Osteocalcin is a compensatory hormone to protect the nervous system in ovariectomized mice Authors: Yiting Kang, Yaoyao Yao, Zeguo Feng, Xiaohang Gao, Yifei Song, Qianyu Zhang, Xiaohui Di, Lei Huang, Wei Han, Jianbao Zhang Published: 2026 ABSTRACT AbstractUndercarboxylated osteocalcin (ucOCN), a bone-derived hormone, modulates neurocognition and emotional regulation. Its concentration declines with aging but rises secondarily in females with estrogen deficiency caused by ovarian decline. Using an ovariectomized (OVX) mouse model, we found serum ucOCN sharply increased in early OVX stages with normal behavioral performance, whereas prolonged estrogen loss decreased ucOCN and induced cognitive and emotional impairments. Ocn knockout mice failed to exhibit OVX-triggered ucOCN elevation and displayed severe synaptic damage and behavioral deficits at the early stage, which were rescued by exogenous ucOCN supplementation. Furthermore, long-term ucOCN administration ameliorated behavioral impairments induced by sustained estrogen deficiency, upregulated the expression of synaptic proteins, and restored hippocampal synaptic plasticity. Collectively, dynamic alterations in ucOCN are correlated with neurobehavioral changes under estrogen deficiency. As a time-dependent neurocompensatory mediator, ucOCN maintains hippocampal synaptic integrity. Long-term ucOCN intervention may provide a promising therapeutic strategy for estrogen deficiency-related cognitive and emotional disorders.Subject terms: Neuroscience, Physiology Introduction Perimenopause, a pivotal midlife transition phase for women characterized by the gradual decline of ovarian function1. This decline is often accompanied by complex perimenopausal syndromes, including anxiety, memory loss, and other neurological symptoms2. More importantly, these physiological changes are widely regarded as major contributors to the increased risk of neurodegenerative diseases, such as Alzheimer’s disease and depression in women3–5. Menopausal hormone therapy (MHT) exerts a beneficial effect on alleviating estrogen deficiency-associated symptoms in women with ovarian function decline6. However, its clinical application remains controversial and limited by a well-recognized age-dependent and regimen-specific risk-benefit profile: emerging evidence has indicated that potential adverse outcomes such as breast cancer risk may be observed in elderly women or those with long-term MHT use6–8. Meanwhile, the underlying mechanisms by which MHT regulates neural function remain poorly understood, largely due to the relative paucity of preclinical studies using female animal models. Furthermore, systematic investigations based on hormone-deficient female models are even more limited. Therefore, it is urgent to identify safe and effective alternative therapeutic strategies to alleviate the adverse neural outcomes induced by estrogen deficiency in females. A sharp decline in estrogen levels in menopausal women is closely associated with significant impairments in synaptic plasticity and cognitive function. The hippocampus and prefrontal cortex, as core brain regions involved in memory encoding, emotional regulation, and synaptic remodeling, are highly sensitive to estrogen deficiency9–11. Long-term estrogen deprivation reduces dendritic spine density in the hippocampus, disrupts synaptic ultrastructure, and downregulates the expression of key synaptic proteins such as PSD95, synaptophysin (SYN), and brain-derived neurotrophic factor (BDNF), thereby weakening synaptic transmission function and ultimately leading to learning and memory deficits as well as anxiety- and depression-like behaviors9–14. These structural and functional alterations in the hippocampus are considered the core pathological basis of menopause-related cognitive and affective disorders. In addition, aging and menopause per se are characterized by a state of chronic low-grade systemic inflammation, which is further exacerbated by estrogen deficiency. Estrogen exerts a significant inhibitory effect on microglial activation and pro-inflammatory cytokine production. Estrogen withdrawal during menopause can trigger peripheral inflammatory responses and promote neuroinflammation in the hippocampus and cortex, which further impairs synaptic plasticity, suppresses BDNF expression, accelerates neuronal dysfunction, and consequently exacerbates cognitive decline and emotional disorders3,5,10,15. This inflammatory mechanism provides a crucial supplementary link in elucidating the association between estrogen withdrawal and neurological dysfunction. Bone, as a target tissue of estrogen, has gained widespread scientific recognition for its endocrine functions2. Osteocalcin (OCN), an important functional protein secreted by bone, is synthesized by osteoblasts. During perimenopause, the decline in estrogen levels accelerates bone remodeling, prompting osteoblasts to synthesize more OCN2. Among them, part of OCN is deposited in the bone matrix after carboxylation (Carboxylated osteocalcin, cOCN), while the other part is released into the bloodstream due to incomplete carboxylation (undercarboxylated Osteocalcin, ucOCN). In addition, enhanced osteoclast activity forms bone lacunae. Within these lacunae, cOCN deposited in the bone matrix undergoes decarboxylation, and the decarboxylated products are subsequently released into the blood2. Notably, a growing body of evidence has clearly unveiled the close association between estrogen and OCN at the levels of synthesis and secretion16,17. A decrease in endogenous estrogen levels disrupts the balance of bone metabolism, triggering a high bone turnover state and subsequently promoting the release of OCN16,17. Conversely, exogenous estrogen supplementation can effectively restore bone metabolic homeostasis and reverse this change16, indicating an intrinsic regulatory interaction between them. Considering that circulating OCN levels gradually decrease with age and present a secondary peak of circulating OCN levels only during perimenopause18,19, we postulate that OCN has an important role in perimenopause. UcOCN has been revolutionarily discovered to traverse the blood-brain barrier, thereby enhancing cognition and preventing anxiety and depression20–23. It is reasonable to postulate that the elevation of ucOCN levels during perimenopause represents an evolutionary adaptation that provides cerebral protection to females in the face of declining estrogen. Notably, this transient rise in ucOCN may reflect an early compensatory effort by the organism to maintain neural homeostasis following the abrupt drop in estrogen. In fact, as bone mineral reserves become depleted and osteoblastic activity declines with advancing age and prolonged hypoestrogenism24,25, this endogenous protective mechanism wanes, potentially contributing to the heightened vulnerability to neurodegenerative and affective disorders observed in postmenopausal women. To verify these hypotheses, we monitored the circulating ucOCN levels and behavioral changes in ovariectomized (OVX) mice at multiple postoperative time points. As a classic preclinical research model, the OVX model can stably induce systemic estrogen deficiency, enabling us to specifically investigate the independent effects of estrogen depletion on neurobehavioral homeostasis26. This model exhibits excellent reproducibility, providing a reliable foundation for related studies; however, the young OVX model cannot fully replicate the perimenopausal and menopausal characteristics in humans, nor can it simulate age-related estrogen fluctuations and complex neurobiological changes. Nevertheless, this model remains suitable for uncovering the core regulatory mechanisms of ucOCN. Subsequently, we compared the early postoperative behavioral performance and synaptic plasticity changes between wild-type and osteocalcin knockout (Ocn−/−) mice after ovariectomy to evaluate the compensatory protective effect of the elevated endogenous ucOCN following ovariectomy. In addition, we administered long-term exogenous ucOCN supplementation to OVX mice to explore the potential therapeutic value of exogenous ucOCN under the condition of long-term estrogen deficiency. To our knowledge, this study is the first to systematically investigate the early compensatory role of endogenous ucOCN in estrogen-deficient models and clarify its neuroprotective mechanism by integrating genetic knockout and exogenous supplementation strategies. Results Circulating OCN and behavioral changes post-ovariectomy in mice Figure 1A represents our experimental protocol. At the 4th week after ovariectomy, ucOCN and cOCN reached a peak and they are higher than those in the Sham mice during weeks 4 to 8, then followed by a gradual decline. At the 14th and 16th weeks, the level of ucOCN became lower than that in the Sham group. However, for Sham mice, the levels of these markers persistently decreased during experiments (Fig. 1B and C). Meanwhile, the estrogen levels in OVX mice were significantly reduced, and remained consistently lower than those in Sham mice (Fig. 1D). We also calculated the ratios of cOCN/E2 and ucOCN/E2 for each mouse, and the results are shown in Fig. S1. In OVX mice, the cOCN/E2 ratio was significantly elevated as early as 2 weeks after ovariectomy, peaked at 4 weeks, remained high at 8 weeks, and gradually decreased after 12 weeks. The ucOCN/E2 ratio was significantly increased at 2 weeks after ovariectomy, reached its maximum at 4 weeks, and then showed a gradual downward trend. In sham-operated mice, both cOCN/E2 and ucOCN/E2 ratios exhibited a stable physiological trend of slow decline over time. The levels of circulating osteocalcin and behavioral performance of mice at different time points post-ovariectomy. A Schematic of the experimental protocol. B Circulating cOCN. C Circulating ucOCN. D Circulating estradiol (E2). B–D Two-tailed t-test for each time point, n = 6 mice. E Behavioral performance of mice in the OFT. F Behavioral performance of mice in the EPMT. G Time spent immobile in the TST. E–G Two-tailed t-test or Mann–Whitney U Test. n = 10 mice. H Number of platform crossings during the spatial probe test in the MWM, Two-tailed t-test or Mann–Whitney U Test, n = 10 mice. I The escape latency in the four-day place navigation test of the MWM, Repeated measures ANOVA. n = 10 mice. Two-tailed t-test for each time point ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, ns: not significant. Data are represented as means ± SD. Behavioral schematics were drawn using BioRender, and remaining graphs were plotted and assembled using GraphPad Prism 9.0 and Adobe Illustrator 2020. Anxiety-like behavioral, depression-like behavior, and learning and memory capabilities were performed respectively at the 4th, 8th, 12th and 16th weeks after ovariectomy (Fig. 1E–I). At the 4th and 8th weeks, OFT, EPMT, TST and MWM test did not detect behavioral differences between OVX and Sham mice. At the 12th weeks, for OFT, OVX mice showed significantly reduced moving distance in the central zone and time spent in the central zone compared to the peripheral zone, while the number of entries into the central zone was not significantly decreased (Fig. 1E). Furthermore, all the above indicators were significantly reduced at the 16th week (Fig. 1E). For EPMT, at the 12th weeks, OVX mice exhibited shortened open-arm activity distance and a reduced open-to-closed arm time ratio, but there was no difference in the number of entries into the open arms (Fig. 1F). However, at the 16th week, all of these indicators were significantly lower in the OVX mice (Fig. 1F), indicating higher anxiety-like behavior. The results of the TST showed that compared with Sham mice, OVX mice did not exhibit a significant increase in immobility time until the 16th week after ovariectomy (Fig. 1G), indicating obvious depressive-like behavior. In the MWM test, at both 12th and 16th weeks after ovariectomy, the escape latency of OVX mice was prolonged, and the number of platform crossings in the spatial probe test was significantly reduced (Fig. 1H, I). These results show that anxiety-like behavior and impairments in learning and memory abilities did not immediately appear after ovariectomy. This phenomenon may be associated with elevated circulating levels of ucOCN, and previous studies have confirmed that circulating ucOCN is involved in the regulation of cognitive function and emotional behavior20,21. Elevated endogenous OCN protects against behavioral impairments in early post-OVX mice To demonstrate the protective effect of ucOCN after ovariectomy, we compared the behavioral performance and synaptic plasticity at 4 weeks post-ovariectomy between Ocn−/− mice and WT mice. Figure 2A represents the experimental protocol. As shown in Fig. 2B, mice exhibited a significant decrease in serum E2 levels at the 4th week after ovariectomy, while ucOCN supplementation had no effect on this parameter. Ocn−/− mice exhibited several behavioral deficits post-ovariectomy, while WT mice remained unaffected. A Schematic of the experimental timeline. B Circulating E2. One-way ANOVA with Tukey’s post hoc test, n = 6 mice. C Distance in the center zone. D The Central-Peripheral Zone Time Ratio in the OFT. E No. of center zone entries in the OFT. C–E One-way ANOVA with Tukey’s post hoc test, n = 10 mice. F Representative traces of OFT. G Distance in the open arm. H The open-closed arm time ratio in the EPMT. I No. of open arm entries in the EPMT. G–I One-way ANOVA with Tukey’s post hoc test, n = 10 mice. J Representative traces of EPMT. K Time spent in immobile in the TST. One-way ANOVA with Tukey’s post hoc test, n = 10 mice. L Number of platform crossings during the spatial probe test in the MWM. One-way ANOVA with Tukey’s post hoc test, n = 10 mice. M The escape latency in the four-day place navigation test of the MWM. Repeated measures ANOVA, n = 10 mice. ****P < 0.0001, ***P < 0.001,**P < 0.01, *P < 0.05, ns: not significant. Data are represented as means ± SD. Figures were plotted and assembled using GraphPad Prism 9.0 and Adobe Illustrator 2020. Based on the finding in Experiment 1 (Fig. 1E–I) that serum ucOCN peaked at 4 weeks post-OVX, the present experiment selected 4 weeks post-OVX as the observation time point to verify whether the peak elevation of endogenous ucOCN exerts compensatory neuroprotective effects. Notably, under sham-operated conditions, Ocn−/− mice exhibited significant baseline differences compared with WT mice. In the OFT, sham Ocn−/− mice showed significantly reduced central zone distance, central/peripheral time ratio, and central zone entries relative to sham WT mice (Fig. 2C–E). In the EPMT, sham Ocn−/− mice also displayed decreased open-arm distance, open/closed arm time ratio, and open-arm entries compared with sham WT mice (Fig. 2G–J). In the TST, sham Ocn−/− mice exhibited significantly longer immobility time than sham WT mice (Fig. 2K). In the MWM, sham Ocn−/− mice showed prolonged escape latency and reduced platform crossing numbers compared with sham WT mice (Fig. 2L, M). These observations indicate that global osteocalcin deficiency alone is sufficient to induce anxiety-like behaviors, depression-like behaviors, and spatial cognitive impairment under physiological estrogen conditions. At the 4th week after ovariectomy, WT-OVX mice showed no behavioral differences compared with the Sham mice in the OFT, EPMT, TST and MWM (Fig. 2C-M). However, only 4 weeks after ovariectomy, Ocn−/− mice showed significantly reduced central zone activity distance, central/peripheral zone time ratio and number of central zone entries in OFT (Fig. 2C–F), significantly decreased open arm activity distance, open/closed arm time ratio and number of open arm entries in EPMT(Fig. 2G–J), significantly prolonged immobility time in TST(Fig. 2K), and delayed platform location learning as well as reduced crossing times in MWM (Fig. 2L, M). These behavioral impairments were ameliorated following ucOCN supplementation. This suggests that when the compensatory protection of endogenous ucOCN is lacking, estrogen deficiency can directly trigger anxiety- and depression-like behaviors and spatial cognitive impairments in mice within 4 weeks. Taken together, this study confirms that the elevation of endogenous ucOCN in the early stage after ovariectomy has a compensatory protective effect against anxiety, depression, and cognitive dysfunction in mice. Elevated endogenous ucOCN maintains hippocampal dendritic spine density and synaptic number in early post-ovariectomy mice Synapses, as critical connections for interneuronal information transmission, exhibit plasticity, a dynamic capacity for morphological and functional changes that plays a pivotal role in cognitive and emotional regulation9. Dendritic spines, serving as essential sites for synaptic formation, are indispensable for neural information processing. To explore the mechanisms underlying behavioral alterations in OVX mice, this study employed Golgi staining to analyze hippocampal dendritic spine density and transmission electron microscopy to visualize synaptic ultrastructure. As shown in Fig. 3A–C, Golgi staining results demonstrated that at the 4th week after ovariectomy, there were no significant differences in hippocampal neuronal morphology or dendritic spine density between the OVX group and the Sham group. Transmission electron microscopy results (Fig. 3D, E) indicated that the synaptic structures of Sham group mice were intact, with normal synaptic vesicle density and postsynaptic membrane thickness, and no significant differences were observed between the OVX group and the Sham group. However, at the 4th week after ovariectomy, Ocn−/− mice exhibited a significant decrease in hippocampal dendritic spine density and a marked reduction in hippocampal synaptic number, and these impairments could be significantly reversed by ucOCN supplementation. These findings collectively indicate that estrogen depletion exacerbates the loss of hippocampal dendritic spines and synapses in the absence of endogenous ucOCN; additionally, the elevation of endogenous ucOCN in the early stage post-ovariectomy is crucial for maintaining hippocampal dendritic spine density and synaptic number, and this loss effect can be rescued by exogenous ucOCN supplementation. Ocn−/− mice showed impaired synaptic plasticity post-ovariectomy, while WT mice remained unaffected. A Representative images of coronal sections of the hippocampal from Golgi stain, Scale bars: 500 µm. B Representative images of selected dendritic segments from the CA1 region. Scale bars: 10 µm. C The average dendritic spine density in the CA1 region of hippocampus. One-way ANOVA with Tukey’s post hoc tes. n = 15 dendritic segments form 3 mice. D Representative images of synaptic structures in the CA1 region of the hippocampus observed by transmission electron microscopy (TEM). Scale bars: 1.0 µm. E The number of synapses per 100 µm2. One-way ANOVA with Tukey’s post hoc test. 5 random fields per slide, 1 slide per animal, and 3 animals per group (n = 15). ns: not significant. Data are represented as means ± SD. Figures were plotted and assembled using GraphPad Prism 9.0 and Adobe Illustrator 2020. G protein-coupled receptor 158 (GPR158), the central nervous system receptor for ucOCN, binds ucOCN in hippocampal neurons to form a GPR158-Gαq complex that promotes BDNF synthesis and secretion21. Based on this mechanism, we quantified hippocampal expression of GPR158, BDNF and synaptic plasticity proteins in mice. As shown in Fig. 4A, B, at 4 weeks post-ovariectomy, there were no significant differences in the protein expression levels of GPR158 in the hippocampal tissues among all groups of mice. In contrast, the expression levels of BDNF and PSD 95 in the hippocampal tissues of Ocn−/− mice were significantly decreased, whereas no significant differences were observed in the above two indicators in wild-type mice (Fig. 4A–C). However, ovariectomy did not induce a significant downregulation of synapsin (SYN) protein expression in either WT or Ocn−/− mice (Fig. 4A, E). Notably, exogenous supplementation with ucOCN significantly promoted the expression of BDNF, PSD-95, and SYN proteins in the hippocampus of ovariectomized Ocn⁻/⁻ mice. Ocn−/− mice showed reduced expression of synaptic plasticity-related proteins in the hippocampus after ovariectomy, while WT mice remained unaffected. A Representative Western blot images. B–E The relative expression of Gpr158, BDNF, PSD 95 and SYN in the hippocampus. One-way ANOVA with Tukey’s post hoc test. n = 4 biologically independent samples. ****P < 0.0001, ***P < 0.001,**P < 0.01, *P < 0.05, ns: not significant. Data are represented as means ± SD. Figures were plotted and assembled using GraphPad Prism 9.0 and Adobe Illustrator 2020. In addition, sham Ocn−/− mice exhibited significantly lower hippocampal dendritic spine density and fewer synapses per unit area than sham WT mice (Fig. 3A–E). Meanwhile, sham Ocn−/− mice also showed significantly decreased protein expression levels of BDNF, PSD95, and SYN in the hippocampus compared with sham WT mice, while GPR158 expression was unchanged (Fig. 4A–E). These results indicate that osteocalcin deficiency alone impairs hippocampal synaptic structure and reduces the expression of key synaptic plasticity–related proteins under physiological estrogen conditions. Long-term exogenous supplementation of ucOCN ameliorates behavioral deficits in OVX mice According to the results of Experiment 1 showing significantly decreased ucOCN levels and severe neurobehavioral deficits at 16 weeks post-OVX, a 16-week exogenous ucOCN administration regimen was applied to investigate its therapeutic effects on long-term estrogen deficiency-induced neural injury. Figure 5A shows the experimental flow chart. As illustrated in Fig. 5B, long-term exogenous ucOCN supplementation exerted no significant impact on the reduced E2 levels in OVX mice, indicating that its mechanism is independent of estrogen regulation. Behavioral analyses revealed that compared with OVX controls, ucOCN-treated mice exhibited significant increases in OFT parameters: central zone activity distance, the ratio of central-to-peripheral activity duration and number of central zone entries (Fig. 5C–F). In the EPMT, these mice also showed enhanced open arm activity distance and a higher open-to-closed arm activity ratio (Fig. 5G, H), However, there was no significant difference in the number of open-arm entries among the various groups (Fig. 5I). Collectively, these results suggest that the anxiety-like behaviors of OVX mice were ameliorated following exogenous supplementation of ucOCN. Findings from the TST further demonstrated that ucOCN supplementation markedly shortened immobility time (Fig. 5K), reflecting improved depressive symptoms. Concurrently, MWM results showed that ucOCN-treated mice displayed a significantly reduced escape latency during spatial exploration and increased platform crossings in the place navigation task, confirming restoration of impaired spatial learning and memory in OVX mice (Fig. 5L, M). Long-term exogenous ucOCN supplementation ameliorates behavioral impairments in OVX mice. A Schematic of the experimental timeline. B Circulating E2. One-way ANOVA with Tukey’s post hoc test, n = 6 mice. C Distance in the center zone. D The Central-Peripheral Zone Time Ratio in the OFT. E No. of center zone entries in the OFT. C–E Kruskal-Wallis test for Distance in the center zone and Center zone time. One-way ANOVA with Tukey’s post hoc test for No. of center zone entries. n = 10 mice. F Representative traces of OFT. G Distance in the open arm. H The open-closed arm time ratio in the EPMT. I No. of open arm entries in the EPMT. G–I One-way ANOVA with Tukey’s post hoc test for Distance in the open arm and Time spent in open arm. Kruskal-Wallis test for No. of open arm entries in the EPMT. J Representative traces of EPMT. K Time spent in immobile in the TST. One-way ANOVA with Tukey’s post hoc test. n = 10 mice. L The escape latency in the four-day place navigation test of the MWM. Repeated measures ANOVA. n = 10 mice. M Number of platform crossings during the spatial probe test in the MWM. One-way ANOVA with Tukey’s post hoc test. n = 10 mice. ****P < 0.0001, ***P < 0.001,**P < 0.01, *P < 0.05, ns: not significant. Data are represented as means ± SD. Figures were plotted and assembled using GraphPad Prism 9.0 and Adobe Illustrator 2020. Long-term exogenous ucOCN supplementation ameliorates reductions in hippocampal dendritic spine density and synaptic number in OVX mice As shown in Fig. 6A, B, the number of hippocampal neurons in OVX mice was reduced compared with the Sham group, while long-term ucOCN supplementation alleviated hippocampal neuronal loss. Quantitative analysis in Fig. 6C showed that the dendritic spine density in the hippocampal CA1 region of the OVX group was significantly lower than that of the Sham group, and this index in the ucOCN supplementation group significantly rebounded compared with the OVX group, approaching the level of the Sham group, suggesting that ucOCN has a reparative effect on castration-induced dendritic spine loss. Transmission electron microscopy observation of hippocampal synaptic ultrastructure (Fig. 6D) showed that the synaptic pre- and post-membranes in the Sham group were intact, the synaptic vesicles in the presynaptic membrane were densely arranged, the postsynaptic density was uniform in thickness, and the synaptic cleft was clear. In contrast, synapses in the OVX group showed obvious degeneration. By comparison, the synaptic structure of the OVX+ucOCN group was significantly improved, with increased number of vesicles, restored postsynaptic density thickness, and narrowed synaptic cleft, indicating that ucOCN could reverse castration-induced synaptic degeneration. Quantitative analysis in Fig. 6E further confirmed that the number of synapses per unit area in the hippocampus of the OVX group was significantly lower than that of the Sham group, while the OVX+ucOCN group showed a significant increase in the number of synapses, suggesting that ucOCN may exert a neuroprotective effect by maintaining synaptic structural integrity and promoting synaptogenesis. Long-term ucOCN supplement prevented decrease in hippocampal spine density and synaptic number in OVX mice. A Representative images of coronal sections of the hippocampal from Golgi stain, Scale bars: 500 µm. B Representative images of selected dendritic segments from the CA1 region. Scale bars: 10 µm. C The average dendritic spine density in the CA1 region of hippocampus. One-way ANOVA with Tukey’s post hoc test. n = 15 dendritic segments form 3 mice. D Representative images of synaptic structures in the CA1 region of the hippocampus observed by transmission electron microscopy (TEM). Scale bars: 1.0 µm. E The number of synapses per 100 µm2. One-way ANOVA with Tukey’s post hoc test. 5 random fields per slide, 1 slide per animal, and 3 animals per group (n = 15). ns not significant. Data are represented as means ± SD. Figures were plotted and assembled using GraphPad Prism 9.0 and Adobe Illustrator 2020. Consistent with these results, as shown in Fig. 7A–E, at the 16th week after ovariectomy, OVX mice exhibited significantly lower hippocampal expression of GPR158, BDNF, PSD95 and SYN versus those in sham-operated controls. These findings aligned with dendritic spine density changes observed via Golgi staining and transmission electron microscopy analysis of synaptic ultrastructure. Long-term ucOCN supplementation prevents ovariectomy-induced reduction of synaptic plasticity-related proteins in the hippocampus. A Representative Western blot images. B–E The relative expression of Gpr158, BDNF, PSD 95 and SYN in the hippocampus. One-way ANOVA with Tukey’s post hoc test. n = 4 biologically independent samples. ****P < 0.0001, ***P < 0.001,**P < 0.01, *P < 0.05, ns not significant. Data are represented as means ± SD. Figures were plotted and assembled using GraphPad Prism 9.0 and Adobe Illustrator 2020. Discussion To the best of our knowledge, this is the first study to investigate the neuroprotective role of endogenous ucOCN in estrogen-deficient mouse models. Our results demonstrated that early elevation of endogenous ucOCN levels following ovariectomy suppressed the development of anxiety-like and depression-like behaviors as well as learning-memory impairments in OVX mice, while preserving hippocampal dendritic spine density and synaptic number. These findings suggest that ucOCN acts as an early compensatory protective factor for the nervous system in the setting of estrogen decline. However, with the progressive reduction in ucOCN levels, OVX mice exhibited significant behavioral deficits by the 16th week post-ovariectomy. Long-term exogenous ucOCN supplementation enhanced the expression of key synaptic plasticity-related proteins, improved synaptic function in OVX mice, and thereby mitigated these behavioral impairments. Estrogen plays an indispensable role in cognitive function and emotional regulation10,11. Notably, in the ovariectomized mouse model, behavioral deficits do not emerge immediately after the abrupt drop in estrogen levels but develop progressively11,27, indicating that the body may possess endogenous compensatory mechanisms to buffer the adverse consequences of hormonal fluctuations. Of relevance, multi-species studies have shown that circulating OCN levels gradually decrease with age during early life but rise again in females with estrogen deficiency induced by ovarian function decline18,19,28,29. In line with these observations, our present study demonstrated that serum ucOCN levels peaked at 4 weeks after OVX and then gradually declined, accompanied by corresponding changes in the cOCN/E2 and ucOCN/E2 ratios. In sham-operated mice, serum ucOCN levels as well as cOCN/E2 and ucOCN/E2 ratios declined slowly with age, reflecting the physiological homeostasis of bone metabolism during skeletal maturation. Specifically, during early adulthood, female mice undergo peak bone mass consolidation, during which osteoblast activity shifts toward matrix mineralization accompanied by enhanced γ‑carboxylation efficiency30,31. Concurrently, vitamin K-dependent γ-carboxylation efficiency matures, increasing the conversion of ucOCN to cOCN, which is subsequently deposited in the bone matrix, leading to reduced circulating ucOCN32. Consistent with our data, circulating ucOCN levels naturally decline substantially in female C57BL/6 J mice during early adulthood (2–9 months), paralleling reduced bone formation rate29,30. Upon reaching full skeletal maturity at approximately 6 months, bone formation and resorption become tightly coupled, allowing ucOCN levels to stabilize with a mild adaptive rebound to preserve bone turnover sensitivity31–33. In contrast, OVX elicited a transient early increase in ucOCN and a prominent elevation in the ucOCN/E2 ratio at 2–4 weeks, representing an acute compensatory response to estrogen deprivation34,35. The abrupt loss of estrogen markedly accelerates bone turnover and activates osteoblasts, which release large amounts of ucOCN before complete γ‑carboxylation36. This surge of ucOCN may partially counteract metabolic disturbances via its endocrine actions and help maintain skeletal and systemic homeostasis. However, with prolonged estrogen deficiency, bone remodeling becomes progressively uncoupled, osteoblastic activity shifts from compensatory activation to gradual attenuation, and both ucOCN levels and the ucOCN/E2 ratio decline, indicating that this compensatory response is transient and limited. A similar pattern observed in the cOCN/E2 ratio further supports the involvement of total osteocalcin secretion in this adaptive process, while ucOCN likely plays a more prominent regulatory role due to its unique hormone‑like activity. Collectively, these dynamic temporal changes confirm that ucOCN acts as an important endogenous compensatory factor activated in response to estrogen deficiency, and its fluctuations reflect the adaptive capacity of the bone‑endocrine axis. This pattern may be related to the release of ucOCN triggered by rapid bone loss during the perimenopausal period37,38. Given that ucOCN can cross the blood‑brain barrier to modulate brain development, cognition, and behavior20, we hypothesized that ucOCN might serve as a compensatory factor enabling the body to counteract the detrimental effects of estrogen reduction on the nervous system. To fully elucidate the compensatory protective effect of ucOCN during the early post-ovariectomy period, we utilized osteocalcin knockout (Ocn−/−) mice and compared their behavioral phenotypes with those of wild-type (WT) mice at the early stage of ovariectomy (4 weeks post-surgery). Our results showed that ovariectomized Ocn−/− mice displayed prominent anxiety-like and depression-like behaviors, along with significantly aggravated impairments in spatial learning and memory abilities. In contrast, these abnormal behavioral phenotypes were not observed in ovariectomized WT mice during the same period. These findings indicate that Ocn−/− mice are unable to initiate the compensatory elevation of endogenous ucOCN, thus failing to mitigate the adverse effects of E2 deficiency on the nervous system and ultimately leading to severe behavioral and cognitive deficits. Further experiments involving exogenous ucOCN supplementation confirmed that ucOCN administration could partially reverse the affective and cognitive deficits in ovariectomized Ocn−/− mice. Taken together, these results strongly support our central hypothesis: the elevation of ucOCN levels during perimenopause constitutes a critical mechanism for protecting brain function against estrogen deficiency-induced damage, and this perimenopause-associated compensatory increase in ucOCN is a necessary condition for counteracting the neurobehavioral impairments caused by E2 reduction. Consistent with our findings, adult Runx2 heterozygous knockout (Runx2+/−) mice exhibit reduced circulating levels of bioactive ucOCN due to bone resorption defects, which subsequently leads to cognitive impairments and anxiety-like behaviors28, further validating the core role of ucOCN in neuroprotection. Of note, Sham-operated Ocn−/− mice exhibited significant anxiety-like behaviors, depression-like behaviors, and spatial cognitive deficits compared with sham WT mice in the absence of ovariectomy. Meanwhile, these mice also showed lower hippocampal dendritic spine density, fewer synaptic numbers, and decreased expression of BDNF, PSD95, and SYN relative to sham WT mice. These baseline differences demonstrate that endogenous osteocalcin is required for maintaining normal emotional function, cognitive performance, and hippocampal synaptic integrity under physiological estrogen conditions. Previous studies have confirmed that osteocalcin can cross the blood–brain barrier, regulate hippocampal synaptic plasticity, and improve anxiety- and depression-like behaviors through the GPR158–BDNF signaling pathway20,21. Consistent with these reports, our present findings confirm that osteocalcin deficiency alone leads to impaired synaptic structure, reduced synaptic protein expression, and obvious neurobehavioral abnormalities, which further supports the indispensable role of osteocalcin in regulating central nervous system function. Notably, the synaptic damage and neurobehavioral deficits in sham Ocn−/− mice were further exacerbated after OVX, indicating that the compensatory increase of ucOCN induced by estrogen deficiency is critical for resisting neural damage. Together, these baseline differences strengthen our conclusion that ucOCN acts as an endogenous neuroprotective hormone against estrogen deficiency–related neural dysfunction. Behavioral changes are closely associated with synaptic plasticity, which is characterized by dynamic alterations in dendritic spine density and neuronal synaptic architecture. As a critical regulatory hub for neuronal communication, synaptic plasticity directly modulates brain functions including learning, memory, and emotional regulation12–14,39,40. In the present study, we observed that hippocampal dendritic spine density and synaptic number per unit area remained unaltered in WT mice at 4 weeks post-ovariectomy. This finding was consistent with their behavioral phenotype, wherein no significant anxiety-like/depression-like behaviors or spatial learning-memory deficits were detected, suggesting the absence of pathological alterations in synaptic structure during the early post-ovariectomy period. In contrast, ovariectomized Ocn−/− mice exhibited evident hippocampal synaptic damage as early as 4 weeks post-surgery, manifested by decreased dendritic spine density, reduced synaptic number, and significantly downregulated expression of synaptic plasticity-related proteins. This synaptic structural damage occurred concomitantly with exacerbated anxiety-like/depression-like behaviors and impaired spatial learning-memory abilities in Ocn−/− mice. Notably, exogenous ucOCN supplementation could partially reverse both the synaptic structural damage and the corresponding behavioral deficits. These results indicate that ucOCN may delay the onset of anxiety-like/depression-like behaviors and cognitive dysfunction following ovariectomy by preserving hippocampal synaptic plasticity. To further explore the molecular mechanism underlying ucOCN-mediated improvement of synaptic plasticity, we examined the expression levels of key central receptors of ucOCN and core synaptic plasticity-related proteins in the hippocampus. Emerging evidence suggestsGPR158 may serve as a functional receptor for ucOCN in the central nervous system. Previous studies have confirmed that GPR158 can promote the expression of BDNF and its vesicular transport to synaptic sites in the hippocampus21. As a core member of the neurotrophin family, BDNF plays a pivotal role in regulating synaptic plasticity41; it modulates neuronal survival, differentiation, functional status, and neural plasticity via high-affinity binding to tropomyosin receptor kinase B (TrkB)42,43. The present study found that exogenous ucOCN supplementation significantly upregulated the protein expression levels of GPR158 and BDNF in the hippocampus of OVX mice, suggesting that ucOCN may initiate the regulatory program of synaptic plasticity by activating theGPR158-BDNF signaling axis. PSD95 and SYN are well-recognized markers of synaptic plasticity: SYN is involved in presynaptic vesicle exocytosis and neurotransmitter release44,45; PSD95 acts as a postsynaptic scaffold protein that maintains the clustering of postsynaptic receptors, ensuring efficient synaptic signal transmission46,47. Our results showed that ucOCN supplementation significantly increased the protein expression levels of PSD95 and SYN in the hippocampus of OVX mice. Combined with the morphological findings that ucOCN supplementation reversed the decrease in hippocampal dendritic spine density and synaptic number in OVX mice, we speculate that ucOCN may promote the expression of core synaptic plasticity proteins such as PSD95 and SYN by upregulating the GPR158/BDNF signaling pathway, thereby facilitating the repair of hippocampal synaptic structure and the improvement of synaptic function. Notably, recent studies have demonstrated that exercise can significantly increase circulating osteocalcin levels and exert neuroprotective effects in VCD-induced menopausal animal models48. Combined with the compensatory protective mechanism of osteocalcin identified in this study, future research could explore the synergistic effects of exercise intervention and osteocalcin supplementation. Particularly in the early stage of menopause, when endogenous osteocalcin levels are still relatively high, timely exercise intervention may further enhance this compensatory protective effect. This provides a new perspective for the development of lifestyle intervention-based menopausal health management strategies. It should be noted that the current study has certain key limitations regarding the animal model utilized. We used 12-week-old young female mice to establish the OVX model, which only mimics the pathological state of estrogen deficiency in the absence of age-related senescent changes. Specifically, human estrogen decline during the menopausal transition is a physiological process accompanied by multiple age-associated alterations in the central nervous system, such as oxidative stress, chronic low-grade inflammation, and cumulative cellular damage, which are the combined effects of aging and hormonal deficiency1,49. By contrast, 12-week-old mice have not developed age-related physiological deterioration, and thus the OVX model in young mice cannot replicate the complex neurobiological characteristics of natural human menopause. Accordingly, the neurocompensatory role of ucOCN observed in this study only reflects the regulatory mechanism of the bone-brain axis under the single factor of estrogen deficiency, and the results cannot be directly extrapolated to the natural menopause population. To address this issue, future studies will optimize the animal model by using middle-aged female mice to establish the OVX model, which combines the dual factors of aging and estrogen deficiency, to further explore the neuroprotective effect and mechanism of ucOCN in a more clinically relevant context of estrogen decline. Additionally, future research will also incorporate the physiological characteristic of estrogen fluctuation during the perimenopausal stage to improve the clinical translatability of the findings. Beyond the model age, different mouse strains (C57BL/6 J in Experiment 1 and 129/B6 mixed background in Experiments 2 and 3) were used in this study, which was a strategic design to adapt to research purposes and gene knockout mouse background. While some studies have reported subtle differences in baseline behavioral phenotypes between mice of different genetic backgrounds50–52, others have shown no significant differences in anxiety-related and cognitive behaviors between these two strains53. Notably, the conclusions of the three experiments are mutually supportive and complementary rather than directly cross-compared, and thus do not affect our core finding that ucOCN exerts compensatory neuroprotective effects. Nonetheless, we acknowledge that genetic background may exert mild effects on certain behavioral indices. Therefore, in future studies, we will establish estrogen deficiency models on a single, uniform genetic background for more systematic comparative analyses to further validate the stability and generalizability of our findings. Finally, this study focuses on ucOCN’s neuroprotective effect and mechanism, without exploring its temporal changes’ specific regulatory relationship with post-OVX bone metabolism. Future studies will supplement bone metabolism indicators to clarify potential interactions between ucOCN’s bone metabolic compensation and neuroprotective effects, providing a more comprehensive basis for bone-brain axis crosstalk. In conclusion, the present study demonstrates that ucOCN acts as a compensatory factor to protect brain function during the early stage of estrogen deficiency following ovariectomy in female mice. However, this compensatory effect is time-dependent, as it gradually weakens over time after ovariectomy with the decline in ucOCN levels. Furthermore, long-term exogenous ucOCN supplementation can enhance synaptic plasticity and alleviate anxiety-like, depression-like behaviors and cognitive impairments in OVX mice by activating the GPR158/BDNF signaling pathway and upregulating the expression of key synaptic plasticity-related proteins, including PSD95 and SYN. These findings highlight the potential therapeutic value of ucOCN in the treatment of estrogen deficiency-related mood disorders and cognitive decline, providing a novel target and strategy for the prevention and management of neurological diseases associated with perimenopause. Methods Animals Female C57BL/6 J mice were purchased from Charles River Laboratories (Beijing, China) and used in Experiment 1 for characterizing the temporal dynamics of ucOCN and behavioral changes post-OVX, owing to their stable genetic background and consistent phenotypic response to OVX in neurobehavioral and bone metabolism studies. Osteocalcin-deficient (Ocn−/−) mice (RRID: IMSR_JAX:034070) were obtained from The Jackson Laboratory and maintained on a mixed 129/B6 genetic background54. Wild-type (WT) mice with the identical 129/B6 mixed background were employed in Experiments 2 and 3 to eliminate confounding effects of genetic background mismatch when comparing WT and Ocn−/− mice. The mice were housed in a standard Specific Pathogen Free (SPF) laboratory animal facility at the Animal Center of Xi’an Jiaotong University. During the experiment, the mice had free access to autoclaved rodent-specific feed and sterile drinking water. The entire experimental protocol was formally approved by the Biomedical Ethics Committee of the Medical School of Xi’an Jiaotong University (approval number: 2020-625) and authorized by the Animal Care and Use Committee. All animal operations were strictly performed in accordance with the ethical principles and the guidelines outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals (8th ed.; National Academies Press, 2011). Animal groups and experimental design Three sequential experiments were performed to (1) characterize the temporal dynamics of circulating ucOCN and behavioral changes following OVX; (2) assess the necessity of endogenous OCN for early neurobehavioral adaptation to estrogen loss; and (3) evaluate the therapeutic potential of chronic exogenous ucOCN supplementation. All three experiments were performed using independent mouse cohorts with no overlapping animals, and the observation time points (4 weeks and 16 weeks post-OVX) were set based on the key temporal characteristics of ucOCN identified in Experiment 1. Experiment 1: Temporal profiling of ucOCN and behavior post-OVX. Eighty 12-week-old female C57BL/6 J mice were randomly assigned to either OVX or sham surgery (n = 40 per group). Each group was further divided into four time-point cohorts (n = 10/cohort). Blood samples were collected at 2, 6, 10, and 14 weeks post-surgery, followed by a second blood draw and behavioral testing at 4, 8, 12, and 16 weeks, respectively (i.e., each cohort underwent testing at one specific postoperative interval: 4, 8, 12, or 16 weeks). This design was chosen to avoid potential confounding effects, including physiological stress caused by frequent blood sampling and learning or habituation effects induced by repeated behavioral testing, which could bias measurements of anxiety‑related, depression‑related, and cognitive performance. Using independent groups ensures that data at each time point are obtained from naïve animals, thereby improving the accuracy and reliability of time‑dependent changes after estrogen deficiency. Experiment 2: Role of endogenous OCN in early post-OVX neuroadaptation. Twenty 12-week-old female WT and thirty Ocn−/− mice were assigned to the following groups (n = 10 per group): WT Sham group, WT OVX group, Ocn−/− Sham group, Ocn−/− OVX group, Ocn−/− OVX + ucOCN group. Starting on postoperative day 4, mice in the ucOCN-treated group received daily intraperitoneal (i.p.) injections of ucOCN (10 μg/kg, H-6552, Bachem, dissolved in sterile saline); all other groups received equivalent volumes of saline40. Treatment lasted 4 weeks, after which behavioral tests were performed, followed by terminal blood and brain tissue collection under isoflurane anesthesia. Experiment 3: Long-term effects of exogenous ucOCN supplementation. Thirty 12-week-old female WT mice were randomized into three groups (n = 10/group): Sham, OVX, and OVX + ucOCN. Surgical procedures and drug administration were identical to Experiment 2, except that ucOCN or saline was administered daily for 16 weeks. Behavioral assessments were initiated after the final injection, followed by sample collection as above. Ovariectomy After anesthetizing the mice via inhalation with 2% isoflurane delivered in oxygen, all animals achieved a stable surgical plane of anesthesia with complete loss of righting reflex and pain response throughout the operation, maintaining unconsciousness and analgesia during the surgical procedure. Subsequently, a dorsal double-incision method was employed to remove the bilateral ovaries of the mice. For the Sham mice, the ovarian structures of the mice were preserved, and only adipose tissue of equivalent volume adjacent to the ovaries was removed as a control. All mice underwent surgery at 12 weeks of age. Blood collection and serum analysis Serum samples were collected using two methods: from the submandibular venous plexus and from the abdominal aorta. All mice were anesthetized with 2% isoflurane prior to blood collection. For submandibular venous plexus sampling, the procedure was performed only after the mice reached a stable anesthetic state with complete unconsciousness and absence of pain responses. This strategy prevented struggling-related sampling failure, tissue injury and serum contamination, thereby minimizing experimental errors. Gentle compression was applied for hemostasis, and mice were then returned to their home cages for spontaneous recovery. During abdominal aorta blood collection, continuous inhalation of 2% isoflurane was delivered throughout the entire procedure to sustain stable unconsciousness and analgesia until sampling was fully completed. Commercial ELISA kits were used to assess estradiol (E2, Cloud-Clone), Gla-osteocalcin (cOCN, Takara), and Glu-osteocalcin (ucOCN, Takara) levels, strictly following the kit instructions. Absorbance values were measured using a Bio-Rad Model 680 microplate reader. The detection limits are as follows: 12.35 pg/mL for E2, 0.5 ng/mL for cOCN and 0.25 ng/mL for ucOCN. Behavioral tests Behavioral assessments were conducted after at least 30 min of environmental adaptation in the behavioral room for all mice. The open field test (OFT) was performed in a 50 × 50 × 40 cm white box. Mice were placed in one corner facing the center and allowed 6 min of free exploration in dim light (~10 lux). The box was sanitized with 30% ethanol before each trial. The experimenter left the room during trials, which were recorded by a top-view camera and analyzed using Labmaze V3.0 (Beijing Zhongshi Dichuang). Key metrics included distance traveled in the central 25 × 25 cm and time spent there to assess anxiety. The elevated plus maze test (EPMT) apparatus consisted of two open arms (30 × 6 cm) and two enclosed arms (30 × 6 × 15 cm), perpendicularly connected by a central 6 × 6 cm platform, with the entire maze elevated 65 cm above the ground. Mice were initially placed on the central platform facing an open arm and permitted to explore freely for 6 min. Open-arm entry frequency, movement distance and residence time were automatically analyzed to reflect anxiety levels. The maze was disinfected with 30% ethanol after each test. The tail suspension test (TST) was applied to assess depressive-like behaviors. The distal one-third of the mouse tail was gently fixed with adhesive tape and suspended on a bracket, keeping the head approximately 10 cm above the desktop. Each test lasted 6 min and was captured by a front-facing camera. Immobility time during the last 4 min (3–6 min) was calculated using Labmaze V3.0 for subsequent statistical analysis. Spatial learning and memory functions were measured via the Morris water maze (MWM), with reference to a previously reported protocol55. The test was carried out in a 100 cm-diameter circular white pool filled with 40 cm-deep water at 23 ± 1 °C, and water opacity was achieved using white emulsion. A 6 cm-diameter escape platform was submerged 1 cm below the water surface. Visible visual cues were fixed around four quadrants of the pool as spatial references. The paradigm included four consecutive days of acquisition training and one day of probe testing. During training, mice received four daily trials with a maximum duration of 60 s to locate the hidden platform. Upon successful platform arrival or manual guidance, mice remained on the platform for 10 s before returning to home cages, with an inter-trial interval of no less than 5 min. The platform position remained fixed, while the starting point varied randomly. The average escape latency of four trials was recorded daily. On day 5, the platform was removed, and a 60 s probe trial was performed; the number of crossings over the original platform site was quantified to evaluate spatial memory retention. Transmission electron microscopy (TEM) Mice were deeply anesthetized to a surgical plane via inhalation of 2% isoflurane with complete loss of righting reflex and peripheral pain response, followed by euthanasia via transcardial perfusion. Animals were then perfused transcardially with 2.5% glutaraldehyde and 2% paraformaldehyde prepared in 0.1 M sodium cacodylate buffer (pH 7.4). Immediately after perfusion, brain tissues were rapidly dissected and immersed in ice-cold glutaraldehyde solution, and coronal slices containing the hippocampus were prepared for subsequent experiment. The CA1 region was dissected, cut into 1 mm³ blocks, and soaked overnight in 2.5% glutaraldehyde at 4 °C. The next day, samples were then rinsed three times with 0.1 M sodium cacodylate buffer, fixed in 1% osmic acid for 2 h, and dehydrated by sequential immersion in 30%, 50%, 70%, 80%, 90%, 95%, 100%, and 100% ethanol solutions (20 min each). After dehydration, tissues were infiltrated with gradient acetone-Epon 812 embedding media (1:1 for 2 h and 1:2 overnight), then transferred to pure Epon 812 for 5–8 h. Embedding media were poured into molds with oriented samples, pre-polymerized at 37 °C overnight, and finally polymerized at 60 °C for 48 h. Ultrathin sections (70 nm) were obtained using an ultramicrotome, double-stained with 2% uranyl acetate in ethanol (15 min) and lead citrate (15 min), and air-dried at room temperature. Synaptic ultrastructure was observed and imaged using a transmission electron microscope. Golgi staining Mice were deeply anesthetized by inhalation of 2% isoflurane until complete loss of righting reflex and peripheral pain response, and then euthanized via rapid decapitation. Brain tissues were immediately dissected out and rinsed with PBS. Rinse with PBS, then immerse in Golgi staining fixative (G1069-15ML, Servicebio) for 24 h. Replace with fresh dye, submerge the tissue, and store in a cool, dark place at 26 °C for 14 days. After staining, replace with tissue treatment solution (G1069-3, Servicebio), refrigerate at 4 °C for 3 days. Glue the tissue to a microtome tray, immerse in treatment solution, and slice at 60 μm. Mount slices on slides, apply treatment solution, let dry, and rinse with ultrapure water. Immerse slides in developer solution (G1069-2, Servicebio) for 30 min, remove excess water, seal with glycerin gelatin, and store in a light-protected box. Use a slide scanner and pannoramic scanning software (Pannoramic MIDI, 3DHISTECH) to scan the slides in bright-field mode, with 30 scanning layers and 10 steps each. Utilize CaseViewer2.4 software (CaseViewer2.4, 3DHISTECH) to observe and capture images. Randomly select 15 secondary or tertiary dendritic segments in the hippocampal CA1 region of each mouse brain and analyze the dendritic spine density with ImageJ. The photography and analysis were conducted blindly, with personnel unaware of the experimental groupings. Western blotting At the experimental endpoint, mice were deeply anesthetized via inhalation of 2% isoflurane until complete loss of righting reflex and peripheral pain response. After blood collection from the abdominal aorta, mice were euthanized by transcardial perfusion with ice-cold PBS. Subsequently, brain tissues were rapidly dissected, and the hippocampus was isolated for subsequent Western blotting analysis. The Western blot was conducted as previously established48. Hippocampal proteins were extracted using RIPA lysates and their total concentration was determined using a BCA quantification kit. Subsequently, equal quantities of total protein (30 μg) were resolved on polyacrylamide gels with a concentration gradient ranging from 8 to 15%, and were then transferred onto polyvinylidene fluoride membranes (Millipore) utilizing the Trans-Blot Turbo system. Membranes were blocked with 5% defatted milk in 0.1% TBST for 1 hour, then incubated with primary antibodies at 4 °C overnight and secondary antibodies at room temperature for 1 h Blots were visualized with ECL-plus reagent (Millipore) and analyzed with Image J. The primary antibodies used were: PSD95 (1:2000, RRID: AB_2687961), Synaptophysin (1:10,000, RRID: AB_2271365), BDNF (1:1000, RRID: AB_2818984) all from Proteintech, and GPR158 (1:1000, RRID: AB_3662585) from Abmart. Statistical analysis Results are presented as mean ± SD. All mouse experimental data were included in the analysis, with no data excluded during the analysis process. The determination of sample size was based on protocols established in previous experiments49. Data normality was checked by Kolmogorov-Smirnov test. When the data followed a normal distribution, two-tailed t-test (for two groups) or one-way ANOVA (for multiple groups with Tukey’s post hoc test) was applied. Non-parametric tests (Kruskal-Wallis or Mann–Whitney U) were used for non-normal data. Repeated measures ANOVA was used for MWM place navigation test. Behavioral data were compared only between groups at the same time point. Longitudinal comparisons across different ages were not performed to avoid confounding effects of batch and environmental differences. Statistical significance was indicated by p-values less than 0.05, 0.01, 0.001, and 0.0001. All analyses were done using SPSS 25.0 and GraphPad Prism 9.0. For data analysis, preliminary statistical analyses were conducted by researchers who were blinded to the experimental grouping. GraphPad Prism 9.0 and Adobe Illustrator 2020 were used for graph plotting, figure assembly, flow chart drawing, and the drawing of schematics. Supplementary information Supplementary Figures Acknowledgements We would like to acknowledge the financial support from the National Natural Science Foundation of China (Grant No. 11872297). Author contributions Yiting Kang: Writing—review & editing, Writing—original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Yaoyao Yao: Writing—review & editing, Software, Methodology, Formal analysis, Data curation. Zeguo Feng: Writing—review & editing, Methodology, Investigation, Formal analysis, Data curation. Xiaohang Gao: Writing—review & editing, Software, Visualization, Methodology. Yifei Song: Writing—review & editing, Methodology, Investigation. Qianyu Zhang: Writing—review & editing, Methodology. Xiaohui Di: Writing—review & editing, Methodology, Investigation. Lei Huang: Writing—review & editing, Investigation. Wei Han: Writing—review & editing, Supervision, Methodology, Conceptualization. Jianbao Zhang: Writing—review & editing, Writing—original draft, Visualization, Supervision, Project administration, Funding acquisition, Conceptualization. Data availability The datasets generated and analyzed in this study are not publicly available due to ongoing follow‑up research and further mechanistic investigations, but are available from the corresponding author on reasonable request. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Wei Han, Email: hanwei@mail.xjtu.edu.cn. Jianbao Zhang, Email: zhangjb@xjtu.edu.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41514-026-00410-9. References Associated Data Supplementary Materials Supplementary Figures Data Availability Statement The datasets generated and analyzed in this study are not publicly available due to ongoing follow‑up research and further mechanistic investigations, but are available from the corresponding author on reasonable request.