Bone marrow mesenchymal stromal/stem cells (BMSCs) constitute a central cellular reservoir that sustains osteoblast-lineage supply and defines the osteogenic differentiation of the microenvironment. In osteoporotic settings, BMSCs commonly display a convergent phenotype—reduced clonogenicity and osteogenic output (e.g., CFU-F frequency, ALP activity, mineralized nodule formation) accompanied by increased adipogenic commitment and marrow adipose expansion—paralleling declines in dynamic histomorphometric indices of formation (e.g., MAR/BFR) and trabecular integrity (e.g., BV/TV) in vivo (62, 118, 119) (Figure 1A). Mechanistically, lineage allocation reflects a shift in dominant transcriptional and signaling modules: attenuation of osteogenic pathways (Wnt/β-catenin, BMP–RUNX2/OSX) alongside reinforcement of adipogenic programs (PPARγ-driven networks), with additional tuning by inflammatory signaling (e.g., NF-κB-linked cytokine milieu) and mechano-metabolic inputs (integrin–FAK/YAP–TAZ, mTOR/AMPK) (56–61). Notably, whether lineage drift is primarily cell-intrinsic (senescence/DDR) or imposed by niche cues (inflammation, ECM mechanics, perfusion) likely varies by skeletal compartment and remains incompletely resolved, highlighting the need for spatially anchored, human-relevant validation.

Osteoblast-lineage cells (including osteoblasts and osteocytes) act as the core coupling regulators in bone remodeling, by providing permissive regulatory signals and rate-limiting regulatory signals for osteoclast differentiation within the basic multicellular unit. The receptor activator of nuclear factor-κB ligand-receptor activator of nuclear factor-κB-osteoprotegerin (RANKL–RANK–OPG) signaling axis represents a canonical coupling node in bone remodeling. Cell-type-targeted perturbation experiments have confirmed that modulating the expression levels of RANKL/OPG in osteoblast-lineage cells is sufficient to reprogram the number and bone surface coverage of osteoclasts, thereby altering bone turnover status and trabecular architecture in vivo (41–43). Beyond the RANKL–OPG signaling axis, osteoblast-lineage cells can also regulate the priming of osteoclast precursors, the lifespan of mature osteoclasts, and the recruitment efficiency of osteoblasts via contact-dependent and paracrine mediators. This enables the transition from bone resorption to bone formation, ultimately maintaining bone turnover balance rather than simply suppressing the resorption process (44). In osteoporosis-relevant pathological conditions such as estrogen deficiency, aging, and chronic inflammation, this bone remodeling coupling set-point undergoes maladaptive shifts—characterized by exaggerated osteoclastogenesis concurrent with impaired osteoblast function. This effectively “uncouples” high bone resorption from compensatory bone formation, thereby accelerating net bone loss (45, 46) (Figure 1B). Notably, the relative regulatory contributions of coupling signals derived from osteoblasts versus osteocytes may vary across different skeletal sites and disease contexts. This highlights the necessity of conducting compartment-resolved, human-aligned validation studies on the hierarchical organization of bone coupling.

Osteoclasts originate from myeloid progenitors, which share overlapping developmental and functional regulatory programs with macrophages (120). Together, these two cell types form the macrophage-osteoclast continuum, through which the body’s inflammatory status and local tissue microenvironment jointly determine the ultimate level of bone resorptive function (121, 122). Osteoclast precursors are not a homogeneous cell population but exhibit detectable heterogeneity in recruitment, priming, and fusion capacity (123) (Figure 1C). Niche cues (e.g., the bioavailability of macrophage colony-stimulating factor/receptor activator of nuclear factor-κB ligand, pro-inflammatory cytokines, etc.) can reprogram the responsiveness of these precursors by altering their receptor expression profiles, intracellular metabolic regulatory pathways, and cell-cell fusion propensity. This, in turn, modulates the formation efficiency of multinucleated osteoclasts and the magnitude of bone resorption (e.g., the number/surface area of tartrate-resistant acid phosphatase-positive osteoclasts, resorption pit area) (123, 124). Single-cell transcriptomic and lineage-tracing studies further support that distinct precursor subsets expand or become preferentially osteoclastogenic under pathological conditions, providing a mechanistic basis for inter-individual and context-dependent variation in “high-turnover” phenotypes observed in osteoporosis and inflammaging-related bone loss (125).

Osteocytes are long-lived, matrix-embedded regulatory cells that integrate mechanical and endocrine signals to coordinately regulate bone remodeling processes within the cellular niche, acting as the regulatory hub of bone tissue. Through the lacunar-canalicular network, osteocytes sense mechanical strain and fluid shear stress and convert these stimulatory signals into regulatory output signals. These signals, in turn, modulate the activity of osteoblasts and the process of osteoclastogenesis, thereby tuning bone turnover toward adaptive bone formation or resorption depending on the mechanical loading status (92). Evidence from mechanical unloading/loading model experiments and osteocyte-targeted genetic perturbation assays indicates that impairment of osteocyte mechanosensing function and its downstream signaling pathways is sufficient to attenuate the skeletal anabolic response to mechanical loading, alter dynamic bone formation indices, and affect osteoclast-related parameters, ultimately leading to the remodeling of bone microstructure (54, 126). In osteoporosis-relevant contexts such as aging and estrogen deficiency, osteocyte dysfunction—including impaired mechanosensitivity and maladaptive coupling outputs—can “lock in” a catabolic set-point where resorption is amplified while formation fails to compensate, accelerating net bone loss (127, 128) (Figure 1D).

The extracellular matrix (ECM) in bone tissue serves as an instructive microenvironment that integrates structural organization and biochemical signal presentation functions. The composition profile and post-translational modifications of the ECM can regulate integrin binding, growth factor sequestration and release, as well as the mechanical processes of mineralization nucleation, thereby affecting the adhesion, proliferation and maturation of osteoblasts (93, 94). Matrix mechanical properties and mechanical loading further impose upstream constraints on osteogenic processes, converting stiffness and strain into transcriptional regulation via activation of mechanotransduction modules. Osteogenesis-related cells can recognize and integrate these mechanical cues through integrin–FAK/focal adhesion signaling, mechanosensitive ion channels, and mechanotranscriptional regulators (e.g., YAP/TAZ-associated programs), which ultimately converge on the expression of osteogenic genes (60, 95, 96). At the tissue scale, the lacunar-canalicular system (LCS) of osteocytes amplifies the aforementioned processes: strain-induced interstitial fluid flow generates shear stress within the canalicular network, enabling osteocytes to sense loading and transmit signals through the interconnected network, thus coordinately regulating bone remodeling (92, 97). Experimental validation of loading conditions and combined computational-imaging analyses further demonstrate that the structural characteristics of the LCS (e.g., lacunar density, canalicular connectivity) are correlated with mechanical sensitivity and its downstream remodeling outcomes (97, 98). In the context of osteoporosis and aging, microstructural degeneration and osteocyte dysfunction can impair LCS-mediated signal transmission, leading to insufficient adaptive osteogenesis and coupling imbalance (98, 99). However, it remains challenging to distinguish the effects of ECM components from covarying mechanical factors (100, 101). Therefore, an integrated research strategy combining matrix composition profiling, mechanical testing and spatial validation is required to obtain more reliable mechanistic interpretations (102).

Osteogenesis is tightly coupled to vascular supply: endothelial cells not only deliver nutrients and oxygen but also maintain osteoprogenitor cells and coordinate remodeling dynamics through angiocrine signaling. Studies on osteogenesis-vascular coupling have demonstrated that disruption of vascular signals can restrict osteogenic processes; conversely, enhancement of angiogenesis can boost osteoprogenitor cell activity in vivo and improve osteogenesis-related outcomes (103, 104). H-type vessels have attracted attention as a spatially specific vascular component. Typically composed of a population of endothelial cells highly expressing CD31 and Endomucin, these vessels are enriched in osteogenically active regions and distributed in close proximity to osteoprogenitor cell populations, thus correlating with higher levels of bone formation (103, 105). Mechanistically, endothelial cells can promote the osteogenic microenvironment by supporting the recruitment, proliferation and differentiation efficiency of progenitor cells, thereby enhancing the coupling between angiogenesis and osteogenesis (104, 106). In osteoporosis-relevant contexts, reduced perfusion and loss of H-type vessels may suppress osteogenesis and amplify the net bone loss effect induced by enhanced osteoclastogenesis (105, 107).

Chronic inflammation is widely recognized as an important contributor to osteoporosis and can markedly impair the maintenance of bone mass and bone mineral density (BMD) (108, 109). Under persistent inflammatory conditions, circulating and local levels of cytokines such as interleukin-1 (IL-1), IL-6, and tumor necrosis factor-α (TNF-α) are elevated; these mediators promote bone resorption by enhancing osteoclast differentiation and activity (Figure 2A). Clinical and experimental studies have reported a direct association between higher inflammatory cytokine burden and reduced BMD (110). In particular, IL-6 not only promotes osteoclastogenesis but can also suppress osteoblast function, thereby uncoupling bone formation from resorption and accelerating net bone loss. Accordingly, chronic disorders characterized by sustained inflammation—such as diabetes, rheumatoid arthritis, and chronic kidney disease—are frequently associated with lower BMD and increased fracture risk (111). Beyond direct effects on bone cells, chronic inflammation may also disrupt systemic regulators of mineral metabolism; for example, it can impair vitamin D activation (112), thereby compromising calcium absorption and utilization.

Inflammatory cytokines are key drivers of osteoclast activation and bone resorption in osteoporosis (47, 48). Pro-inflammatory mediators such as TNF-α and IL-1 can enhance osteoclastogenesis by upregulating RANKL expression in osteoblast-lineage and stromal cells, thereby strengthening pro-resorptive signaling (49, 50). These cytokines may also reduce osteoprotegerin (OPG) production, shifting the RANKL/OPG balance toward RANKL dominance and further amplifying bone resorption. In parallel, chronic inflammation activates nuclear factor-κB (NF-κB) signaling and promotes the sustained production of additional inflammatory mediators, creating a feed-forward cycle that reinforces osteoclast activity and progressive bone loss. IL-6 exemplifies this dual role: it contributes directly to resorption while also serving as a hallmark mediator in multiple chronic inflammatory diseases. Collectively, these mechanisms drive continuous loss of bone mass and deterioration of BMD. Therefore, therapeutic strategies that block key cytokines or modulate their downstream signaling pathways may help slow osteoporosis progression and offer clinically actionable intervention targets (51–53).

Oxidative stress arises from an imbalance between reactive oxygen species (ROS) generation and antioxidant defense systems and plays a substantial role in the pathogenesis of osteoporosis (23, 63). Excess ROS can induce macromolecular damage, trigger apoptosis, and amplify inflammatory signaling. Accumulating evidence indicates that oxidative stress directly promotes osteoblast apoptosis while enhancing osteoclast differentiation and activity, thereby accelerating bone loss (64–66). Consistent with this, patients with osteoporosis often exhibit elevated circulating or tissue markers of oxidative stress, supporting a close link between redox imbalance and dysregulated bone metabolism.

Mechanistically, oxidative stress influences bone remodeling through multiple pathways. ROS can activate stress- and inflammation-related signaling cascades, including NF-κB and c-Jun N-terminal kinase (JNK), which promote the production of pro-inflammatory mediators and further stimulate osteoclastogenesis (67–69). Oxidative stress also impairs osteogenic potential at the progenitor level: beyond inducing apoptosis in osteoblasts, it can compromise bone marrow mesenchymal stem/stromal cell (BMSC) function and shift lineage commitment toward adipogenesis at the expense of osteogenesis, reducing overall bone-forming capacity (70, 71) (Figures 2A, B). Emerging evidence suggests that alleviating oxidative stress may serve as an adjunctive strategy for the prevention and management of osteoporosis. Interventional approaches that reduce reactive oxygen species (ROS) burden—such as antioxidant supplementation or strategies to enhance the endogenous antioxidant system—may help restore redox homeostasis and preserve skeletal health (Figure 2C).

Cellular senescence is increasingly recognized as a core hallmark of organismal aging and a key biological process contributing to age-associated tissue dysfunction, including musculoskeletal decline (131, 132). In the bone microenvironment, the accumulation of senescent cells—particularly within bone marrow stromal cells (BMSCs), osteoblast-lineage cells, and osteocytes—has been implicated in osteoporosis pathogenesis and age-related bone loss (133–135). With advancing age, persistent stressors such as oxidative stress and mitochondrial dysfunction activate DNA damage responses (DDR) and enforce stable cell-cycle arrest, thereby reducing the proliferative capacity and osteogenic potential of BMSCs and osteoblasts (136–138).Recent high-resolution single-cell atlases have begun to resolve aging-associated niche remodeling at cellular and intercellular-communication levels, identifying senescent-like mesenchymal subclusters alongside age-biased immune populations and altered cell–cell signaling within cranial skeletal stem cell niches, supporting a direct link between senescence programs and niche dysfunction during skeletal aging (139). A defining feature of senescent cells is the acquisition of the senescence-associated secretory phenotype (SASP), a pro-inflammatory and matrix-modulating secretome that includes cytokines such as interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), and interleukin-1β (IL-1β) (133, 140). Mechanistically, SASP output is maintained by inflammatory transcriptional programs, with nuclear factor-κB (NF-κB) serving as a central hub that amplifies cytokine production and sustains a chronic inflammatory milieu (55, 141). Integrative analyses combining scRNA-seq with bulk transcriptomics and cell–cell communication inference further suggest that aging bone marrow is characterized by coordinated shifts in stromal and myeloid compartments (e.g., expansion of BMSCs and macrophages), enrichment of fibrotic and immune-inflammatory programs, and directional paracrine signaling from aged BMSCs that can promote myeloid aging-like phenotypes, collectively reinforcing inflammaging and impaired osteogenesis (142). Consistent with this concept, mechanistic work has uncovered a mechanoinflammatory autocrine loop in BMMSCs whereby loss of Piezo1 signaling enhances Ccl2/CCR2 activation, triggers NF-κB–dependent Lcn2 production, and biases BMSCs toward adipogenesis at the expense of osteogenesis, providing a concrete pathway linking inflammation, cell-fate drift, and osteoporotic bone loss (143).

In bone, a SASP-enriched environment promotes osteoclastogenesis by enhancing RANKL-dependent signaling and disrupting the OPG/RANKL balance, thereby increasing bone resorption (135, 144). Concurrently, SASP-associated inflammation impairs osteoblast differentiation and function and can suppress osteoanabolic pathways (e.g., via inhibition of Wnt signaling), culminating in uncoupled remodeling characterized by increased resorption and decreased formation (145) (Figure 3). Beyond local effects, circulating SASP factors contribute to systemic “inflammaging,” potentially exacerbating osteoporosis progression and increasing susceptibility to age-related comorbidities (132, 146). Importantly, targeting senescent cells has shown therapeutic promise in skeletal aging; for example, clearance of senescent cells prevented age-related bone loss in vivo, supporting senescence as an actionable driver rather than merely a correlate of aging (135). Recent syntheses further consolidate these mechanistic links and summarize emerging senotherapeutic strategies across skeletal pathophysiology (147). Together, senescent cells and their SASP represent key mechanistic mediators of osteoporotic bone loss and compelling targets for future interventions.

A prominent hallmark of skeletal aging is the lineage shift of bone marrow mesenchymal stromal/stem cells (BMSCs): the osteogenic differentiation capacity declines, accompanied by enhanced adipogenic differentiation and expansion of bone marrow adipose tissue (148, 149). At the transcriptional regulatory level, osteogenic modules such as RUNX2/OSX, which are driven by the Wnt/β-catenin and BMP pathways, become less amenable to activation. In contrast, the adipogenic regulatory network centered on PPARγ and C/EBP family members predominates, thereby lowering the threshold for adipogenic differentiation and raising the barrier to osteogenic initiation (61, 150). This cellular reprogramming is not entirely governed by cell-intrinsic factors but is subject to persistent regulation by the aging microenvironment. Inflammatory signals enriched in the senescence-associated secretory phenotype (SASP) and the NF-κB-dependent cytokine milieu can suppress pro-osteogenic pathways and enhance the sensitivity of precursor cells to adipogenic cues. Concurrently, mechano-metabolic inputs—including reduced mechanical loading, altered ECM mechanical properties, and changes in the activity of pathways such as AMPK/mTOR and YAP/TAZ—further skew the differentiation fate of BMSCs from osteogenesis toward adipogenesis (57, 60, 151). Mounting evidence also indicates that BMSC populations themselves exhibit inherent heterogeneity. Aging can amplify the lineage shift by altering the proportion and plasticity of distinct precursor subpopulations (e.g., expansion of certain subpopulations, pre-activation of adipogenic programs, or loss of osteogenic potential). This provides a mechanistic explanation for the interindividual variability in bone marrow fat accumulation and osteogenic decline observed in aging populations (152, 153). In terms of functional outcomes, the osteogenic-to-adipogenic drift limits the supply of osteoblast-lineage cells and favors a “resorption-over-formation” remodeling pattern. This accelerates net bone loss and impairs the integrity of bone microstructure (148, 149).