Abstract

   Neuroendocrine regulation is essential for maintaining metabolic
   homeostasis. However, whether neuroendocrine pathway influence bone
   metabolism and skeletal senescence is unelucidated. Here, a central
   neuroendocrine circuit is identified that directly controls
   osteogenesis. Using virus based tracing, this study is identified that
   melanin concentrating hormone (MCH) expressing neurons in the lateral
   hypothalamus (LH) are connected to the bone. Chemogenetic activation of
   MCH neurons in the LH induces osteogenesis, whereas inhibiting these
   neurons reduces osteogenesis. Meanwhile, MCH is released into the
   circulation upon chemogenetic activation of these neurons. Single cell
   sequencing reveals that blocking MCH neurons in the LH diminishes
   osteogenic differentiation of bone marrow stromal cells (BMSCs) and
   induces senescence. Mechanistically, MCH promotes BMSC differentiation
   by activating MCHR1 via PKA signaling, and activating MCHR1 by MCH
   agonists attenuate skeletal senescence in mice. By elucidating a
   brain‐bone connection that autonomously enhances osteogenesis, these
   findings uncover the neuroendocrinological mechanisms governing bone
   mass regulation and protect against skeletal senescence.

   Keywords: BMSC, lateral hypothalamus, melanin concentrating hormone,
   neuroendocrinology, osteogenesis, skeletal senescence
     __________________________________________________________________

   This study reveals a new neuroendocrine circuit connecting the lateral
   hypothalamus to bone, where MCH neurons enhance bone formation and
   protect against skeletal aging. Activating these neurons increases MCH
   levels, promoting bone marrow stromal cell differentiation and
   osteogenesis via MCHR1‐PKA signaling, suggesting potential therapeutic
   targets for osteoporosis and skeletal senescence.

   graphic file with name ADVS-11-2309951-g009.jpg

1. Introduction

   Bone undergoes continuous dynamic metabolism, which is traditionally
   thought to be regulated by secretion of humoral hormones.^[ [42]^1 ^]
   In senile or pathophysiological conditions, disruption of bone
   metabolic homeostasis leads to the development of osteoporosis, which
   is a common marker for skeletal senescence.^[ [43]^2 ^] Although
   various therapeutic interventions have been developed to alleviate
   osteoporosis, including intermittent parathyroid hormone (PTH), Vitamin
   D, bisphosphonates, and monoclonal antibodies, a complete cure for
   osteoporosis is still lacking due to limited understanding of the key
   regulatory factors of osteogenesis. Recent evidence suggests that the
   brain has a crucial role in the delicate regulation of bone metabolism.
   The bone is robustly innervated by sensory nerve which transmits signal
   to the brain.^[ [44]^3 ^] The brain can sense the status of bone
   metabolism through skeletal interception, providing feedback and
   monitoring the internal state of the skeleton to regulate
   osteogenesis.^[ [45]^4 ^] In skeletal interoception, bone mass is
   sensed through nonconscious mechanisms, like the sensing of blood
   glucose fluctuations and gut motility.^[ [46]^5 ^] Several studies have
   revealed the significant role of the sympathetic nervous system in bone
   metabolism.^[ [47]^6 ^] It regulates both bone formation and resorption
   by influencing osteoblastic and osteoclastic cells.^[ [48]^7 ^]
   Hormonal interactions, particularly involving leptin, are also key to
   this process.^[ [49]^4 , [50]^8 ^] The balance and timing of
   sympathetic activity are crucial for maintaining bone health.
   Additionally, research in microgravity conditions offers insights into
   the adaptability of bone metabolism in different environments.^[ [51]^9
   ^] Overall, the sympathetic nervous system's role in bone health is
   complex, involving hormonal, cellular, and environmental interactions.
   However, the mechanisms by which the neuroendocrinological circuitry
   from bone to the brain controlling bone metabolism remain unclear.

   The hypothalamus acts as a central control center, transmitting
   information from the forebrain to regulate whole‐body metabolism and
   maintain homeostasis. By orchestrating the autonomic nervous system and
   regulating hormonal activity, the hypothalamus plays a dominant role in
   the maintenance of metabolic homeostasis, including body temperature,
   thirst, hunger, and sleep‐wake cycles.^[ [52]^10 ^] The lateral
   hypothalamus (LH), a crucial part of the hypothalamus, governs energy
   metabolism including feeding and sleeping by secreting hormones like
   orexin and melanin concentrating hormone (MCH).^[ [53]^11 ^]
   Interestingly, orexin has been found to regulate skeletal homeostasis
   and bone mass through the activation of orexin receptor 2 (OX2R) in the
   LH, promoting bone formation^[ [54]^12 ^] and reversing bone loss in
   hypoxia.^[ [55]^13 ^] However, the function and effect of MCH on bone
   mass regulation have not been elucidated.^[ [56]^12 , [57]^14 ^]
   Additionally, the hypothalamic action of leptin and ghrelin, have been
   found to regulate bone metabolism by affecting sympathetic tone.
   Neuropeptide Y (NPY), secreted by the hypothalamus, has also been
   implicated to regulate bone formation through central actions.^[
   [58]^15 ^] However, the neuroendocrinological circuit that integrates
   skeletal interoception and the molecular mechanisms of these hormones
   on bone tissue aside from sympathetic system remain largely unknown.^[
   [59]^5 ^]

   To address this issue, we initially employed virus‐based tracing^[
   [60]^16 ^] to confirm that MCH neurons in LH, but not orexin neurons,
   are directly connected to bone. We observed that the activity of MCH
   neurons in LH exhibited a strong correlation with bone metabolic
   status. Specifically, the activity of MCH neurons in LH was
   significantly reduced in cases of osteoporosis. Furthermore, plasma
   concentration of MCH was significantly lowered in mice with
   osteoporosis. To further investigate the function of MCH neurons in LH
   in the regulation of bone mass, we utilized chemogenetic viruses and
   transgenic mouse models to manipulate this specific group of neurons.
   Chemogenetic stimulation of MCH neurons in LH resulted in robust
   osteogenesis and accelerated bone tissue regeneration. Single cell
   analysis revealed that BMSC senescence was aggravated in chemogenetic
   inhibition of MCH neurons in LH. Subsequent in vitro experiments
   demonstrated the expression of MCHR1 in BMSCs and revealed that MCH
   induces osteogenesis through the activation of MCHR1 and the PKA
   pathway. Collectively, by elucidating this novel neuroendocrinological
   pathway, our study uncovers the role of MCH neurons in LH in sensing
   bone metabolism and controlling osteogenesis by secreting MCH to
   stimulate BMSC differentiation. Targeting MCH pathway in LH could be an
   effectively therapeutic intervention for skeletal senescence.

2. Results

2.1. Bone is Connected to MCH Neurons in LH

   To elucidate the central neuronal populations implicated in bone mass
   regulation, we first identified neurons with direct connections to
   bone. We utilized a novel approach involving the injection of Herpes
   Simplex Virus (HSV‐129), engineered to encode Enhanced Green
   Fluorescent Protein (EGFP), directly into the bone marrow of adult male
   mice (Figure [61]1A). This technique enabled us to anterogradely and
   trans‐synaptically label central neurons receiving bone‐derived
   signals. Seven days post‐injection, consistent observation revealed a
   discrete cluster of EGFP‐positive neurons within a specific region of
   the lateral hypothalamus, precisely located between −2.00 mm bregma,
   ±1.37 mm lateral to the midline, and at a depth of −4.65 mm
   (Figure [62]1B).

Figure 1.

   Figure 1
   [63]Open in a new tab

   Bone is connected to MCH neurons in LH. A) Schematic of the injection
   method for anterograde tracing of the bone. B) Representative image
   showing HSV co‐stained with MCH. C) Representative image showing HSV
   co‐stained with MCH (upper), Orexin (middle), and Vgat (lower) in the
   LH. D) The Venn diagram showing the overlap between HSV fluorescence
   and MCH (left), Orexin (middle), and Vgat (right) in the LH.

   Further investigations were conducted to determine the specific
   subtypes of neurons directly connected to bone. Co‐staining of these
   EGFP‐positive neurons in the LH with major protein markers including
   Melanin‐Concentrating Hormone (MCH), orexin, and Vesicular Glutamate
   Transporter 2 (Vglut2) was performed. Remarkably, the majority of
   EGFP‐positive neurons demonstrated co‐localization with MCH (based on
   all fluorescence signals, all fluorescence signals include virus green
   fluorescence and neuron marker red fluorescence. The yellow
   fluorescence generated by the superposition of the two in the cell
   (Figure [64]1C). The percentage of yellow fluorescence in all
   fluorescence signals is the standard (Figure [65]1D), and the overlap
   of MCH, Orexin, Vgate with the virus is confirmed to be 24.5%, 2.3%,
   and 3.9%, respectively. In contrast, only a minimal percentage of these
   neurons exhibited co‐staining with orexin or Vglut2. These findings
   collectively indicate a trans‐synaptic connection between bone and a
   specific population of MCH neurons in the lateral hypothalamus,
   indicating a novel neural pathway in bone mass regulation.

2.2. Activity of MCH Neurons in LH Correlates with Osteogenesis

   To explore whether the activity MCH neurons in LH are connected to the
   bone and correlates with osteogenesis, we first used c‐Fos (an
   immediate early gene upon neuronal stimulation) staining to examine the
   rapid activation of MCH neurons in LH upon bone stimulation as
   previously described.^[ [66]^17 ^] Capsaicin or saline was used in bone
   marrow to rapidly stimulate central reaction. If the HSV positive cell
   were also c‐Fos positive, they were regarded to be responsible for bone
   stimulation. 4 h after injection, the brain sliced were harvested for
   c‐Fos staining. We found that the capsaicin stimulation group had a
   significantly higher c‐Fos and HSV double positive cell count compared
   with control group (Figure [67]2A,B). There were nearly no HSV and
   c‐Fos double positive cells before capsaicin stimulation. However, 55%
   cells were HSV c‐Fos positive 4 h after capsaicin stimulation. To
   further examine the activities of MCH neurons in LH at different bone
   mass, we used age induced osteoporotic mice to record the spontaneous
   inhibitory postsynaptic currents (sIPSCs) of MCH neurons in LH
   (Figure [68]2C) by patch clamp. Since the sIPSCs refer to the flow of
   ions across the postsynaptic membrane that leads to a decrease in
   spontaneously fired neuronal activity, it could be used to indicate
   neuronal inhibition status.^[ [69]^18 ^] We first raised pMCH mice and
   injected the EGFP AAV virus (PT‐4250 ScAAV‐hSyn‐DIO‐EGFP‐WPRSaav2/9)
   into the LH region. Our observations revealed a notable increase in the
   amplitude of sIPSCs within the osteoporotic group compared to the
   control group. This difference between the two groups was found to be
   statistically significant (Figure [70]2D).

Figure 2.

   Figure 2
   [71]Open in a new tab

   Activity of MCH neurons in LH correlates with osteogenesis. A)
   Representative image and the Venn diagram showing the merge of HSV and
   c‐fos positive neurons in LH in control state (n = 3). B)
   Representative image and the Venn diagram showing the merge of HSV and
   c‐fos positive neurons in LH 4 h after capsaicin stimulation. C)
   Representative sIPSCs of MCH neurons in LH in control (upper) and
   osteoporotic (lower) groups. D)The sIPSC frequency is increased in
   osteoporosis group. control: n = 12 from 3 mice, p = 0.0015, t (22) =
   3.620 compared with osteoporosis group (n = 12 from 3 mice), unpaired t
   test. Quantitative analyses of the sIPSC amplitude from 2 groups.
   control: n = 12 from 3 mice, p = 0.0409, t (22) = 2.173 compared with
   osteoporosis group (n = 12 from 3 mice), unpaired t test. ^* p < 0.05,
   ^** p < 0.01. E) Schematic of the injection method for chemogenetic
   virus in LH and injection of CNO. F) Concentrations of MCH in plasma
   after injection of CNO in chemogenetic activation, inhibition, and
   control group (n = 7). G) Concentrations of MCH in mouse CSF of
   osteoporosis mice and normal controls (n = 12). H) Concentrations of
   MCH in human plasma of osteoporosis patients and normal controls (n =
   6).

   Since the hypothalamus is the biggest neuroendocrine organ where
   hormones disseminate to circulation, we tested whether MCH was released
   into circulation upon electrical stimulation. To precisely manipulate
   MCH neurons in LH, we introduced pMCH‐Cre mice, which were specifically
   engineered to target MCH neurons in LH. The use of pMCH‐Cre mice
   provides a controlled model to observe the effects of stimulating or
   inhibiting MCH neurons on hormone dissemination into circulation. To
   precisely manipulate MCH neurons in LH in pMCH‐Cre mice, we used
   chemogenetics technique by which neurons can be stimulated or inhibited
   for extended periods of time upon chemical signals.^[ [72]^19 ^] After
   injection of adenovirus (AAV) in specific area of interest, neurons are
   allowed to express designer receptors exclusively activated or
   inhibited by designer drugs (DREADDs), which are activated by
   clozapine‐N‐oxide (CNO).^[ [73]^20 ^] The viruses used for activation,
   inhibition, control were rAAV‐hSyn‐hM3D(Gq)‐EGFP‐WPRE‐hGH polyA,
   rAAV‐hSyn‐hM4D(Gi)‐EGFP‐WPRE‐hGH polyA and rAAV‐hSyn‐EGFP‐WPRE‐hGH
   polyA, respectively (Figure [74]2E). We found a burst of MCH released
   into circulation ≈2 h after chemogenetic stimulation of MCH neurons in
   LH with the highest plasma concentration at 200 pg mL^−1. However,
   chemogenetic inhibition and control group did not show significant
   changes of MCH in circulation (Figure [75]2F). This result indicated
   that MCH was released into circulation upon electrical stimulation of
   the MCH neuron in LH.

   To further delineate the action of MCH in bone metabolism in the
   periphery, we quantified the MCH concentrations in the cerebrospinal
   fluid (CSF) of osteoporotic mice. Our data indicate that osteoporotic
   mice exhibited significantly reduced MCH levels in the CSF when
   compared to control mice, with levels dropping from ≈90 to ≈40 pg mL^−1
   (Figure [76]2G). This reduction in CSF MCH levels implies a potential
   association between diminished MCH activity and the development of
   osteoporosis. In parallel, we assessed the plasma MCH levels in
   osteoporotic patients and observed a marked decrease in MCH
   concentrations relative to healthy controls, with levels decreasing
   from ≈30 to ≈20 pg mL^−1 (Figure [77]2H). This clinical observation
   reinforces the hypothesis that lower MCH activity is linked to
   decreased bone density and the progression of osteoporosis.

   Taken together, these results suggest that the activity of MCH neurons
   in LH correlate with bone density, and MCH is released into circulation
   once these neurons are stimulated.

2.3. Activating MCH Neurons in LH Induces Osteogenesis and Protect Against
Skeletal Senescence

   To test the role of MCH neurons in LH in osteogenesis and bone
   regeneration, we performed 2 type of animal models, including age
   induced osteoporosis and aged fracture repair. In osteoporosis model,
   we chemogenetically activated MCH neurons in LH when age induced
   osteoporosis was seen in pMCH‐Cre mice in 18 months old (Figure
   [78]3A). Micro‐CT analysis was employed to investigate both the
   trabecular bone and cortical bone of the distal femur, as illustrated
   in Figure [79]3B. The ratio of trabecular bone volume to tissue volume
   (BV/TV) was significantly higher in the activation group (20%) compared
   with control group (15%) and inhibition group (10%) (Figure [80]3C).
   The trabecular number (Tb. N), trabecular separation (Tb.Sp), and
   connectivity (Conn. Dn) were partially restored following chemogenetic
   activation compared with control group and inhibition group
   (Figure [81]3D–F). To demonstrate the resorption and formation status
   after MCH neurons in LH activation, we did Tartrate‐resistant acid
   phosphatase (TRAP) staining and in vivo calcein labeling respectively.
   TRAP staining was performed to evaluate bone resorption, indicating
   less TRAP positive cells of trabecular bone in the activation group
   compared with the control group and inhibition group (Figure [82]3G,H).
   In vivo calcein labeling demonstrated that rates of new bone formation
   were higher in the activation group compared with the control group and
   inhibition group (Figure [83]3I,J).

Figure 3.

   Figure 3
   [84]Open in a new tab

   Activating MCH neurons in LH induces osteogenesis. A) schematic of the
   injection method for chemogenetic virus in LH. B) Representative images
   of Micro‐CT 3D reconstruction of the trabecular and cortical bone of
   activation, inhibition, and control group in OSTEOPOROTIC mice. C–F)
   Micro‐CT analysis of BV/TV, Tb.Sp, Tb.N and Ct.Th of the activation,
   inhibition, and control group in OSTEOPOROTIC mice. (n = 8). G,H)
   Representative image and statistical analysis of the TRAP staining of
   activation, inhibition, and control group in OSTEOPOROTIC mice (n = 4).
   I,J) Representative image and statistical analysis of the calcein
   labeling of activation, inhibition, and control group in OSTEOPOROTIC
   mice (n = 4). K) Representative images of Micro‐CT 3D reconstruction of
   the bone callus of activation, inhibition, and control group in femoral
   fracture mice. L,M) Micro‐CT analysis of BV/TV and Tb.Sp of the
   activation, inhibition, and control group in femoral fracture mice. (n
   = 5). ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001;
   1‐way ANOVA with Bonferroni's correction for multiple comparisons.

   In senile fracture repair model, we used the three‐point bending
   technique to induce a controlled fracture followed by intramedullary
   pinning for internal fixation.^[ [85]^21 ^] We aimed to use the
   fracture model to test the new bone formation and fracture repair
   followed by stimulation of MCH neuron in LH. First, we used micro‐CT to
   analyze the bone callus formation at the fracture healing site 4 weeks
   after fracture. Better fracture repair and larger bone callus formation
   could be seen in MCH activation group compared with control and
   inhibition groups (Figure [86]3K). Consistently with age induced
   osteoporosis model, significantly higher BV/TV and lower Tb. Sp were
   seen following chemogenetic activation compared with control group and
   inhibition group (Figure [87]3L,M), indicating better quality of bone
   callus formation. Taken together, these 2 animal models demonstrate
   that activation of MCH neurons in LH robustly induce osteogenesis and
   protect against skeletal senescence.

2.4. Inhibiting MCH Neurons in LH Induces BMSC Senescence

   To further elucidate the primary cell types and signaling pathways
   involved, we employed chemogenetic inhibition of MCH neurons in LH and
   subsequently conducted single‐cell RNA sequencing (scRNA‐seq).
   scRNA‐seq, a widely acclaimed method, enables the analysis of gene
   expression at the individual cell level, offering a detailed view of
   cellular diversity within populations. After quality control
   procedures, which included the removal of damaged or dead cells and
   potential cell doublets, we successfully retained a total of 18271 cell
   transcriptomes from both the MCH neuron inhibition group and the
   control group for further analysis. We utilized Uniform Manifold
   Approximation and Projection (UMAP) for visualizing transcriptional
   heterogeneity, which identified 10 distinct cellular clusters,
   including bone marrow stromal cells (BMSCs), as shown in Figure [88]4A.
   Notably, we observed a reduction in the proportion of BMSCs following
   the chemogenetic inhibition of MCH neurons in the LH (Figure [89]4B).
   This alteration in cellular cluster distribution, with and without the
   chemogenetic inhibition of MCH neurons in the LH, is depicted in
   Figure [90]4C.

Figure 4.

   Figure 4
   [91]Open in a new tab

   Inhibiting MCH neurons in LH retards BMSC activity. A) Visualization of
   bone marrow cells with identification of 10 clusters by single cell
   sequencing. B) Subpopulation proportion of 10 clusters with or without
   chemogenetic inhibition of MCH neurons in LH. C) Differentiation of
   cell clusters with our without chemogenetic inhibition of MCH neurons
   in LH. D) Dot plots showing average expression of known markers in
   indicated cell clusters. The dot size represents percent of cells
   expressing the genes in each cluster. The expression intensity of
   markers is shown. E,F) Representative pathways of upregulated (E) and
   downregulated (F) genes in the MCH neuron inhibition group compared to
   the control group based on biological pathway classification entries in
   the GO database. G,H) Representative pathways of upregulated (G) and
   downregulated (H) genes in the MCH neuron inhibition group compared to
   the control group based on biological pathway classification entries in
   the KEGG database. (I) Representative β‐Gal staining in the MCH neuron
   inhibition group compared to the control group. I,J) SASP from BMSC
   (left) and bone marrow (right) in the MCH neuron inhibition group
   compared to the control group.

   To delve deeper into the biological processes associated with BMSCs, we
   conducted a gene ontology (GO) analysis focusing on genes
   differentially expressed in BMSCs. The GO analysis indicated an
   upregulation in processes such as cellular senescence, negative
   regulation of osteoblast differentiation, and osteoclast
   differentiation (Figure [92]4E), while processes related to lymphocyte
   differentiation were downregulated (Figure [93]4F).

   Furthermore, we performed a Kyoto Encyclopedia of Genes and Genomes
   (KEGG) pathway analysis to explore the functional pathways in BMSCs
   under the influence of chemogenetic inhibition of MCH neurons in the
   LH, compared to the control group. This analysis revealed an
   upregulation in several pathways associated with bone resorption,
   including osteoclast differentiation and autophagy, in the MCH neuron
   inhibition group (Figure [94]4G). Conversely, pathways such as the
   estrogen signaling pathway and mitophagy were found to be downregulated
   (Figure [95]4H). Upon further analysis of isolated Bone Marrow Stromal
   Cells (BMSCs), we noticed a substantial increase in β‐galactosidase
   (β‐gal) activity in the group where MCH neurons were inhibited
   (Figure [96]4I). β‐gal ’s activity is commonly used as a biomarker for
   cellular senescence, indicating aging cells. Additionally, there was a
   significant upregulation in the expression of the senescence‐associated
   secretory phenotype (SASP) in the inhibition group compared to the
   control group (Figure [97]4J). SASP is characterized by the secretion
   of inflammatory cytokines, growth factors, and proteases, and is known
   to contribute to the aging process and age‐related diseases.

   Collectively, these scRNA‐seq findings suggest that the chemogenetic
   inhibition of MCH neurons in the LH specifically attenuates the
   activity of BMSCs, highlighting the intricate interplay between
   neuronal signaling and bone marrow cellular dynamics.

2.5. MCH Induces Osteogenic Differentiation in BMSC via PKA Signaling

   Since we demonstrated MCH neurons in LH release MCH into circulation
   upon chemogenetic stimulation, we hypothesize that MCH locally induces
   osteogenesis through activating MCH receptors once it is released into
   bone area. First, we tested the expression of MCH receptors on
   different cell types including BMSC and osteoclasts. By immunostaining,
   we found MCH receptor 1 (MCHR1) was specifically expressed on BMSCs
   (Figure [98]5A). To test whether MCH administration influences BMSCs
   proliferation, we performed osteogenic, adipogenic, and chondrogenic
   differentiation with or without MCH. Using CCK8 as an indicator, we
   found 0.5µm MCH administration have slight influence on osteogenic,
   adipogenic, and chondrogenic differentiation of BMSCs (Figure [99]5B).

Figure 5.

   Figure 5
   [100]Open in a new tab

   MCH induces osteogenic differentiation in BMSC via PKA signaling. A)
   Representative image showing MCHR1 is expressed on BMSCs. B) Plots
   showing CCK8 activities with or without MCH administration in
   osteogenic, adipogenic, and chondrogenic differentiation. C)
   Representative image showing Alizarin red (upper) and AKP (lower)
   positive cells in osteogenic induction of BMSCs. D) Representative
   image showing oil red O (upper) and Alcian blue (lower) positive cells
   in adipogenic and chondrogenic induction of BMSCs. E,F) The statistical
   analysis of Alizarin red and AKP positive cells in osteogenic induction
   of BMSCs (n = 4). G) The statistical analysis of oil red O positive
   cells in adipogenic induction of BMSCs (n = 4). H, The statistical
   analysis of Alcian blue positive cells in chondrogenic induction of
   BMSCs (n = 4). I and J, Representative image and statistical analysis
   of PKA signaling with or without MCH administration in BMSCs. K,L)
   Representative image and statistical analysis of PKC signaling with or
   without MCH administration in BMSCs. ^* p < 0.05, ^** p < 0.01, ^***
   p < 0.001, ^**** p < 0.0001; student t‐test.

   Subsequently, we tested the effect of MCH on osteogenic differentiation
   of BMSCs by Alizarin red and AKP staining (Figure [101]5C). The
   osteogenic potential of BMSCs treated with MCH was significantly higher
   than control group, as evidenced by elevated calcium nodule deposition
   in the Alizarin Red (Figure [102]5E) and AKP staining (Figure [103]5F).
   Moreover, to test the adipogenic and chondrogenic potential of BMSCs
   with or without MCH administration, we performed oil red O and Alcian
   blue staning respectively. We found a decrease in the adipogenic
   potential (Figure [104]5G) and an increase in the chondrongenic
   potential (Figure [105]5H) of BMSCs treated with MCH compared to the
   control group. To elucidate the signal transduction pathway of MCH on
   BMSC, we tested PKA pathway following previous studies.^[ [106]^22 ^]
   The western blot analysis showed that phosphorylated PKA was
   significantly upregulated following the MCH administration
   (Figure [107]5I,J). We tested the master osteogenic transcriptional
   factor runx2 signaling was also upregulated with MCH administration
   (Figure [108]5K,L). Since the majority of osteogenic pathways were
   related to PKA signaling, we focused on the PKA signaling in the
   following study. Taken together, these results collectively suggested
   that MCH induced osteogenic differentiation of BMSCs through MCHR1 via
   PKA signaling.

2.6. MCH Activates Osteogenic and Inhibit Senescence Pathways in BMSCs

   To explore the molecular processes driving the osteogenic
   differentiation of BMSCs upon MCH administration, we employed advanced
   high‐throughput mRNA sequencing. For this purpose, BMSCs, both treated
   and untreated with MCH, were subjected to mRNA sequencing following a
   9‐day period of osteogenic induction. The subsequent quantitative
   analysis of gene expression levels revealed notable differences between
   the two groups. A total of 212 genes were found to be upregulated,
   while 31 genes exhibited downregulation in the MCH‐treated BMSCs.
   Noteworthy among the upregulated genes were alkaline phosphatase and
   integrin family proteins, both of which showed significant elevation.
   These genes, known for their roles in bone metabolism and cell adhesion
   respectively, underscore the osteogenic potential of MCH. This
   differential gene expression was eloquently illustrated through volcano
   and waterfall plots (Figure [109]6A,B).

Figure 6.

   Figure 6
   [110]Open in a new tab

   MCH activates osteogenic and inhibits senescence pathways in BMSCs. A)
   Volcano plot depicting the distribution of differentially expressed
   genes in the MCH group compared to the control group in BMSCs. B)
   Correlation of the differentially expressed genes in the MCH group
   compared to the control group in BMSCs. C,D) Representative GO
   classification and KEGG pathways of differentially expressed genes in
   the MCH group compared to the control group in BMSCs. E,F) GO and KEGG
   enrichment analysis of the MCH group compared to the control group in
   BMSCs. G‐I) GSEA analysis of Wnt, PKA and SASP pathways in the MCH
   group compared to the control group in BMSCs.

   Furthermore, Gene Ontology (GO) enrichment analysis delineated a series
   of bone formation pathways significantly influenced by MCH treatment.
   These pathways included skeletal system development, ossification,
   integrin complex formation, bone mineralization, and Wnt‐protein
   binding. In contrast, pathways like negative regulation of RNA
   Polymerase II transcriptional activity and apoptosis were significantly
   subdued in the MCH group (Figure [111]6C). Additional GO analysis
   highlighted the enrichment of processes like collagen biosynthesis,
   tissue regeneration, bone remodeling, and ossification
   (Figure [112]6E). Complementing the GO analysis, Kyoto Encyclopedia of
   Genes and Genomes (KEGG) pathway enrichment analysis, conducted using
   the latest database, categorized the differentially expressed genes
   into representative biological pathways. Predominantly, pathways
   related to bone formation such as TGF‐beta signaling, Wnt signaling,
   Hippo signaling, and parathyroid hormone signaling were markedly
   upregulated in the MCH‐treated group. Conversely, pathways associated
   with cellular senescence and osteoclast differentiation were found to
   be downregulated (Figure [113]6D). The KEGG analysis further
   corroborated the enrichment of the ECM receptor interaction and Hippo
   signaling pathways (Figure [114]6F).

   To refine our understanding and eliminate potential ambiguities from
   the separate analyses of up‐ and down‐regulated genes, we selected
   three pathways closely associated with bone formation for a more
   focused Gene Set Enrichment Analysis (GSEA). This analysis reinforced
   our previous findings, highlighting a significant upregulation of the
   Wnt signaling pathway and PKA signaling, while SASP signaling was
   notably downregulated in the MCH‐treated group (Figure [115]6G–I).

   In summary, our comprehensive study elucidates that MCH administration
   triggers a cascade of biological pathways, fostering osteogenic
   differentiation in BMSCs. This discovery opens new vistas in
   understanding the molecular mechanisms underlying bone formation and
   regeneration.

2.7. Activating MCHR1 Alleviates Skeletal Senescence in Mice

   To investigate the effect and therapeutic potential of MCHR1 signaling
   in skeletal senescence, we intravenously delivered both MCHR1 agonist
   (MCH TFA) and antagonist (ALB‐127158a) or vehicle (saline) for age
   induced osteoporosis in mice (Figure [116]7A). With the development of
   osteoporosis, micro‐CT analysis of distal femurs revealed significant
   bone loss in consistent with previous results. Notably, mice treated
   with MCH agonist, exhibited markedly increased bone mass compared to
   vehicle treated osteoporotic mice (Figure [117]7B). This increase was
   evident in terms of the BV/TV (Figure [118]7C), Tb.N (Figure [119]7D),
   and decreased Tb.Sp (Figure [120]7E). However, there was no difference
   in Ct.Th between the MCH agonist and vehicle group (Figure [121]7F).
   Consistently, the mice treated with MCH antagonist showed significant
   reduction in Tb.N, Ct.Th (Figure [122]7D,F) and increase in Tb.Sp
   (Figure [123]7E). These findings suggest that the MCH agonism
   significantly attenuates bone loss in mouse model of osteoporosis. To
   determine whether MCH agonism activates bone remodeling, we assessed
   markers of bone formation and resorption. Compared to the control mice,
   MCH agonist treated mice exhibited a significant reduction of TRAP
   positive cells, indicating that MCH agonism decreased osteoclast
   activity (Figure [124]7G,H). Consistently, an increased number of OCN
   positive cells, as evidenced by OCN staining on trabecular bone in MCH
   agonist treated mice compared to the vehicle group (Figure [125]7I,J).
   Moreover, in vivo calcein labeling showed increased MAR in MCH agonist
   group compared with the control group (Figure [126]7K,L). Meanwhile, we
   noted the decreased OCN level, decreased MAR, and increased TRAP
   positive cells in the MCH antagonism treated mice, indicating lower
   bone formation and elevated bone resorption compared to the control
   group. To further substantiate the therapeutic efficacy of MCHR1
   signaling in mitigating skeletal senescence, we conducted an analysis
   of β‐gal staining. Consistent with our prior observations, the
   administration of the MCH agonist resulted in a pronounced reduction in
   senescence‐associated β‐gal markers when compared to the vehicle group,
   as depicted in Figure [127]8M. This was further quantified through a
   significant decrease in the prevalence of β‐gal positive cells within
   both the bone marrow and trabecular regions in mice receiving MCH
   agonist therapy (Figure [128]8M,N). Conversely, the administration of
   the MCH antagonist led to a notable increase in β‐gal positive cells,
   indicative of accelerated senescence within the bone tissues.
   Collectively, these results provide compelling evidence that MCH
   agonism attenuates skeletal senescence in mice by activating
   osteogenesis and suppressing osteoclastogenesis.

Figure 7.

   Figure 7
   [129]Open in a new tab

   Activating MCHR1 induces osteogenesis and alleviates osteoporosis in
   mice. A) Schematic of the tail vein injection of MCHR1 agonist and
   antagonist in age‐induced osteoporotic mice. B) Representative images
   of Micro‐CT 3D reconstruction of the trabecular and cortical bone of
   MCHR1 activation, inhibition, and control group in osteoporotic mice.
   C‐F, Micro‐CT analysis of BV/TV, TbN, Tb.N, Tb.Sp and Ct.Th of the
   MCHR1 activation, inhibition, and control group. (n = 10). G and H,
   Representative image and statistical analysis of the TRAP‐positive
   cells of MCHR1 activation, inhibition, and control group in
   osteoporotic mice (n = 3). I and J, Representative image and
   statistical analysis of the OCN positive cells of MCHR1 activation,
   inhibition, and control group in osteoporotic mice (n = 3). K,L)
   Representative image and statistical analysis of the calcein labeling
   of activation, inhibition, and control group in osteoporotic mice (n =
   3). M and N, Representative image and statistical analysis of the β‐gal
   staining of activation, inhibition and control group in osteoporotic
   mice (n = 3).^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p <
   0.0001; 1‐way ANOVA with Bonferroni's correction for multiple
   comparisons.

Figure 8.

   Figure 8
   [130]Open in a new tab

   Working model of neuroendocrine circuit in LH for bone mass regulation
   through MCH signaling.

3. Discussion

   Bone is a specialized organ with various functions. Through getting
   mechanical strength by biological mineralization, bone serves
   mechanical functions such as providing rigidity, shape, protection of
   the body structures and facilitating movement. Despite supporting
   functions, bone undergoes constant remodeling to maintain metabolic
   homeostasis and keep regeneration.^[ [131]^23 ^] The overall mechanical
   properties of the bone, including its rate of turnover, collagen
   matrix, size, structure, geometry, and density, are critical in
   determining its strength and resilience. Any defects in these
   parameters can result in various skeletal disorders, such as
   osteoporosis, osteopetrosis, Paget's disease, and osteogenesis
   imperfecta.^[ [132]^24 ^]

   It has long been hypothesized that both central and peripheral nervous
   system are responsible for controlling bone metabolism and disorders.
   Central neuropeptides are known to play crucial roles in the regulation
   of bone metabolism.^[ [133]^15 , [134]^25 ^] For instance, leptin
   inhibits osteogenesis through activating the sympathetic nervous
   system,^[ [135]^4 ^] and orexin maintains skeletal metabolic
   homeostasis with a dual role in bone mass regulation.^[ [136]^26 ^] For
   centuries, the existence of local nerve fibers supplying and regulating
   the activity of bone metabolism has been recognized.^[ [137]^3 ^]
   Meanwhile, there is increasing evidence showing that different types of
   neurons play distinct roles in the regulation of bone metabolism and
   homeostasis. Sensory nerves play a critical role in detecting
   mechanical stress and strain on bones, which is necessary for bone
   remodeling and adaptation.^[ [138]^6b ^] Sympathetic nerves regulate
   bone resorption and formation by releasing neurotransmitters such as
   norepinephrine and neuropeptide Y.^[ [139]^6 , [140]^15 ^] Although
   these molecular and genetic studies have provided valuable insights
   into the neuropeptides that regulate bone mass, studies on the central
   neural regulation of bone metabolism and skeletal senescence at the
   neuroendocrinological level have been very limited.

   Skeletal senescence is a complex process leading to reduced bone mass
   and increased fragility, often resulting in conditions like
   osteoporosis.^[ [141]^27 ^] This process is driven by changes at the
   cellular level, particularly in osteoblasts and osteoclasts responsible
   for bone formation and resorption. As these cells age, there's a
   decline in bone formation and an increase in bone resorption,
   contributing to overall bone mass loss.^[ [142]^28 ^] This cellular
   aging is influenced by systemic factors such as hormonal changes and
   local factors like oxidative stress, which impair bone regeneration and
   increase fragility.^[ [143]^29 ^] Additionally, skeletal aging involves
   alterations in bone architecture and quality, significantly impacting
   the risk of fractures in the elderly. Key molecular pathways, such as
   p53/p21 and p16INK4a/Rb, are activated in response to stressors like
   DNA damage, leading to cellular senescence.^[ [144]^30 ^] Genetic
   factors also play a crucial role in bone aging, with specific genes and
   transcription factors regulating mesenchymal stem cell functions and
   influencing bone density and quality during aging.^[ [145]^31 ^]
   Moreover, the interaction between the immune system and bone tissue
   becomes increasingly important in the context of aging. Senescent
   immune cells release factors that contribute to the acceleration of
   bone aging, underscoring the complex relationship between these
   systems.^[ [146]^32 ^] However, whether targeting neuroendocrinological
   way is effective for skeletal senescence is still largely unknown.

   As an anabolic hormone, MCH promotes increased food intake and energy
   storage,^[ [147]^33 ^] which supports its role in bone formation by
   enhancing osteoblast activity and reducing osteoclast activity. This
   contrasts with leptin, which inhibits bone osteogenesis and promotes
   energy expenditure. Unlike MCH, orexin positively influences bone
   formation and is crucial for maintaining bone health.^[ [148]^26 ^]
   Moreover, a Chinese traditional drug called Zuogui Wan improves
   trabecular bone microarchitecture in ovariectomy‐induced osteoporosis
   rats by regulating orexin‐A and orexin receptor.^[ [149]^34 ^] MCH is
   known to be released from the hypothalamus into the circulation,
   exerting its effects both centrally and peripherally. Studies have
   demonstrated that MCH may cross the blood‐brain barrier (BBB) and enter
   the CSF before making its way into the bloodstream.^[ [150]^35 ^] This
   dual release mechanism allows MCH to impact a variety of physiological
   processes, including energy balance, feeding behavior, and bone
   metabolism. One primary mechanism through which MCH enters the
   circulation involves tanycytes, specialized glial cells in the
   hypothalamus. MCH may utilize receptor‐mediated transcytosis to cross
   the endothelial cells of the BBB. This process involves MCH binding to
   specific receptors, being internalized, and then released on the other
   side, allowing it to enter the CSF and subsequently the bloodstream.^[
   [151]^36 ^] Furthermore, MCH can diffuse through circumventricular
   organs (CVOs), regions of the brain with naturally more permeable BBBs,
   facilitating direct exchange between the blood and the brain.

   MCH exerts dual actions through its neuroendocrine release into the
   circulation and its interactions with the sympathetic nervous system
   (SNS). Centrally, MCH influences feeding behavior, energy expenditure,
   and adiposity by acting on MCH receptors in the hypothalamus and other
   brain regions. Peripherally, MCH affects various metabolic processes,
   including bone formation, by interacting with receptors in peripheral
   tissues. The neuroendocrine action of MCH involves its release into the
   bloodstream, where it can interact with distant organs. Simultaneously,
   MCH can modulate the SNS, affecting bone metabolism by altering
   sympathetic tone. This dual mechanism allows MCH to coordinate complex
   physiological responses. Recent research by Saito et al.^[ [152]^37 ^]
   supports these dual roles, demonstrating how MCH can influence both
   central and peripheral systems through its actions on the circulation
   and the SNS.^[ [153]^12 ^]

   Here we identified a neuroendocrinological circuit underlying skeletal
   interoception to regulate bone metabolism. The circuit is based on
   several lines of evidence. First, we used virus‐based tracing strategy
   to identify MCH neurons but not orexin neurons in LH are
   trans‐synaptically connected to the bone. Second, we found the
   electrical activity of MCH neurons in LH were highly correlated with
   bone metabolism status. To investigate whether MCH was secreted into
   circulation upon stimulation, we used chemogenetic tools to confirm
   that MCH was secreted into plasma upon electrical stimulation by MCH
   neurons in LH. Third, to further investigate the effect of MCH neurons
   in LH in bone mass regulation, we used chemogenetic viruses and
   transgenic mice model to manipulate this specific group of neurons.
   Chemogenetic stimulation of MCH neurons in LH results in robust
   osteogenesis. Fourth, to elucidate the specific cell type responsible
   for MCH action of osteogenesis, we performed scRNA‐seq. The results
   suggested that BMSCs were specifically influenced by chemogenetic
   inhibition of MCH neurons in LH. Fifthly, we performed in vitro
   experiments demonstrate that MCHR1 is expressed BMSCs and MCH induces
   osteogenesis through MCHR1 activation of PKA pathway. Lastly, we test
   that MCHR1 agonist is capable of attenuating osteoporosis in mice.

   Based on our results, we propose a working model where MCH neurons in
   LH is the center of skeletal interoception and neuroendocrine
   regulation of bone mass (Figure [154]8). First, the MCH neurons
   trans‐synaptically receive the signal from bone through sensory nerve
   for bone density surveillance. At the same time, when stimulated by
   sensory signals, the MCH neurons in LH secret MCH into circulation to
   accelerate BMSC osteogenic differentiation through MCHR1 via PKA
   signaling. We found a central peptide MCH in LH has been functionally
   demonstrated to control skeletal interoception and locally induce
   osteogenesis upon ascending nervous signals. In the near future, if
   this mechanism is confirmed in human, targeting MCH neurons in LH shows
   therapeutic potential for various skeletal diseases.

4. Experimental Section

Animals

   The study utilized adult C57BL/6 mice, both male and female, obtained
   from the animal facility center at Central South University. The p‐MCH
   Cre mice were purchased from the Jackson Lab. The age‐induced
   osteoporotic mice were 18 months old, while the control mice were 8
   weeks old. The mice were kept in a controlled environment with a
   temperature range of 22–25 °C, humidity between 40% and 60%, and
   subjected to a 12 h light/dark cycle. They had unrestricted access to
   food and water. All animal care and experimental procedures were
   approved by the Research Committee of Xiangya Hospital, Central South
   University. The study included an equal number of male and female mice.

Tail Vein Injection

   For the tail vein injections, MCH agonists (MCH, TFA) were administered
   at a dose of 18 mg kg^−1 twice daily (bid) and were obtained from
   MedChemExpress (HY‐P1205A). MCH antagonists (ALB‐127158(a)) were
   administered at a dose of 10 mg kg^−1 twice daily (bid) and were
   sourced from MedChemExpress (HY‐111398). Saline was used as a control
   and was administered at a volume of 0.1 ml per day. The total duration
   of these treatments was 30 days.

Staining

   For immunohistochemistry staining of bone samples, the samples were
   dehydrated using serial ethanol solutions (70%, 1 h × 2; 95%, 2 h × 2;
   100%, 1 h × 2) and cleared with xylene (1 h × 2) before embedding them
   in paraffin (Surgipath Paraplast, Leica). Slices of 5–7 µm thickness
   were obtained using a microtome (RM2235, Leica) and transferred to
   adhesion glass slides. The slides were deparaffinized by washing with
   xylene (3 min × 2) and serial ethanol solutions (100%, 3 min × 2; 95%,
   3 min × 2; 70%, 3 min × 2). The slices were incubated with primary
   antibodies, including anti‐MCH (1:100, Abcam), anti‐orexin (1:100,
   Abcam), anti‐VGAT (1:100, Abcam), anti‐EGFP (1:200, Sigma), anti‐c‐Fos
   (1:100, CST) at 4 °C overnight. After washing, the sections were
   labeled with fluorescent secondary antibodies. For secondary antibody
   staining, Alexa Fluor 488 or 594‐conjugated antibodies (1:100; Jackson
   ImmunoResearch, PA) were used. Nuclei were counterstained with DAPI,
   and the samples were then mounted for image acquisition.

Western Blotting

   Cells or tissues were homogenized in ice‐cold RIPA buffer (50 mM
   Tris‐HCl, pH 7.4, 150 mM NaCl, 1% NP‐40, 0.5% sodium deoxycholate, 0.1%
   SDS) with protease and phosphatase inhibitors. The homogenate was
   centrifuged at 12 000 rpm for 15 min at 4 °C, and the supernatant
   protein concentration was measured using a BCA assay. Equal amounts of
   protein (20–30 µg) were mixed with 4x Laemmli buffer, heated at 95 °C
   for 5 min, and loaded onto a 10% SDS‐PAGE gel. Electrophoresis was
   conducted at 100 V until the dye front reached the gel's bottom.
   Proteins were transferred to a PVDF or nitrocellulose membrane at 100 V
   for 60–90 min. Membranes were blocked in 5% non‐fat milk or BSA in TBST
   (Tris‐buffered saline with 0.1% Tween‐20) for 1 h at room temperature,
   then incubated overnight at 4 °C with primary antibodies in blocking
   buffer (anti‐PKA, PA5‐28160; anti‐pPKA, ZENBIO347334; anti‐Runx2,
   Abmart T55395F;). After three TBST washes, membranes were incubated
   with HRP‐conjugated secondary antibodies for 1 h at room temperature.
   Bands were visualized using an enhanced chemiluminescence (ECL) system
   and captured on X‐ray film or a digital imaging system.

Virus Based Tracing

   To trace the anterograde pathways from the bone marrow to the lateral
   hypothalamus, HSV‐129 EGFP (BrainVTA, [155]H01001) was used for
   multi‐site injections in the bone marrow of adult mice. The virus was
   administered at multiple injection sites, specifically targeting the
   femoral and tibial bone marrow cavities, with each site receiving 1 µL
   of virus. The total volume of virus injected per mouse was 10 µL. After
   the injections, the mice were allowed to survive for 7 days to enable
   adequate viral transport and expression of EGFP. Following this period,
   the mice were perfused, and tissues were collected from the brain and
   spinal cord for analysis.

Chemogenetics

   Adult mice aged 6–8 weeks were anesthetized intraperitoneally using
   sodium pentobarbital (100 mg kg^−1) and gently secured to a stereotaxic
   frame (RWD Life Science, Shenzhen, China). A total volume of 200 nL of
   the virus suspension was loaded into a 1 ml‐Hamilton syringe fitted
   with a needle made of glass electrode. The AAVs used were
   rAAV‐hSyn‐DIO‐hM3D(Gq)‐EGFP‐WPRE‐hGH polyA, rAAV‐hSyn‐DIO‐EGFP‐WPRE‐hGH
   polyA rAAV‐hSyn‐DIO‐hM4D(Gi)‐EGFP‐WPRE‐hGH polyA, obtained from
   BrainVTA, Wuhan, China. The viral titer for all AAVs was maintained at
   2.32‐5.72E+12 PFU/mL. Mice(C57BL/6‐Tg(Pmch‐cre)1Rck/J) carrying genes
   specifically expressing CRE in pMCH cells come from JAX lab, The
   injection was performed at a rate of 20 nL min^−1 into specific target
   locations (−2.00 mm bregma, ±1.37 mm lateral to the midline, and at a
   depth of −4.65 mm). The surgical procedure adhered to the protocol
   outlined by Lowery and Majewska (reference 54). Following the surgery,
   a recovery period of at least 1 week was provided for the mice before
   commencing any experimental procedures. For chemogenetic activation and
   inhibition, clozapine‐N‐oxide (CNO) at a dosage of 1 mg kg^−1 (MCE,
   34233‐69‐7) was administered intraperitoneally three weeks after viral
   expression. The injections were repeated every 48 h for a duration of
   four weeks before serum and bone samples were collected for subsequent
   analysis. The AAV9 viruses were prepared in‐house, ensuring their
   quality and consistency.

Calcein Labeling

   In this investigation, mice were subjected to intraperitoneal
   injections of calcein at doses of either 20 mg kg^−1, which were
   dissolved in a 2% sodium bicarbonate solution. Calcein labeling was
   conducted over a period of 5 days preceding the sacrifice of the mice.
   The femur bones were fixed in a solution composed of 70% ethanol and
   subsequently embedded in methylmethacrylate for sectioning. Analysis of
   the bone surface and quantification of the bone formation rate (BFR)
   per unit of bone surface (BS). Furthermore, the mineral apposition rate
   (MAR) was determined by measuring the distance between the midpoints of
   two labels and dividing it by the time interval between the midpoints.

Micro‐CT

   The femurs of mice were harvested and fixed in 4% paraformaldehyde
   before undergoing micro‐CT scanning with a SkyScan 1076 system.
   Scanning parameters included an isotropic voxel size of 11.53 mm, a
   voltage of 48 kV, a current of 179 mA, and an exposure time of 1800 ms.
   3D reconstructions of the scans were performed using SkyScan NRecon
   software with a voxel size of 8.66 mm. The datasets were reoriented
   using SkyScan DataViewer, and morphological parameters were quantified
   using CTAn software.

   For trabecular bone analysis, a region of interest (ROI) was manually
   delineated from 2D image slices, specifically 0.1–0.9 mm distal to the
   proximal tibia growth plate, excluding the cortical area. Binarization
   was conducted using a global thresholding approach (gray level range
   70–255) to differentiate mineralized tissue. A Gaussian filter with a
   radius of 1 was applied for 3D reconstruction. Quantitative analysis
   covered all bone areas within the ROI. Morphometric parameters included
   total volume (TV), bone volume (BV), bone fraction (BV/TV), trabecular
   thickness (Tb. Th), trabecular number (Tb. N), trabecular separation
   (Tb. SP), and bone mineral density (BMD). BMD calibration was performed
   using hydroxyapatite phantoms with known densities.

   For cortical bone analysis, the ROI was selected from 2D images,
   6.5–7.2 mm proximal to the distal femur growth plate. A threshold range
   of 85–255 was used for mineralized tissue identification. A Gaussian
   filter radius of 1 was applied for 3D reconstruction. Quantitative
   analysis included bone volume (BV) and bone mineral density (BMD)
   within the ROI.

Electrophysiology

   In this experiment, mice underwent deep anesthesia followed by
   transcardial perfusion with an ice‐cold cutting solution enriched with
   95% oxygen and 5% carbon dioxide, using N‐methyl‐D‐glucamine (NMDG) as
   the primary component. Post‐perfusion, the brain was quickly removed
   and submerged in the same NMDG‐based solution at low temperature.
   Coronal brain slices of approximately 300 µm were prepared from the
   lateral hypothalamus (LH) using a vibratome. These slices were
   initially recovered in the cutting solution at a slightly higher
   temperature and then transferred to oxygenated artificial cerebrospinal
   fluid (ACSF) for further preparation before electrophysiological
   recordings.

   For recordings, the slices were continuously perfused with oxygenated
   ACSF and visualized using an upright microscope with infrared optics.
   MCH neurons in the LH were identified by their fluorescence and
   morphology through EGFP expression, marking them as MCH‐positive. To
   study spontaneous inhibitory postsynaptic currents (sIPSCs), the
   intracellular recording solution contained cesium chloride (CsCl),
   isolating GABAergic currents by blocking AMPA and NMDA receptors. The
   holding potential was set at 0 mV. Neuronal stimulation was not
   applied, allowing for the observation of spontaneous events.

   Recorded signals were amplified, filtered, and sampled using a
   MultiClamp 700B amplifier and appropriate software. Data validity was
   contingent on the series resistance remaining within a specific range.
   The number of neurons recorded varied and is detailed in the
   corresponding figure descriptions.

BMSC Differentiation

   For the induction of osteogenic differentiation, BMSCs were seeded at a
   density of 4 × 104 cells per well in 24‐well culture plates. After 24 h
   of incubation, the culture medium was replaced with osteogenic
   induction medium (OIM). As a negative control, cells were incubated in
   OIM without scaffold extracts but containing the same
   osteogenesis‐inducing factors. Throughout the differentiation process,
   the medium was refreshed every two days.

   To induce differentiation into adipocyte‐like cells, BMSCs were
   cultured until reaching confluence. Two days after confluence, the
   medium was replaced with DMEM supplemented with 10% fetal bovine serum
   (FBS; Gibco), IBMX (0.5 mM; Sigma‐Aldrich), dexamethasone (0.25 µM;
   Sigma‐Aldrich), and bovine insulin (10 g/mL). After 72 h, the medium
   was further replaced with DMEM supplemented with 10% FBS and 10 µg/mL
   bovine insulin (Sigma‐Aldrich, I6634). The medium was refreshed every 2
   days during the differentiation process. To assess the osteogenic
   differentiation potential of BMSCs, Alizarin Red S staining was
   performed. Additionally, oil red staining was used to evaluate the
   adipogenic differentiation potential.

Single Cell Sequencing

   The tissue dissociation and sample preparation involved several steps.
   Ascites samples were filtered through 40‐micron strainers and
   centrifuged to separate cells from the supernatant. The cell pellet was
   resuspended in PBS and treated with erythrocyte lysis buffer.
   Concurrently, fresh surgical tissues were preserved, washed, minced,
   and enzymatically digested. Post‐digestion, samples were centrifuged,
   and pellets resuspended in PBS, with cell viability assessed using
   Trypan Blue staining. For scRNA‐seq, single‐cell suspensions were
   processed using a microwell chip and barcoding beads for mRNA capture.
   Reverse transcription generated cDNA, followed by PCR amplification,
   cDNA fragmentation, and adapter ligation. scRNA‐seq libraries, prepared
   using a specific kit, were pooled and sequenced on an Illumina NovaSeq
   6000 platform for 150 bp paired‐end reads. Raw sequencing reads were
   processed with CeleScope v1.5.2, using default parameters for gene
   expression profiling. Barcodes and UMIs were extracted and corrected
   from R1 reads, while R2 reads, trimmed of adapters and poly A tails,
   were aligned to the transcriptome using STAR and FeatureCounts. The
   resulting gene expression matrix was used for further analysis. Quality
   control, dimensionality reduction, and clustering analyses were
   conducted using Scanpy v1.8.1 in Python 3.7. The expression matrix
   underwent filtering to exclude cells with extreme gene or UMI counts
   and high mitochondrial content. Post‐filtering, normalization and
   logarithmic transformation were applied. The top 2000 variable genes
   were identified using the “seurat” method, and PCA was performed. Cells
   were clustered using the Louvain algorithm and visualized using UMAP.
   Then, the KEGG and GO analysis were performed.^[ [156]^38 ^]

RNA Sequencing

   BMSCs treated with MCH or control were collected for RNA sequencing.
   After obtaining the raw sequencing data, quality control evaluation and
   comparative analysis of the sequencing data were performed. Sample
   expression levels were quantified, and differential expression analysis
   of genes and transcripts between the two groups was conducted using the
   R package Ballgown. Differentially expressed genes were identified
   using a threshold of a 0.5‐fold difference, a p‐value of ≤ 0.05, and a
   mean FPKM (reads per kilogram of transcript per million mapped reads)
   within the group ≥0.5. Fold change (FC) was calculated as FC =
   2^(test_FPKM – control_FPKM), and the p‐value was determined using the
   F‐test for homogeneity of variance in the R package. For enrichment
   analysis of differentially expressed genes, the Entrez IDs of
   up‐regulated and down‐regulated genes were separately input into the R
   package cluster Profiler. The built‐in function enrich KEGG was used to
   perform KEGG pathway enrichment analysis. Additionally, all expressed
   genes' Entrez IDs and logFC data were input into the cluster Profiler
   package. The built‐in function gse KEGG was employed to conduct Gene
   Set Enrichment Analysis (GSEA) on each KEGG pathway identified in the
   previous step. The gseaplot2 function was used to visualize the GSEA
   results, where the p‐value was obtained through empirical
   phenotype‐based permutation tests to evaluate the nominal p‐value.

Statistical Analysis

   All experimental results were reported as mean ± standard error of the
   mean (SEM) with n indicating the number of animals. Statistical
   analysis was conducted using the appropriate unpaired Student's t‐test
   to compare control and treatment groups. For 3 or more group analysis,
   1‐way ANOVA with Bonferroni's correction for multiple comparisons was
   used. The statistical software Prism version 8 (GraphPad Software, San
   Diego, CA, USA) was employed for all analyses. Statistical significance
   was determined when the probability (P) values were less than 0.05,
   denoting a significant difference between the groups.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   B.G., Y.Z., and S.L. contributed equally to this work. J.Z. designed
   the experiments. B.G. and J.Z. conducted virus tracing experiments.
   B.G., Z.R., H.L., and C.W. conducted chemogenetics experiment. B.G.,
   S.L., and X.C. conducted the electrophysiological experiments. B.G.,
   Y.Z., and X.Y. performed osteoporosis and fracture mice models. B.G,
   Y.Z., and Z.R. performed cellular experiments including BMSC
   differentiation and collection of the samples for high throughout put
   sequencing. J.Z. and B.G. wrote the manuscript with the help of all
   authors.

Acknowledgements