Bone aging: a paradigm of multiscale degeneration and targeted rejuvenation Authors: Kangpeng Li, Chenyuan Gao, Ti Zhang, Cheng Wang, Xiao Geng, Xinguang Wang, Hua Tian Published: 2026 ABSTRACT AbstractBone aging is not merely a process of bone mass loss but rather a multi-level, networked degenerative phenomenon spanning molecular, cellular, tissue, and system levels. With the accelerating pace of global population aging, bone degenerative diseases have emerged as a major public health challenge. Consequently, there is an urgent need to elucidate the underlying mechanisms from an integrated perspective and to develop innovative intervention strategies. This review systematically synthesizes the comprehensive landscape of bone aging. At the macroscopic level, bone aging is characterized by declines in bone mass, microstructure, and biomechanical properties, ultimately leading to increased fragility. At the microscopic level, its pathogenesis involves dysfunction across multiple layers. Molecular mechanisms include genetic and epigenetic alterations, oxidative stress, the senescence-associated secretory phenotype, and dysregulation of key signaling pathways. Cellular level changes involve telomere attrition, mitochondrial and endoplasmic reticulum dysfunction, impaired proteostasis, altered nutrient sensing, and cellular senescence. At the microenvironmental and systemic levels, contributing factors include chronic inflammation, hormonal changes, and extracellular matrix remodeling. Correspondingly, intervention strategies include non-pharmacological approaches, such as dietary optimization, physical activity, and avoidance of alcohol and tobacco, as well as pharmacological therapies, including antiresorptive agents, anabolic bone-forming drugs, and senolytics. Emerging therapeutic avenues include regenerative medicine, gene therapy, and tissue engineering. Future research should focus on constructing a multi-omics-integrated bone aging atlas, developing bone organoids and computational models, establishing multimodal risk assessment systems, advancing precision therapies that target aging-related pathways, and promoting prevention-centered public health strategies. This review aims to provide a systematic framework for understanding the complex pathophysiology of bone aging and to provide key insights and directions for future interdisciplinary collaborations, translational research, and clinical interventions.Subject terms: Metabolic disorders, Bone Introduction The skeletal system is not merely an inert structural scaffold but a living organ characterized by a highly active metabolism and continuous dynamic remodeling. Central to this process is bone remodeling, a tightly regulated and precisely coupled sequence of bone resorption mediated by osteoclasts and bone formation driven by osteoblasts. This process is essential not only for maintaining mineral homeostasis but also for repairing microdamage, adapting to changes in the mechanical environment, and preserving skeletal integrity and biomechanical function. Disruption of the balance between bone resorption and bone formation directly compromises bone quality and strength (Fig. 1). Bone aging: a multiscale degenerative cascade and targeted rejuvenation. Taking a femur in the process of repair as a life metaphor, this figure summarizes the multi-level degenerative cascade of bone aging and the targeted regeneration strategy. The composition uses a contrast narrative between the left and right: the left side depicts the multi-scale destructive forces from macroscopic bone loss, microscopic cellular aging to systems-driven. On the right side, the multi-level recovery strategy from basic nutrition and exercise (lifestyle factors appear here in a conceptual/foundational role, not as mechanistic drivers or detailed intervention targets, see Figs. 5 and 6 for those distinct perspectives), precision medicine intervention to cutting-edge regenerative therapies are systematically displayed. The focal repair zone of the central skeleton symbolizes the central role of targeted intervention, highlighting the dynamic balance of damage and repair and the transformation potential. The illustration is conceptual and metaphorical, not a literal anatomical rendering; vascular elements are stylized to represent hierarchical therapeutic strategies rather than physiological vessel caliber With advancing age, the regulatory capacity of this finely tuned system inevitably declines, resulting in progressive and systemic skeletal deterioration, a process referred to as bone aging. Bone aging is characterized not only by a reduction in bone mass but also by spatiotemporal dysregulation of bone remodeling and destructive changes in bone microstructure. The most immediate clinical consequence is a markedly increased risk of osteoporosis. According to the International Osteoporosis Foundation, an osteoporotic fracture occurs every three seconds worldwide.1 In the context of accelerated global population aging, the rising incidence of disability and mortality associated with fragility fractures, particularly of the hip and vertebrae, along with the substantial medical and social care burden they impose, represents a major public health challenge. Beyond its clinical implications, the study of bone aging also offers a valuable model for understanding coordinated aging processes across multiple organ systems. Against this background, this review seeks to move beyond the traditional dualistic framework of bone resorption-formation imbalance and to systematically delineate bone aging as a multi-level, networked degenerative process. We first summarize the phenotypic manifestations of bone aging, ranging from declines in macroscopic mechanical properties to deterioration of microstructural integrity. We then examine the core mechanisms underlying the spatiotemporal dysregulation of bone remodeling units at the molecular and cellular levels. Finally, we integrate recent advances in therapeutic and preventive strategies targeting bone aging. By synthesizing current concepts and emerging evidence, this review aims to provide an integrated framework for understanding the pathophysiology of bone aging and to offer a solid theoretical foundation and clear research directions for the development of innovative strategies to prevent and treat age-related skeletal disorders and restore bone homeostasis. Macroscopic characteristics: progressive deterioration of bone mass, structure, and biomechanics Bone mass Bone mass refers to the total mineral content within bone tissue. Dynamic changes in bone mass throughout life are governed by bone remodeling, a tightly coordinated process involving osteoblast-mediated bone formation and osteoclast-mediated bone resorption. This process underlies the characteristic trajectory of bone mass across the lifespan—growth, peak, maintenance, and loss (Fig. 2). Macroscopic deterioration in bone aging. a Age-related trends in bone mineral density in men and women, particularly the rapid bone loss in women after menopause, are supplemented by a simulated comparison of trabecular bone mass. b The degeneration process of cancellous bone from a young thick plate-like network to an old sparse broken rod-like structure through a sequence diagram, and the change of cortical bone with a thinner cross-section and increased porosity. c The upper stress-strain curve illustrates decreased elastic modulus and increased brittleness in aged bone compared with young bone. The inferior diagram depicts age-related shifts in bone material composition: young bone maintains a balanced organic/inorganic matrix, whereas elderly bone exhibits increased mineralization and reduced collagen quality, leading to increased brittleness and fracture susceptibility Bone mass increases rapidly from birth, with peak bone mass achieved between 25 and 30 years of age, showing significant sex differences: males reach peak later (around 30 years) and at a higher level (lumbar spine BMD ~ 1.2 g/cm²) than females, who peak 2–3 years earlier at ~1.0 g/cm². After age 40, both sexes lose bone, but women experience an accelerated phase during menopause, with annual bone loss reaching 3%–5%—approximately twice that of age-matched men.2 Cumulative trabecular bone loss from 20 to 80 years is ~40% in women versus ~25% in men.3 Earlier studies indicate that in women, vertebral bone loss begins in early adulthood and progresses linearly, whereas appendicular atrophy typically starts around age 50, accelerates between 51 and 65 years, and slows thereafter. Lifetime vertebral loss reaches ~47%, compared with ~30% in the mid-radius and ~39% in the distal radius. In men, age-related bone loss at both axial and appendicular sites is comparatively modest or statistically insignificant.4 Substantial interindividual variability in bone aging rates suggests a strong genetic component.5 Despite the widespread acceptance of these general trends, several areas remain debated. For instance, the relative contributions of estrogen decline versus other age-related factors to the menopausal acceleration of bone loss are not fully resolved, and cross-sectional estimates of cumulative loss may obscure substantial individual variability in longitudinal trajectories. Furthermore, whether age-related bone loss in men is truly modest or simply underdetected due to less frequent study of male cohorts at advanced ages remains contested. From a therapeutic standpoint, these uncertainties complicate decisions about when to initiate pharmacologic intervention in perimenopausal women versus older men, as most clinical trial evidence derives from postmenopausal women, with limited guidance for sex-specific or frailty-stratified approaches. Identifying these knowledge gaps is essential for prioritizing future research into personalized fracture prevention. Bone microstructure Bone microstructure, a central determinant of bone quality, refers to the microscopic organization and structural characteristics of bone tissue.6 At the time peak bone mass is achieved, bone microarchitecture is relatively stable. With advancing age, however, progressive deterioration occurs, particularly within the trabecular bone. Age-related increases in relative or absolute bone resorption lead to a reduction in trabecular number, thickness, and connectivity. Trabeculae become thinned, perforated, and ultimately fragmented, transforming the original interconnected plate-like network structure into isolated rod-like structures. This architectural disruption markedly compromises mechanical integrity and load-bearing capacity, particularly in cancellous bone-rich regions such as the vertebral bodies and hip.7 Age-related changes in cortical bone are equally consequential. Cortical thickness gradually decreases while porosity increases, driven by enhanced endocortical resorption and insufficient periosteal bone formation, along with enlargement of Haversian and Volkmann canals and increased intramedullary fat accumulation.8–10 This cortical porosification reduces bone density and substantially weakens resistance to bending and torsional forces.6 Concurrently, aging dysregulates bone remodeling and microdamage repair: resorption cavities are inadequately refilled, leading to net bone loss per remodeling cycle, while declining repair capacity permits microdamage accumulation. Within a compromised microarchitectural framework, these microdamages coalesce into microcracks, ultimately precipitating fragility fractures.11 While the above microarchitectural deterioration model is widely accepted, several important uncertainties remain. First, causal relationships between specific microstructural parameters and fracture risk are often inferred from cross-sectional histomorphometry or high-resolution imaging, but longitudinal evidence tracking the same individuals over time remains sparse. Second, although dysregulated remodeling and microdamage accumulation are implicated in age-related fragility, the threshold at which microdamage becomes clinically relevant—and whether it can be modified independently of bone mineral density—remains poorly defined. Critically, the majority of microstructural data derive from postmenopausal women, limiting extrapolation to men, younger populations, or those with chronic diseases, where bone microarchitecture may follow distinct, non-classical aging trajectories. Bone biomechanics The biomechanical deterioration of aged bone extends beyond simple declines in BMD, involving complex alterations in material properties and fracture behavior. While it is well established that aging reduces bone strength, stiffness, and energy absorption capacity, recent advances have shifted focus toward understanding the multiscale origins of brittleness. Studies examining age-related changes in vertebral trabecular bone biomechanics have shown that between 20 and 80 years of age, BMD decreased by approximately 48%-50%, whereas vertical stress resistance, stiffness, and energy absorption capacity decreased by 75%-80%. Biomechanical performance in both vertical and horizontal orientations exhibits a strong positive correlation with BMD.12 In cortical bone, bending strength decreases by approximately 15%-20% between 35 and 70 years of age, while compressive strength decreases by nearly 50%. In parallel, fracture energy absorption is markedly reduced reflecting increased bone brittleness.13 These changes are associated with alterations in bone composition. Under normal conditions, the organic matrix and inorganic mineral components are well-balanced; however, aging shifts the balance toward higher mineral content and reduced collagen quality, thereby increasing bone stiffness at the expense of toughness.14 Two emerging concepts are worthy of attention. First, the spatiotemporal heterogeneity of tissue mineralization—rather than mean mineral content—has been identified as a critical determinant of fracture resistance. Studies reveal that aged bone exhibits increased mineral heterogeneity at the lamellar level, creating local stress concentrators that facilitate microcrack propagation.15–17 Second, the quality and cross-linking profile of type I collagen are now recognized as equally important as mineral content. Age-related accumulation of advanced glycation end products (AGEs) induces non-enzymatic cross-links that restrict collagen fibril sliding, a key energy-dissipation mechanism.18,19 A critical unresolved question is whether current anti-osteoporotic therapies—which primarily target bone mass—meaningfully restore the organic matrix quality and toughness of aged bone. Addressing this gap will require integrating biomechanical testing with longitudinal clinical cohorts and advanced imaging biomarkers. Microscopic features: from molecules and cells to the microenvironment Molecular level Genetics The genetic mechanisms underlying bone aging are often portrayed as a complex, multi-dimensional network integrating aging biology with skeletal regulatory processes (Fig. 3). However, direct genetic studies focused explicitly on bone aging remain limited, and most current knowledge is extrapolated from transcriptomic analyses of mixed aging models.20–22 While features such as DNA damage accumulation, metabolic reprogramming, and senescence-associated secretory phenotype (SASP) expression are consistently described,23,24 whether these events drive bone aging or merely accompany it remains unclear. Functional evidence is largely confined to mouse models carrying deficiencies in DNA repair genes (e.g., XPD, ATM, ERCC1-XPF), which produce premature aging and osteoporosis-like phenotypes,25–28 but the translational relevance of these severe genetic disruptions to human physiological bone aging is questionable. Conversely, enhanced DNA repair via Ku70 or DNA-PKCs has been shown to protect bone mass in animal models, yet human validation is lacking.29 Molecular network of bone aging. The radial network Figure illustrates the core molecular mechanism network driving bone aging. The central nucleus highlights genomic instability with epigenetic reprogramming. Four interconnected pathways were identified: 1) oxidative stress leads to biomacromolecule damage and inhibits osteogenic Wnt/β-catenin signaling; 2) aging-associated secretory phenotypes triggered by the DNA damage response disrupt the bone microenvironment by releasing proinflammatory cytokines; 3) imbalance of key signaling pathways, including decreased Wnt activity, increased RANKL/OPG ratio, and inhibited Hippo/YAP pathway; 4) dysregulated nutrition sensing Genome-wide association studies have repeatedly implicated the Wnt signaling pathway (LRP4/5, SOST, WNT16/4) in BMD and fracture risk,30–32 but these associations explain only a small fraction of heritable variation, and the causal variants at most loci remain unidentified. Similarly, transcriptomic profiling has nominated genes such as ACSL3, DKK1, and SOST, while the SenMayo gene set has confirmed upregulation of CDKN1A/p21Cip1 and SASP factors in aged human bone at single-cell resolution.33 These descriptive inventories, however, do not establish causality, nor do they clarify whether the same senescence markers are consistently expressed across different skeletal sites, cell types, or disease states (e.g., diabetes, chronic kidney disease). Furthermore, proposed links between bone aging and lipid metabolism remain correlational,34 with no interventional evidence that modifying lipid pathways alters skeletal aging. A critical unmet need is therefore to move from transcriptomic cataloging toward functional validation—ideally in human tissue—and to determine which of these genetic and molecular features are amenable to therapeutic targeting. Epigenetics Epigenetic regulation—including DNA methylation, histone modifications, non-coding RNAs, and chromatin remodeling—is widely invoked as a central mechanism in bone aging.35 DNA methylation studies have reported associations between accelerated epigenetic aging and fracture risk in twin cohorts,36 and epigenome-wide association studies have identified age-related differentially methylated loci such as NOP56, TNXB, PDE8A, and AKAP12.37 Yet these associations are largely derived from bulk tissue analyses, which cannot distinguish whether methylation changes drive bone aging or merely reflect it. Tissue-specific comparisons between bone and blood have revealed different methylation patterns at key metabolic regulators (MEPE, SOST, WIF1, DKK1),38 highlighting a critical but unresolved question: to what extent can blood-based epigenetic markers serve as valid proxies for skeletal aging? The proposed development of tissue-specific epigenetic clocks39 is a reasonable response to this problem, but such clocks remain technically challenging and not yet validated in independent cohorts. Non-coding RNA studies, predominantly in ovariectomized mouse models, have implicated miR-29a-3p, miR-29c-3p, and miR-874-3p in suppressing osteogenic differentiation.40,41 Whether these findings translate to human physiological bone aging—or to male skeletal aging—is unclear, as most rodent studies use acute hormone withdrawal rather than gradual age-related decline. Histone modifiers such as KDM3A, KDM4C, and BRD9 have been linked to mesenchymal stem cell senescence and osteogenesis,42,43 but the majority of evidence comes from in vitro or overexpression models, with little understanding of how these modifications integrate with other epigenetic layers or vary across skeletal sites. Finally, SIRT3 S-sulfhydration has been shown to counteract BMSC senescence in animal models,44 but whether this represents a tractable therapeutic target in aged humans remains speculative. Collectively, the epigenetic literature on bone aging provides a rich descriptive foundation, but causal validation, tissue-specific benchmarking, and translational evidence lag considerably behind. Oxidative stress Bone aging is closely associated with oxidative stress, with reactive oxygen species (ROS) serving as a core driver. Excessive ROS directly damage cellular components and activate DNA damage response pathways, leading to cell-cycle arrest and senescence. Oxidative stress-induced senescence of BMSCs has been identified as a significant factor in postmenopausal osteoporosis.45 BMSCs isolated from OVX mice exhibit elevated ROS levels and premature senescence, both of which can be rescued by the phytoestrogen genistein.45 ROS exerts dual effects within the skeletal system. At physiological levels, osteoclast-derived ROS are essential for bone remodeling and promote osteoclast differentiation via activation of RANKL signaling.46 However, excessive ROS disrupts skeletal homeostasis by inhibiting osteogenic and mineralization in osteoblasts47 while simultaneously enhancing osteoclastogenesis through mechanisms such as increasing the RANKL/OPG ratio,48 ultimately leading to net bone loss. Similar ROS-driven mechanisms have been invoked in osteoarthritis progression,49 but whether therapeutic ROS reduction would benefit both bone and cartilage or exert site-specific effects is unknown. Therapeutic strategies targeting oxidative stress have demonstrated promising effects. Pyrroloquinoline quinone alleviates age-related osteoporosis by activating the Nrf2 antioxidant pathway, reducing Rankl expression in osteoblasts, and inhibiting osteocyte senescence.50 Ganoderic acid D delays BMSC senescence by enhancing antioxidant enzyme activity and reducing advanced glycation end product accumulation.51 Modulation of deacetylases such as SIRT1 and SIRT6 has also shown protective effects against chondrocyte senescence.52,53 However, virtually all of these findings are at the pre-clinical stage, with no large-scale human trials demonstrating fracture reduction or long-term safety. Senescence-associated secretory phenotype The SASP comprises a complex, diverse array of bioactive factors secreted by senescent cells and functions as a potent signaling system that reshapes the tissue microenvironment. SASP components include pro-inflammatory cytokines, chemokines, growth factors, and proteases. Beyond the general features, two additional dimensions of SASP biology warrant particular attention in the context of bone aging: its tissue-specific signature and age-dependent dynamics. In the skeletal system, the SASP profile differs markedly from that of other aging tissues. For instance, while senescent fibroblasts in skin and lung predominantly secrete IL-6 and IL-8,54,55 senescent osteocytes and bone marrow macrophages show heightened expression of osteoclastogenic factors such as RANKL, M-CSF, and IL-11,56 alongside canonical inflammatory cytokines. This pro-resorptive bias of the bone SASP reflects the unique evolutionary pressure on the skeleton as a load-bearing, metabolically active organ. SASP production is also cell-type specific. Senescent osteocytes and bone marrow macrophage-lineage cells represent major sources of SASP, whereas osteoblasts and their progenitors contribute relatively less.57,58 This heterogeneity complicates SASP identification at the single-cell level, highlighting the utility of standardized tools such as SenMayo aids for precise characterization.33 Equally important is the biphasic nature of SASP across the lifespan.56 In youth, low-level, transient SASP serves homeostatic functions—recruiting immune cells for tissue repair, coordinating remodeling after microdamage, and maintaining hematopoietic stem cell niche dynamics. With advancing age, however, the accumulation of senescent cells and persistent DNA damage transforms SASP into a chronic, self-amplifying inflammatory network.59 This age-related SASP escalation in bone directly correlates with cortical porosity, trabecular fragmentation, and fracture risk, independent of bone mineral density. Understanding these tissue- and age-specific SASP features is not merely academic; it carries direct therapeutic implications. Suppression of SASP signaling using Janus kinase inhibitors (JAKi) effectively attenuates age-related bone loss.60 Similarly, the tyrosine kinase inhibitor cabozantinib prevents estrogen deficiency-induced bone loss in mice by inhibiting SASP secretion from senescent osteoblasts and their progenitors.61 A senolytic or senomorphic strategy effective in clearing senescent cells or suppressing SASP in bone may not translate to the brain or kidney, and vice versa. Therefore, future studies must systematically map the SASP atlas across organs and ages to enable precision targeting. Key signaling pathways Bone aging is governed by a complex and interconnected network of signaling pathways that collectively regulate skeletal homeostasis and degeneration. The Wnt/β-catenin pathway is widely recognized as a central regulator of skeletal development and maintenance, essential for osteoblast lineage commitment. While overall Wnt signaling appears downregulated in the bones of elderly women,62 certain pathway members exert protective effects: osteoblast-derived WNT16 inhibits osteoclast RANK signaling, and WNT4 promotes bone formation while suppressing resorption and inflammation.63,64 These findings have prompted therapeutic targeting of Wnt inhibitors, most notably anti-sclerostin antibodies, which have shown positive clinical results.65 However, key uncertainties remain regarding the relative contribution of canonical versus non-canonical signaling in human bone aging, the cardiovascular safety of pathway activation, and whether interactions with oxidative stress or age-related decline of Gαi1/3 are causally significant in human skeletal aging.66,67 The RANKL/RANK/OPG axis serves as the principal regulator of osteoclast differentiation and bone resorption. Aging is associated with a two- to threefold increase in the RANKL/OPG ratio, significantly enhancing osteoclast activation, particularly under estrogen-deficient conditions. RANKL binding to RANK activates NF-κB and MAPK signaling via TRAF6, inducing NFATc1 expression and driving osteoclastogenesis.68 Therefore, this axis represents a key therapeutic target for intervening in bone resorption. The Hippo signaling pathway, a key sensor of mechanical forces and cellular stress, regulates aging in a cell type-dependent manner. Activity of its downstream effectors, YAP and TAZ typically declines with age. YAP is essential for maintaining osteogenic potential in mesenchymal stem cells and suppressing adipogenic differentiation.69 Moreover, YAP/TAZ exert senomorphic effects by suppressing the cGAS-STING pathway and display extensive crosstalk with Wnt and TGF-β signaling.70,71 Fibroblast Growth Factor (FGF) signaling exhibits diverse and context-dependent effects on skeletal aging. FGF2 promotes bone growth and repair via activation of Wnt/β-catenin and BMP2 pathways.72 FGF19 exerts a protective effect on bone density by regulating bile acid and lipid metabolism.73,74 Conversely, excessive FGF23 disrupts phosphate homeostasis and reduces bone density.75,76 The relationship between FGF21 and bone density remains controversial.77,78 Cell level Telomeres Telomeres are protective nucleoprotein structures whose progressive shortening is a core hallmark of cellular senescence. Within the skeletal system, dysregulation of telomere homeostasis is considered a significant cellular and molecular mechanism driving bone aging, although existing evidence remains heterogeneous (Fig. 4). Cellular senescence in the skeletal niche. This figure focuses on the aging status of key cell populations and their subcellular lesions in the bone microenvironment. In the context of an aging bone marrow with increased steatosis and decreased vascularity, four types of cells are highlighted: mesenchymal stem cells/osteoblasts, osteoclasts, osteocytes, and immune cells. The close-up of three magnmagnies further reveals common molecular lesions such as telomere-mitochondrial axis disorders, autophagic flow disorders, and ER-mitochondrial calcium homeostasis imbalance, which jointly drive decreased bone formation and increased bone resorption Several observational studies have reported positive associations between leukocyte telomere length and BMD or reduced fracture risk,79,80 but an approximately equal number of well-powered studies—including a regression analysis of 2 750 community-dwelling older adults,81 a cohort of 1 867 older Chinese adults,82 and UK Biobank data83—have found no association with BMD, osteoporosis, fracture risk, or other skeletal parameters. This inconsistency cannot be dismissed as mere sampling variation; rather, it suggests that the telomere–bone relationship, if real, is likely modest in effect size and heavily modulated by population characteristics, skeletal site, and unmeasured confounders such as chronic infection or inflammation.84 Mendelian randomization analyses have further failed to support a causal link between telomere length and sarcopenia,85 raising the possibility that observed associations in observational studies reflect confounding rather than causality. Experimental models provide mechanistic evidence that telomere dysfunction can drive bone aging under certain conditions—radiation-induced models show telomere dysfunction-induced foci (TIFs) in osteocytes,34 and genetically engineered mice with disruptions in telomere maintenance genes (Wrn, Terc) exhibit impaired osteoblast differentiation and osteoporosis.86,87 However, the translational relevance of these models is questionable: radiation-induced damage and complete gene knockouts represent extreme perturbations that differ substantially from the gradual, physiological telomere attrition that occurs in human aging. Whether the TIFs and senescence markers observed in these models appear in normally aging human bone—and if so, at what frequency—remains largely unknown. Thus, while telomere dysfunction can cause bone pathology in experimental settings, its contribution to typical human skeletal aging remains unproven, and the field would benefit from prospective studies that directly measure telomere dynamics in bone tissue rather than relying on leukocyte proxies or extreme genetic models. Mitochondria Mitochondria function as cellular powerhouses and play a dual role in maintaining skeletal homeostasis. They generate energy via oxidative phosphorylation while also producing ROS as crucial signaling molecules. To fully elucidate how mitochondrial dysfunction drives bone aging, three interconnected mechanistic pathways should be distinguished, along with the unresolved controversies. First, the ROS-DDR-p53/p21 axis represents the most classical pathway. Excessive mitochondrial ROS directly induce DNA double-strand breaks, activating the DNA damage response and downstream p53/p21-mediated cell cycle arrest.88 However, emerging evidence challenges the unidirectional model: moderate ROS elevation can paradoxically promote osteoblast differentiation via Nrf2-mediated antioxidant gene activation, suggesting a hormetic dose-response relationship that shifts from beneficial to deleterious with age.89 Moreover, whether ROS acts primarily as an initiator or a perpetuator of senescence remains debated, as senescent cells themselves produce ROS, creating a self-reinforcing loop. Second, the mitophagy-mitochondrial dynamics axis garnered increasing attention. While impaired mitophagy clearly accelerates bone aging—as evidenced by PINK1 deficiency leading to excessive mitochondrial fusion and senescent cell accumulation—the role of mitochondrial fragmentation versus hyperfusion is contested.90 Some studies report that Drp1-mediated fragmentation promotes senescence,91 whereas others demonstrate that elongated, hyperfused mitochondria are characteristic of aged osteoblasts and protect against apoptosis.92 This discrepancy may reflect cell-type specificity or the duration of stress exposure, with transient fragmentation facilitating mitophagy but chronic fragmentation overwhelming clearance capacity. Third, the mitochondrial-ER crosstalk via mPTP and calcium signaling has emerged as a critical but understudied mechanism. The recent discovery that STING promotes mPTP opening through the HK2-VDAC1 axis provides a molecular link between innate immunity and mitochondrial dysfunction in bone aging.93 Nevertheless, divergent viewpoints exist regarding the temporal order of events: some argue that mPTP opening is a late-stage consequence of severe mitochondrial damage, while others posit it as an early trigger of senescence. Endoplasmic reticulum The endoplasmic reticulum (ER) is a crucial organelle responsible for protein synthesis, folding, modification, and calcium storage. Disruption of ER homeostasis, commonly referred to as ER stress, is a significant mechanism underlying cellular senescence and skeletal degeneration. Functional defects in ER function can directly lead to bone disease; for instance, mutations in TMEM38B disrupt ER calcium homeostasis, impair type I collagen folding and secretion, and result in osteogenesis imperfecta.94–96 In bone aging, ER stress exerts a biphasic effect on skeletal homeostasis. Moderate or transient ER stress may promote osteogenesis, whereas prolonged or excessive stress shifts cellular responses toward catabolism.97 For instance, mild ER stress induced by curcumin enhances osteogenic differentiation of mesenchymal stem cells via activation of the ATF6 pathway,98 whereas intracellular accumulation of AGEs induces severe ER stress, leading to osteoblast apoptosis.99 Multiple pathological stimuli associated with aging can trigger ER stress in bone. Glucocorticoids inhibit bone marrow stem cell proliferation and induce osteoblast apoptosis by activating the unfolded protein response, particularly the PERK-NRF2 pathway, which is a central mechanism underlying glucocorticoid-induced osteoporosis. Pharmacological inhibition of ER stress with 4-phenylbutyric acid mitigates this apoptotic effect. Long-term use of proton pump inhibitors and the immunosuppressant cyclosporine A has also been linked to increased osteoporosis risk through ER stress activation and ROS accumulation.100,101 Metabolic stressors such as the saturated fatty acid palmitate induce lipid droplet accumulation and senescence in nucleus pulposus cells via ER stress pathways.102 Furthermore, ER stress impairs myogenic responses to anabolic stimuli and alters bone cell mechanosensitivity, thereby contributing to both sarcopenia and osteopenia.103 In intervertebral disc degeneration, deficiency of site-1 protease disrupts protein transport, leading to ER dilation, abnormal ER-mitochondrial calcium flux, and ultimately mitochondrial dysfunction and a senescent phenotype.104 ER stress further drives skeletal catabolism and inflammation through activation of downstream signaling pathways. Studies indicate it promotes inflammation, apoptosis, and autophagy via the p38 MAPK signaling pathway, leading to bone loss.105 In BMSCs, ER stress enhances inflammation-induced bone loss by activating the ROCK1 signaling.106 Additionally, specific non-coding RNAs are involved in ER stress regulation; for example, miR-1260b targets ATF6β to modulate ER stress and attenuate periodontal bone loss.107 Proteostasis Protein homeostasis, or proteostasis, is fundamental to the maintenance of bone cell function. In the context of bone aging, failure of proteostasis directly impairs three critical processes: the secretory capacity of osteoblasts for type I collagen and other matrix proteins; the survival and mechanosensing function of long-lived osteocytes; and the tightly regulated resorptive activity of osteoclasts. Consequently, age-related decline in proteostasis is not merely a general cellular phenomenon but a specific driver of skeletal degeneration. For the chaperone system (HSP70/HSP90), in aged bone, the imbalance between HSP70 and HSP90 tips the osteoblast-osteoclast coupling toward resorption, as HSP70 can promote osteoclastogenesis by activating the RANKL pathway via TLR signaling108 by cooperating with macrophage colony-stimulating factor (M-CSF) to induce transdifferentiation of dendritic cells into osteoclasts.109 In contrast, HSP90 negatively regulates RANKL expression and osteoclast formation.110 For the ubiquitin-proteasome system, age-related decline in proteasome function in osteoclast precursors leads to the accumulation of pro-resorptive signaling molecules, exacerbating bone loss.111 Accordingly, proteasome inhibitors such as bortezomib and carfilzomib can reduce bone resorption by disrupting key signaling pathways involving p62 and TRAF6, offering potential therapeutic strategies for osteoporosis.112 Additionally, gasdermin D (GSDMD) has been identified as a lysosomal checkpoint protein with anti-osteoclast activity that protects against bone loss.113 For the autophagy-lysosome system, in aged osteocytes, autophagic failure is particularly detrimental, leading to the accumulation of damaged organelles, oxidative stress, and senescence. AMPK promotes osteoblast differentiation and mineralization by inducing autophagy114 and protects osteoblasts from apoptosis by clearing damaged organelles and limiting ROS accumulation.115 Age-related decline in autophagic capacity, particularly in long-lived osteocytes, leads to intracellular waste accumulation and a senescent phenotype, thereby disrupting bone metabolic balance. Deregulated nutrient sensing Deregulated nutrient sensing during aging reflects impaired cellular capacity to detect and respond appropriately to nutritional signals such as glucose and amino acids. This dysfunction represents a core mechanism linking systemic metabolic abnormalities to skeletal degeneration and is driven primarily by an imbalance among the insulin/IGF-1, mTORC1, AMPK, and sirtuin signaling pathways, which collectively reshape the metabolic microenvironment of the aging skeleton.116 The insulin/IGF-1 signaling pathway is a potent anabolic signal in bone. As an important target organ for insulin, bone exhibits features of insulin resistance during aging and in metabolic bone disorders.117 Activation of this pathway promotes osteoblast proliferation, differentiation, and survival via the PI3K/AKT signaling pathway, whereas insulin resistance blunts osteoblast anabolic responses. Impaired osteoblast function also disrupts the production and activation of osteocalcin, a hormone that regulates systemic insulin sensitivity. This dysregulation establishes a vicious cycle within the bone-pancreas axis, exacerbating both local skeletal and systemic metabolic disturbances.118 The mTORC1 pathway integrates growth factor, energy, and amino acid signals and undergoes significant functional changes with aging.119 While physiological mTORC1 activity supports bone formation, chronic overactivation during aging has deleterious consequences. It induces premature senescence in osteoprogenitor cells, suppresses their differentiation, and inhibits autophagy, leading to the accumulation of toxic intracellular substances, impaired osteoblast function, and increased apoptosis.120 Furthermore, mTORC1 promotes osteoclast differentiation and survival, and its excessive activation accelerates bone resorption. Genetic disruption of negative mTORC1 regulators in osteoclasts results in severe age-related bone loss.121 AMPK functions as a central cellular energy sensor and generally antagonizes mTORC1 signaling. Its activity declines with age,122 leading to multiple adverse effects. Reduced AMPK signaling shifts osteoblast metabolism from efficient oxidative phosphorylation toward glycolysis, insufficiently supporting the energy demands of bone formation. As a crucial initiator of autophagy, AMPK deficiency impairs cellular clearance functions, accelerating organelle damage and functional deterioration. Furthermore, diminished AMPK activity weakens antioxidant defenses mediated by FOXO, increasing susceptibility to oxidative stress and accelerating senescence.123 The sirtuin family of NAD⁺-dependent deacetylases, particularly SIRT1, SIRT3, and SIRT6,124 plays a crucial role in bone aging. Age-related declines in NAD⁺ levels reduce sirtuin activity. Loss of SIRT1 impairs osteogenic differentiation and favors adipogenesis; deficiency of the mitochondrial SIRT3 leads to mitochondrial dysfunction and excessive ROS production in osteoblasts; and SIRT6 loss disrupts osteoblast metabolism and genomic stability, producing a premature aging-like skeletal phenotype.125 Cellular senescence Cellular senescence is a state of irreversible cell cycle arrest accompanied by specific phenotypic alterations.126–128 It represents a fundamental mechanism driving the degeneration and functional decline of bone tissue.129 Multiple cell types within bone accumulate senescent cells with advancing age. Senescence of the osteoblastic lineage. Senescence of osteoblasts and their progenitors, including osteoprogenitor cells and MSCs, directly contributes to reduced bone formation. Due to replicative senescence and telomere attrition, the limited proliferative capacity of MSCs and osteoprogenitors leads to progressive telomere shortening, ultimately triggering the DDR-p53/p21CIP1-p16INK4a/Rb pathways, inducing cell cycle arrest.130 Senescent osteoblasts exhibit significantly reduced proliferation, differentiation capacity, and mineralized nodule formation.131 Senescence of osteoclasts. Osteoclasts originate from hematopoietic stem cells, and research on their senescence is relatively limited. Nevertheless, evidence suggests that the senescent bone marrow microenvironment affects the differentiation and function of osteoclast precursors. Senescence-associated inflammatory factors may abnormally stimulate osteoclastogenesis, and senescent osteoclasts themselves may exhibit dysregulated resorptive activity.132 Senescence of osteocytes. Osteocytes are the longest-lived cells in the bone matrix and serve as sensors for mechanical stress and regulators of bone remodeling. Aging osteocytes undergo network degradation and increased apoptosis but also acquire a senescent phenotype. Altered SASP profiles in senescent osteocytes may transmit aberrant pro-resorptive signals to osteoclast precursors while inhibiting osteoblast activity, thereby disrupting bone homeostasis.5 Senescence of BMSCs. BMSCs are common precursors to osteoblasts and adipocytes. Senescent BMSCs exhibit reduced osteogenic potential and a bias toward adipogenic differentiation, leading to increased marrow adipose tissue (marrow adiposity). This shift further compromises the hematopoietic and osteogenic microenvironment.133 Vascular senescence in bone. Angiogenesis is tightly coupled to osteogenesis. With aging, bone endothelial cells (BECs) acquire a senescent phenotype, characterized by telomere shortening and increased p16INK4a expression.134 Senescent BECs exhibit impaired capacity to form H-type vessels, which are specialized capillaries that couple angiogenesis with osteogenesis by secreting factors such as SLIT3 and RANKL antagonists.135 Age-related loss of H-type vessels directly reduces perivascular osteoprogenitor cells and impairs bone formation.135 Neural senescence in bone. In the peripheral nervous system, sympathetic nerve density in bone decreases with age, while the remaining nerves exhibit signs of dysfunction, including reduced norepinephrine reuptake capacity and altered neuropeptide Y expression.136 This dysregulation leads to an unbalanced sympathetic tone, promoting osteoclast activity via β2-adrenergic receptors on osteoblasts and osteoclast precursors. Sensory nerves, which release calcitonin gene-related peptide (CGRP) and substance P with pro-osteogenic and vasodilatory effects, also show age-dependent degeneration.137 Microenvironment level Chronic inflammation Chronic low-grade inflammation is a fundamental and persistent feature of bone aging (Fig. 5). The senescence of hematopoietic stem cells (HSCs) constitutes a central component of immunosenescence, and the aging bone marrow microenvironment plays a pivotal role in driving this process. With advancing age, red bone marrow is progressively replaced by adipose tissue, resulting in diminished nutritional and regulatory support. Concurrently, the accumulation of senescent cells and pro-inflammatory mediators collectively impairs HSC function.138,139 Under inflammatory stress, HSC metabolism shifts from glycolysis toward oxidative respiration. The accumulated ROS can trigger DNA damage, senescence, or apoptosis.140 The CXCR4/CXCL12 signaling axis has been shown to protect HSCs by attenuating mitochondrial oxidative stress.108 Notably, the transcription factor Bcl11a exhibits a dual role: while exerting cell-intrinsic protective effects, it simultaneously promotes bone marrow inflammation, thereby exacerbating HSC dysfunction.141 Inflammation also induces a myeloid bias in HSC differentiation, favoring myeloid over lymphoid lineage commitment and contributing to immune imbalance. Interestingly, skull bone marrow appears more resistant to this aging-related shift than femoral bone marrow.142 Systemic drivers and microenvironmental disruption in bone aging. This figure illustrates the multi-system drivers of bone aging and their synergistic disruption of the local bone marrow microenvironment. The center of the figure shows the characteristics of aging bones: fatty red bone marrow, depletion of hematopoietic stem cells, and chronic low-grade inflammation. It is surrounded by six driver modules, including: 1) osteogenesis activated by chronic inflammation; 2) multiple hormone axis changes; 3) multiple organ-bone axis disorders; 4) extracellular matrix hardening and fragility; 5) dysbiosis of gut microbiota leading to translocation of endotoxin and reduction of beneficial metabolites; 6) unhealthy lifestyles increase oxidative stress and inflammation. All systemic blows eventually converge and disrupt local bone homeostasis Senescent immune cells and the pro-inflammatory factors they secrete are the direct executors disrupting bone homeostasis. Aging bone marrow exhibits increased numbers and activity of senescent macrophages, T lymphocytes, and other immune cells, resulting in sustained secretion of factors such as TNF-α, IL-1β, and IL-6.143,144 These cytokines are potent stimulators of osteoclast differentiation: through activation of NF-κB and MAPK signaling pathways, they upregulate RANKL expression in osteoblasts and activated T cells, thereby promoting excessive osteoclastogenesis, prolonged lifespan, and hyperactive bone resorption.145 Senescent skeletal stem cells further contribute to the formation of an inflammatory degenerative niche, characterized by diminished bone formation potential and enhanced ability to secrete pro-inflammatory and pro-resorptive cytokines, further skewing normal differentiation within the skeletal and hematopoietic systems.146 Additionally, senescent fibroblasts exacerbate inflammation and osteoclast differentiation via activation of mTOR signaling,147 while aging disrupts the normal circadian regulation of macrophage function.148 This chronic inflammatory state ultimately leads to bone loss and increased skeletal fragility through multiple intersecting mechanisms. Pro-inflammatory cytokines not only directly stimulate bone resorption but also impair bone formation by inducing osteoblast apoptosis and pyroptosis and by disrupting autophagic processes, thereby disrupting the osteoblast-osteoclast coupling.143–145 Interventions targeting this inflammatory axis show therapeutic promise. For instance, IL-11 inhibitors have been shown to improve metabolic and muscle function by modulating the ERK-AMPK-mTORC1 axis and extend the lifespan in experimental models.149 Hormone level changes Age-related hormone alterations represent one of the core intrinsic factors driving bone aging. Bone homeostasis is regulated by a complex network, and age-associated declines, dysregulation, and altered interactions among hormones collectively create an endocrine aging environment leading to skeletal homeostatic imbalance. The decline in sex hormones is among the most impactful events. In women, the abrupt reduction in estrogen levels during menopause directly enhances osteoclast activity, increases osteoblast apoptosis, reduces bone mechanosensitivity, and intensifies local inflammation, leading to rapid early postmenopausal bone loss.150 In contrast, men experience a gradual decline in gonadal function. Reduced testosterone not only diminishes androgen receptor-mediated anabolic signaling in bone but also decreases its aromatization to estrogen, resulting in a combined androgen and estrogen deficiency that is closely associated with age-related bone loss in older men.151 The growth hormone/IGF-1 axis also progressively attenuates with age. This axis is critical for the acquisition of peak bone mass during adolescence, and its decline leads to reduced proliferation and differentiation of osteoblast precursors, lower bone formation rates, impaired bone matrix quality, and exacerbation of sarcopenia, thereby reinforcing the vicious cycle of sarcopenia-osteoporosis.152 Dysregulation of the parathyroid hormone (PTH) and vitamin D systems often occurs in parallel during aging. Vitamin D deficiency reduces intestinal calcium absorption, triggering a compensatory increase in PTH levels.153 Chronic mild elevations in PTH promote cortical bone resorption, while vitamin D deficiency also directly further compromises osteoblast function and mineralization.154 Additional hormonal changes also contribute significantly. Declines in endogenous calcitonin secretion may weaken its inhibitory effects on osteoclast activity.155 The widespread use of exogenous glucocorticoids remains a major iatrogenic cause of secondary osteoporosis in older adults.156 Thyroid dysfunction accelerates bone turnover, leading to net bone loss.157 Furthermore, age-related changes in adipose tissue distribution alter the secretion of adipokines such as leptin and adiponectin, which influence bone metabolism in complex and context-dependent manners.158,159 Hormonal changes in bone aging are not isolated events but rather an interconnected network. This network-based dysregulation may explain the limited efficacy of interventions targeting single hormonal pathways. Extracellular matrix changes The bone extracellular matrix (ECM) is not merely a structural scaffold but an active regulator of cell fate, signaling, and tissue mechanics. Dynamic alterations in ECM composition during aging represent a central mechanism driving bone aging.160 With aging, ECM components undergo progressive degradation, directly impairing bone quality and function. The deterioration of the collagen network is particularly critical: type I collagen synthesis declines, while degradation increases, leading to net loss. Moreover, collagen cross-linking shifts from physiological enzymatic cross-links to pathologic AGE-mediated cross-links, rendering collagen fibers brittle and inflexible and forming a structural basis for increased bone fragility.161 Concurrently, collagen fiber alignment becomes disorganized, and the expression and modification of non-collagenous proteins that influence hydration and signaling, such as osteocalcin, osteopontin, and glycosaminoglycans, are dysregulated.162 These changes impair ECM mineralization control, growth factor storage, and cell-matrix communication. Altered ECM mechanics profoundly affect cellular mechanosensing. At the macroscopic level, aging ECM exhibits reduced elasticity and increased brittleness.160 At the nanoscale, changes in stiffness and elasticity transmit dysregulated mechanical signals to cells via adhesion molecules such as integrins. This disruption of mechanotransduction impairs osteogenic differentiation, skewing mesenchymal stem cells toward adipogenic differentiation and also causing osteocyte dysfunction.163 As the core mechanosensors in bone, the lacunar-canalicular network of osteocytes becomes distorted and obstructed due to pathological changes in the ECM, reducing the efficiency of their ability to sense and transmit mechanical signals. This, in turn, leads to an imbalance in their secretion of key regulatory factors like sclerostin and RANKL, exacerbating bone resorption.164 As a regulatory hub for cellular signaling, ECM aging also disrupts key biological signal transduction. The ECM serves as a natural reservoir for pro-osteogenic growth factors such as TGF-β, BMPs, and IGF-1. Changes in its composition and structure affect binding, presentation, and release of these factors, potentially leading to their abnormal retention or degradation, thereby weakening osteogenic signals.165 Furthermore, osteoblasts and osteocytes bind to the ECM via integrins. Alterations in ligands within the aging ECM result in insufficient activation of integrin-mediated downstream signaling, subsequently impairing cell survival, proliferation, and differentiation capacities.166 Intervention and treatment Non-pharmacological interventions Lifestyle interventions are the cornerstone of bone health across the lifespan. However, their effectiveness depends critically on timing, dose, adherence, and individual baseline characteristics. Three principles guide their implementation: (1) primary prevention should begin early (childhood to young adulthood) to maximize peak bone mass; (2) secondary prevention in older adults focuses on slowing bone loss and reducing fall risk; (3) interventions show dose-response relationships, and “more is not always better”—excessive exercise or calcium supplementation may have diminishing returns or adverse effects. The following subsections provide specific, quantifiable recommendations and, where evidence permits, comparative efficacy estimates (Fig. 6). Therapeutic strategies targeting bone aging at multiple levels. This figure systematically shows the multi-level therapeutic strategies for targeted bone aging, and the layout reflects the evolution of intervention concepts from the basic to the cutting-edge. The left column is the basic lifestyle intervention, emphasizing the central role of balanced nutrition and mechanical stimulation (see Fig. 5 for pathological mechanisms). In the middle column, drugs targeting the pathological aspects of aging, including anti-resorptive drugs, anabolic drugs, and anti-aging drugs targeting the core mechanisms of aging. The right column shows cutting-edge regenerative strategies, including stem cell and exosome therapy, bone-targeted gene therapy, and tissue engineering. The bottom timeline summarizes the strategic development path from prevention and symptomatic treatment to targeted mechanisms and tissue regeneration Balanced diet A balanced intake of macronutrients is fundamental to skeletal health. Adequate, but not excessive, protein consumption is essential for the synthesis of bone matrix collagen; insufficient protein intake compromises both muscle and bone integrity. The quality and balance of dietary fatty acids are equally crucial. Increased consumption of foods rich in omega-3 polyunsaturated fatty acids helps mitigate inflammation and inhibit osteoclast activity, whereas intake of saturated and trans fats should be reduced.167 Sufficient calcium and vitamin D intake are direct prerequisites for preventing accelerated bone aging, requiring special attention for supplementation in older adults and postmenopausal women.168 However, the latest evidence published in the BMJ indicates that calcium, vitamin D, or combined supplements offer little or no benefit in preventing fractures and falls.169 Vitamin K, which is required for the activation of proteins such as osteocalcin that facilitate appropriate calcium deposition in bone, can be obtained from leafy green vegetables and fermented foods.170 Furthermore, micronutrients such as magnesium, potassium, zinc, and vitamin C are extensively involved in bone metabolism and matrix synthesis, and their deficiencies can indirectly harm skeletal health.171 The regulatory role of dietary phytochemicals in bone health is increasingly recognized. Polyphenols, carotenoids, and other compounds abundant in fruits and vegetables possess potent antioxidant and anti-inflammatory activities, protecting bone by inhibiting osteoclastogenesis and promoting osteogenic signaling pathways. Evidence from research on specific berries supports these effects. Early-life consumption of blueberries has been shown to prevent osteoblast aging and bone loss in adulthood,172,173 while plums, sour cherries, bilberries, and other fruits have demonstrated efficacy in ameliorating age-related bone loss.174–180 These benefits may be mediated through modulation of the gut-bone axis and antioxidant mechanisms.181 Population-based studies further indicate that higher fruit intake during childhood is associated with greater femoral neck BMD in adulthood,182 and that dietary patterns rich in fruits, vegetables, and whole grains among middle-aged and older adults can reduce fracture risk (female HR: 0.86, 95% CI 0.76-0.98, male HR: 0.83, 95% CI 0.64-1.08).183 Based on current evidence,184 the following daily intakes are recommended for older adults: calcium 1 000–1 200 mg/day (dietary priority over supplements; 500–600 mg per meal maximizes absorption); vitamin D 800–1 000 IU/day (higher doses, e.g., >2 000 IU/day, show no additional bone benefit185); protein: 1.0–1.2 g/kg body weight/day (higher, 1.2–1.5 g/kg, for frail older adults with sarcopenia). vitamin K (90–120 μg/day) from leafy green vegetables supports osteocalcin activation, while magnesium (320–420 mg/day) and zinc (8–11 mg/day) are cofactors for bone metabolism. Among dietary patterns, the Mediterranean diet (rich in fruits, vegetables, whole grains, and olive oil) shows the strongest evidence for fracture prevention, compared to calorie restriction (pooled hazard ratios = 0.95, 95% CI 0.93-0.96), low-carbohydrate/ketogenic diets, and higher-protein diets.186 Moderate exercise Regular physical activity is an effective non-pharmacological intervention for maintaining skeletal health and combating bone aging. The benefits of exercise include increases in BMD, improvements in bone quality, and reductions in fracture risk. The effects of exercise on BMD are modality-dependent.187 Weight-bearing exercises and resistance training are particularly effective. Weight-bearing exercises stimulate bone through ground reaction forces, whereas resistance training imposes tensile loads through muscle contraction, both of which elicit adaptive responses in bone. Regular engagement in moderate- to high-intensity forms of these activities can effectively slow age-related BMD loss at clinically important sites such as the lumbar spine and femoral neck, with particularly pronounced benefits observed in postmenopausal women.188 Beyond BMD, exercise enhances bone quality by improving trabecular microarchitecture and increasing geometric parameters such as cross-sectional area, thereby strengthening resistance to bending and torsion.189 In older adults with established osteoporosis, supervised, individualized, and progressive intensified strength and balance training, although potentially producing only modest increases in BMD, plays a critical role in improving muscle strength, coordination, and postural stability. These functional gains substantially reduce the risk of falls and, consequently, fragility fractures.190 For BMD gains, the following exercise prescriptions are recommended based on meta-analytic evidence191: resistance training at 60%–80% of one-repetition maximum, 8–12 repetitions × 2–3 sets, 2–3 sessions per week; weight-bearing aerobic exercise (walking, jogging, stair climbing) for 30–45 minutes per session, 3–5 sessions per week; combined (multicomponent) training incorporating both resistance and weight-bearing elements. For fall prevention, balance training (tai chi, standing exercises, dynamic balance tasks) should be performed at least 2 sessions per week. For frail or deconditioned older adults, a “start low, go slow” approach is recommended, with intensity progression every 4–6 weeks as tolerated. High-impact exercise (jumping, hopping) produces larger BMD gains but is contraindicated in individuals with established osteoporosis or high fracture risk. Comparative efficacy: A meta-analysis of 74 RCTs (n = 5 112) compared the effects of different exercise modalities on lumbar spine BMD in postmenopausal women: combined training > resistance training > weight-bearing aerobic, with effect sizes (SMD) of 0.42, 0.40, and 0.26, respectively.192 For fall prevention, a 2025 meta-analysis found Multicomponent physical activity (combining strength, balance, and aerobic training) and mind-body exercises (tai chi, yoga) demonstrated the most consistent benefits.193 Regarding the effect of different training frequency, Zitzmann et al. found that training at least twice per week has a significantly greater impact on lumbar BMD compared to a lower frequency (<2 times per week) (SMD = 0.55, 95% CI = 0.20-0.90).194 Restriction of alcohol and tobacco use Tobacco use and excessive alcohol intake are independent yet synergistic risk factors for bone health. Smoking adversely affects bone health through direct cellular toxicity, disruption of sex hormone balance, induction of oxidative stress and inflammation, impairment of vascular function, and interference with vitamin D and calcium metabolism.195,196 Excessive alcohol intake further compromises skeletal integrity by suppressing osteoblast activity, promoting secondary osteoporosis, increasing fall risk, and delaying fracture healing.197–199 For adolescents and young adults who have not yet achieved peak bone mass, abstaining from smoking and limiting alcohol intake are essential for building maximal skeletal reserve. In perimenopausal women and middle-aged men, these measures represent a first-line defense against age-related physiological bone loss. In individuals with osteoporosis, continued smoking and alcohol consumption can reduce the therapeutic efficacy of anti-osteoporotic medications such as bisphosphonates, denosumab, and teriparatide.200–202 Nicotine and alcohol interfere with the microenvironment of the target cells of these agents, potentially leading to a suboptimal therapeutic response. Therefore, smoking cessation and moderation of alcohol intake are essential prerequisites to ensure that pharmacological treatments achieve their intended effects. Pharmacotherapy Antiresorptive agents Bisphosphonates: Oral bisphosphonates, such as alendronate, risedronate, and zoledronate, remain the most cost-effective interventions. They selectively bind to the mineralized bone matrix and are internalized by osteoclasts, where they disrupt the mevalonate pathway, inducing apoptosis and thereby potently inhibiting bone resorption.203 Long-term studies have demonstrated significant reductions in vertebral and hip fracture risk. However, prolonged use requires careful monitoring because of rare but serious adverse events, including osteonecrosis of the jaw and atypical femoral fractures.204 Contemporary real-world data from large cohort studies (n > 26 000) have refined the risk estimates: atypical femoral fractures occur at a rate of approximately 1.93 per 1 000 person-years after three years of continuous use, rising to 2.50 per 1 000 person-years after five years, whereas osteonecrosis of the jaw remains exceedingly rare at less than 0.07 per 1 000 person-years205 (Table 1). Table 1Pharmacotherapy for bone agingDrug CategoryRepresentative DrugsMechanismIndicationsAdverse ReactionsSpecial ConsiderationsAntiresorptive AgentsBisphosphonatesAlendronatefarnesyl pyrophosphate synthase inhibition → osteoclast apoptosisFirst-line for postmenopausal OP; male OP; glucocorticoid-induced OP; Paget’ s diseaseAFF; ONJ; EsophagitisDrug holiday after 3–5 years; most cost-effectiveZoledronateConvenient once-yearly; preferred after hip fractureRisedronateFaster onset than alendronateRANKL Monoclonal AntibodyDenosumabRANKL neutralizationPostmenopausal OP; male OP; glucocorticoid-induced OPONJ; AFF; rebound vertebral fractures; hypocalcemiaNo drug holiday; never discontinue without transition to bisphosphonate; rebound bone loss within 6 monthsSelective Estrogen Receptor ModulatorsRaloxifeneEstrogen agonist in bonePostmenopausal OP; prevention in women with high invasive breast cancer riskVTE; hot flashes; leg crampsNo hip fracture benefit; favorable breast safety profileHormone Replacement TherapyEstradiolestrogen receptor on osteoblasts → reduces osteoclast activityYounger postmenopausal women (age ≤60 yrs) with menopausal symptoms and high fracture riskStroke; VTE; breast cancer (long-term combined therapy)Not recommended for women >60 yrs or without menopausal symptoms; reserve for those with moderate-to-severe vasomotor symptomsAnabolic AgentsPTH/PTHrp AnalogsTeriparatidePTH1R activationVery high fracture riskHypercalcemia; leg cramps; dizzinessMax 24 months lifetime; must follow with anti-resorptiveAbaloparatidePTH1R activation (selective)Sclerostin Monoclonal AntibodyRomosozumabSclerostin inhibition → Wnt activationCV events; ONJ; AFFMax 12 months; must follow with anti-resorptive; assess CV risk before useSenolytic TherapiesSenolytic AgentsDasatinib/ QuercetinBCL-2/HSP90/SCF inhibition → selective senescent cell apoptosisInvestigational for age-related bone loss; postmenopausal OPGenerally well-toleratedHigh senescence burden subgroup showed BMD gains; biomarker-driven therapySASP ModulatorsRuxolitinibJAK-STAT inhibition → suppression of SASPInvestigational for age-related bone loss; preclinical modelsTo be determinedNot yet tested in clinical trials for boneRapamycinmTOR inhibition → suppression of SASPTo be determinedKnown immunosuppressive effects limit long-term useFisetinMultiple pathwaysGenerally recognized as safe (food-derived)Natural flavonoid; senolytic activity demonstrated in preclinical models; human bone data lacking AFF atypical femoral fracture, ONJ osteonecrosis of the jaw, VTE venous thromboembolism, CV cardiovascular, HR hazard ratio, SASP senescence-associated secretory phenotype, OP osteoporosis Pharmacotherapy for bone aging AFF atypical femoral fracture, ONJ osteonecrosis of the jaw, VTE venous thromboembolism, CV cardiovascular, HR hazard ratio, SASP senescence-associated secretory phenotype, OP osteoporosis RANKL inhibitors: Denosumab is a fully human monoclonal antibody that binds to RANK. By preventing the interaction between RANKL and its receptor RANK on osteoclast precursors and mature osteoclasts, denosumab profoundly suppresses osteoclast differentiation, activation, and survival.206 Administered via subcutaneous injection, it is convenient and has been shown to continuously increase BMD across multiple skeletal sites while significantly reducing fracture risk. However, abrupt discontinuation may result in a rebound increase in bone turnover markers and rapid bone loss207—a phenomenon now well-quantified. Following denosumab withdrawal, approximately 10% of patients experience multiple vertebral fractures, particularly those with a prior vertebral fracture history.208 Although rare (<0.5%), rebound hypercalcemia can also occur, typically between three and eleven months after the last dose. These complications are generally reversible upon re-treatment with either denosumab or bisphosphonates.209 Selective estrogen receptor modulators (SERMs): Raloxifene exerts estrogen-like effects on bone, inhibiting bone resorption and reducing vertebral fracture risk, without stimulating breast or endometrial tissues. These properties make it suitable for the prevention and treatment of postmenopausal osteoporosis.210 Calcitonin, due to insufficient evidence supporting long-term efficacy and concerns regarding potential cancer risk, now plays a secondary role and is primarily used for short-term relief of acute bone pain.211 Hormone replacement therapy (HRT): Estrogen-based hormone replacement therapy is a highly effective anti-resorptive agent for younger postmenopausal women (age ≤60 years) with high fracture risk, and low baseline risk for adverse malignant and thromboembolic events. HRT acts primarily by binding to estrogen receptors on osteoclasts, inducing their apoptosis and thereby suppressing bone resorption. This mechanism leads to sustained improvements in bone mineral density across all skeletal sites.212 Large-scale randomized controlled trials and their long-term follow-up studies have demonstrated that HRT significantly reduces the risk of vertebral, hip, and non-vertebral fractures.213 However, HRT is not recommended as a first-line therapy for older postmenopausal women without menopausal symptoms due to the increased absolute risks of stroke, venous thromboembolism, and, with long-term combined therapy, breast cancer. When prescribed, the lowest effective dose for the shortest duration necessary to control symptoms is recommended, with annual re-evaluation of risks and benefits. Anabolic agents PTH/PTHrP analogs: Teriparatide, a recombinant human PTH, preferentially stimulates bone formation when administered intermittently at low doses via subcutaneous injection. This anabolic effect is mediated primarily through activation of the PKA signaling pathway in osteoblasts. Teriparatide significantly improves BMD and bone microstructure and reduces the risk of both vertebral and non-vertebral fractures.214 The duration of teriparatide therapy is typically limited to 24 months, a precaution originally based on an increased incidence of osteosarcoma observed in high-dose, near-lifetime rat studies.215 However, long-term clinical trials and post-marketing surveillance extending up to 15 years have not demonstrated any increased risk of osteosarcoma in humans treated with teriparatide.216 Large observational studies using claims data linked to cancer registries have shown no increase in osteosarcoma incidence compared with unexposed populations or expected background rates, leading to the removal of the boxed warning from regulatory labels.217 Another PTHrP analog is Abaloparatide. Compared with teriparatide, it shows superior anti-fracture efficacy in major osteoporotic fractures (upper arm, wrist, hip, or clinical vertebrae) and has a lower risk of hypercalcemia.218 Sclerostin monoclonal antibody: Romosozumab inhibits sclerostin, a negative regulator of bone formation. By releasing inhibition of the Wnt/β-catenin signaling pathway, romosozumab robustly stimulates bone formation while simultaneously exerting moderate anti-resorptive effects.219 Treatment can substantially increase BMD and reduce fracture risk within a relatively short period (typically 12 months). Initial concerns about a potential increase in major adverse cardiovascular events (MACE), raised by a pivotal phase 3 trial, have been reassessed in subsequent real-world studies. A large cohort study using Japan’s national claims database (251 219 romosozumab users vs. 500 445 teriparatide users) found no statistically significant difference in MACE risk between romosozumab and teriparatide (adjusted HR 1.00; 95% CI 0.94–1.06), regardless of prior MACE history.220 Meta-analyses of randomized controlled trials have similarly shown no significant differences in cardiovascular events between romosozumab and placebo.221 Nevertheless, given the persistent clinical uncertainty, careful assessment of cardiovascular risk remains prudent before use, particularly in patients with a recent history of myocardial infarction or stroke. Sequential treatment strategies The availability of both anti-resorptive and anabolic agents has expanded therapeutic options for osteoporosis but has also introduced complexity regarding the optimal sequence of treatment. Given that anabolic agents achieve their maximal fracture-reduction benefits within 12–24 months and that their effects are not sustained after discontinuation, while anti-resorptive agents provide durable fracture protection with extended use, sequential therapy has emerged as the standard of care for patients at very high fracture risk. From anabolic to anti-resorptive therapy (the anabolic-first strategy): Contemporary evidence strongly supports initiating treatment with an anabolic agent followed by an anti-resorptive agent in patients with very high fracture risk—defined as those with recent or multiple fragility fractures, very low BMD (T-score < −3.0), or fractures despite ongoing anti-resorptive therapy.222 This sequential approach capitalizes on the rapid and robust bone formation induced by anabolic agents and subsequently maintains the accrued bone mass with anti-resorptive therapy. A landmark meta-analysis of six randomized controlled trials enrolling 17 872 participants demonstrated that, compared with anti-resorptive monotherapy or placebo, initial anabolic therapy followed by anti-resorptive agents reduced vertebral fracture risk by 57%, clinical fracture risk by 38%, and hip fracture risk by 35%.223 From anti-resorptive to anabolic therapy: In patients who have been on long-term bisphosphonate therapy and remain at very high fracture risk, switching to an anabolic agent is recommended. However, bone turnover may be significantly suppressed after prolonged bisphosphonate use, potentially attenuating the initial response to anabolic therapy. In such cases, a washout period of several months before initiating an anabolic agent may be considered.222 Transition from denosumab: Denosumab discontinuation presents a unique clinical challenge due to rapid bone loss and increased vertebral fracture risk. Therefore, denosumab should never be discontinued without transition to an alternative anti-resorptive agent. When denosumab is stopped, a bisphosphonate should be administered promptly, typically within six months of the last denosumab injection, to mitigate rebound bone loss. For patients transitioning from denosumab to an anabolic agent, the optimal timing and sequence remain uncertain; current clinical practice often initiates the anabolic agent while continuing denosumab or shortly after the last denosumab dose to avoid an unprotected interval.224 Senolytic therapies Senolytic agents: Senolytics represent an emerging therapeutic strategy targeting a core mechanism of bone aging: the accumulation of senescent cells.225 Senescent cells disrupt bone homeostasis through the secretion of SASP factors. Senolytic agents selectively induce apoptosis in senescent cells by targeting their anti-apoptotic pathways.226 Among the earliest studied combinations is dasatinib and quercetin (D + Q), which has been shown in aged mice and radiation-induced bone injury models to eliminate senescent osteoblasts, osteocytes, and other cell types, thereby restoring skeletal structure and function.58,60,227 The first phase 2 RCT of intermittent senolytic therapy (D + Q) in postmenopausal women (n = 60) was published in Nature Medicine (2024).228 The primary endpoint—change in the bone resorption marker CTX at 20 weeks—did not differ between groups (D + Q -4.1% vs. control -7.7%; P = 0.611). However, the bone formation marker P1NP increased significantly in the D + Q group at both 2 weeks (+16%, P = 0.020) and 4 weeks (+16%, P = 0.024). In exploratory analyses, women with high senescent cell burden (top tertile of T cell p16INK4a mRNA) showed concomitant increases in P1NP (+34%, P = 0.035) and reductions in CTX (−11%, P = 0.049) at 2 weeks, and significant radius BMD gains (+2.7%, P = 0.004) at 20 weeks. SASP Modulators (Senomorphics): In addition to direct senescent clearance, senomorphics exert therapeutic effects by suppressing the secretion of SASP factors. For instance, JAK inhibitors such as ruxolitinib can mitigate age-related osteoporosis by inhibiting cytokines such as IL-6.58 Rapamycin and its analogs have also been reported to suppress SASP activity.229 Additionally, the natural flavonoid fisetin exhibits senolytic activity and has demonstrated potential benefits in improving osteoarthritis.230 Emerging therapies Beyond established pharmacotherapies, several innovative strategies are emerging to target the core mechanisms of bone aging. These approaches—stem cell therapy, exosome therapy, gene therapy, and smart biomaterials—differ substantially in their mechanism of action, stage of development, efficacy profile, and safety considerations. Rather than describing each in isolation, this section synthesizes the current evidence across strategies to guide future research priorities. In terms of efficacy comparison, although direct head-to-head comparisons are lacking, preclinical studies using similar animal models allow for approximate effect size estimation. Gene therapy, particularly Wnt-pathway modulating gene silencers delivered via bone-targeted AAV vectors, has demonstrated the most robust bone regeneration, achieving >50% defect healing and systemic reversal of bone loss in osteoporotic mice following a single administration.231 Stem cell therapy has shown consistent but moderate efficacy, with allogeneic BMSC transplantation achieving approximately 40%–50% defect healing and improved BMD in osteoporotic models.232,233 Exosome therapy, while generally less potent than cell-based or gene-based approaches, offers distinct advantages in safety and scalability. For instance, MSC-derived exosomes carrying miR-29 reduced fracture healing time by approximately 30% in animal studies without the tumorigenic risk associated with stem cells.40 On the path to clinical translation, among these strategies, stem cell therapy has advanced farthest into clinical translation, with allogeneic BMSCs successfully used for segmental bone defect repair in humans, and adipose-derived stem cells entering early-phase trials for osteoporotic fractures.234 Gene therapy has achieved regulatory approval for osteoarthritis,235 and bone-targeted AAV-based approaches are now in late-stage preclinical development.236 Exosome therapy and smart biomaterials remain predominantly at the preclinical stage, though both have demonstrated promising safety profiles in small and large animal models.237,238 Regarding the trade-off between safety and effectiveness, a consistent pattern emerges across strategies: interventions with greater efficacy often carry higher safety risks. Gene therapy offers durable, single-administration efficacy but raises concerns about insertional mutagenesis, immunogenicity, and off-target gene editing. Stem cell therapy faces challenges of post-transplantation cell survival, batch-to-batch variability, and potential tumorigenicity. Exosome therapy exhibits the most favorable safety profile (low immunogenicity, no live cells, no viral vectors) but requires optimization of targeting efficiency and half-life extension in circulation.239 Smart biomaterials provide structural support and controlled local delivery but lack the biological complexity to fully recapitulate native bone tissue.240 Challenges and future directions Ongoing challenges Depth and integration of mechanistic elucidation Current research on bone aging largely focuses on changes within individual cell types or isolated signaling pathways. However, a defining feature of bone aging is the systemic failure of communication networks. Several fundamental questions remain inadequately addressed: How does the functional decline of osteocytes, as mechanosensors and endocrine regulators, precisely affect osteoblast-osteoclast coupling? What is the causal relationship between impaired osteogenic-adipogenic lineage commitment of BMSCs and the accumulation of marrow adipose tissue? How is crosstalk between bone and other organs, including muscle, brain, gut, and the immune system, altered during aging? Existing studies often offer a fragmented view, lacking a unified framework that temporally and spatially integrates molecular, cellular, tissue, and systemic levels. The gap between research models and human bone aging Bone aging research relies predominantly on animal models, particularly mice. However, substantial differences exist between murine and human skeletal physiology and aging trajectories. Humans exhibit extensive Haversian remodeling, a prolonged period of skeletal maturation, and distinct patterns of age-related bone loss. Conventional aged mouse models fail to replicate the decades-long process of bone accrual and gradual degeneration observed in humans. Furthermore, in vitro cell culture systems cannot adequately reproduce the complex three-dimensional microenvironment, mechanical loading, and systemic hormonal regulation present in vivo. These limitations contribute to the frequent failure of therapeutics that show efficacy in preclinical models to translate successfully into human clinical trials. Scarcity and heterogeneity of bone aging biomarkers The premise of precision medicine is accurate assessment. Yet, clinical evaluation of bone aging relies heavily on BMD, a delayed and incomplete indicator. BMD does not capture critical factors determining bone fragility, such as bone turnover rate, microarchitectural deterioration, and declines in mechanical properties. There is an urgent need to identify multimodal biomarker panels capable of capturing the biological process of bone aging in an early, dynamic, and comprehensive manner. Moreover, the molecular basis underlying the significant individual heterogeneity and sexual dimorphism in bone aging, and how to achieve risk stratification and personalized intervention, remains poorly understood. Bottlenecks in intervention strategies Current first-line anti-osteoporotic therapies primarily work by potently suppressing bone resorption, raising concerns about excessive inhibition of bone remodeling during long-term use. Although anabolic agents can increase bone mass, their application is limited by restricted treatment duration and high cost. More fundamentally, most existing therapies are symptomatic rather than causal, failing to address the underlying drivers of bone aging. Emerging interventions based on the biology of aging, such as interventions targeting cellular senescence, mitochondrial dysfunction, defective autophagy, and stem cell exhaustion, offer considerable promise. However, they face significant challenges, including targeted delivery efficiency, tissue specificity, and long-term safety. Sociomedical issues Skeletal aging represents a classic “silent epidemic”: bone loss progresses asymptomatically until a fragility fracture occurs, often with devastating consequences. This delayed clinical manifestation contributes to insufficient public awareness, delayed policy responses, and chronic underinvestment in preventive strategies. Although fragility fractures impose enormous direct and indirect economic burdens through medical costs, rehabilitation, and loss of productivity, these downstream consequences have not been effectively translated into incentives for upstream prevention. Furthermore, the rate and impact of bone aging are profoundly influenced by social determinants of health, including socioeconomic status, gender and age discrimination, and geographic disparities, resulting in significant health inequities. Future directions Establishing multi-omics integrated research paradigms Future research must transcend the perspective of isolated pathways and shift towards constructing a comprehensive bone aging atlas. This requires integrating genomics, epigenomics, proteomics, metabolomics, and single-cell sequencing technologies to systematically characterize dynamic interaction networks among different cell types within bone tissue during aging. A key focus will be elucidating how bone, as an endocrine organ, communicates with and is influenced by distant organs over time. Identifying central regulatory hubs and critical network nodes will provide a conceptual foundation for the development of multi-target, mechanism-based interventions. Developing organoid and computational models To address the limitations of traditional animal models, future efforts should prioritize the development of advanced in vitro systems and digital models that more closely mimic human physiology. Human induced pluripotent stem cells can be used to generate three-dimensional organoids or “bone-on-a-chip” platforms incorporating bone, bone marrow, vasculature, and immune components, thereby simulating the complex microenvironment and aging process of human bone. In parallel, integrating clinical imaging data, biomechanical measurements, and multi-omics datasets will facilitate the development of digital twins and artificial intelligence-driven predictive models of bone aging. These approaches have the potential to significantly enhance the efficiency and precision of translating basic discoveries into clinical applications and reduce the failure rate in drug development. Establishing multimodal, dynamic precision risk assessment systems Moving beyond the reliance on BMD alone, future efforts should establish precision assessment systems that integrate structural, functional, and molecular information. This includes the use of advanced imaging techniques such as high-resolution peripheral quantitative computed tomography to quantitatively assess bone microstructure; developing novel biomarker panels based on blood or urine that can dynamically reflect bone turnover, cellular senescence status, and metabolic signatures; and incorporating genetic risk scores and clinical risk factors. Ultimately, artificial intelligence algorithms will be employed to integrate multidimensional data, enabling early warning, precise stratification, and dynamic monitoring of the bone aging rate, fracture risk, and therapeutic response. Developing therapies targeting core mechanisms of bone aging Future therapeutic strategies should directly target the fundamental biological mechanisms driving skeletal aging. These include the development of bone-targeted senolytic drugs to selectively eliminate senescent cells and ameliorate the pro-inflammatory bone microenvironment. At the stem cell level, strategies include restoring the regenerative potential of aged stem cells through epigenetic reprogramming, mitochondrial function enhancement, or targeted drug delivery. Systemic metabolic and immune modulation will focus on interventions against age-related chronic inflammation and metabolic disorders to systemically improve skeletal health. Concurrently, the development of smart delivery systems, including nanocarriers, is essential to ensure these novel therapies act precisely and safely on bone tissue. Promoting equity-centered public health strategies focused on prevention Future public health initiatives should prioritize the implementation of comprehensive lifespan-oriented bone health action plans. These should incorporate bone density screening and risk assessment into routine health check-ups. Additionally, digital health and wearable should be leveraged for community- and household-based remote monitoring and personalized exercise and nutritional guidance. Crucially, health equity must be placed at the core. Through targeted screening, health education, and the allocation of medical resources, special attention must be provided to reduce skeletal health disparities caused by socioeconomic differences, ensuring that intervention measures benefit all populations. Summary Bone aging represents a progressive disruption of skeletal homeostasis at macroscopic, microscopic, and systemic levels. Macroscopically, it manifests as bone loss, deterioration of bone microstructure, and a decline in biomechanical properties. Microscopically, it involves genetic alterations, epigenetic dysregulation, oxidative stress, dysfunction of key signaling pathways, and the emergence of the senescence-associated secretory phenotype. At the systemic level, it is closely linked to chronic inflammation, hormonal changes, and alterations in the extracellular matrix. Current intervention strategies encompass lifestyle modifications, anti-resorptive and anabolic agents, and emerging strategies such as senolytic agents and regenerative medicine approaches. However, challenges remain, including insufficient integration of underlying mechanisms, limitations of existing models, scarcity of sensitive and specific biomarkers, restricted therapeutic precision, and persistent sociomedical barriers. Future research necessitates the development of multi-omics integrated aging atlases, organoid and computational models, multimodal risk assessment systems, precision therapies targeting core mechanisms, and the promotion of prevention-centered public health strategies. These efforts are crucial for providing the scientific foundation and clinical translation pathways to achieve healthy skeletal aging. Acknowledgements This study was supported by the National Key Research and Development Program of China (2024YFB3814700); National Natural Science Foundation of China (82372450, 82402828); Natural Science Foundation of Beijing Municipality (L254010, L234062, L256033); China Postdoctoral Science Foundation General Fund (2025M782894). We are profoundly grateful to Dr. Lingxue Luo for her invaluable intellectual contributions and technical guidance in the preparation of the figures. Furthermore, we appreciate the indispensable assistance provided by the graphic design software Adobe Illustrator (Adobe Inc.) and the scientific illustration platform BioRender (BioRender.com). Author contributions K.P. Li, C.Y. Gao and T. Zhang contributed to writing the original draft and preparing the figures. C. Wang, X. Geng and X.G. Wang contributed to funding acquisition and review. H. Tian contributed to conceptualization, project administration, funding acquisition, review, and editing. All authors have reviewed and approved the final manuscript. Competing interests The authors declare no competing interests. Footnotes These authors contributed equally: Kangpeng Li, Chenyuan Gao, Ti Zhang Contributor Information Kangpeng Li, Email: likangpeng@bjmu.edu.cn. Hua Tian, Email: tianhua@bjmu.edu.cn. References