induced anxiety.

VMHdm/c SF1⁺ neurons are essential for triggering defensive behaviors ranging from avoidance to freezing and activity bursts, depending on photostimulation intensity and frequency³². For instance, a threshold intensity of 0.65 mW/mm² induces freezing, while 5.25 mW/mm² induces both freezing and activity bursts. High-intensity, high-frequency stimulation (10.5 mW/mm² at 100 Hz) evokes activity bursts without prior freezing³³. In this study, we utilized a blue light intensity of 1.2 mW/mm² (20 Hz, 20 ms pulse width) to bilaterally activate VMHdm/c αCaMKII⁺ neurons. This stimulation did not induce obvious freezing or activity bursts during EPM or RTPP tests. However, a higher intensity of 2.5 mW/mm² occasionally elicited freezing or jumping behavior.

Our in vivo Ca²⁺ imaging studies combined with the EPM test reveal that VMHdm/c αCaMKII⁺ neurons exhibit an acute activity increase when mice enter the open arms of the maze, though this response is attenuated under stressed conditions (RRS exposure) compared to baseline. Increased c-Fos expression in VMHdm/c αCaMKII⁺ neurons shortly after RRS exposure confirms their activation in response to stress. Intriguingly, optogenetic or chemogenetic activation of these neurons increases open-arm exploration time without significantly altering overall motor function, indicating that activating these neurons alleviates anxiety-related behavior. We propose that VMH αCaMKII⁺ neuron activation plays an adaptive role in response to stress, helping maintain a balanced level of anxiety and arousal necessary for survival. Following RRS exposure, these neurons may transition from an adaptive to a maladaptive state due to GHS-R1a upregulation, which suppresses their excitability and contributes to heightened anxiety. Increased Ca²⁺ activity during open-arm exploration in stressed mice may initially act as a compensatory mechanism to manage anxiety but is eventually overridden by GHS-R1a-mediated inhibition. Experimental activation via photostimulation or CNO may bypass this inhibition, restoring normal function and reducing anxiety. Further studies are needed to unravel the circuit mechanisms underlying VMH αCaMKII⁺ neurons’ regulation of anxiety-related behavior, particularly their downstream inhibitory pathways.

SF1⁺ neurons in the VMHdm/c play a complex role in regulating anxiety-related behavior. Ablation of SF1⁺ neurons alleviates anxiety-related behavior and eliminates fear response to predators³², consistent with findings that conditional knockout of the excitatory output of VMH SF1⁺ neurons reduces anxiety-related behavior⁵⁵. However, mice with CNS-specific knockout of SF1 exhibit increased anxiety-like behavior compared to wild-type littermates⁵⁶. Previous studies report that unilateral activation of SF1⁺ neurons with blue light (20 Hz, 20 ms) at 2.5–5 mW/mm² (equivalent to bilateral stimulation at 1.25–2.5 mW/mm²) reduces immobility and heart rate, whereas lower intensities (0.8-1.7 mW/mm²) produce the opposite effect³⁰. In this study, we found that activating VMHmd/c αCaMKII⁺ neurons reduced immobility time during the TS test, supporting their role in alleviating anxiety-related behavior. Activating somatostatin (SOM)⁺ neuronal inputs from the bed nucleus of the stria terminalis (BNST) to the VMHdm inhibits SF1⁺/Vglut2⁺ neuron firing, leading to increased anxiety-related behavior and bone loss under both baseline and chronic stress conditions⁵⁷. Chemogenetic inhibition of SOM⁺ neurons in the BNST disinhibits SF1⁺ neurons in the VMH, relieving stress-induced anxiety and bone loss⁵⁷. Similarly, we observed that increasing αCaMKII⁺ neurons activity in the VMHdm/c reduced anxiety-like behavior, while suppressing their activity enhanced anxiety-like behavior in both unstressed and stressed mice. Importantly, we specifically targeted stress-driven αCaMKII⁺ neurons (i.e., c-Fos⁺αCaMKII⁺ neurons) rather than a random population of VMHdm/c αCaMKII⁺ neurons to investigate stress-induced excitability changes. We also identified that increased GHS-R1a expression in VMHdm/c αCaMKII⁺ neurons suppressed their excitability, driving stress-induced anxiety-related behavior. We propose that a subset of αCaMKII⁺ neurons in the VMHdm/c are rapidly activated in respond to stress and function adaptively to prevent excessive anxiety. This mechanism may involve feedback or feedforward modulation, warranting further investigation.

We observed that VMHdm neurons displayed at least three different firing patterns—tonic firing, burst-firing, and silent (i.e., the absence of firing). Chronic stress increases the proportion of burst-firing SF1⁺ neurons (from 33% – 51%) through Cav3.1 upregulation, which is critical for anxiety-like behavior⁴⁶. Under our experimental conditions, most VMHdm neurons exhibited a tonic-firing pattern, and we did not observe an increase in burst-firing neurons in stressed mice. The photostimulation frequency used in our study (20 Hz) to activate VMH neurons significantly differs from the 0.1 Hz used to specifically activate burst-firing neurons without affecting tonic-firing or silent neurons⁴⁶. Interestingly, stress-activated αCaMKII⁺ neurons (αCamk2⁺/mKate⁺) in the VMHdm/c displayed higher tonic-firing frequency shortly after stress exposure compared to 6 d later, when anxiety-like behavior was fully developed. This contrasts with burst-firing neurons, which show increased excitability alongside stress-induced anxiety enhancement. In contrast, tonic-firing neurons exhibited reduced excitability as stress-induced anxiety-like behavior progressed. It would be valuable to investigate whether burst-firing neurons express GHS-R1a and how stress alters its expression in these neurons. A very recent study report that activating αCaMKII⁺ neurons in the VMH induces anxiety-like behavior, while inhibition reduces³⁴. Our results are inconsistent with these findings, possibly due to the smaller viral volume (60 nl vs. 120 nl in our study) and higher photostimulation intensity (5–10 mW/mm² vs. 1.2 mW/mm² in our study) used in the prior study, which may have targeted a subpopulation of αCaMKII⁺ neurons. Additionally, previous studies have shown that activation of SF1⁺ neurons can drive distinct, even opposite, behavioral responses depending on the specific stimulus parameters^30,32. These differences highlight the complexity of VMH neuron activity in regulating stress and anxiety-like behavior.

The VMH is only one component of a broader neural network involved in stress and anxiety regulation. VMHdm/c neurons project to distinct brain regions including the BNST, anterior hypothalamic nucleus, lateral hypothalamus, medial amygdala, and dorsolateral periaqueductal gray, among others. Future studies are still needed to clarify the specific roles of the VMH and these connected regions in anxiety-like behavior. Additionally, it is essential to investigate whether circulating ghrelin contributes to stress-induced anxiety in the VMH and to uncover the mechanisms driving GHS-R1a upregulation following stress exposure. In conclusion, our findings demonstrate that GHS-R1a signaling promotes stress-induced anxiety by shaping the activity of VMH αCaMKII⁺ neurons. This suggests that GHS-R1a may represent a therapeutic target for treating stress-induced anxiety disorders, including PTSD.

Methods

Animals

Male C57BL/6 J mice were purchased from the Vital River Laboratory Animal Technology Co. (Beijing, China). The Ghsr1a^fl/fl mice (project no. EGE-ZYY-043-A flox mice) were generated by Biocytogen Boston Corp (Massachusetts, USA), based on CRISPR/Cas9 technique to conditionally knockout the exon 1-2 of mouse Ghsr gene (NCBI ID 208188) by Cre-loxP system. A 5’ loxP site is located upstream of exon 1 and 3’ loxP site is inserted into intron 2. The Camk2a-cre mice (B6.Cg-Tg(Camk2a-cre)T29-1Stl/J) were purchased from the Jackson Laboratory. The Ghsr1a^−/− mice with genetic deletion of exon1-2 (Ghsr^tm1Smoc) in C57BL/6 background were purchased from Shanghai Research Center for Model Organisms¹⁷. All F1 heterozygous mice were backcrossed with C57BL/6 J mice for more than seven generations before expanding our own colony. Mice were group-housed under a 12:12 h light/dark cycle with free access to water and food. Only adult male littermates with the age of 3–6 months old were used in experiments. Mice were randomly assigned to experimental or control groups with blockrandomization method. All behavioral experiments were carried out during light cycle from 9:00 AM–6:00 PM. All animal protocols used in this study were approved by the Chancellor’s Animal Research Committee at Qingdao University, in accordance with National Institutes of Health guidelines.

Repeated restraint stress (RRS)

Restraint device was a plexiglas cylindrical holder with front air holes for breathing and an adjustable backboard to minimize animal movements. Mice were temporarily anesthetized with 1.5% isoflurane, and quickly enclosed into the restrainer. The awake mice were then restrained from voluntary movement (without any difficulty to breath) in the holder for 1 h/d, and for consecutive 4 d. Mice returned to home cage after daily restraint.

Virus injection in the VMH and Optical fiber implantation

Mice were anesthetized with 2% isoflurane and head-fixed on a stereotaxic apparatus (RWD Life Science, China). Ophthalmic ointment was applied to prevent eye drying and body temperature was maintained at 37°C using a heating pad. High titer AAV virus (10 ^∧13 GC/ml) were bilaterally delivered into the VMHdm/c through a glass needle powered by a Nanoliter 2000 microinjector (World Precision Instruments Inc., USA) at a flow rate of 50 nl/min. The VMH coordinates were AP -1.45 mm, ML ± 0.3 mm, DV -5.5 mm relative to bregma. GHS-R1a-expressing virus AAV-Camk2a-hGhsr1a-GFP and control virus AAV-Camk2a-GFP, Cre-dependent GHS-R1a-expressing virus AAV-Camk2a-DIO-hGhsr1a-GFP and control AAV-Camk2a-DIO-GFP, chemogenetic virus AAV-Camk2a-hM3D(Gq)-2A-hGhsr1a-GFP and control AAV-Camk2a-hM3D(Gq)-2A-mCherry, AAV-Camk2a-hM4D(Gi)-2A-hGhsr1a-GFP and control AAV-Camk2a-hM4D(Gi)-2A-mCherry, rAAV-RAM-d2TTA::TRE-NLS-mKate2-WPRE-pA, Cre-expressing virus AAV-Camk2a-Cre, and optogenetic virus AAV-Camk2a-hChR2-EYFP and control AAV Camk2a-EYFP were prepared by OBiO Technology (Shanghai, China). AAV-Camk2a-GCaMP6f virus were prepared by Taitool Bioscience Co. Ltd (Shanghai, China). The glass needle was left in original place for an additional 10 min after injection to ensure optimal diffusion. A custom-made fiber (FOC-C-1.25-200-0.37-5.5) prepared by Inper Technology Co. Ltd (Hangzhou, China) was placed 0.2 mm above viral injection site in the VMH. Fiber was embedded unilaterally (for photometry) or bilaterally (for optogenetic stimulation). Mice were recovered for 3 to 4 weeks after virus injection and fiber implantation.

In vivo Ca2⁺ fiber photometry

A single-channel fiber photometry (Thinker Tech Nanjing Biotech Co., Ltd.) was used to record in vivo Ca²⁺ signals in the VMH during mice exploring the elevated plus maze (EPM). The original signal was first subjected to unified denoising processing using MATLAB function ‘wden’. The fluorescence signals were then normalized by (F– F₀)/ F₀, where F₀ is the baseline fluorescence. Mice were allowed to habituate the system for 3 days before starting behavioral test. dF/F reflects the real-time, active state of VMH neuron population during mice exploring predefined areas (open or closed arms). Average dF/F were then calculated by the following equation: Average (dF/ F(X)) = sum (dF/ F(X)) / Time(X), X refer to open or closed arms.

In vivo and ex vivo optogenetic stimulation

Photostimulation parameters used for in vivo experiment were 473 nm wavelength, 1.2 mW/mm² intensity, 20 Hz frequency, 20 ms pulse width. Each stimulation lasted for 20 s with a 10 s interstimulus interval (ISI). Light stimulation was delivered using a manual closed-loop protocol, where the laser was switched on by the observer immediately when the mouse left the closed arms, and was switched off as soon as the mouse re-entered the closed arms. Specifically, as for a 10 min EPM test, light stimulation was delivered only when mice entered center area or open arms during the 2^nd phase of 5 min test, not during the 1^st phase of test. We also conducted control experiments where light stimulation was applied either when mice entered the closed arms or across all zones of the EPM. As for 10 min tail-suspend (TS) test, light was operated alternatively between two testing days. In the real-time place preference (RTPP) test, light stimulation was delivered to a pseudo-random side for 10 min. A TTL-driven light-emitting diode (Lumen Dynamics) was used to generate photostimulation (473-nm, 20 Hz, 5 ms duration) to activate ChR2-expressing VMH neurons in brain slices.

Behavioral tests

Behavioral tests were done between 9:00 am and 6:00 pm by the same investigators unaware of experimental design. Animal behaviors were video-tracked and analyzed by two independent investigators with Noldus EthoVision XT software.

Elevated Plus Maze (EPM)

The EPM test was used to evaluate anxiety-like behavior. The maze has two open arms and two closed arms (with wall height of 16.5 cm), and each arm is 30 cm long and 6 cm wide. Mice were released from the center and allowed to explore the maze for 10 min. Time spent in open and closed arms, number of arm entries, and total travel distance in the maze were analyzed.

Tail Suspension (TS)

Mice was gently suspended upside down with tail (2 cm from the tip) being fixed with adhesive tape to a horizontal bar located 30 cm above the table. Immobility was defined as hanging passively without any voluntary movement except breath. The total immobility time and the latency to first immobility were analyzed during a 10 min TS test.

Forced Swimming (FS)

The FS test was used to evaluate despair-like behavior. Mice were gently released into a transparent plastic cylinder (25 cm height × 10 cm diameter) filled with water (24.5 ± 0.5 °C) up to a depth of 15 cm. Total immobility time and the latency to first immobility were analyzed during a 5-min FS test.

Social Interaction (SI). Individual mouse was first allowed to freely explore the test box (30 cm × 60 cm × 25 cm) for 5 min. After habituation, it was introduced to an unfamiliar, ovariectomized female mouse enclosed in the center of the test box. The social activity, i. e. sniffing the female mouse within close proximity, was then recorded during the following 10-min exploration.

Real-Time Place Preference (RTPP). The RTPP test box was divided into two equal, separated parts with an partition in the middle, and a small door to allow mice freely travel between two sides. First, mice were allowed to explore freely in the box for 10 min, and side preference was analyzed. Then, a train of 20 Hz (20 ms) photostimulation was initiated once mice entered a randomly defined side of stimulation (SOS), and photostimulation was automatically switched off after 20 s or once the mice left the SOS. Movement trajectory was recorded during a 10-min test. The percentage time and travel distance that mice spent in SOS in the presence and absence of photostimulation were analyzed.

Immunohistology and cell counting

Mice were deeply anaesthetized with 2% isoflurane and then transcardially perfused with 4% paraformaldehyde (PFA). The brain were removed and post-fixed with 4% PFA for 2-6 h, and then transferred to 30% sucrose solution. Brains were sliced coronally (40 μm thick) with a vibratome. The primary antibodies used included rabbit anti-FOS antibody (Cell Signaling Technology, 1:3000), mouse anti-αCaMKII antibody (Cell Signaling Technology, 1:100), and chicken anti-GFP antibody (Cell Signaling Technology, 1:1000). The secondary antibodies used were goat anti-rabbit IgG Alexa 488 (Cell Signaling Technology, 1:500), goat anti-mouse IgG Alexa 568 (1:500, Cell Signaling Technology), goat anti-chicken IgG Alexa 488 (Cell Signaling Technology, 1:500). The slices were counter-stained using 40, 60-diamidino-2-phenylindole (DAPI, 1:1000). Fluorescence images were acquired with a laser confocal microscope (FV500, Olympus) equipped with Fluoview 2000 software. The objective lens used was 20X. At least three representative coronal sections spaced equally along the AP axis were adopted for quantifications.

Fluorescence in situ hybridization

Fluorescence in situ hybridization (FISH) was performed with RNAscope Multiplex Fluorescent Reagent Kit V2 (ACD) according to the manufacturer’s instruction and our previous studies¹⁷. Briefly, mice were anaesthetized with 2% isoflurane, and brains were immediately collected and frozen in isopentane. Coronal brain slices (14 μm in thickness) were fixed in 4% PFA for 15 min at 4°C, and then dehydrated sequentially in 50%, 70%, and 100% ethanol, slices were finally air-dried at room temperature. RNA probes for Camk2a, Ghsr1a and Nr5a1 mRNA were purchased from ACD. TSA Plus fluorescein, TSA Plus Cyanine 3, and TSA Plus Cyanine 5 (Akoya Biosciences) were used to develop fluorescent signals that can be captured by a Leica LAS-X confocal microscope with a 63 × oil-immersion objective lens. Gain, threshold, and black levels kept unchanged during individual experiment. We delineated individual neuron boundaries in FISH image data using fluorescence intensity thresholds combined with morphological criteria. Thresholding techniques were applied to isolate neurons based on their fluorescence signals, followed by manual adjustments to align with observed morphology. These methods were applied consistently across samples to ensure accuracy.

Quantitative reverse transcription PCR

Mice were deeply anaesthetized with 2% isoflurane and the brains were quickly dissected out. The VMH tissue was freshly isolated with a Harris Micropuncher (World Precision Instruments, USA) from frozen coronal slices (1 mm thickness) prepared using Mouse Brain Matrix (World Precision Instruments, USA). Isolated VMH samples were quickly collected in enzyme-free EP tube and stored at -80°C until use. The whole process was finished in -20 °C Leica cryostat chamber.

Total RNA was extracted from the VMH of RRS mice and controls using the PureLink RNA Mini Kit (Thermo Fisher Scientific, USA) following manufacturer’s instructions. RNA quantity and quality were measured with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). Single-stranded cDNA was synthesized from 1 µg of total RNA with SuperScript III reverse transcriptase (Invitrogen). The reaction condition was as follows: 25 °C for 10 min, 50 °C for 30 min, 85 °C for 5 min. PCR-based quantification of Ghsr1a transcripts was performed using a Thermal Cycler Dice Real Time System (Roche, USA) and QuantiFast SYBR Green PCR kit (Qiagen). The PCR cycling parameters were: 95 °C for 10 min followed by 40 cycles of PCR reaction at 95 °C for 15 s, 60 °C for 1 min, and 72 °C for 1 min. PCR primer sequences (Thermo Fisher Scientific, USA) were as follow: Ghsr1a-F GAAAATGCTGGCTGTAGTGGTG, Ghsr1a-R GACAAAGGACACGAG-GTTGC; Gapdh-F TGACGTGCCGCCTGGAGAAAC, Gapdh-R CCGGCATCGAAGGTGGAAGAG. 2^−ΔΔCT method was used to normalize CT values against housekeeping gene Gapdh and to quantify relative expression of Ghsr1a. Triplicates were done for each sample.

DREADDs and CNO injection

The hM3Dq- or hM4Di-DREADD is activated by clozapine-N-oxide (CNO, from Tocris, USA). CNO was initially dissolved in DMSO as 5 mg/ml stock solution, which was then diluted 500 times with 0.9% saline for in vivo experiment, and 1500 time with ACSF for ex vivo experiment. Specifically, animals received intraperitoneal (i.p.) injection of CNO (1 mg/kg of body weight) or same amount of normal (0.9%) saline 45 min before behavioral tests¹⁷.

Pharmacological experiments

GHS-R1a antagonist YIL781 (Tocris, USA) was dissolved in 0.9% saline to make the final concentration of 0.3 mg/ml. YIL781 or vehicle was bilaterally microinjected into the VMH through a 28-gauge infusion cannula with the guidance of a preimplanted 22-gauge guide cannula (RWD Life Science, China). The infusion cannula was connected via a PE20 tubing to a Hamilton microsyringe driven by a microinfusion pump (Stoelting Co., USA) with an infusion rate set at 0.1 μl/min for total 0.5 μl/side. The infusion cannula was left in position for an additional 5 min before withdrawal. Drug dosages and the timing of administration were chosen based on our previous studies or preliminary results⁵⁸.

VMH slices preparation and electrophysiological recordings

Mice were deeply anesthetized with 2% isoflurane and then perfused with carbogenated (95% O₂/5% CO₂) ice-cold cutting solution (pH 7.4) containing (in mM): 2.5 KCl, 26 NaHCO₃, 1 NaH₂PO₄, 7 MgSO₄, 1 CaCl₂, 30 Glucose, 119 Choline chloride, 3 Sodium pyruvate, 1 Kynurenic acid, and 1.3 sodium L-ascorbate. Coronal VMH slices (300 μm in thickness) were freshly prepared using a Leica VT-1000 vibratome in oxygenated (95% O₂/5% CO₂) ice-cold cutting solution.Slices were quickly transferred to recovery solution containing (in mM) 85 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 0.5 CaCl₂, 4 MgCl₂, 24 NaHCO₃, 25 glucose, and 50 sucrose to recover for 30 min at 32°C, and then for at least 1 h at room temperature before recording. Slices were then transferred to the submerged recording chamber and continuously perfused with 32°C artificial cerebrospinal fluid (ACSF) containing (in mM): 120 NaCl, 3.5 KCl, 1.25 NaH₂PO₄, 2.5 CaCl_2, 1.3 MgSO_4, 26 NaHCO₃, and 10 glucose.

Spontaneous firing of VMH neurons were measured with both cell-attached voltage-clamp and whole-cell current-clamp recordings. Glass electrodes (4–6 MΩ) were pulled by a micropipette puller (P-1000, Sutter instrument). Cell-attached electrodes were filled with ACSF (pH 7.3-7.4, 295-305 mOsm). Whole-cell recording electrodes were filled with an internal solution containing (in mM): 125 potassium gluconate, 20 KCl, 10 HEPES, 2 MgCl₂, 4 ATP, and 1 EGTA (pH 7.3-7.4, 285-295 mOsm). Only neurons with a resting membrane potential (RMP) smaller than −55 mV were included in data analyses. Data were acquired using digidata 1440 A and pCLAMP 10.0 software (Molecular Devices, USA) with a sampling rate of 10 kHz¹⁷.

For ex vivo whole-cell recordings on stress-driven αCaMKII⁺ neurons identified with the Robust Activity Marking (RAM) system. Considering the characteristics of RAM virus expression, we fed mice food with doxycycline (40 mg/kg) 24 h before injection viruses mixed AAV-Camk2a-GFP and rAAV-RAM-d2TTA::TRE-NLS-mKate2-WPRE-pA. In order to capture more neurons that were activated after RRS (1 h * 4 d), we replace doxycycline with regular food after two days of RRS, and restored doxycycline feed on the 3^rd to 5^th day of RRS. One group resumed a doxycycline diet for 90 min, while the other group resumed a doxycycline diet for 6 d. Subsequently, EPM detected anxiety levels in mice before recoding the excitability changes of the activated neurons (mKate2⁺GFP⁺) in the VMH.

Statistical analyses

Data were expressed as mean ± S.E.M. Statistical analyses were performed with GraphPad Prism 6.0 software. ANOVAs or t-tests were used for statistical comparisons between groups as described in figure legends. Sample sizes were determined based on prior studies conducted in the lab. Sample sizes were also provided in figure legends. Normal distributions and equal variances were assured before performing parametric statistical analyses. Multiple comparisons were done as recommended by the software, using either Tukey’s or Dummett’s multiple comparisons tests. p < 0.05 indicates a significant difference.

Reporting summary

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

Data availability

The data that support the findings of this study are available in the Supplementary Data file. All other data are available from the corresponding author on request. Source data underlying the figures in the main and supplemental figures are provided as supplementary data. Source data are provided with this paper.

References

Penninx, B. W., Pine, D. S., Holmes, E. A. & Reif, A. Anxiety disorders. Lancet (London, England) 397, 914–927 (2021).
Article PubMed Google Scholar
Du Preez, A. et al. The type of stress matters: repeated injection and permanent social isolation stress in male mice have a differential effect on anxiety- and depressive-like behaviours, and associated biological alterations. Transl. Psychiatry 10, 325 (2020).
Article PubMed PubMed Central Google Scholar
Koskinen, M. K. & Hovatta, I. Genetic insights into the neurobiology of anxiety. Trends Neurosci. 46, 318–331 (2023).
Article CAS PubMed Google Scholar
de Kloet, E. R., Joëls, M. & Holsboer, F. Stress and the brain: from adaptation to disease. Nat. Rev. Neurosci. 6, 463–475 (2005).
Article PubMed Google Scholar
Sharma, S., Powers, A., Bradley, B. & Ressler, K. J. Gene × environment determinants of stress- and anxiety-related disorders. Annu. Rev. Psychol. 67, 239–261 (2016).
Article PubMed 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
Collaborators, C.-M. D. Global prevalence and burden of depressive and anxiety disorders in 204 countries and territories in 2020 due to the COVID-19 pandemic. Lancet (London, England) 398, 1700–1712 (2021).
Article Google Scholar
Daly, M. & Robinson, E. Depression and anxiety during COVID-19. Lancet (London, England) 399, 518 (2022).
Article CAS PubMed Google Scholar
Yanagi, S., Sato, T., Kangawa, K. & Nakazato, M. The homeostatic force of Ghrelin. Cell Metab 27, 786–804 (2018).
Article CAS PubMed Google Scholar
Lutter, M. et al. The orexigenic hormone ghrelin defends against depressive symptoms of chronic stress. Nat. Neurosci. 11, 752–753 (2008).
Article CAS PubMed PubMed Central Google Scholar
Huang, H. J. et al. Ghrelin alleviates anxiety- and depression-like behaviors induced by chronic unpredictable mild stress in rodents. Behav. Brain Res. 326, 33–43 (2017).
Article ADS CAS PubMed Google Scholar
Yousufzai, M., Harmatz, E. S., Shah, M., Malik, M. O. & Goosens, K. A. Ghrelin is a persistent biomarker for chronic stress exposure in adolescent rats and humans. Transl. Psychiatry 8, 74 (2018).
Article PubMed PubMed Central Google Scholar
Guan, X. M. et al. Distribution of mRNA encoding the growth hormone secretagogue receptor in brain and peripheral tissues. Brain Res. Mol. Brain Res. 48, 23–29 (1997).
Article CAS PubMed Google Scholar
Zigman, J. M., Jones, J. E., Lee, C. E., Saper, C. B. & Elmquist, J. K. Expression of ghrelin receptor mRNA in the rat and the mouse brain. J. Comp. Neurol. 494, 528–548 (2006).
Article CAS PubMed PubMed Central Google Scholar
Li, N. et al. Selectively increasing GHS-R1a expression in dCA1 excitatory/inhibitory neurons have opposite effects on memory encoding. Mol. Brain 14, 157 (2021).
Article CAS PubMed PubMed Central Google Scholar
Hornsby, A. K. et al. Short-term calorie restriction enhances adult hippocampal neurogenesis and remote fear memory in a Ghsr-dependent manner. Psychoneuroendocrinology 63, 198–207 (2016).
Article CAS PubMed PubMed Central Google Scholar
Li, N. et al. GHSR1a deficiency suppresses inhibitory drive on dCA1 pyramidal neurons and contributes to memory reinforcement. Cereb. Cortex. 33, 2612–2625 (2023).
Article PubMed Google Scholar
Spencer, S. J., Emmerzaal, T. L., Kozicz, T. & Andrews, Z. B. Ghrelin’s role in the hypothalamic-pituitary-adrenal axis stress response: implications for mood disorders. Biol. Psychiatry 78, 19–27 (2015).
Article CAS PubMed Google Scholar
Lu, Y. et al. Acute but not chronic calorie restriction defends against stress-related anxiety and despair in a GHS-R1a-dependent manner. Neuroscience 412, 94–104 (2019).
Article CAS PubMed Google Scholar
Zhang, F. et al. Ghrelin/GHS-R1a signaling plays different roles in anxiety-related behaviors after acute and chronic caloric restriction. Biochem. Biophys. Res. Commun. 529, 1131–1136 (2020).
Article CAS PubMed Google Scholar
Kojima, M. et al. Ghrelin is a growth-hormone-releasing acylated peptide from stomach. Nature 402, 656–660 (1999).
Article ADS CAS PubMed Google Scholar
Ge, X. et al. LEAP2 is an endogenous antagonist of the ghrelin receptor. Cell Metab. 27, 461–469.e466 (2018).
Article MathSciNet CAS PubMed Google Scholar
Xiao, X. et al. A new understanding of GHSR1a-independent of ghrelin activation. Ageing Res. Rev. 64, 101187 (2020).
Article CAS PubMed Google Scholar
Wyrwicka, W. & Dobrzecka, C. Relationship between feeding and satiation centers of the hypothalamus. Science (New York, N.Y.) 132, 805–806 (1960).
Article ADS CAS PubMed Google Scholar
Garfield, A. S. et al. A parabrachial-hypothalamic cholecystokinin neurocircuit controls counterregulatory responses to hypoglycemia. Cell Metab. 20, 1030–1037 (2014).
Article CAS PubMed PubMed Central Google Scholar
Sun, J. S. et al. Ventromedial hypothalamic primary cilia control energy and skeletal homeostasis. J. Clin. Invest. 131, e138107 (2021).
Article CAS PubMed PubMed Central Google Scholar
Anderson, D. J. Circuit modules linking internal states and social behaviour in flies and mice. Nat. Rev. Neurosci. 17, 692–704 (2016).
Article CAS PubMed Google Scholar
Hashikawa, K. et al. Esr1(+) cells in the ventromedial hypothalamus control female aggression. Nat. Neurosci. 20, 1580–1590 (2017).
Article CAS PubMed PubMed Central Google Scholar
Kim, D. W. et al. Multimodal analysis of cell types in a hypothalamic node controlling social behavior. Cell 179, 713–728.e717 (2019).
Article CAS PubMed PubMed Central Google Scholar
Wang, L., Chen, I. Z. & Lin, D. Collateral pathways from the ventromedial hypothalamus mediate defensive behaviors. Neuron 85, 1344–1358 (2015).
Article CAS PubMed PubMed Central Google Scholar
Engelke, D. S. et al. A hypothalamic-thalamostriatal circuit that controls approach-avoidance conflict in rats. Nat. Commun. 12, 2517 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Kunwar, P. S. et al. Ventromedial hypothalamic neurons control a defensive emotion state. Elife 4, e06633 (2015).
Article PubMed PubMed Central Google Scholar
Kennedy, A. et al. Stimulus-specific hypothalamic encoding of a persistent defensive state. Nature 586, 730–734 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Li, Y. et al. CaMKIIa neurons of the ventromedial hypothalamus mediate wakefulness and anxiety-like behavior. Neurochem. Res. 48, 2463–2475 (2023).
Article CAS PubMed Google Scholar
Singh, O. et al. Ghrelin-responsive mediobasal hypothalamic neurons mediate exercise-associated food intake and exercise endurance. JCI Insight 8, e172549 (2023).
Article PubMed PubMed Central Google Scholar
Guo, L. et al. GHS-R1a deficiency alleviates depression-related behaviors after chronic social defeat stress. Front. Neurosci. 13, 364 (2019).
Article PubMed PubMed Central Google Scholar
Zhang, M. et al. Increased GHS-R1a expression in the hippocampus impairs memory encoding and contributes to AD-associated memory deficits. Commun. Biol. 7, 1334 (2024).
Article CAS PubMed PubMed Central Google Scholar
Thiagarajan, T. C., Piedras-Renteria, E. S. & Tsien, R. W. alpha- and betaCaMKII. Inverse regulation by neuronal activity and opposing effects on synaptic strength. Neuron 36, 1103–1114 (2002).
Article CAS PubMed Google Scholar
Yasuda, R., Hayashi, Y. & Hell, J. W. CaMKII: a central molecular organizer of synaptic plasticity, learning and memory. Nat. Rev. Neurosci. 23, 666–682 (2022).
Article CAS PubMed Google Scholar
Lein, E. S. et al. Genome-wide atlas of gene expression in the adult mouse brain. Nature 445, 168–176 (2007).
Article ADS CAS PubMed Google Scholar
Dawson, G. R. & Tricklebank, M. D. Use of the elevated plus maze in the search for novel anxiolytic agents. Trend Pharmacol. Sci. 16, 33–36 (1995).
Article CAS 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
Roth, B. L. DREADDs for neuroscientists. Neuron 89, 683–694 (2016).
Article CAS PubMed PubMed Central Google Scholar
Boyden, E. S., Zhang, F., Bamberg, E., Nagel, G. & Deisseroth, K. Millisecond-timescale, genetically targeted optical control of neural activity. Nat. Neurosci. 8, 1263–1268 (2005).
Article CAS PubMed Google Scholar
Sørensen, A. T. et al. A robust activity marking system for exploring active neuronal ensembles. eLife 5, e13918 (2016).
Article PubMed PubMed Central Google Scholar
Shao, J. et al. Cav3.1-driven bursting firing in ventromedial hypothalamic neurons exerts dual control of anxiety-like behavior and energy expenditure. Mol. Psychiatry 27, 2901–2913 (2022).
Article PubMed PubMed Central Google Scholar
Zhang, X. Y. et al. Targeting presynaptic H3 heteroreceptor in nucleus accumbens to improve anxiety and obsessive-compulsive-like behaviors. Proc. Natl Acad. Sci. USA 117, 32155–32164 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Gebara, E. et al. Mitofusin-2 in the nucleus accumbens regulates anxiety and depression-like behaviors through mitochondrial and neuronal actions. Biol. Psychiatry 89, 1033–1044 (2021).
Article CAS PubMed Google Scholar
Saenz, C. et al. Growth hormone secretagogue receptor and cannabinoid receptor type 1 intersection in the mouse brain. Brain Struct. Funct. 230, 15 (2024).
Article PubMed PubMed Central Google Scholar
Mende, F. et al. Translating biased signaling in the ghrelin receptor system into differential in vivo functions. Proc. Natl Acad. Sci. USA 115, E10255–e10264 (2018).
Article ADS CAS PubMed PubMed Central Google Scholar
Harmatz, E. S. et al. Central Ghrelin resistance permits the overconsolidation of fear memory. Biol. Psychiatry 81, 1003–1013 (2017).
Article CAS PubMed Google Scholar
Spencer, S. J. et al. Ghrelin regulates the hypothalamic-pituitary-adrenal axis and restricts anxiety after acute stress. Biol. Psychiatry 72, 457–465 (2012).
Article CAS PubMed Google Scholar
Stone, L. A., Harmatz, E. S. & Goosens, K. A. Ghrelin as a stress hormone: implications for psychiatric illness. Biol. Psychiatry 88, 531–540 (2020).
Article CAS PubMed Google Scholar
Meyer, R. M., Burgos-Robles, A., Liu, E., Correia, S. S. & Goosens, K. A. A ghrelin-growth hormone axis drives stress-induced vulnerability to enhanced fear. Mol. Psychiatry 19, 1284–1294 (2014).
Article CAS PubMed Google Scholar
Cheung, C. C. et al. Sex-dependent changes in metabolism and behavior, as well as reduced anxiety after eliminating ventromedial hypothalamus excitatory output. Mol. Metab. 4, 857–866 (2015).
Article CAS PubMed PubMed Central Google Scholar
Zhao, L. et al. Central nervous system-specific knockout of steroidogenic factor 1 results in increased anxiety-like behavior. Mol. Endocrinol. (Baltimore, Md.) 22, 1403–1415 (2008).
Article CAS Google Scholar
Yang, F. et al. A GABAergic neural circuit in the ventromedial hypothalamus mediates chronic stress-induced bone loss. J. Clin. Invest. 130, 6539–6554 (2020).
Article CAS PubMed PubMed Central Google Scholar
Li, N. et al. Blocking constitutive activity of GHSR1a in the lateral amygdala facilitates acquisition of conditioned taste aversion. Neuropeptides 68, 22–27 (2018).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

This work was supported by NNSFC (Grant no. 32071141, 91732110 and 32371211 to Yu Zhou), and NSF of SD province (Grant no. ZR2019ZD34 to Yu Zhou). We thank Ms. Jennifer Li and Mr Vincent Li for native language editing.

Author information

Jie Yang
Present address: Dongying No.1 Middle School, Dongying, Shandong, 257000, China
Xiaomin Sun
Present address: Qingdao West Coast New Area No. 9 Senior High School, Qingdao, Shandong, 266500, China
Minglu Niu
Present address: Qingdao Endocrine diabetes Hospital, Qingdao, Shandong, 266000, China
Meng Zhang
Present address: College of Agriculture and Bioengineering, Heze University, Heze, Shandong, 274000, China

Authors and Affiliations

School of Life Sciences and Health, University of Health and Rehabilitation Sciences, Qingdao, Shandong, 266113, China
Meng Zhang, Ming Yu, Nan Li & Yu Zhou
Department of Physiology and Pathophysiology, School of Basic Medical Sciences, Qingdao University, Qingdao, Shandong, 266071, China
Meng Zhang, Liu Yang, Xue Mi, Gonghui Hu, Yingchang Lu, Chen Wang, Jie Yang, Xiaomin Sun, Minglu Niu, Xianchao Li, Sihan Wang, Jingsai Zhang, Hanbing Yu & Yu Zhou
Institute of Brain Sciences and Related Disorders, Qingdao University, Qingdao, Shandong, 266071, China
Meng Zhang, Liu Yang, Xue Mi, Gonghui Hu, Yingchang Lu, Chen Wang, Jie Yang, Xiaomin Sun, Minglu Niu, Xianchao Li, Sihan Wang, Jingsai Zhang, Hanbing Yu & Yu Zhou
Department of Rehabilitation Medicine, Affiliated Hospital of Qingdao University, Qingdao, Shandong, 266000, China
Yuyang Wang & Yu Zhou

Authors

Meng Zhang
View author publications
Search author on:PubMed Google Scholar
Liu Yang
View author publications
Search author on:PubMed Google Scholar
Xue Mi
View author publications
Search author on:PubMed Google Scholar
Gonghui Hu
View author publications
Search author on:PubMed Google Scholar
Yingchang Lu
View author publications
Search author on:PubMed Google Scholar
Chen Wang
View author publications
Search author on:PubMed Google Scholar
Jie Yang
View author publications
Search author on:PubMed Google Scholar
Xiaomin Sun
View author publications
Search author on:PubMed Google Scholar
Minglu Niu
View author publications
Search author on:PubMed Google Scholar
Xianchao Li
View author publications
Search author on:PubMed Google Scholar
Sihan Wang
View author publications
Search author on:PubMed Google Scholar
Jingsai Zhang
View author publications
Search author on:PubMed Google Scholar
Hanbing Yu
View author publications
Search author on:PubMed Google Scholar
Yuyang Wang
View author publications
Search author on:PubMed Google Scholar
Ming Yu
View author publications
Search author on:PubMed Google Scholar
Nan Li
View author publications
Search author on:PubMed Google Scholar
Yu Zhou
View author publications
Search author on:PubMed Google Scholar

Contributions

Y.Z. designed and supervised the experiments. M.Z., X.M., G.H.H., C.W., J.Y., M.L.N. and S.H.W. performed behavioral experiments. L.Y. and N.L. performed electrophysiological experiments. M.Z., G.H.H., C.W. and X.M.S. performed biochemical experiments. X.C.L., J.S.Z. and H.B.Y. perform genotyping of transgenic mice. M.Z., Y.C.L., L.Y., Y.Y.W. and M.Y. helped with data analysis. M.Z., X.M., L.Y. and G.H.H. constructed figures. Y.Z., M.Z. and Y.C.L. wrote the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Yu Zhou.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Haijiang Cai, Ki Goosens 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 (download DOCX )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source data (download XLSX )

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

Zhang, M., Yang, L., Mi, X. et al. GHS-R1a signaling drives anxiety-related behavior by shaping excitability of ventromedial hypothalamic neurons. Nat Commun 16, 3858 (2025). https://doi.org/10.1038/s41467-025-59116-3

Download citation

Received: 29 November 2023
Accepted: 10 April 2025
Published: 24 April 2025
Version of record: 24 April 2025
DOI: https://doi.org/10.1038/s41467-025-59116-3