Authors: Wenhui Liu, Xiekai Shen, Yuxiang He, Zhihao Ding, Jin Zhou, Jingzhen Guo, Xiang Lin, Liangzhi Zhang, Peihua Yuan, Yihang Wu, Jiashuo Guo, Xinghao Rong, Lehui Sun, Enyu Yu, Yuan Shi, Jiye He, Yun Ji, Tao Li, Jinwu Wang, Tao Wu
Categories: Review Article, Bone regeneration, Brain–bone axis, Central nervous system, Clinical translation, Peripheral nervous system, Smart materials
Source: Journal of Orthopaedic Translation
Authors: Wenhui Liu, Xiekai Shen, Yuxiang He, Zhihao Ding, Jin Zhou, Jingzhen Guo, Xiang Lin, Liangzhi Zhang, Peihua Yuan, Yihang Wu, Jiashuo Guo, Xinghao Rong, Lehui Sun, Enyu Yu, Yuan Shi, Jiye He, Yun Ji, Tao Li, Jinwu Wang, Tao Wu
In recent years, the concept of neuro–skeletal crosstalk, highlighting the reciprocal interactions between the nervous and skeletal systems, has opened new avenues for understanding the pathogenesis and intervention strategies of complex diseases. This review summarizes the roles of molecular networks such as neurotransmitters, endocrine factors, immune mediators, and extracellular vesicles in bone metabolism, repair, and neurodegenerative diseases, with an emphasis on recent advances regarding bone–derived signals—including the Piezo1 channel and osteocalcin—in neural regulation. Building on this foundation, we focus on advances in frontier materials such as nanomaterials and hydrogels for modulating the brain–bone microenvironment and facilitating coordinated tissue regeneration, as well as new strategies for targeted drug delivery and immune microenvironment modulation. Empowered by next–generation technologies—including multi–omics, artificial intelligence, and organ–on–a–chip systems—the investigation of the fundamental mechanisms and personalized interventions of the brain–bone axis is entering a new era of opportunity. We hope that this review will provide a theoretical basis and valuable reference for future mechanistic studies and innovation in this interdisciplinary field.
By elucidating bidirectional regulatory networks, this review underscores the significant translational potential of targeting the brain–bone axis (BBA) for the treatment of skeletal disorders and neurodegenerative comorbidities. Therapeutic strategies harnessing neurotransmitters (e.g., norepinephrine, serotonin) and neuropeptides (e.g., CGRP) can directly modulate osteoblastic/osteoclastic activity and immune responses, thereby orchestrating fracture repair and metabolic homeostasis. The integration of functional materials—such as stimuli–responsive hydrogels, nanomaterials, and bioelectronic devices—enhances the spatiotemporal precision of signal modulation and facilitates drug delivery across biological barriers, including the blood–brain barrier (BBB). However, challenges regarding low cross–organ targeting efficiency, the complexity of dynamic pathological microenvironments, and physiological discrepancies between animal models and humans necessitate further optimization. Advances in multi–omics analysis, AI–driven network modeling, and intelligent biomimetic delivery systems hold promise for bridging these gaps, offering scalable solutions for clinical translation. This work highlights neuro–skeletal modulation as a transformative dual–targeting strategy for complex diseases, yet its implementation remains contingent upon the refinement of precise intervention technologies and rigorous clinical validation.
The concept of the "brain–bone axis" originates from breakthrough insights into the interactive mechanisms between the nervous system and the skeletal system. Traditionally, bones were regarded as static supporting structures, while the brain was considered the central organ directing systemic regulation. However, recent studies have revealed that the skeleton is not only a metabolically active endocrine organ but also regulates neural activity in a feedback manner by secreting bone–derived factors, such as osteocalcin and fibroblast growth factor 23 [1]. Meanwhile, the nervous system modulates bone metabolism and repair through sympathetic signaling, metabolic byproducts, and extracellular vesicles (EVs) [2]. The discovery of this bidirectional communication offers a novel perspective for understanding the comorbidity mechanisms of age–related diseases, such as Alzheimer's disease and osteoporosis. For instance, epidemiological data show that patients with Alzheimer's disease (AD) have significantly reduced bone mineral density, while those with osteoporosis exhibit an increased risk of cognitive impairment [3], suggesting that both conditions may share a molecular basis involving brain–bone axis dysfunction. Further research has demonstrated that brain–derived EVs from AD patients carry miR–483–5p, which inhibits the expression of Igf2 in bone tissue, resulting in impaired osteogenic differentiation and excessive adipogenesis of bone marrow mesenchymal stem cells (BMSCs), thereby exacerbating osteoporosis [4]. Moreover, bone–derived EVs have been shown to cross the blood–brain barrier and deliver neuroprotective proteins that modulate β–amyloid (Aβ) deposition in the brain [5], establishing a bidirectional information transfer network between bone and brain.
The brain–bone axis is essentially a dynamic, bidirectional regulatory network, centered on cross–organ communication between the nervous and skeletal systems via multiple signaling pathways, such as neurotransmitters, endocrine factors, immune mediators, and extracellular vesicles (EVs) [5]. The introduction of this concept originated from early studies on the “hypothalamus–sympathetic nerve–bone axis,” in which the sympathetic nervous system releases norepinephrine (NE) to activate β2–adrenergic receptors on bone cells, thereby suppressing osteoblast activity and promoting osteoclast differentiation [10]. This mechanism is particularly prominent in stress–induced osteoporosis. However, as bone has been redefined as both an “endocrine organ” and a “mechanosensory organ,” the connotation of the brain–bone axis has been further expanded. For example, osteocalcin (OCN), a typical bone–derived hormone, not only regulates bone metabolism but also improves motor impairment and dopaminergic neuronal loss in Parkinson's disease (PD) models by modulating gut microbiota metabolism and promoting the production of propionate [1], thereby establishing a cross–system regulatory pattern known as the “bone–brain–gut axis.” Meanwhile, osteocytes can sense mechanical stress or metabolic changes via mechanosensitive channels such as Piezo1, and thereby secrete EVs that transmit signals to the brain [11]. Studies have shown that bone cell–derived EVs from young individuals can carry anti–Aβ molecules, such as clusterin, into the brain parenchyma to slow the progression of Alzheimer's disease (AD) [5], whereas EVs from aged individuals lose neuroprotective function due to alterations in their cargo. Notably, there are direct ossified vascular channels, such as skull–marrow channels, connecting the cranial bone marrow and meninges. These allow the bidirectional migration of immune cells and metabolic products, playing a crucial role in neuroinflammation and neurodegenerative diseases. This interaction is not isolated, but rather is deeply integrated with the systemic neuroendocrine–immune network. For instance, inflammatory factors such as TNF–α and IL–6 secreted by osteoclasts in the bone microenvironment can cross the blood–brain barrier and exacerbate neuroinflammation [12], forming a vicious “bone–brain–immune” cycle. Additionally, hypothalamic leptin signaling regulates systemic energy balance through its effects on bone metabolism [10]. In summary, research on the brain–bone axis has not only broadened the theoretical framework for inter–organ communication, but also provided a scientific basis for developing interdisciplinary therapeutic strategies targeting EVs, microbial metabolites, or mechanotransduction pathways.
Neural regulation of the Brain–bone Axis (BBA) is stratified into three distinct the Central Level, the Peripheral Level, and the Local Level. These levels engage in synergistic interplay via specific signaling pathways and distinct temporal scales, thereby constituting a multi–scale regulatory network that spans from macro– to micro–domains and from holistic to local perspectives. This hierarchical regulatory model not only elucidates the sophisticated mechanisms by which the nervous system fine–tunes bone metabolism but also provides a theoretical foundation for understanding the pathogenesis of skeletal disorders and for developing targeted therapeutic strategies (Fig. 1).Fig. 1Schematic diagram depicting multi–mechanistic neural–bone interactions regulating bone homeostasis. a)–PNS regulates bone via neurotransmitters and sensory neurons [6]. b) Leptin controls bone through the hypothalamic–sympathetic axis [6]. c) PGE2 modulates bone sense and metabolism via EP4 and hypothalamic signaling [7]. d) Metal cations promote bone formation via the immune–neural axis [7]. e) Wnt signaling is a therapeutic target for bone diseases [8]. f) (A–B) αCGRP in periosteal cells post–NE. (C–D) αCGRP/CGRP in mice post–fracture/TBI [9].Panels highlight key pathways and their effects on bone formation and resorption. Detailed mechanisms are described in Sections 2.1–2.5 and summarized in Table 1.
Functioning as the supreme regulatory center of the BBA, the brain—primarily through regions such as the hypothalamus and brainstem—integrates peripheral metabolic information to systemically modulate bone metabolic balance via neuroendocrine pathways. Key regulatory regions including the paraventricular nucleus (PVN), arcuate nucleus (ARC), and ventromedial hypothalamus (VMH) act as critical hubs. These nuclei integrate systemic energy metabolism by directly or indirectly receiving hormonal signals secreted by adipose tissue, the gastrointestinal tract, and bone tissue itself (e.g., leptin, insulin, osteocalcin), subsequently transmitting control signals to bone tissue via the sympathetic (SNS) and parasympathetic nervous systems (PNS) [7,6].
The central sympathetic axis represents a classic pathway for regulating bone metabolism. Leptin secreted by adipocytes crosses the blood–brain barrier (BBB) and binds to leptin receptors (ObRb) in the hypothalamic ARC. This interaction activates cocaine– and amphetamine–regulated transcript (CART) neurons, relaying signals through PVN neurons to the sympathetic nervous system, thereby triggering the release of norepinephrine (NE) from sympathetic terminals,NE subsequently targets β2–adrenergic receptors (β2–AR) on bone marrow stromal cells and osteoblasts, activating the cAMP/PKA signaling cascade. This downregulates key osteogenic transcription factors, c–Myc and Runx2, ultimately reducing bone formation while upregulating RANKL expression to promote bone resorption [7].Studies confirm that leptin–deficient (ob/ob) mice exhibit obesity and high bone mass; notably, bone mass increases further following sympathetic blockade, underscoring the sympathetic nervous system as the critical neural pathway mediating leptin–dependent regulation of bone metabolism.
The regulatory role of 5–hydroxytryptamine (5–HT) in bone metabolism highlights the importance of distinguishing peripheral from central serotonergic signaling. Serotonin originates from two major peripheral 5–HT is synthesized by tryptophan hydroxylase 1 (TPH1) in enterochromaffin cells, whereas central 5–HT (accounting for approximately 5% of total body 5–HT) is synthesized by TPH2 in the raphe nuclei. Peripheral 5–HT has been reported to inhibit osteoblast differentiation and bone formation by activating 5–HT1B receptors on osteoblast precursors and interfering with LRP5–dependent Wnt/β–catenin signaling; under certain conditions, it may also indirectly influence bone metabolism through effects on bone resorption. In contrast, centrally synthesized 5–HT acts through serotonergic projections from the raphe nuclei to the ventromedial hypothalamus (VMH), where activation of 5–HT2C receptor–expressing neurons suppresses sympathetic tone and reduces norepinephrine (NE) release, thereby relieving β–adrenergic inhibition of osteoblast activity and favoring bone formation [9,13].
The peripheral nervous system serves as the nexus connecting the CNS to bone tissue, transducing central regulatory signals into local metabolic responses via three major sympathetic, parasympathetic, and sensory nerves. Sympathetic Nerves: Characterized by a high density of tyrosine hydroxylase–positive (TH^+^) fibers, these nerves exhibit extensive infiltration throughout the periosteum and bone marrow. Terminals release NE to transmit inhibitory signals via β2–AR. Activation of β2–AR elevates osteoblastic cAMP levels and phosphorylates the transcription factor ATF4, which downregulates osteocalcin (OCN) and collagen type I (Col1a1) expression, reducing the bone formation rate by approximately 40% [6].
Parasympathetic Nerves: Although choline acetyltransferase–positive (ChAT^+^) fibers are less dense, they facilitate osteogenic differentiation by releasing acetylcholine (ACh) to activate M3 muscarinic receptors (M3R) on osteoblasts, subsequently upregulating Wnt10b expression [14]. Recent studies reveal that butyrate, a gut microbiota metabolite, upregulates osteoblastic Wnt10b expression by enhancing regulatory T cell (Treg) function—a process contingent upon the integrity of the parasympathetic nervous system [14].
Sensory Nerves: Calcitonin gene–related peptide–positive (CGRP^+^) fibers originate from the dorsal root ganglia (DRG). Under physiological conditions, CGRP enhances bone matrix synthesis by activating the CALCRL/RAMP1 receptor complex on osteoblasts, thereby upregulating Runx2 and Osterix expression [8]; However, in chronic pain states (e.g., bone cancer pain), DRG neurons undergo pathological sensitization, causing massive CGRP release into the dorsal horn of the spinal cord, which triggers central sensitization and hyperalgesia. This constitutes a "peripheral healing–promoting vs. central pain–inducing" dichotomous phenotype, implying limited therapeutic windows for targeting this pathway [15].
The local microenvironment functions as the executive apparatus of neuro–skeletal regulation, comprising the highly structured "neuro–vascular–bone unit."
Mechanosensation–Neural Coupling: Osteocytes serve as mechanosensors equipped with Piezo1 ion channels. Upon mechanical stress, these channels open to release ATP and prostaglandin E2 (PGE2). These signaling molecules diffuse to adjacent nerve terminals, activating purinergic P2X3 receptors and prompting sensory neurons to release CGRP, thereby establishing a positive feedback "mechanical stress—osteocyte—neuropeptide—osteogenic response" [11].
Hierarchical Integration and Feedback: Bone–derived osteocalcin (OCN) can cross the BBB to activate Gpr158 receptors on hypothalamic neurons. This interaction upregulates the expression of brain–derived neurotrophic factor (BDNF), forming a "bone–brain–bone" closed–loop feedback mechanism (Fig. 2) [16].Fig. 2Tripartite regulatory framework of the brain–bone axis. The framework integrates three scales of Central Level: The hypothalamus (PVN, ARC, and VMH) integrates systemic signals like Leptin and central 5–HT to modulate bone mass via the autonomic nervous system. Peripheral Level: Signal transmission occurs through sympathetic (NE), parasympathetic (ACh/Butyrate), and sensory (CGRP) pathways, directly affecting bone cell activity. Local Level: Immediate regulation is driven by osteocyte mechanosensation (Piezo1–ATP/PGE2 path) and bone–derived OCN feedback to the brain.
NE triggers the canonical G α s–cAMP–PKA signaling cascade by activating β2–adrenergic receptors (β2–AR) on osteoblasts. Upon activation, PKA phosphorylates multiple downstream targets, including the transcription factor ATF4, thereby reducing its binding affinity for the Runx2 promoter. Concomitantly, it downregulates c–Myc and Cyclin D1 expression, impeding the G1/S phase transition in the osteoblast proliferation cycle [6]. In traumatic brain injury (TBI) mouse models, the unfractured contralateral femur exhibits downregulated osteogenic gene expression and a significantly reduced bone formation rate (BFR/BS), attributed to elevated plasma NE concentrations resulting from high sympathetic tone [17].
NE stimulates bone marrow stromal cells (BMSCs) to express RANKL, which binds to RANK receptors on osteoclast precursors, thereby promoting osteoclast differentiation. Simultaneously, NE inhibits osteoprotegerin (OPG) expression, elevating the RANKL/OPG ratio from a physiological range of 0.5–1.0 to >2.0, thus exacerbating osteoclastic activity [6].
Following TBI, NE signaling exhibits spatiotemporal heterogeneity during fracture healing. In the early phase (days 0–14), moderate TBI–induced β2–AR activation upregulates VEGFA expression in periosteal cells, promoting the formation of Type H vessels (CD31^hi^Emcn^hi^) and accelerating endochondral ossification. However, persistent sympathetic hyperactivity leading to chronic high NE levels disrupts the osteoblast–osteoclast coupling balance, resulting in delayed callus remodeling [17].
Central Regulation: High Npy expression in the hypothalamic arcuate nucleus exerts a bone–protective effect by inhibiting sympathetic outflow via Y2 receptors (Y2R). Global Y2R knockout mice exhibit increased cortical bone thickness, suggesting a regulatory role for NPY in stress–induced bone loss [18].
Peripheral Regulation: Osteoblasts secrete NPY, which activates the Wnt/ β–catenin pathway via autocrine or paracrine signaling. Conversely, NPY overexpression impairs osteoblast mineralization; this "dose–dependent inverse effect" may be attributable to receptor desensitization [19].
Collectively, sensory afferent inputs may relay peripheral skeletal information to hypothalamic nuclei, where NPY– and 5–HT–related pathways participate in the central regulation of bone metabolism [20]. Consistent with the central serotonergic mechanism described above, centrally synthesized 5–HT acts on 5–HT2C receptor–expressing neurons in the VMH to suppress sympathetic outflow, thereby reducing β–adrenergic inhibition of osteoblasts and favoring bone formation [9,13]. Together, these pathways support a sensory nerve–hypothalamus–autonomic axis that contributes to systemic regulation of bone remodeling [18,20]. To improve the structural clarity of the neural regulatory mechanisms discussed above, a summary of representative central, peripheral, and local pathways is provided in Table 1.Table 1Central, peripheral, and local neural regulation of the Key mediators and mechanisms.Table 1Regulatory LevelRepresentative Mediator/PathwayPrimary Anatomical SourceMajor Mechanism in BoneKey Orthopaedic RelevanceCentralLeptin–hypothalamus–SNS axisAdipose tissue→ARC/PVN/VMHActivates sympathetic outflow; NE on β-AR inhibits osteoblast activityOsteoporosis,stress-induced bone lossCentralCentral 5-HT (5-HT2C) signalingRaphe nuclei → VMHSuppresses sympathetic tone, reduces NE release, favors bone formationSystemic bone homeostasisCentralNPY-related hypothalamic regulationArcuate nucleusModulates autonomic output and energy–bone balanceStress/metabolic bone remodelingCentralSFO–PTH bidirectional controlSubfornical organ (SFO)Central sensing regulates PTH secretion and bone massEndocrine–neural integration in bone mass regulationPeripheralSympathetic NE signalingSympathetic fibers in periosteum/marrowβ-AR→cAMP/PKA pathway; inhibits osteoblast, promotes resorptionOsteoporosis, fracture healingPeripheralParasympathetic cholinergic signalingCholinergic fibersACh on muscarinic receptors promotes Wnt10b and osteogenesisBone remodeling,gut–bone axisPeripheralSensory CGRP signalingDRG sensory fibersCALCRL/RAMP1 → Runx2 upregulation; inhibits osteoclastogenesisFracture repair, neurogenic inflammationPeripheralSubstance P/neurogenic inflammationSensory nervesModulates macrophage activity and early inflammatory responseEarly fracture healing, bone painLocalPiezo1–ATP/PGE2–sensory couplingOsteocytes + nerve terminalsMechanosensation couples loading to osteogenic signalingMechanotransduction, bone adaptationLocalNeurovascular–bone unitLocal bone microenvironmentCoordinated nerve–vessel–osteoprogenitor signalingFracture healing and regeneration
These pathways collectively illustrate the hierarchical and interactive nature of neural regulation in skeletal homeostasis and repair.
The dorsal root ganglia (DRG) function peripherally, serving as the initiation point for afferent signals that trigger the hypothalamic–sympathetic axis. Upon bone tissue injury, nociceptors (C fibers and A δ fibers) transmit signals via the DRG to the spinal dorsal horn, projecting to the hypothalamic paraventricular nucleus (PVN). This excites CRH neurons, generating sympathetic output. Studies reveal that following sciatic nerve transection, local CGRP expression in bone tissue decreases, whereas hypothalamic NE levels rise approximately 2.3–fold compared to controls, ultimately leading to delayed callus formation [21].
It is imperative to distinguish the site–specific effects of CGRP [15,22,23].
Peripheral Anabolic Effect: At fracture sites, locally produced CGRP (∼10^−9^ M) stimulates the CALCRL/RAMP1 receptor complex on osteoblasts, upregulating Runx2 to promote bone matrix synthesis and repair, independent of nociceptive transmission.
Central Sensitization: In chronic bone pain, DRG neurons undergo pathological remodeling, with a 3– to 5–fold upregulation of TRPV1 channels and CGRP expression. Massive presynaptic release of CGRP in the spinal dorsal horn activates second–order neurons, triggering central sensitization.
Conclusion: An ideal therapeutic strategy should achieve spatially selective modulation characterized by "peripheral CGRP enhancement plus central CGRP blockade."
Inflammatory Phase (Days 0–7): The nervous system responds rapidly, with sensory nerves releasing Substance P (SP) and CGRP. These neuropeptides induce vasodilation to increase blood supply and modulate immune cell activity. SP activates NK–1 receptors on macrophages, inducing polarization toward the M1 phenotype and stimulating the secretion of TNF–α and IL–1 β to clear necrotic tissue, thereby laying the foundation for repair [24].
Reparative Phase (Days 7–21): Neuropeptides stimulate the differentiation of mesenchymal stem cells into osteoblasts and induce VEGFA expression, facilitating the construction of the "neuro–vascular–bone unit." Immunofluorescence analysis reveals a significantly higher density of nerve fibers surrounding Type H vessels compared to ordinary vessels, indicating a synergistic promotion of callus mineralization [6,24,25].
Remodeling Phase (>21 Days): The nervous system guides callus remodeling by balancing osteoblastic and osteoclastic activities. Sensory nerves detect changes in the mechanical environment (e.g., microstrain induced by weight–bearing) and transmit signals to the CNS. The CNS subsequently modulates neuropeptide release via feedback mechanisms, enabling bone tissue structure to adapt to mechanical demands [6].
Notably, increased peripheral nerve density does not invariably correlate with improved bone healing. Clinical observations have reported aberrantly high nerve density in non–union tissues. Although early reinnervation may facilitate vascularization and tissue repair, persistent or dysregulated nerve ingrowth during later phases of healing may instead indicate maladaptive remodeling. Importantly, this phenomenon should not be interpreted as evidence that excessive CGRP directly promotes osteoclast activation or callus resorption. Rather, CGRP has been reported to inhibit osteoclastogenesis and to participate in bone repair, suggesting that impaired healing in the setting of aberrant innervation more likely reflects complex neuroinflammatory dysregulation, sustained nociceptive signaling, and disruption of the local reparative microenvironment [13,20]. In addition, pathological nerve sprouting is often associated with overactivation of the NGF/TrkA pathway, which may perpetuate an inflammation–pain cycle and further compromise tissue regeneration.
TBI patients often exhibit accelerated fracture healing but an increased risk of heterotopic ossification, a phenomenon driven by unique mechanisms [26]:
Central Mechanisms: Post–injury elevation of serum leptin levels reduces hypothalamic NPY, thereby lifting sympathetic inhibition on bone. Concurrently, TBI–induced high catecholamine concentrations and a sympathetic alert state enhance local vascularization and bone formation at the fracture site via potentiated β2–AR signaling, despite the potential for systemic bone loss [17].
Peripheral Effects: Brain–resident microglia activated by injury release exosomes carrying miR–133a–3p into the circulation. These exosomes target the fracture site, where they inhibit RhoA to promote osteogenic differentiation [27].
The bone interior is richly innervated by A δ and C fibers. These fibers sense microenvironmental pH changes (e.g., acidic inflammatory environments) and high intraosseous pressure via ASICs and TRPV1 channels, transmitting nociceptive signals to the CNS [28].
The analgesic efficacy of teriparatide exhibits distinct characteristics, primarily observed with intermittent administration [29,30]:
Intermittent Low–Dose (Analgesic): This regimen upregulates NGF expression in bone tissue to promote normal sensory nerve repair while simultaneously attenuating nociceptive signal transmission and alleviating chronic bone pain by inhibiting spinal glial activation.
Clinical Observation (Initial Pain Exacerbation): Some patients experience exacerbated bone pain during the initial phase of teriparatide therapy. This may be linked to increased micro–damage resulting from rapid bone turnover or transitional neural remodeling, which subsequently shifts to a significant analgesic effect, explaining the biphasic pain response observed in clinical settings.
Mechanistically, norepinephrine released by the sympathetic nervous system and epinephrine secreted by the adrenal glands act as key catecholamines. By activating α– and β–adrenergic receptors on immune cells, they exert pro– or anti–inflammatory effects at different stages of the immune response. These effects are primarily mediated through β–adrenergic receptors, which suppress immune cell function and modulate the ability of dendritic cells (DCs) to direct the differentiation of Th1 and Th2 cells(Fig. 3).Fig. 3Schematic diagram of the molecular regulatory mechanism of the Neuro–Immune–Skeletal Axis The sympathetic nervous system releases norepinephrine, which reduces the differentiation of DCs into Th1 cells and thus reduces inflammatory cytokines. The release of acetylcholine from the parasympathetic nervous system promotes the polarization of DC and T cells towards Th2 cells through muscarinic receptors, increases anti-inflammatory factors, and inhibits pro-inflammatory factors by immune cells through a7nachs. Harm receptors release substance P to activate immune cells to secrete pro-inflammatory cytokines. They release VIP to inhibit inflammatory cytokines and release CGRP to induce mast cell degranulation and transfer of Langerhans cells, together promoting the differentiation of Th2 cells.
Th1 cells participate in cell–mediated immune responses and promote inflammation, whereas Th2 cells primarily regulate humoral immune responses.
Acetylcholine produced by the parasympathetic nervous system influences immune cells. It exhibits anti–inflammatory properties by suppressing the activation of macrophages, basophils, and mast cells through α7–nicotinic acetylcholine receptors (α7nAChRs). Acetylcholine can also promote the polarization of dendritic cells (DCs) and CD4^+^ T cells toward the Th2 lineage via activation of muscarinic acetylcholine receptors(Fig. 3).
Nociceptors, when exposed to noxious stimuli or inflammation, release neuropeptides such as calcitonin gene–related peptide (CGRP), substance P, and vasoactive intestinal peptide (VIP), which interact with various immune cell types. Substance P, as a pro–inflammatory neuropeptide, can activate multiple immune cell populations. VIP and CGRP tend to favor a Th2 cytokine profile; VIP, in particular, suppresses the production of inflammatory cytokines from DCs and macrophages, while promoting the differentiation, survival, and migration of Th2 cells. CGRP can induce mast cell degranulation and trigger the migration of Langerhans cells, thereby facilitating Th2 cell differentiation(Fig. 3).
Experimental evidence shows that [31] chrna7–deficient mice produce significantly higher levels of inflammatory cytokines—tumor necrosis factor–α (TNFα), interferon–γ (IFNγ), and interleukin–6 (IL–6)—than wild–type mice when immunized with ovalbumin and the Gram–negative bacterium Bordetella pertussis. This phenomenon is attributed to the loss of α7 nicotinic acetylcholine receptors (α7nAChRs) in macrophages. Normally, acetylcholine (ACh) binds to α7nAChRs on macrophages to suppress TNFα production. In vitro, specific engagement of α7nAChRs on macrophages attenuates the production of inflammatory cytokines.
Efferent vagus nerve stimulation experiments have shown that [32] stimulation of the efferent vagus nerve suppresses the synthesis of the pro–inflammatory cytokine TNF–α and inhibits the elevation of serum TNF–α levels. In macrophages stimulated by lipopolysaccharide (LPS), activation of α7 nicotinic acetylcholine receptors (α7nAChRs) suppresses TNF–α synthesis, indicating that acetylcholine (ACh) is involved in the efferent anti–inflammatory reflex pathway. In this reflex circuit [33], efferent preganglionic vagus nerve stimulation activates postganglionic cholinergic neurons that directly innervate macrophages. The ACh released from nerve terminals binds to α7nAChRs on macrophages, thereby suppressing TNF–α synthesis.
In clinical studies, activation of the sympathetic nervous system has been observed to correlate with disease activity in certain autoimmune diseases, such as rheumatoid arthritis, and may contribute to the control of inflammation and disease progression [34]. The interaction between nociceptors and macrophages links inflammation with pain; neuropeptides and inflammatory mediators released by nociceptors and immune cells modulate neuroimmune crosstalk, thereby regulating apical periodontitis [35].
Neurotransmitters and neuropeptides regulate immune cell function via specific receptors, influencing inflammatory responses and immune cell differentiation. These findings provide new mechanistic insights and potential therapeutic targets for understanding neuro–immune interactions and inflammation regulation in bone–related therapies.
The gut microbiota (GM) modulates the immune system by producing molecules such as short–chain fatty acids (SCFAs), which possess immunomodulatory and anti–inflammatory properties and can influence immune cell activity. SCFAs activate G protein–coupled receptors such as Gpr109A (also known as HCA2); when stimulated by butyrate in the immune system, this pathway promotes immune tolerance and supports the development of regulatory T cells (Tregs) [36].
SCFAs act on osteoclasts via Gpr43 to inhibit bone resorption. Gpr41/43–dependent signaling pathways influence the differentiation of mesenchymal stem cells into either osteoblasts or adipocytes. However, a systemic increase in SCFA levels may be sufficient to activate Gpr41/43 signaling on bone cells, thereby promoting adipogenic differentiation (Fig. 4).Fig. 4Schematic diagram of the mechanism of gut microbiota on the brain–bone axis SCFAs activate G protein coupled receptors to promote the balance between regulatory T cells and helper T cells, reduce pro-inflammatory cytokines from regulatory T cells, regulate the differentiation of macrophages into osteoblasts and osteoclasts at different differentiation time points, enhance the differentiation of mesenchymal stem cells into osteoblasts, inhibit the differentiation into adipocytes, and promote the inhibition of bone resorption. The 5-HT produced by the gut microbiota activates the 5-HTR1B receptor located on pre osteoblasts, inhibits the Wnt/β - catenin pathway, and hinders the differentiation of pre osteoblasts into mature osteoblasts.
Serotonin (5–HT) produced by the gut microbiota acts as a hormone and inhibits bone formation by activating 5–HTR1B receptors on pre–osteoblasts, thereby reducing osteoblast proliferation. In contrast, 5–HT synthesized in the central nervous system (CNS) functions as a neurotransmitter that promotes bone growth by increasing sympathetic nervous tone via activation of 5–HTR2C receptors on neurons (Fig. 4) [37].
Using animal models, one research team [38] compared the expression patterns of Gpr41 and Gpr43 across various mouse tissues and in both undifferentiated and differentiated primary bone cells from C57Bl6/J wild–type mice. Moreover, Gpr41 was found to be expressed in both osteoblasts and osteoclasts at different stages of differentiation. Further controlled experiments demonstrated that Gpr41/43–dependent signaling pathways influence the differentiation of mesenchymal stem cells into osteoblasts or adipocytes. In addition, the role of SCFAs as bone resorption inhibitors acting through Gpr43 on osteoclasts was also validated.
A clinical trial involving patients with osteoporosis demonstrated that SCFAs can prevent bone loss and enhance bone strength in osteoporotic animal models. SCFAs hold promise as a therapeutic approach to reduce bone loss and fracture risk in patients with osteoporosis [39].
Another clinical study found that the diversity of the gut microbiota is closely associated with skeletal health in the host. For example, the intake of certain probiotic strains has been linked to increased bone mineral density, possibly due to microbial production of molecules such as SCFAs and serotonin (5–HT) [40].
Through the production of SCFAs and 5–HT, the gut microbiota regulates both immune responses and skeletal health, offering new strategies for the prevention and treatment of osteoporosis.
The tumor microenvironment (TME) is a highly organized ecosystem. Tumors achieve aggressive growth and metastatic spread by establishing mutually reinforcing interactions among angiogenesis, inflammation, and fibrosis. Bone serves as a critical organ in the study of disseminated tumor cells (DTCs), and bone metastasis represents a central focus in clinical research on the bone tumor microenvironment.
In bone metastasis, a “vicious cycle” mechanism has been identified. Tumor cells secrete transforming growth factor–β (TGF–β) to activate osteoclasts, which in turn release growth factors—such as insulin–like growth factor–1 (IGF–1) and calcium ions (Ca^2+^)—that promote tumor proliferation, forming a positive feedback loop [41].
The contribution of the nervous system to the pathogenesis of malignant tumors has emerged as a critical component of the tumor microenvironment (TME). Neural factors such as adrenergic signaling, β2–adrenergic receptor (β2–AR), β3–adrenergic receptor (β3–AR), and norepinephrine influence tumor progression and dissemination through distinct pathways, thereby affecting bone metastasis.
Denervation studies have shown that deletion of β2–AR and β3–AR reduces tumor growth and dissemination in prostate cancer. Vagal nerve transection inhibits gastric tumorigenesis by suppressing the muscarinic acetylcholine receptor M3 (CHRM3) via Wnt signaling. Elevated norepinephrine levels reduce pancreatic tumor initiation, while vagotomy alters pancreatic tumor morphology and shortens survival [42].
Studies on neurotransmitters and tumor progression have shown that sympathetic neurotransmitters and β2–adrenergic receptor (β2–AR) signaling promote the progression of breast and ovarian cancers. Activation of the parasympathetic nervous system reduces the expression of programmed death receptor–1 (PD–1) on CD4^+^ T cells and the expression of PD–1 ligands in tumor tissues, thereby suppressing tumor growth and metastasis [42].
Increasing clinical evidence suggests that neural regulation significantly influences tumor progression and bone metastasis, prompting exploration of denervation or neural signaling modulation as novel therapeutic strategies for bone tumors [43].
Neuro–immune interactions within the tumor microenvironment play a pivotal role in tumor progression and bone metastasis, inspiring new perspectives for bone tumor treatment. The complexity of the bone tumor microenvironment offers opportunities for multidisciplinary therapeutic approaches, with the multi–target regulatory effects of traditional Chinese medicine (TCM) presenting novel directions for bone tumor research and treatment.
Traumatic brain injury (TBI) is defined as structural damage and functional dysfunction of the brain induced by external mechanical forces. Research indicates that TBI accelerates fracture healing by activating the sympathetic nervous system (SNS). This mechanism primarily involves the sympathetic release of norepinephrine (NE), which subsequently modulates bone marrow hematopoiesis and immune cell function via β2– and β3–adrenergic receptor signaling pathways. Specifically, TBI significantly elevates NE levels in the serum and bone marrow of both patients and mice. In fracture models, β3–AR signaling mediates the proliferation of hematopoietic stem cells (HSCs) and drives their differentiation into anti–inflammatory myeloid lineages (e.g., M2 macrophages). Concurrently, this signaling promotes M2 macrophage infiltration and Type H vessel formation within the callus, while β2–AR signaling directly mediates macrophage polarization toward the M2 phenotype, collectively accelerating the healing process [44].
Although TBI typically accelerates fracture healing, it frequently results in heterotopic ossification (HO) and compromised callus quality. Clinical data reveal that while the incidence of HO in patients with isolated TBI is approximately 20%, this rate surges to over 50% when TBI is comorbid with femoral fractures. The disruption of the blood–brain barrier (BBB) induced by severe traumatic brain injury (S–TBI) serves as the initiating event, facilitating the leakage of CNS–resident osteogenic macromolecules (e.g., BMPs) into the systemic circulation. Thrombin directly promotes osteoblast proliferation and inhibits apoptosis by activating protease–activated receptor–1 (PAR–1). Simultaneously, it indirectly stimulates bone healing by inducing cyclooxygenase–2 (COX–2) and downstream growth factors (e.g., VEGF, FGF). Furthermore, levels of calcitonin gene–related peptide (CGRP) and substance P (SP) released by sensory nerves are elevated following S–TBI. CGRP enhances blood supply and promotes osteogenesis, whereas SP not only exacerbates inflammatory responses but also directly stimulates the osteogenic differentiation of mesenchymal stem cells. Leptin levels in serum and cerebrospinal fluid (CSF) rise significantly post–brain injury, accelerating both bone formation and HO. The intense inflammatory response triggered by the recruitment of immune cells, such as mast cells, by the aforementioned factors (e.g., SP, CGRP) establishes a microenvironment conducive to HO formation [45].
Clinical observations of accelerated fracture healing, heterotopic ossification, and enhanced callus formation following TBI are regarded as the most direct evidence for the existence and functional regulation of the "Brain–bone Axis."
Traditional Chinese medicine (TCM) has been shown to effectively treat osteoporosis through multiple mechanisms. TCM regulates the gut microbiota balance and modulates intestinal serotonin (5–HT) levels, thereby contributing to osteoporosis management.
Based on the TCM theory that “the brain is the sea of marrow,” insufficiency of the marrow sea is believed to result in osteoporosis. In TCM, the kidneys are responsible for producing bone marrow and governing the skeleton; deficiency of kidney essence leads to depletion of the marrow sea, resulting in brittle and fragile bones. Modern research has confirmed that neural signaling within the brain–bone axis is closely associated with bone metabolism. The brain regulates osteoblast and osteoclast activity—and thereby influences bone mineral density—by releasing neurotransmitters such as serotonin and norepinephrine. In addition, the hypothalamic–pituitary–adrenal (HPA) axis mediates communication between the central nervous system and the skeletal system, regulating bone growth and remodeling. These findings further support the scientific validity of the “brain as the sea of marrow” theory. Kidney–tonifying herbal medicines improve bone mineral density and alleviate osteoporotic symptoms by modulating the neuro–osteogenic network, central neurotransmitter concentrations, glutamatergic signaling, hypothalamic protein kinase C activity, and bone morphogenetic proteins and their associated signaling pathways [46].
Kidney–tonifying herbal medicines also modulate the hypothalamic–pituitary–ovarian (HPO) axis, restore ovarian function, and improve estrogen levels in postmenopausal women, thereby influencing bone metabolism. In addition, they upregulate neuropeptide gene expression, further contributing to the regulation of bone metabolism.
In experimental studies on secoisolariciresinol–rich lignans (SWRH), SWRH injection significantly inhibited bone loss and improved bone microarchitecture. This effect was associated with the downregulation of tryptophan hydroxylase 1 (TPH1) mRNA and protein expression, thereby suppressing 5–HT synthesis in RBL–2H3 rat basophilic leukemia cells. In studies on ACP, a polyphenol derived from Areca catechu seeds, ACP was found to modulate the Lrp5/TPH1/5–HT signaling pathway, suppressing 5–HT production and activating the osteoblastic Wnt/β–catenin pathway, thereby alleviating osteoporosis. In experiments on Bushen Zhuanggu granules, the formulation was shown to reduce gut–derived 5–HT synthesis, increase bone mineral density, and ameliorate osteoporotic symptoms [47]. After 3 months of intervention with Bushen Zhuanggu granules, ovariectomized osteoporotic rats exhibited increased bone density along with elevated serum levels of growth hormone (GH) and insulin–like growth factor–1 (IGF–1), compared with controls. Bushen Zhuanggu granules may improve the expression of GH, IGF–1, and their receptors, thereby preventing further bone loss and enhancing bone mineral density [48].
Kidney–tonifying herbal medicines are widely applied in clinical management of osteoporosis. They exert therapeutic effects by modulating the neuro–immune–skeletal axis through multiple pathways, thereby alleviating osteoporotic symptoms. Clinical studies have shown that these herbal therapies significantly improve bone mineral density and bone metabolism. Further clinical trials in specific populations (e.g., postmenopausal women) are needed to confirm their efficacy and safety [49].
Traditional Chinese medicine treats osteoporosis through multiple mechanisms, significantly enhancing bone density and relieving symptoms. Among them, kidney–tonifying herbal medicines offer promising new strategies for osteoporosis therapy.
The metaphysis is located at both ends of a long bone, between the diaphysis and the epiphysis. It is highly vascularized and rich in trabecular bone, serving as a key site for bone growth and remodeling. Cortical bone is characterized by high strength and stiffness, providing structural support and protection, and playing a crucial role in maintaining skeletal morphology and mechanical integrity. The vertebral body is the principal component of a vertebra and plays a vital role in protecting the spinal cord and nerve roots. It consists of trabecular (spongy) bone internally, enclosed by a thin layer of cortical bone on the surface.
Experimental studies have shown that parathyroid hormone (PTH) can increase the thickness of both trabecular and cortical bone(Fig. 5e) [50],an effect attributed to TGF–β signaling. PTH stimulates osteoblasts to secrete the chemokine MCP–1 (CCL2), thereby enhancing TGF–β levels, which in turn facilitates the recruitment of mesenchymal stem cells to active bone surfaces(Fig. 5b). It also coordinates local signaling molecules, including Wnts, bone morphogenetic proteins (BMPs), and insulin–like growth factor–1 (IGF–1), to regulate bone remodeling [51]. Due to structural differences, the metaphysis, cortical bone, and vertebral body exhibit distinct responses to neural signals. The metaphysis is highly responsive to bone growth signals; cortical bone requires fine–tuned regulation to maintain mechanical strength; and the vertebral body must balance responsiveness with its role in protecting the spinal cord and nerve roots.Fig. 5Schematic diagram of PTH and leptin releasing neural signals to regulate different bone tissuesa) Strain and strain region of the distal femur metaphysis. Reproduced with permission [52]. Copyright 2024, Elsevier. b) A model for the role of the chemokine, MCP–1. Reproduced with permission [53]. Copyright 2007, Elsevier. c) NPY‐mediated mechanism. Reproduced with permission [54]. Copyright 2021, Wiley–Adv Sci. d) Cortical bone and spongiosa mineral density induced by leptin. Reproduced with permission [52]. Copyright 2024, Elsevier. e) Histomorphometric analysis of trabecular bones of PPR. Reproduced with permission [55]. Copyright 2021, Aging.Fig. 6Schematic Diagram of Application Mechanism of Nanomaterials and Hydrogels a) Targeted Delivery: Nanomaterials transport therapeutics across biological barriers. b) Controlled Release: Stimuli–responsive systems trigger drug release. c) Cartilage Repair: A sustained–release exosome hydrogel for synergistic cartilage regeneration.
In experimental studies using leptin–deficient mouse models, leptin deficiency was found to increase trabecular bone volume and trabecular thickness, but resulted in reduced cortical bone strength (Fig. 5a,d) [52]. Moreover, leptin deficiency significantly reduced cortical bone mass in the vertebral body [56]. Leptin signaling promotes osteogenic differentiation while inhibiting adipogenic differentiation [57], a process that involves neuropeptide Y (NPY), epinephrine, and β–adrenergic receptors. Loss of neuropeptide Y results in increased trabecular bone volume and cortical bone thickness(Fig. 5c) [58]. Leptin deficiency reduces sympathetic nervous tone, and genetic or pharmacological ablation of adrenergic signaling leads to high bone mass despite leptin resistance. β–adrenergic receptors regulate osteoblast proliferation and mitigate the negative regulation of bone formation by the sympathetic nervous system (SNS) [4].
Further clinical studies are needed to validate the effects of parathyroid hormone (PTH) and leptin on bone remodeling in humans.Additionally, the specific roles of neuropeptide Y, epinephrine, and β–adrenergic receptors in human bone metabolism remain to be elucidated. The mechanisms and safety of PTH and leptin signaling pathways in humans, as well as the issue of cortical bone weakening due to leptin deficiency, require further investigation. Precisely regulating neuropeptide Y, epinephrine, and β–adrenergic signaling to optimize bone remodeling remains a significant challenge.
PTH and leptin influence bone architecture through the modulation of neural signaling pathways, highlighting the critical role of neural inputs in bone remodeling and offering new therapeutic targets for skeletal disorders.
Several mouse models were employed in the studies cited in this review.
In studies investigating the role of acetylcholine signaling in pathological immune responses [32], mice with a genetic deficiency in the chrna7 gene were used to assess the function of this pathway. The results demonstrated that α7 nicotinic acetylcholine receptor (α7nAChR) signaling effectively suppresses inflammatory cytokine production in mice. Parallel studies in human sepsis patients indicated that high chrna7 expression correlates with reduced inflammatory status and improved clinical outcomes. These findings support the conclusion that the ACh–α7nAChR pathway plays a crucial role in controlling inflammation in both mice and humans, demonstrating good cross–species comparability.
In studies on the regulation of different bone tissues by leptin signaling [52], a leptin–deficient mouse model was used. The findings revealed that leptin deficiency increased trabecular bone volume and trabecular thickness at the distal metaphysis of the femur, while cortical bone was comparatively softer. However, the study included only female mice, with no sex–based comparisons conducted, and lacked supporting clinical research—thus limiting its overall comparability.
Moreover, chrna7–deficient mice differ from humans in terms of gene expression. In human sepsis patients, high chrna7 expression is associated with reduced inflammation, whereas in mice, α7nAChR signaling suppresses inflammatory cytokines. Although leptin–deficient mouse models demonstrate bone structural alterations, there are notable metabolic differences between these mice and humans.
In summary, while animal models provide valuable foundational data in brain–bone axis research, caution must be exercised when interpreting and extrapolating these results to humans. Researchers must account for interspecies differences and refine models through validation and adaptation to improve the reliability and clinical relevance of their findings.
Notably, current animal models have limitations in addressing sex–specific differences. These limitations underscore the need to incorporate age stratification and sex specificity into human studies to advance mechanistic understanding of the brain–bone axis.
Relevant studies have shown that [59] men aged ≥50 years have higher body mass index (BMI), total fat mass (FM), percentage of total FM, trunk FM, and appendicular fat mass (AFM) than those aged <50 years, but exhibit lower body weight and lumbar spine bone mineral density (BMD). In postmenopausal women, both total fat mass (FM) and total lean mass (LM) increase significantly, while bone mineral density (BMD) generally declines. In men, BMD typically peaks during the second decade of life (ages 10–20) and declines gradually thereafter. In women, BMD continues to increase until perimenopause, followed by a sharp decline after menopause. Additionally, appendicular lean mass (ALM) begins to decline markedly in both men and women after the age of 50.
Another study found that [60] serotonin (5–HT) levels decline with age in both women and men. In women, 5–HT levels are negatively correlated with body weight, body mass index (BMI), and fat mass. In men, 5–HT is positively correlated with height and lean body mass. Among premenopausal women, 5–HT is negatively associated with lumbar spine bone mineral density (BMD), whereas this relationship reverses in postmenopausal women. These findings partially reflect the influence of sex and age on the brain–bone axis.
Following the decline in estrogen levels, on one hand, reduced expression of estrogen receptors in bone cells leads to decreased osteoblast activity and increased osteoclast activity, resulting in bone resorption exceeding bone formation; on the other hand, diminished estrogen signaling in the central nervous system disrupts the regulation of the hypothalamic–pituitary–adrenal (HPA) axis and autonomic nervous system, further exacerbating bone metabolic imbalance [61].
These findings suggest that sex and age have significant impacts on skeletal health, and should therefore be carefully considered in the development of personalized strategies for clinical treatment and prevention. Further studies are needed to elucidate the specific mechanisms of the brain–bone axis and to develop effective skeletal health interventions tailored to different sex and age groups. However, advancing our understanding and therapeutic modulation of the brain–bone axis, and achieving clinical translation, remains challenged by a range of technical and mechanistic barriers.
This subsection primarily concerns peripheral neural regulation rather than central brain-bone-axis signaling.
The clinical translation of brain–bone axis research requires a clearer distinction between central brain–bone-axis mechanisms and peripheral neuro–skeletal regulation. In the context of bone tumors, denervation or modulation of peripheral neural signaling has emerged as a potential neuro–skeletal therapeutic strategy rather than a direct example of central brain–bone-axis intervention [62]. By contrast, other translational approaches may involve broader neuroendocrine or gut–brain–bone interactions. For example, modern studies on traditional Chinese medicine (TCM) suggest that certain herbal compounds may modulate the gut microbiota, influence serotonin-related pathways, and engage neuroendocrine signaling, thereby contributing to the treatment of osteoporosis [47]. In addition, based on the classical concept in Huangdi Neijing that “the brain is the sea of marrow,” kidney-tonifying herbal medicines have been reported to regulate neuro-osteogenic interactions through multiple pathways, thereby alleviating osteoporosis and improving bone metabolism [46].
Despite these emerging translational directions, the field still faces major challenges in experimental methodology, disease modeling, and data integration. Neural regulation of bone exhibits marked context- and tissue-specific variability, and factors such as parathyroid hormone (PTH) and leptin may exert differential effects on trabecular and cortical bone, highlighting the need for more precise mechanistic and safety evaluation in humans [63]. Although animal models have provided important foundational insights, their translational value remains constrained by interspecies differences, insufficient sex-based comparisons, and limited clinical validation. Aging and sex further shape brain–bone postmenopausal women experience accelerated bone loss, and serotonin-related regulation of skeletal homeostasis may also be sex-dependent, supporting the need for more individualized intervention strategies. Future studies should therefore focus on refining mechanistic models, improving clinically relevant experimental systems, advancing personalized therapeutic strategies, and integrating emerging technologies with both conventional and traditional medical approaches.
These challenges underscore the need for advanced research methodologies capable of dissecting the multi-level mechanisms of brain–bone and neuro–skeletal interactions. The following sections therefore focus on how molecular and cellular approaches—including gene editing, RNA interference, and electrophysiological techniques—may help overcome current research bottlenecks and accelerate both mechanistic discovery and clinical translation in this field.
Bidirectional signal transmission within the Brain–bone Axis (BBA) relies on precise molecular modulation. However, traditional bone–targeting therapeutics fail to synchronously regulate the central nervous system (CNS), thereby limiting therapeutic efficacy for BBA–associated disorders such as Alzheimer's disease (AD) comorbid with osteoporosis.
Recently developed dual–targeting nanocarriers represent a significant
RVG–Exosome System: The RVG–exosome system designed by Alvarez–Erviti et al. leverages rabies viral glycoprotein (RVG) to specifically bind neuronal acetylcholine receptors, facilitating the transport of siRNA across the blood–brain barrier (BBB) to target and silence BACE1 (thereby reducing amyloid–β generation). Concurrently, these exosomes are internalized by osteoclasts, where the delivered miR–29b inhibits bone resorption. This system demonstrated dual therapeutic efficacy in an AD mouse model [64].
Ferrite–Based Composite Nanosystem: The superparamagnetic iron oxide/silica/carbon composite nanoparticles ("earthicles") developed by Shein et al. achieved dual–site targeting to both the brain and bone. Their enrichment in these tissues was significantly higher than that of non–targeted systems, effectively resolving the "brain drug vs. bone drug selectivity paradox" [65].
The inherent piezoelectric effect of bone tissue naturally couples with neural electrical activity; furthermore, conductive materials can simultaneously modulate the functions of both neurons and osteocytes.
Polypyrrole (PPy) Hydrogels: Under electrical stimulation (100 mV/mm), bone marrow stromal cells (BMSCs) cultured on PPy–gelatin hydrogels exhibited synchronized expression of osteogenic markers (ALP, OCN) and nerve growth factor (NGF). Following implantation in cranial defects, the nerve fiber density in newly formed bone reached 89% of that in native bone, significantly outperforming non–conductive scaffolds [66].
Carbon Nanotube Composite Scaffolds: Designed to mimic neural axons, these scaffolds provide longitudinal conductive pathways. They can record and transmit transient membrane potential changes (± 50 mV) generated during osteoblast differentiation to adjacent undifferentiated cells, thereby establishing an "electrically coupled osteoblast network" [67].
Ideal materials for BBA regulation should possess the capability to sense and respond to physiological changes occurring within the central nervous system.
Dopamine–Responsive Hydrogels: These hydrogels utilize cross–linked networks containing boronic ester bonds. When local dopamine concentrations rise (e.g., following Parkinson's disease treatment), the catechol structure of dopamine competitively binds with the boronic ester, inducing gel swelling and subsequent drug release. In a PD model, this material improved both motor function and bone density [68].
pH–Gradient Dual–Control Systems: Addressing the temporal characteristics of complex wound healing (acidic inflammatory phase → alkaline reparative phase), a bilayer pH–responsive hydrogel was constructed. It releases anti–inflammatory/antioxidant nanoparticles (e.g., morin) during the acidic phase and sequentially releases angiogenic factors (e.g., bFGF) during the alkaline phase. This precise realization of an "early anti–inflammatory/late pro–regenerative" physiological rhythm offers technical insights for modulating the fracture microenvironment (Fig. 6) [69].
Although functional materials demonstrate potential in modulating the BBA, clinical translation remains hindered by two core bottlenecks.
Bidirectional signal transmission in Alzheimer's disease (AD) requires precise molecular regulation; however, no current method effectively mimics natural cross–barrier transport. For instance, bone–derived Sclerostin can cross the BBB and target the neuronal Wnt/ β–catenin pathway, promoting neuronal apoptosis and memory deficits [70], yet existing artificial nanocarriers cannot replicate this efficient cross–barrier delivery. Furthermore, as the mechanisms by which brain–derived extracellular vesicles (EVs) mediate osteogenic/adipogenic disorders in AD remain incompletely understood, targeted interventions using AD brain–derived EVs currently suffer from significant off–target effects.
BBA pathology involves complex intersecting factors such as oxidative stress and metabolic reprogramming, with the fracture microenvironment undergoing continuous spatiotemporal evolution. However, traditional hydrogel materials typically fail to achieve spatiotemporal responsiveness to pathological microenvironments. Moreover, they often cannot simultaneously satisfy the divergent mechanical and degradation requirements inherent in the repair of multi–organ BBA injuries [71].
To overcome these limitations, future material design must integrate multidisciplinary technologies, focusing on the following aspects.
Combining brain imaging technologies with intelligent algorithms allows for the design of materials with spatiotemporal dynamic responsiveness. For example, novel nanocomposite hydrogels (e.g., MPDA@DEX@gel) have achieved precise controlled drug release in traumatic brain injury [72]. Future strategies should extend this to the delivery of bone–derived factors (e.g., SOST inhibitors) to achieve synchronous intervention in bidirectional "brain–bone" signaling.
The integration of non–invasive brain stimulation with material delivery systems represents a growing trend. For instance, transcranial alternating current stimulation (tACS) can modulate cortical theta waves (4–8 Hz) to enhance neuroplasticity. When combined with engineered exosomes delivering miR–483–5p, this approach could generate synergistic effects to ameliorate AD–associated bone metabolic abnormalities [73].
Functional materials should be developed to target the energy metabolic characteristics of the BBA. For example, metabolic activation materials (such as sustained–release 3–hydroxybutyrate scaffolds) can regulate BBA energy metabolism by enhancing ATP synthesis and the TCA cycle. Simultaneously, pleiotropic molecules targeting the gut–brain–bone axis (e.g., SLC39A8 pathway modulators) can not only alleviate inflammation in osteoarthritis but also improve concurrent cognitive decline [71].
Investigating the molecular mechanisms of the brain–bone axis relies on precise gene manipulation and the construction of multicellular interaction models. The rapid advancement of molecular and cellular biology techniques has established a solid technical foundation for dissecting the molecular regulation of the brain–bone axis, particularly by enabling more accurate tools for genome editing and the analysis of intercellular communication.
Recent studies have employed CRISPR–Cas9 technology to selectively knock out parathyroid hormone receptor (PTH1R) in the subfornical organ (SFO), providing crucial molecular evidence for the regulation of the "brain–parathyroid–bone axis" [74]. Moreover, CRISPR–Cas9 is no longer limited to gene knockout applications; dual base–editing systems such as CRISPR–DdCBE have been used to repair mitochondrial DNA in damaged cells, significantly improving neuronal and osteoblastic functions. This offers a novel perspective on mitochondria–related signaling crosstalk within the brain–bone axis [75,76]^,^ [77]. These findings suggest the existence of a mitochondria–associated regulatory pathway linking brain and bone functions; however, further investigation is required to confirm this hypothesis.
In recent years, breakthroughs in integrative single–cell multi–omics and spatiotemporally resolved omics have provided critical tools for uncovering cross–scale interactions between the nervous system and bone metabolism. By integrating single–cell transcriptomic (scRNA–seq) and chromatin accessibility (scATAC–seq) data, researchers for the first time identified a bidirectional communication circuit between Sema3A–expressing nerve fibers and osteoblasts in bone tissue, mediated via the Nrp1 Sema3A signaling promotes osteogenic differentiation, while osteoblast–derived BMP–2 can retrogradely regulate axonal growth [78,79]. This finding provides direct molecular evidence supporting the bidirectional regulation hypothesis of the brain–bone axis. Furthermore, panoramic analysis of fracture healing using spatial transcriptomics (10x Visium) enabled precise localization of mechanosensitive zones—such as strain–dependent expression of Wnt pathway genes and cell adhesion molecules—and, through 3D spatial modeling, established for the first time a dynamic correlation between spatial transcriptomic profiles and local mechanical stress fields in bone tissue. This technological breakthrough revealed the spatiotemporal coupling among mechanosensation, osteogenic differentiation, and neural regulation, and identified novel mechanotransduction targets such as Ccn2 [80]. These multidimensional omics tools have established a comprehensive framework bridging single–cell interactions and tissue–scale regulation, offering a new systems biology approach to brain–bone axis research.
Emerging bioengineering technologies offer transformative toolkits for dissecting the bidirectional regulation between the brain and bone. Magnetically responsive liquid–metal multi–electrode arrays (MEAs) have been innovatively integrated into brain organoids, whose biomimetic Young's modulus and magnetically actuated deformation properties allow for non–invasive, three–dimensional recording of dynamic neural network activity [81]. Complementing this system, next–generation bone–on–a–chip platforms employ microfluidics and bioprinting to construct vascularized bone tissue models, further enhanced by dynamic mechanical loading modules that simulate the in vivo bone microenvironment [82]. The integrated application of these two systems enables bidirectional functional the MEA–equipped brain organoid captures real–time neural electrical signals, while the bone–on–a–chip simultaneously monitors osteogenic responses under mechanical stimuli (e.g., activation of the RUNX2 pathway), thereby enabling, for the first time, decoding of mechano–electrical coupling in cross–organ regulation within a biomimetic system. Building a cross–validated, multiscale technological pipeline further enhances research depth—for instance, by using three–photon microscopy to trace the transmission pathways of brain–derived signaling molecules into bone tissue [83]. This integrated [organoid–organ–on–chip–in vivo imaging] research framework not only enables experimental validation of the hypothesis that brain–derived signals directly regulate skeletal function, but also offers a systematic approach to unraveling the multi–dimensional communication mechanisms underlying the brain–bone axis.
Animal models hold an irreplaceable central role in brain–bone axis research, offering the unique advantage of preserving the dynamic interaction between the nervous and skeletal systems under physiological conditions. Compared with in vitro models such as brain organoids or bone–on–a–chip systems, in vivo models retain the spatiotemporal coupling between the neuroendocrine network, autonomic innervation, and the skeletal microenvironment. Moreover, they allow for the application of genetic manipulation, neural modulation, or mechanical loading paradigms to dissect the causal pathways underlying bidirectional brain–to–bone signaling.
The combined application of optogenetics and chemogenetics enables targeted modulation of specific neural circuits and receptor signaling, systematically revealing the brain's direct regulatory influence on peripheral skeletal tissue. For example, using a dual–pathway manipulation approach to precisely activate AgRP neurons in the hypothalamic arcuate nucleus, researchers found that this activation significantly elevated cortisol secretion and inhibited osteocalcin synthesis in osteoblasts, thereby reducing the rate of bone formation [84]. This finding is significant in that it establishes a complete causal chain—[brain region activation → hormone release → skeletal metabolic response]—through spatiotemporally specific neural circuit interventions, providing critical experimental evidence for the central hypothesis that brain–derived signals directly regulate bone tissue.
The breakthrough application of trans–synaptic viral tracing systems has provided indispensable evidence for elucidating long–range functional connections between the central nervous system and the skeletal system(Fig. 7A). For instance, retrograde monosynaptic tracing techniques can precisely label key brain regions involved in bone metabolism regulation (e.g., the hypothalamus and dorsal raphe nucleus) along with their upstream neural networks. By identifying neural circuits that directly or indirectly project to bone tissue, this approach allows researchers to determine whether specific nuclei—such as the lateral habenula—regulate bone remodeling via sympathetic or vagal pathways(Fig. 7B) [85]. This methodology thus provides critical evidence for mapping the multi–synaptic hierarchical architecture of the brain–bone axis.Fig. 7Integrated transcriptomic characterization and chronic electrophysiology. a. Transcriptomic profiling of rabies–labeled inputs. (A) Workflow for monosynaptic rabies tracing in V1. Cre–dependent helper virus and EnvA + rabies (RVdG–H2B–mCherry) were sequentially injected into specific excitatory neuron driver lines. mCherry + nuclei were isolated via FANS for snRNA–seq. (B) Hierarchical taxonomy used to classify input neurons by class, subclass, and subtype [86]. Copyright 2024, Patiño M b. Chronic recording interface. Schematics of a miniaturized docking and payload system utilizing dual Neuropixels 2.0 probes, designed for stable, high–density recording in freely moving animals [87]. Copyright 2025, Bimbard C c. Long–term tracking of neural ensembles. Analysis of unit stability over extended periods (up to 106 days). Panels show unit presence rasters (A), consistent depth tracking (B), stable waveforms (C), and Inter–Spike Interval (ISI) histograms (D), demonstrating the system's ability to monitor neural dynamics over months [87]. Copyright 2025, Bimbard C.
Wireless implantable multichannel recording technologies, such as Neuropixels 2.0, overcome the limitations of traditional brain–bone axis research in capturing dynamic signals, thanks to their ultrahigh throughput (>5000 channels) and high spatiotemporal resolution(Fig. 7C). These tools have become central to decoding the temporal dynamics underlying central–to–peripheral skeletal regulation. In freely moving animal models, this technology enables millisecond–level synchronous recording of neural ensemble activity across multiple brain regions—including the prefrontal cortex, basal ganglia, and hypothalamus—allowing for the first time the construction of a dynamic closed–loop map linking environmental stimuli, brain network integration, and skeletal adaptation responses(Fig. 7D) [87]^,^ [88]. Compared with viral tracing techniques, the unique advantage of this approach lies in its capacity to embed temporal resolution into brain–bone axis studies. It enables the validation of the brain's rapid, remote regulation of bone metabolism via neural activity, offering preclinical data to support the development of neuroelectrical modulation–based bone–targeting therapies, such as using deep brain stimulation to optimize bone remodeling cycles.
System–level analysis of the brain–bone axis, enabled by multimodal data integration and causal inference, has begun to uncover the deep regulatory logic of inter–organ interactions in physiological processes. Scientifically, this approach integrates dynamic inter–organ metabolite profiles, molecular functional networks, and temporally resolved epigenetic data to construct a multiscale framework linking metabolic interactions to physiological functions—thereby overcoming the spatial and mechanistic fragmentation inherent in single–organ studies. By constructing cross–organ causal network models, it becomes possible to systematically distinguish core drivers of metabolic signal transmission from secondary or coincidental effects, addressing the limitations of conventional correlation–based analyses and reducing the risk of spurious causality. This research paradigm offers a dynamic framework for validating comorbidity mechanisms between neurodegenerative disorders and bone metabolic abnormalities, and—through a closed–loop workflow of “metabolic interaction modeling–key node identification–intervention simulation”—advances precision medicine from single–target interventions toward inter–organ homeostatic regulation. This marks a paradigm shift in complex disease research, from localized mechanistic dissection to integrated systemic interpretation.
Untargeted metabolomics enables comprehensive profiling of metabolites in body fluids or tissues, revealing the regulatory patterns of metabolic signal transmission between the brain and the skeletal system. Compared to traditional single–biomarker approaches, its high–throughput capacity—for example, the simultaneous analysis of lactate, ceramides, and other metabolites—can overcome physiological barriers to elucidate the dynamic inter–organ transport of signaling molecules [89]. By integrating isotope tracing techniques, such as ^13^C–glucose labeling experiments, researchers have precisely tracked the transport of metabolites from the cerebrospinal fluid circulation or vascular interstitium into bone tissue, thereby systematically uncovering the biological basis of brain–to–bone information exchange via metabolic intermediates [90]. The core value of metabolomics lies in its ability to identify key signaling molecules within complex metabolic networks and to elucidate the molecular mechanisms by which the brain communicates with the skeletal system through metabolic intermediates.
Graph neural networks (GNNs) are well–suited for analyzing complex metabolic interaction data between the brain and skeletal system [91], offering precise resolution of dynamic relationships among diverse biomolecules. The resulting “brain–bone metabolic interaction map,” constructed using GNNs, provides a core analytical framework for integrating multi–source datasets and reconstructing inter–organ metabolic networks. This framework has enabled three major advances in critical research areas such as drug toxicity evaluation, disease mechanism interpretation, and dynamic metabolic First, it facilitates the investigation of inter–organ metabolite transport patterns; second, it enables identification of aberrant metabolic regions during disease progression; and finally, it offers a novel approach for personalized drug target identification, accelerating translational applications of bidirectional brain–bone regulatory mechanisms.
While Sections 5.1–5.3 delineated the technological arsenal for dissecting neuro–skeletal mechanisms, the ultimate imperative is to translate these high–dimensional datasets and in vitro models into actionable clinical strategies. Bridging the gap between mechanistic discovery and clinical application requires a translational framework that leverages these methodologies to address patient heterogeneity and interspecies discrepancies.
Clinical heterogeneity in skeletal disorders often stems from uncharacterized variations in neuroendocrine profiles. The "metabolic interaction maps" constructed via multi–omics (Section 5.3) provide a methodological foundation for precision diagnosis [89]^,^ [90].
Liquid Biopsy for BBA Profiling: By translating untargeted metabolomics into targeted clinical assays, specific "neuro–skeletal metabolite signatures"—such as distinct ratios of serum serotonin derivatives or brain–derived exosomal miRNAs—can be identified. This allows for the stratification of osteoporotic patients into "central–driven subtypes" (requiring hypothalamic–pituitary modulation) versus "peripheral–metabolic subtypes" (requiring local skeletal intervention), thereby guiding the choice of therapeutic agents with greater precision than traditional bone mineral density (BMD) measurements [92,93].
A major bottleneck in drug development is the physiological divergence between rodent models and humans. The organoid and organ–on–a–chip platforms described in Section 5.1 offer a robust translational bridge [82]^,^ [83].
Pre–Clinical "Avatar" Screening: Utilizing patient–derived induced pluripotent stem cells (iPSCs) to construct personalized neuro–skeletal organoids enables the creation of "clinical avatars." Before administering neuro–active agents (e.g., novel β–blockers or neuropeptide analogues), these agents can be screened on the patient's own organ–on–a–chip system. This approach not only predicts individual drug sensitivity but also assesses blood–brain barrier (BBB) permeability and potential off–target effects on neural tissue, significantly de–risking the transition to Phase I clinical trials.
The precise mapping of neural circuits achieved through viral tracing and optogenetics (Section 5.2) lays the groundwork for next–generation bioelectronic medicine [87]^,^ [86].
Closed–Loop Neuro–Skeletal Modulation: While optogenetics remains largely restricted to animal research, its functional maps guide the placement of clinical neuromodulation devices. Emerging technologies, such as focused ultrasound or vagus nerve stimulation (VNS), can be calibrated to target the specific "skeletal–projecting" autonomic pathways identified in preclinical mapping. Furthermore, integrating the chronic recording capabilities of extensive electrode arrays (as seen in Section 5.2) into implantable devices could enable closed–loop therapeutic systems. These systems would dynamically adjust stimulation parameters in response to real–time fluctuations in skeletal biomarkers (e.g., osteocalcin levels), thereby maintaining homeostatic neuro–skeletal tone without the systemic side effects of pharmacological interventions.
By systematically applying these advanced methodologies to clinical challenges, we can transition from a descriptive understanding of the brain–bone axis to an engineered, predictable, and personalized therapeutic framework.
Technological innovations in brain–bone axis research are shifting from unidirectional causal exploration toward dynamic network using molecular editing as a "causal scissor" to dissect critical pathways, in vivo models as a "spatiotemporal magnifier" to capture dynamic interactions, and multi–omics combined with artificial intelligence as a "system decoder" to integrate cross–scale data. Moving forward, overcoming three major barriers—organ–level isolation, temporal resolution, and interspecies differences—will be essential for constructing a comprehensive “dialogue map” of brain–bone interactions.
The discovery of the brain–bone axis (BBA) has revolutionized the traditional framework of skeletal biology, revealing the bidirectional and dynamic interplay between the central nervous system and the skeletal system. Its core mechanisms include.(1)**Neuroendocrine ** Leptin modulates bone metabolism bidirectionally through both the sympathetic nervous system and the CART pathway. This dynamic balance plays a vital role in maintaining skeletal homeostasis and responding to pathological stress [94,95].(2)**Direct peripheral neural ** Calcitonin gene–related peptide (CGRP) released by local neurons promotes osteogenic differentiation by activating the cAMP–CREB1–SP7 signaling axis in periosteal–derived stem cells (PDSCs), offering direct mechanistic support for neural regulation of bone repair [96].(3)**Central–to–skeletal signaling ** Under chronic inflammatory conditions, spinal IL–1β activates astrocytes, leading to the secretion of various pro–inflammatory cytokines and chemokines. These mediators further influence neuronal and glial activity, sustaining the chronic inflammatory state [97]. For instance, in osteoarthritis, inflammation induces articular cartilage degeneration, while spinal IL–1β exacerbates the inflammatory response, ultimately contributing to bone loss [98].
The aforementioned neural regulatory mechanisms provide a molecular foundation for understanding the bidirectional pathogenesis of related diseases. Notably, dysregulation of neural–skeletal interactions not only results in bone metabolic abnormalities, but also contributes to the formation of complex pathological networks involving various neurodegenerative diseases.(1)Alzheimer's disease (AD): There exists a complex bidirectional relationship between AD and skeletal health. On one hand, AD affects skeletal health through multiple mechanisms, including pathogenic proteins (e.g., Aβ peptides and Tau), genetic risk factors (e.g., APOE and TREM2), neurohormones (e.g., estrogen and FSH), neuropeptides (e.g., NPY and Kisspeptin), peripheral nerves and neurotransmitters (e.g., acetylcholine, norepinephrine, and glutamate), as well as brain–derived extracellular vesicles (EVs). For example, Aβ peptides promote osteoclast differentiation, and the AD risk genes APOE and TREM2 are involved in bone metabolic regulation. On the other hand, the skeleton actively participates in AD pathogenesis. Bone–derived hormones (e.g., osteocalcin and lipocalin–2), bone marrow–derived cells (including immune cells and mesenchymal stem cells), bone–derived EVs, and inflammatory responses are all linked to AD pathology and cognitive function. These findings offer crucial insights into the intrinsic connection between AD and skeletal homeostasis [99,100].(2)Parkinson's Disease: Studies have shown that Parkinson's disease (PD) is also closely linked to skeletal health. Recent evidence suggests that dopaminergic signaling, particularly of peripheral origin, can inhibit RANKL-induced osteoclast differentiation in RAW264.7 cells (a murine monocyte–macrophage cell line) via p–eIF2α signaling. Additionally, they stimulate osteogenic activity in MC3T3–E1 pre–osteoblasts, enhancing alkaline phosphatase (ALP) activity by 80% and increasing osteocalcin (OCN) mRNA expression by 2.8–fold [101]. While the loss of dopaminergic neurons in the substantia nigra is the pathological hallmark of PD motor dysfunction, the associated skeletal decline likely stems from a broader systemic dopaminergic deficiency and metabolic alterations rather than direct central neural projection alone. For instance, L–DOPA metabolism via catechol–O–methyltransferase (COMT) produces elevated levels of homocysteine (Hcy), which impairs bone collagen cross–linking and promotes osteoclast differentiation(Fig. 8) [102].Fig. 8This image illustrates the complex and bidirectional relationship between Parkinson's disease (PD) and bone health, highlighting several key ** PD–related brain changes lead to bone loss. Conversely, the skeleton influences brain health, and conditions like osteoporosis can worsen PD through inflammation.Fig. 9Schematic of Multi–dimensional Therapeutic Strategies for Promoting Bone Regeneration under the Regulation of the Central Nervous System.** Pharmacological (Pink): Tocilizumab (IL–6 pathway) for anti–inflammation; β–blockers (PERN/AKT pathway) for osteogenesis. Non–Pharmacological (Green): Acupuncture (FAK/ERK/MAPK) and PEMF (Wnt/β–catenin) promote osteogenic differentiation of BMSCs. Tissue Engineering (Blue): 3D printed scaffold releases NGF to promote osteogenesis and nerve repair.
To address the critique regarding the need for a unified framework for clinical relevance, we categorize emerging interventions into pharmacological and non–pharmacological modalities, explicitly identifying current barriers hindering bench–to–bedside translation.
Current pharmacological strategies primarily leverage the molecular overlap between neural and skeletal regulation.(1)Neuro–targeting Agents: The non–selective β–blocker "Propranolol" shows promise in promoting bone healing by downregulating osteoclast proliferation and upregulating collagen synthesis via the sympathetic nervous system [103].However, clinical application requires careful dose titration, as high doses may yield counterproductive results, attenuating anabolic effects [104].(2)Immunomodulators: Given the "neuro–immune–skeletal" triad, agents such as Tocilizumab (an IL–6 receptor antagonist) offer dual benefits. By inhibiting IL–6 signaling, they not only alleviate neurogenic inflammation (as seen in rheumatoid arthritis) but also maintain immune homeostasis conducive to bone repair [105].
Integrative medicine and engineering approaches offer targeted modulation devoid of systemic side effects.(1)Acupuncture Therapy: "Bone–touching" acupuncture techniques (e.g., at the Baihui point) align with the Traditional Chinese Medicine concept of "stimulating the skeleton to tonify the kidney and benefit the brain." These mechanical stimuli activate periosteal peptidergic nerve fibers, potentially triggering osteocalcin release to modulate CNS function, although robust clinical trial data are still accumulating (Fig. 9) [106].(2)Electromagnetic Therapy: Low–intensity pulsed electromagnetic fields (PEMF) harness the piezoelectric properties of bone. By activating the Wnt/ β–catenin pathway, PEMF promotes the osteogenic differentiation of bone marrow mesenchymal stem cells (BMSCs), serving as a non–invasive bridge connecting neural signaling and bone formation (Fig. 9) 107,108Tissue Engineering: Next–generation 3D–printed scaffolds transcend mere structural support. By incorporating nerve growth factor (NGF) or conductive materials, these scaffolds create a "neuro–osteogenic microenvironment," guiding the synchronous regeneration of nerve fibers and bone matrix (Fig. 9) [109].
Despite encouraging preclinical data, the transition to clinical practice is impeded by three critical barriers.(1)Interspecies Metabolic and Anatomical Discrepancies: The majority of BBA research relies heavily on rodent models. However, humans exhibit significant divergence in cortical–to–trabecular bone ratios, metabolic rates, and neural complexity [110]. For instance, the specific effects of leptin on cortical bone differ distinctively between mice and humans. Furthermore, the "brain drug vs. bone drug" paradox agents beneficial to the skeleton may fail to effectively traverse the blood–brain barrier (BBB), whereas neuroactive drugs may induce deleterious off–target effects on bone mass [111].(2)Patient Heterogeneity: Current interventions frequently presume a uniform mechanism of action. However, clinical reality dictates that postmenopausal women with osteoporosis and young males with traumatic brain injury possess fundamentally distinct neuroendocrine profiles. A "one–size–fits–all" approach fails to account for age–related declines in neural plasticity or sex–specific variations in serotonin regulation [112].(3)Absence of Closed–Loop Delivery Systems: Current therapies are predominantly "open–loop," administering constant drug dosages irrespective of physiological feedback. There remains a paucity of intelligent systems capable of sensing real–time neural or metabolic fluctuations (e.g., cortisol bursts or osteocalcin decline) and dynamically adjusting therapeutic release [113].
To bridge the gap between mechanistic discovery and clinical application, future research must pivot from descriptive phenomenology to functional integration.
While Piezo1 has been identified as a pivotal mechanotransducer, its downstream epigenetic regulation remains a "black box." [114,115]. Furthermore, regarding the central transmission of bone–derived signals, it is imperative to transcend Gpr158 to identify the full spectrum of central receptors responsive to osteocalcin and other bone factors, thereby elucidating their specific roles in neurodevelopment and neurodegeneration.
1Organoid Modeling: Microfluidic technology enables the construction of human–derived neuro–bone organoid systems to simulate osteoporotic microenvironments, offering a novel platform for studying neural–skeletal interactions and drug screening. Future work should focus on enhancing physiological relevance and high–throughput screening capacity [116].2AI–Assisted Applications:AILarge–scale AI models show significant potential in image–based analysis and protein–level prediction, and may be used to assess skeletal fragility, fracture risk, and disease classification. For example, AI–driven analysis of CT or MRI images can improve the accuracy of bone density and microarchitecture assessment, thereby enabling better fracture risk prediction [[117], [118], [119]].3Cross–Disciplinary Applications: Neuro–bone tissue engineering aims to address large bone defect repair and highlights the synergistic role of nerves in bone regeneration. Although various scaffold materials (e.g., polymers, inorganic compounds, and composites), seed cells (e.g., bone marrow mesenchymal stem cells, Schwann cells, and endothelial cells), and signaling factors have been applied in skeletal repair, key challenges remain in designing an optimal neuro–bone microenvironment and in elucidating the signaling pathways of different seed cells and regulatory factors [109,120].
(1)Early Warning Systems: While Substance P (SP) is a ubiquitous mediator of general pain and neurogenic inflammation, its limited skeletal specificity precludes its use as a primary screening tool [121]. However, the integration of SP with bone-specific turnover markers (e.g., CTX-1, OC) into a multi-parametric panel may offer enhanced predictive value for identifying high-risk populations [122].(2)Stratified Therapeutic Strategies:
While the classification of patients into Sympathetic-Dominant or Parasympathetic-Dominant subtypes represents an intriguing conceptual framework, validated clinical diagnostic criteria and standardized protocols for such stratification are currently lacking. Based on this theoretical model.1Sympathetic–Dominant Patients: In individuals with sympathetic overactivity, β–adrenergic blockers can effectively inhibit neurotransmitter–induced β–receptor activation, reducing heart rate and blood pressure while simultaneously improving bone metabolism [104].2Parasympathetic–Dominant Patients: A combination of vagus nerve stimulation (VNS) and low–intensity pulsed ultrasound (LIPUS) may synergistically enhance bone formation. VNS promotes skeletal repair by modulating inflammatory responses, improving circulation, and inhibiting osteoclast activity [123], while LIPUS accelerates new bone formation by enhancing osteoblast activity, such as in MC3T3–E1 cells [115].(3)Targeted Interventions for the Neuro–Bone Signaling Network:1Smart Drug Delivery Systems: Patient–derived hematopoietic stem cells can be induced in vitro to differentiate into immature dendritic cells, which are genetically engineered to produce RVG peptide–displaying targeted exosomes (RVG–Exo). Therapeutic siRNAs are then loaded into these exosomes via advanced delivery technologies. These engineered exosomes enable cross–system, precision delivery of siRNAs to modulate the function of osteoblasts, immune cells, and neurons—simultaneously suppressing inflammatory cytokine release and pathological bone resorption [[124], [125], [126]].2CRISPR/dCas9–Based Targeted Editing of the SOST Gene: The catalytically inactive dCas9 protein can be utilized as a modular epigenetic editing tool. By guiding it with SOST–specific gRNAs to the promoter or regulatory regions, the Wnt/β–catenin signaling pathway can be selectively activated. This approach relieves SOST–mediated suppression of osteogenesis, thereby enhancing bone formation and regeneration capacity [127,128].However, this approach remains highly theoretical. A formidable challenge lies in the targeted delivery of large CRISPR complexes into the dense, mineralized matrix of bone. Furthermore, given the current clinical success of FDA-approved anti-Sclerostin (e.g., Romosozumab) and anti-RANKL (e.g., Denosumab) antibodies, dCas9 editing may not serve as a primary therapy, but rather as a potential future alternative for patients who develop resistance or severe adverse effects to existing biologicals.
The exploration of the Brain–bone Axis (BBA) has fundamentally dismantled the traditional paradigm of studying the skeletal and nervous systems in isolation, unveiling a complex network orchestrated by neuroendocrine pathways and bone–derived feedback. However, advancing this field from localized repair strategies to holistic system–level remodeling requires moving beyond the mere cataloging of signaling molecules to confront three critical mechanistic specifically, decoding the precise neural circuits and "instruction sets" that mediate retrograde bone–to–brain signaling; establishing closed–loop therapeutic systems capable of dynamically sensing and responding to real–time physiological feedback to supersede current static interventions; and dissecting the profound heterogeneity introduced by age and sex. Addressing these challenges is imperative for transitioning the field from descriptive phenomenology to engineered, precision, and adaptive interventions, ultimately treating the patient as an integrated system rather than a collection of isolated organs.
Wenhui Liu: Conceptualization, Writing – review & editing. Xiekai Shen: Investigation, Visualization, Writing – original draft, Writing – review & editing. Yuxiang He: Investigation, Visualization, Writing – original draft. Zhihao Ding: Investigation, Visualization, Writing – original draft. [Other Middle Authors, e.g., Jin Zhou, Jingzhen Guo]: Investigation, Writing – review & editing. Tao Wu: Conceptualization, Supervision, Funding acquisition, Writing – review & editing. All authors have read and agreed to the published version of the manuscript.
During the preparation of this work the author(s) used Gemini in order to improve the readability and language quality of the manuscript. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article.
10.13039/501100012166National Key Research and Development Program of China (2023YFC2411304), 10.13039/100014717National Natural Science Foundation of China (82272486, and 82572771), 10.13039/100020740Natural Science Foundation of Hunan Province (2023JJ30749), and 10.13039/501100002858China Postdoctoral Science Foundation (2023M7427373 and 2025M781698). Pudong New Area Postdoctoral Innovation Practice Base Program.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
All authors have read and approved the final manuscript. The authors alone are responsible for the content and writing of the paper.