Natural bioactive compounds in estrogen deficiency-induced osteoporosis: Classification and an integrated gut microbiota-metabolite-immune-bone framework

Abstract

Postmenopausal osteoporosis is primarily driven by estrogen deficiency, while gut dysbiosis, inflammation, microbial metabolites, and immune reprogramming may modify bone loss, although current evidence remains largely preclinical or associative. Natural bioactive compounds, probiotics, fermented foods, and medicinal-food resources have emerged as potential interventions, but existing research is fragmented and rarely distinguishes direct skeletal effects from microbiota-dependent actions. This review proposes a hypothesis-generating microbiota-dependent classification, rather than a validated clinical system, comprising phytoestrogens; polyphenols, flavonoids, and terpenoids; polysaccharides and dietary fibers; probiotics and postbiotics; fermented foods and dietary-pattern interventions; and complex traditional herbal formulations. These interventions may act directly on bone cells or indirectly through microbial biotransformation, intestinal barrier-mmune signaling, endocrine and estrobolome pathways, and epigenetic regulation. We additionally propose a microbiota-etabolite-xtracellular vesicle microRNA-one marrow cell fate axis as an unvalidated regulatory model. Preclinical evidence is strongest for short-chain fatty acid-related and gut barrier-mmune mechanisms, whereas the skeletal roles of bile acids, tryptophan-derived metabolites, microbiota-associated vitamins, and extracellular vesicle microRNAs remain uncertain, particularly in humans. Future studies should prioritize causal validation and microbiota-stratified clinical investigations.

Key Words: Natural bioactive compounds; Postmenopausal osteoporosis; Gut microbiota; Phytoestrogens; Short-chain fatty acids; Th17/Treg; Microrna; Extracellular vesicles; Gut-bone axis

Core Tip: This minireview introduces a microbiota-dependent classification of natural bioactive compounds for estrogen deficiency-induced osteoporosis and proposes an integrated gut-metabolite-immune-bone framework. Rather than acting solely on bone cells, these interventions may reshape microbial metabolites, intestinal barrier function, estrogen availability, immune balance, and post-transcriptional regulation. This framework highlights priorities for causal multi-omics studies, microbiota-stratified trials, and the clinical evaluation of individualized nutrition-based interventions as potential adjunctive strategies.

INTRODUCTION

Osteoporosis (OP) is a systemic skeletal disorder characterized by reduced bone mineral density (BMD), deterioration of bone microarchitecture, impaired bone quality, and increased fracture risk. Screening and diagnostic gaps remain, including delayed recognition of vertebral fractures[1,2]. Postmenopausal OP is primarily driven by declining ovarian function and estrogen production. Estrogen deficiency increases the RANKL/OPG ratio, inflammatory cytokine production, and osteoclast survival while impairing osteoblast and osteocyte survival, thereby disrupting bone remodeling[3-5].

Beyond direct endocrine effects, estrogen deficiency-induced bone loss involves intestinal dysbiosis, impaired gut barrier function, immune activation, altered microbial metabolites, bone marrow immune reprogramming, and cellular senescence[6,7]. The gut microbiota may influence bone through mineral absorption, short-chain fatty acids (SCFAs), bile acid and tryptophan metabolism, vitamin-related pathways, barrier integrity, and immune responses[3-6], although causality in humans remains incompletely established.

Experimental studies link estrogen deficiency-associated dysbiosis to barrier dysfunction, systemic immune activation, and osteoclastogenic signaling, whereas human studies mainly report associations among microbial composition, metabolites, BMD, and bone-turnover markers[8-10].

Natural bioactive compounds, including phytoestrogens, flavonoids, polyphenols, polysaccharides, terpenoids, probiotics, fermented foods, and traditional Chinese medicinal formulations, are being investigated as adjunctive interventions and have improved bone microarchitecture and metabolism in ovariectomized models[11-14]. Their clinical efficacy, safety, dosing, and interactions remain uncertain, and the causal chain linking microbial changes and metabolites to immune regulation and bone-cell function is often incomplete.

Current reviews usually classify natural bioactive compounds by chemical structure, which does not explain their differing dependence on microbial biotransformation or fermentation. A microbiota-dependent classification may better account for interindividual responses to phytoestrogens and the skeletal effects of poorly absorbed polysaccharides.

Existing natural-product reviews generally organize interventions by chemical class and summarize compound-specific osteoprotective effects, whereas gut-bone-axis reviews mainly describe the microbial, metabolic, barrier, and immune pathways involved in OP. In contrast, our framework links the identity and form of each intervention to its degree of microbiota dependence and explicitly distinguishes direct bone-cell effects from effects mediated by microbial biotransformation, fermentation, or other microbiota-associated processes. It further considers evidence strength and translational readiness and places extracellular-vesicle microRNAs within a testable, hypothesis-generating gut-metabolite-immune-bone network.

Natural bioactive compounds may act directly on bone cells or indirectly through microbial biotransformation and systemic mediators. Some combine both routes, whereas poorly absorbed polysaccharides may depend more strongly on microbial metabolism[15,16]; this distinction underpins our framework.

MicroRNAs regulate bone-cell differentiation, osteoclastogenesis, bone marrow mesenchymal stem cell fate, and immune polarization. Host fecal microRNAs can shape microbial composition, while extracellular vesicle-associated microRNAs may connect intestinal, immune, and bone marrow compartments[17,18] and serve as candidate biomarkers or regulators of bone metabolism[19-22].

In ovariectomized rats, a microbiota intervention was associated with BMSC-derived exosomal miR-151-3p and miR-23b-3p, BMSC differentiation, and bone loss[23]. This finding is hypothesis-generating and does not establish causality.

This review therefore develops a microbiota-dependent classification, distinguishes direct bone-cell effects from microbiota-dependent effects, compares mechanisms across compound classes, and considers how microRNAs and extracellular vesicles may integrate the gut-metabolite-immune-bone network. The resulting framework is intended to guide mechanistic research and clinical translation in estrogen deficiency-induced OP.

A structured literature search was conducted in PubMed, EMBASE, Web of Science and Scopus from database inception to August 2026. Search terms encompassed three conceptual domains: Estrogen deficiency or bone loss, gut microbiota and the gut-bone axis and natural bioactive compounds and microbiota-targeted interventions. Human studies, estrogen-deficiency or ovariectomized animal models and mechanistic in vitro studies were included, whereas irrelevant publications, conference abstracts and reports with insufficient information were excluded. Given study heterogeneity, evidence was synthesized narratively using predefined qualitative criteria without formal GRADE assessment or meta-analysis. Evidence strength reflected study design, consistency, directness of skeletal outcomes, and causal validation. Three descriptive categories were applied: (1) Strong preclinical causal evidence, where experimental perturbations such as fecal microbiota transplantation (FMT), microbiota depletion, defined-microbe intervention, or metabolite rescue support causal contributions to skeletal outcomes; (2) Moderate preclinical evidence, comprising multiple supportive studies or convergent mechanisms without sufficient validation of microbiota-dependent causality or clinical efficacy; and (3) Preliminary or hypothesis-generating evidence, based mainly on limited preclinical studies, in vitro experiments, omics associations, or unvalidated mechanistic hypotheses. Human evidence was assessed separately by design and clinical endpoints, distinguishing observational associations from intervention findings. These categories are a qualitative narrative appraisal, not a validated grading instrument; strong experimental evidence does not establish clinical efficacy. Translational readiness was assessed separately through human data availability, clinical endpoint relevance, reproducibility, and unresolved safety or implementation issues.

CONCEPTUAL BASIS

Estrogen deficiency is the principal upstream driver of postmenopausal OP, directly disrupting the balance between bone resorption and bone formation. Intestinal dysbiosis, altered microbial metabolism, impaired barrier function, immune activation, bone marrow stromal-cell dysfunction, and cellular senescence are proposed interacting modifiers rather than equivalent primary causes. Experimental evidence further suggests that estrogen loss may alter intestinal epithelial function, microbial composition, metabolite profiles, and systemic inflammatory signaling. The gut-bone axis is therefore considered here as an interacting network rather than as a simple linear pathway. Within this network, the gut microbiota provides a metabolic and immunological interface between diet, natural bioactive interventions, and bone. Microbial functions include the production of SCFAs, secondary bile acids, tryptophan-derived indoles, TMA/TMAO, selected vitamins, and enzymes involved in intestinal estrogen deconjugation[24]. Although estrobolome-associated β-glucuronidase and sulfatase activities can influence enterohepatic estrogen recycling, their contribution to BMD, fracture risk, and treatment response in postmenopausal women remains incompletely established. Microbiota-associated immune signaling may also involve Th17/Treg balance, inflammatory cytokines, macrophage activation, and RANKL-mediated osteoclastogenesis. Changes in these immune pathways have been associated with more favorable bone microarchitecture in selected preclinical intervention studies.

Natural bioactive interventions may exert direct effects on bone cells or indirect effects involving microbial biotransformation, fermentation, ecological remodeling, and microbial products; these mechanisms are overlapping and should not be interpreted as uniformly causal. The evidence reviewed here is predominantly preclinical, whereas limited human evidence is available from selected observational and intervention studies, including an 8-prenylnaringenin-standardized hop-extract trial and a breath-metabolomics study[25,26]. Ovariectomy models are useful for mechanistic exploration but do not fully reproduce gradual reproductive aging, long-term dietary exposure, multimorbidity, medication use, or the heterogeneous microbiota of postmenopausal women. Moreover, improvements in BMD, bone microarchitecture, bone-turnover markers, or pathway-related biomarkers should not be interpreted as evidence of fracture prevention or established clinical efficacy.

MICROBIOTA-DEPENDENT CLASSIFICATION OF NATURAL BIOACTIVE COMPOUNDS

In this minireview, interventions were classified according to three predefined criteria: The identity and form of the administered intervention, its chemical or biological composition, and its dominant relationship with the gut microbiota. The first criterion distinguished purified natural compounds, poorly absorbed fermentable macromolecules, microbial preparations, fermented foods, dietary-pattern interventions, and complex multicomponent herbal formulations. The second criterion defined the predominant chemical or biological identity of each intervention. The third criterion determined whether the intervention acted primarily through direct host signaling, microbial biotransformation, microbial fermentation, modulation of gut microbial composition or function, or the administration of live microorganisms, inanimate microbial preparations, or defined microbial products.

Using these criteria, phytoestrogens were defined as plant-derived estrogen-like compounds or their gut-derived metabolites that exert direct or microbiota-dependent estrogen receptor (ER) activity. Polyphenols and flavonoids were defined as phenolic secondary metabolites, with flavonoids treated as a subclass of polyphenols. Terpenoids were treated as structurally distinct plant secondary metabolites but were retained in the same broad category of low-molecular-weight natural compounds used in this review. Polysaccharides, oligosaccharides, and dietary fibers were defined as poorly absorbed, fermentable carbohydrates whose principal gut-mediated effects depend on microbial fermentation. Probiotics were defined as live microorganisms, whereas postbiotics comprised preparations of inanimate microorganisms and/or their components that confer a health benefit on the host. Fermented foods were defined as whole-food matrices transformed by microbial fermentation, whereas dietary patterns represented combined exposure to multiple foods and nutrients. Traditional herbal formulations and chemically heterogeneous natural extracts were classified as complex multicomponent interventions when the administered material could not be adequately represented by a single dominant chemical class. When an intervention fulfilled the definitions of more than one category, its primary category was assigned according to the identity and form of the administered material and its dominant microbiota-dependent mechanism. Secondary mechanisms were cross-referenced rather than used to assign the same intervention to an additional category (Table 1).

Table 1 Microbiota-dependent classification of natural bioactive compounds in estrogen deficiency-induced osteoporosis.

Category	Representative compounds or interventions	Microbiota dependency	Major gut-derived mediators	Dominant bone-related mechanisms	Evidence level	Ref.
Phytoestrogens and estrogen-like compounds	Puerarin, genistein, daidzein, S-equol, calycosin, 8-prenylnaringenin	Mixed: Direct estrogen receptor modulation plus microbial biotransformation	S-equol, SCFAs, estrobilome-related β-glucuronidase/sulfatase activity	ERα/ERβ signaling, reduced osteoclastogenesis, associated with greater intestinal barrier integrity, decreased systemic inflammation	Moderate preclinical evidence; limited human intervention evidence	[15,24,25,27-29]
Polyphenols, flavonoids, and terpenoids	Flavan-3-ols, ginkgolide B, taraxasterol, Moringa oleifera leaf constituents, quercetin, resveratrol, curcumin	Moderate to high; many compounds require microbial metabolism or reshape microbiota	Phenolic metabolites, SCFAs, altered bile acid and indole metabolites	Antioxidant and anti-inflammatory effects, RANKL/OPG regulation, NF-κB/NFATc1 suppression, MAPK and NLRP3 modulation	Moderate preclinical evidence; mostly OVX animal studies with 16S/metabolomicsevidence	[11,12,30-33]
Polysaccharides, oligosaccharides, and dietary fibers	Dendrobium officinale polysaccharides, Polygonatum polysaccharides, algal oligosaccharides, dietary fiber-rich foods	High; generally act as prebiotic substrates rather than directly absorbed bone-active molecules	Butyrate, acetate, propionate, isovalerate, SCFA-producing bacteria	Enhanced tight junction proteins, Treg induction, reduced inflammation, Wnt/β-catenin activation, improved osteoblastogenesis	Moderate to high when FMT or metabolite evidence is included	[16,35,49,51]
Probiotics, postbiotics, and defined bacterial strains	Lactobacillus rhamnosus GG, Lactobacillus acidophilus, Lactobacillus sakei CVL-001, Parabacteroides distasonis, Clostridium sporogenes	Direct microbiota- targeted intervention	Butyrate, indole propionic acid, bacterial components, TLR2/GPCR-related signals	Th17/Treg rebalancing, reduced IL-17/TNF-α, improved gut barrier, suppressed RANKL-induced osteoclastogenesis	Moderate preclinical evidence; limited evidence for clinical skeletal efficacy	[13,36-40]
Fermented foods and dietary-pattern interventions	Fermented soybeans, Bacillus subtilis-rich foods, quinoa, tryptophan-restricted intermittent diet	High; effects depend on dietary substratesand microbialecological remodeling	Biotin, SCFAs, soyisoflavone metabolites, altered tryptophan metabolites	Improved BMD and trabecular microarchitecture, reduced M1 macrophage polarization, induced osteoclast apoptosis, improved metabolic status	Preliminary-to-moderate preclinical evidence	[35,42,43]
Complex herbal formulations and natural extracts	Roucongrong Pill, Jiaru fuling traditional Chinese medicine formulas, deer velvet antler protein extract	Variable; often multi-component and multi-pathway	SCFAs, steroid-related metabolites, taurine/hypotaurine metabolites, lipid and nucleotide metabolites	Multi-target regulation of gut microbiota, immune balance, steroid hormone biosynthesis, osteoclast inhibition, osteoblast support	Preliminary or hypothesis-generating evidence; causal attribution remains insufficient	[14,34,44,45]

Interventions were classified according to their predominant gut microbiota-dependent effects and bone-related mechanisms. Evidence levels were qualitatively assessed based on available in vitro, animal, and human studies; categories may overlap. SCFAs: Short-chain fatty acids; OVX: Ovariectomy; FMT: Fecal microbiota transplantation; BMD: Bone mineral density; ER: Estrogen receptor.

Phytoestrogens: Dual regulation through ERs and microbial biotransformation

Phytoestrogens include soy isoflavones, genistein, daidzein, puerarin, icariin, 8-prenylnaringenin, and S-equol. Their defining feature is structural similarity to 17β-estradiol and the ability to bind ERs. This makes them particularly relevant to estrogen deficiency-induced OP. The ER affinity, potency, and ERα/ERβ selectivity of phytoestrogens vary among individual compounds.

For some phytoestrogens, gut microbial biotransformation represents an additional and compound-specific determinant of biological activity rather than a universal requirement for the entire class. Daidzein requires specific intestinal bacteria to be converted into S-equol, a more bioactive metabolite with higher affinity for ERβ[27,28]. The equol-producer phenotype varies across individuals and may explain inconsistent results in soy isoflavone trials. Therefore, the efficacy of phytoestrogens depends not only on intake dose, but also on microbial metabolic capacity, ER expression, host age, baseline hormonal status, and bone turnover state.

Puerarin illustrates the complexity of this class. Although puerarin has phytoestrogenic properties and has been studied for postmenopausal symptoms and OP, its low oral bioavailability suggests that direct systemic exposure may not fully explain its skeletal benefits. In ovariectomized rats, puerarin improved BMD, increased SCFA levels, repaired intestinal mucosal integrity, reduced systemic inflammation, and modulated gut microbial composition[15]. Thus, puerarin may act through a hybrid mechanism involving both estrogen-like signaling and microbiota-mediated restoration of gut-bone homeostasis.

In a model combining obesity and estrogen decline, genistein supplementation was reported to attenuate bone damage and was accompanied by changes in gut microbial composition, intestinal barrier integrity, inflammatory indices, lipid peroxidation, and bile acid metabolism[29]. This suggests that phytoestrogens may be especially relevant in metabolically stressed postmenopausal states, where estrogen deficiency interacts with obesity, insulin resistance, and chronic inflammation.

Therefore, phytoestrogens should be classified as mixed direct-microbiota-dependent compounds. They may bind ERs and influence osteoblast- and osteoclast-related pathways, whereas their biological effects may also depend on microbial conversion and estrobolome function. The magnitude of these contributions in humans remains uncertain.

Furthermore, clinical findings are also inconsistent: Benefits reported for soy foods or S-equol-producing subgroups cannot be assumed for all individuals or extrapolated to purified high-dose supplements. Differences in compound identity, dose, duration, baseline diet, menopausal stage, estrogen-receptor subtype activity, and equol-producer status may account for divergent outcomes. Concentrated phytoestrogen supplements may also produce tissue-specific estrogenic effects and should be used cautiously in women with estrogen-sensitive malignancies or those receiving tamoxifen or aromatase inhibitors; their long-term endocrine and oncological safety remains insufficiently defined.

Polyphenols, flavonoids, and terpenoids: Antioxidant, anti-inflammatory, and microbial metabolite-mediated effects

Polyphenols, flavonoids, and terpenoids include flavan-3-ols, quercetin, resveratrol, curcumin, ginkgolide B, taraxasterol, Moringa oleifera leaf constituents, and many other plant-derived molecules. These structurally and pharmacokinetically heterogeneous compounds are discussed together for organizational convenience and should not be interpreted as a mechanistically unified class. These compounds have traditionally been interpreted through antioxidant, anti-inflammatory, and osteoregulatory mechanisms. They may promote osteoblastogenesis, inhibit osteoclastogenesis, regulate the RANKL/OPG axis, activate Wnt/β-catenin or MAPK pathways, and reduce oxidative stress.

However, many polyphenols have limited small-intestinal absorption and undergo extensive microbial metabolism in the colon. Their microbial metabolites may be more bioavailable and biologically active than parent compounds. Flavan-3-ols, for instance, largely reach the colon and are metabolized by gut bacteria into phenolic metabolites. In ovariectomized mice, flavan-3-ols improved bone quality and were associated with changes in Clostridia, Ruminococcus, and Lachnospiraceae-related taxa. Transcriptomic analysis suggested involvement of Adrb3, Gdf10/BMP3, Fosb, Cxcl2, and PPARα/PGC-1/UCP1-related signaling[30]. These findings indicate that polyphenol-microbiota interactions may regulate bone metabolism through both systemic metabolism and local bone signaling.

Ginkgolide B was reported to attenuate bone loss in ovariectomized animals, accompanied by greater intestinal barrier integrity, lower inflammatory indices, changes in the Th17/Treg balance, and altered abundances of taxa such as Lactobacillaceae and Prevotellaceae[31]. Taraxasterol was reported to attenuate bone loss, accompanied by reduced NLRP3 inflammasome activity, increased expression of tight-junction proteins, and greater Parabacteroides abundance[32]. Moringa oleifera leaf improved bone parameters and altered gut microbiota while regulating MAPK-related signaling[33]. These studies provide representative compound-specific preclinical examples[34] rather than evidence of a mechanism shared by all polyphenols, flavonoids, and terpenoids.

The limitation is that many studies still rely on correlations between altered bacterial taxa and bone indices. To determine whether microbiota is necessary, future studies should include antibiotic depletion, germ-free models, FMT, and metabolite rescue experiments.

Polysaccharides, oligosaccharides, and dietary fibers: Prototypical microbiota-mediated compounds

Many polysaccharides and dietary fibers are poorly absorbed in the upper intestine and are therefore available for colonic microbial fermentation. This makes the gut microbiota a plausible mediator of their biological effects, but the degree of microbiota dependence is likely to vary among individual interventions and should be established experimentally rather than assumed for the entire class. This process produces SCFAs and alters the abundance of beneficial bacteria, thereby affecting gut barrier function, immune regulation, and bone metabolism.

Dendrobium officinale polysaccharides provide one of the strongest examples. DOP administration attenuated ovariectomy-induced bone loss, improved bone turnover and inflammation markers, activated Wnt/β-catenin signaling, preserved intestinal barrier function, increased SCFA-producing genera such as Allobaculum and Clostridia_UCG-014, and restored butyric and isovaleric acid levels. Crucially, FMT from DOP-treated donors reproduced the skeletal phenotype in recipient mice, providing relatively strong evidence for a causal contribution of the DOP-modulated gut microbiota[16]. However, fecal or intestinal SCFA concentrations do not directly reflect their rates of production or absorption and do not by themselves demonstrate increased systemic or bone-tissue exposure. The FMT findings make DOP a compelling compound-specific preclinical example of a causal microbial contribution; however, this result should not be generalized to all polysaccharides, oligosaccharides, or dietary fibers.

Other dietary and plant-derived interventions also support this mechanism. Quinoa administration was associated with more favorable bone microstructure and bone-turnover indices in ovariectomized rats, together with changes in intestinal barrier function and gut microbial composition[35]. Polysaccharide-rich or fiber-containing interventions have been reported to increase SCFA levels and tight-junction protein expression and to be associated with lower inflammatory indices and a more favorable bone formation/resorption balance.

This class should therefore be considered “microbiota-first” rather than “bone-first”. The key research question is not whether the parent polysaccharide reaches bone, but which microbial communities ferment it, which metabolites are produced, which immune cells respond, and how these systemic signals reshape bone remodeling.

Probiotics, postbiotics, and defined bacterial strains: Toward precision microbial bone therapy

Probiotics and defined bacterial strains represent the most direct microbiota-targeted interventions. Lactobacillus rhamnosus GG, Lactobacillus acidophilus, Lactobacillus sakei, Parabacteroides distasonis, and Clostridium sporogenes have all been investigated in estrogen deficiency-induced bone loss models.

LGG ameliorated OP in ovariectomized rats by improving bone microarchitecture, regulating gut microbiota, restoring intestinal barrier markers, and correcting Th17/Treg imbalance in gut and bone marrow[36]. Lactobacillus acidophilus further refined this immune mechanism by showing that estrogen-deficient inflammatory conditions disrupt tTreg and pTreg balance, and that LA restores RORγT-negative pTreg and Th17 homeostasis in a butyrate-mediated manner[37]. Parabacteroides distasonis treatment was reported to increase bone strength and improve trabecular microarchitecture and was accompanied by higher butyrate levels, altered barrier markers, lower inflammatory cytokine levels, and Th17/Treg rebalancing[38].

Lactobacillus sakei CVL-001 and its metabolites inhibited RANKL-induced osteoclast differentiation through cooperative TLR2 and GPCR-dependent pathways[39]. This finding is important because it shows that probiotic effects may involve not only SCFAs, but also bacterial components and host pattern-recognition receptors. Clostridium sporogenes-derived indole propionic acid suppressed osteoclast formation by activating pregnane X receptor and modulating the PXR/P65 interaction[40]. This moves probiotic research from whole-organism effects toward defined bacterial metabolite mechanisms.

Evidence for probiotic effects on bone is predominantly preclinical, and limited human studies do not yet establish strain-specific fracture prevention or long-term clinical efficacy[13,41]. Their effects are strain-specific, metabolite-specific, and immune-context-dependent. Future research should define the bacterial strain, viability, colonization potential, metabolite output, receptor pathway, dose, and duration associated with beneficial skeletal outcomes.

Fermented foods, dietary patterns, and nutrient restriction: Ecological rather than single-compound interventions

Fermented foods and dietary interventions are more complex than purified compounds because they modify gut ecology through multiple nutritional and microbial inputs. In ovariectomized rats, long-term administration of fermented soybeans rich in Bacillus subtilis was associated with more favorable glucose metabolism, BMD, and bone microstructure, together with altered gut microbial composition[42]. This effect may reflect combined contributions of soy isoflavones, microbial fermentation products, Bacillus species, and altered intestinal ecology.

Tryptophan-restricted intermittent diet provides a different but highly innovative model. Instead of adding a compound, it modifies substrate availability for microbial metabolism. In estrogen deficiency-induced OP, intermittent tryptophan restriction was reported to alter gut microbial composition, increase Akkermansia and Bifidobacterium abundance and biotin levels, reduce M1 macrophage polarization, increase mature osteoclast apoptosis, and produce more favorable bone microarchitecture.

These findings should not be interpreted as support for unsupervised dietary restriction. Evidence for intermittent tryptophan restriction is currently based on a specific preclinical protocol[43], and it is unknown whether comparable skeletal benefits occur in humans. Prolonged or poorly designed restriction could reduce overall protein and micronutrient adequacy, accelerate loss of lean mass, worsen frailty, and indirectly increase fall and fracture risk in older women. Clinical translation should therefore use time-limited, nutritionally complete protocols and monitor body weight, muscle mass, dietary adequacy, and skeletal outcomes.

These studies shift the field from pharmacological thinking toward ecological thinking. Bone health may be influenced not only by specific compounds, but also by long-term dietary patterns that shape microbial metabolic capacity. Future clinical research should therefore integrate dietary assessment, microbiome profiling, metabolomics, and bone outcomes.

Complex herbal formulations and natural extracts: Multi-target potential and mechanistic ambiguity

Traditional Chinese medicine formulations, herbal extracts, fungal-derived products, animal-derived protein extracts, and multi-component natural preparations have shown osteoprotective potential[12,14]. Their advantage lies in multi-target regulation, but their limitation lies in mechanistic ambiguity.

Roucongrong Pill improved bone strength, BMD, bone turnover markers, gut microbiota, and metabolite profiles in ovariectomized rats. Multi-omics, MetOrigin analysis, network pharmacology, molecular docking, and molecular dynamics suggested involvement of taurine and hypotaurine metabolism, steroid hormone biosynthesis, ESR1/ESR2, SHBG, PTGER4, CNR2, CTSK, and androgen receptor-related targets[44]. Isaria felina modulated Bacteroides and Ruminococcus, altered nucleotide and lipid metabolism, restored Th17/Treg balance, and reduced IL-17 and TNF-α[45]. Reported benefits of complex herbal formulations are also difficult to compare because source species, processing, extraction, dose, and batch composition are often incompletely standardized. Positive findings for one preparation therefore cannot be generalized to products bearing the same name. Potential harms include contamination or adulteration, unpredictable systemic exposure, cumulative hepatic or renal toxicity, and interactions with anti-OP agents or medications used for age-related comorbidities. Chemical fingerprinting, batch-release criteria, toxicological assessment, and prospective adverse-event monitoring should accompany mechanistic and efficacy studies. These studies are valuable for hypothesis generation, but complex formulations require stronger causal validation. Network pharmacology and molecular docking can identify plausible targets, but they cannot prove that a specific compound, bacterium, metabolite, or receptor is necessary. High-quality research should deconstruct complex formulations into functional modules: Key compound groups, key microbial functions, key metabolites, key immune pathways, and key bone cell targets. The microbiota-dependent classification of natural bioactive compounds and their bone-protective mechanisms are illustrated in Figure 1.

Figure 1 This figure systematically summarizes the microbiota-dependent classification of natural bioactive compounds in the context of estrogen deficiency, including phytoestrogens, polyphenols and flavonoids, polysaccharides and dietary fibers, probiotics/postbiotics, dietary interventions, and complex herbal formulations or natural extracts. SCFAs: Short-chain fatty acids; FMT: Fecal microbiota transplantation.

Safety, herb-drug interactions, and interindividual variability

“Natural” does not necessarily mean safe, particularly when bioactive compounds are administered as purified molecules, concentrated extracts, or multicomponent herbal preparations. Their safety may vary with dose, treatment duration, product quality, host metabolism, and concomitant medications. Variability in botanical sources, extraction methods, chemical composition, contamination, adulteration, and batch consistency may further complicate the safety assessment of herbal products.

Safety assessment should distinguish among intervention forms. Food-based interventions should be evaluated for nutritional adequacy, adherence, and possible nutrient displacement. Concentrated extracts and supplements require product-specific assessment of dose, bioavailability, organ toxicity, and drug interactions because their exposure cannot be equated with customary food intake. Multi-component traditional formulations additionally require botanical authentication, compositional standardization, contaminant testing, and batch-quality control.

Herb-drug interactions are particularly relevant in postmenopausal women receiving anti-OP agents or medications for age-related disorders. Natural compounds may affect drug absorption, metabolism, transport, or enterohepatic recycling, whereas concomitant medications may alter the gut microbiota and thereby influence microbial biotransformation and systemic exposure[24]. Phytoestrogens and microbiota-derived metabolites such as S-equol may also exert tissue- and ER subtype-dependent effects that vary according to hormonal status, dose, treatment duration, and equol-producer phenotype[27,28]. Particular caution is warranted with concentrated phytoestrogen supplements in patients with estrogen-sensitive malignancies or those receiving tamoxifen or aromatase inhibitors, because evidence from conventional soy-food intake cannot be extrapolated to high-dose supplements and their long-term oncological safety remains uncertain. Multivalent-cation supplements may reduce oral bisphosphonate absorption when taken concurrently, whereas interactions between concentrated natural products and other antiresorptive or anabolic OP medications remain poorly characterized.

Differences in baseline microbiota, diet, age, body mass index, medication exposure, and microbial metabolic capacity may contribute to substantial variability in efficacy and adverse responses. Because long-term safety data for concentrated natural products and traditional herbal formulations remain limited, future trials should use standardized preparations, document concomitant medications, prospectively monitor adverse events and hepatic, renal, hematological, and endocrine safety, and include adequate follow-up.

DIRECT BONE CELL EFFECTS VERSUS MICROBIOTA-MEDIATED INDIRECT EFFECTS

A central question in this field is whether natural bioactive compounds act directly on bone cells or indirectly through gut microbiota-derived signals. This distinction is critical for mechanism, clinical translation, dosing, and patient stratification. We propose four mechanistic patterns (Table 2).

Table 2 Mechanistic modules linking natural bioactive compounds, gut microbiota, and bone remodeling.

Mechanistic module	Upstream trigger or intervention	Key microbial or metabolic mediators	Immune/endocrine/cellular pathway	Skeletal consequence	Main uncertainty	Ref.
Gut barrier-bone module	Estrogen deficiency, dysbiosis, natural compound intervention	LPS, microbial-associated molecular patterns, tight junction proteins	TLR4/NF-κB activation, systemic inflammation, TNF-α and IL-6 release	Increased osteoclastogenesis when barrier is impaired; bone protection when barrier is restored	Whether barrier repair is primary or secondary to microbiota remodeling	[8,15,16,32,36-38]
SCFA-mediated immune-bone module	Prebiotics, polysaccharides, probiotics, fiber-rich diets	Butyrate, acetate, propionate, SCFA-producing bacteria	FFAR2/3 signaling, Treg induction, reduced Th17 response, Wnt/β-catenin activation	Reduced bone resorption and enhanced bone formation	Optimal SCFA profile, dose, and target cell specificity remain unclear	[16,37,38,49,51]
Tryptophan-indole metabolite module	Tryptophan metabolism, defined bacteria, dietary restriction	IPA, IAA, IA, biotin-associated microbial changes	PXR/P65, AhR/NF-κB, AKT/NFATc1/ROS, M1 macrophage polarization, osteoclast apoptosis	Suppressed osteoclast formation and improved trabecular bone microarchitecture	Context-dependent effects of tryptophan restriction and indole metabolites need clarification	[40,43,46,47]
Bile acid-bonemodule	Natural compounds, estrogen deficiency, gut microbiota remodeling	Primary and secondarybile acids, DCA, TGR5-related signals	TGR5/NF-κB, BMSC differentiation, inflammation regulation	Reduced inflammatory bone loss and improved bone remodeling	Human relevance and bile acid subtype-specific effects remain insufficiently defined	[29,49]
Estrobolome-phytoestrogen module	Menopause, phytoestrogen intake, long-term medication	β-glucuronidase/sulfatase activity, S-equol, estrogen recycling metabolites	ERβ signaling, estrogenavailability, intestinal deconjugation, enterohepatic circulation	Potential preservation of bone mass and reduction of bone turnover	Equol-producer status and medication-induced estrobolome changes are under-tested clinically	[24,27,28,41]
TMAO-PERK/mitochondrial stressmodule	Choline/carnitine metabolism, dysbiosis, TMAO exposure	TMAO, TMA-producing bacteria, FMO3 activity	PERK/ATF4/ATF5, ER stress, mitochondrial unfolded protein response, osteoblast proteostasis	Either bone damage or protection depending on dose, source, and metabolic context	Direction of effect is contradictory and likely context-dependent	[48,53]
miRNA/extracellular vesicle module	Gut dysbiosis, natural bioactive compounds, microbial metabolites	Host fecal miRNAs, BMSC-derived exosomal miRNAs, circulating EV-miRNAs	Post-transcriptional regulation of osteogenesis, adipogenesis, immune polarization, and osteoclastogenesis	Potential bridge between microbial signals and bone marrow cell fate	Still emerging; causal links in natural compound studies are largely missing	[17-23,54]

SCFAs: Short-chain fatty acids; IPA: Indole-3-propionic acid; IA: Indoleacrylic acid; IAA: Indole-3-acetic acid; LPS: Lipopolysaccharide.

PATTERN I: DIRECT BONE CELL-TARGETING EFFECTS

In this pattern, the parent compound or an absorbable metabolite enters systemic circulation and directly affects osteoblasts, osteoclasts, osteocytes, or bone marrow mesenchymal stem cells. Direct mechanisms include activation of ERs, Wnt/β-catenin, BMP/Smad, PI3K/Akt, Nrf2/HO-1, or suppression of RANKL/RANK/TRAF6/NF-κB/NFATc1, MAPK, reactive oxygen species (ROS), and NLRP3 signaling.

Phytoestrogens and selected microbial metabolites are most likely to exhibit direct bone cell effects. S-equol can act as an ERβ agonist[27]. Individual tryptophan-derived metabolites should be distinguished because they are not mechanistically interchangeable. Indole-3-propionic acid was reported to suppress osteoclast differentiation through activation of PXR and modulation of PXR/P65 signaling[40]. Indoleacrylic acid was reported to act through the AhR/NF-κB pathway[46], whereas indole-3-acetic acid was associated with inhibition of the AKT/NFATc1 pathway and reduced ROS production[47]. These findings are derived predominantly from cell and animal models, and their physiological exposure and skeletal relevance in postmenopausal women remain uncertain. TMAO can directly affect osteoblast proteostasis through PERK/ATF5 and mitochondrial unfolded protein response pathways[48]. Such findings suggest that selected microbial metabolites may directly influence skeletal cells after entering the circulation, although their physiological exposure and relevance in humans remain uncertain. However, direct in vitro activity alone does not prove that direct action is dominant in vivo. Physiological concentration, bioavailability, metabolism, tissue distribution, and microbiota dependence must be considered.

PATTERN II: MICROBIAL CONVERSION INTO ACTIVE METABOLITES

In this pattern, the parent compound or dietary substrate is converted by gut bacteria into bioactive metabolites that mediate skeletal effects. Examples include daidzein-to-S-equol conversion, dietary fiber-to-SCFA fermentation, tryptophan-to-indole derivative conversion, bile acid transformation, and choline/carnitine-to-TMAO metabolism[28,40,46,47,49]. Daidzein-to-S-equol conversion and dietary fiber fermentation are well-characterized microbial processes, although their contribution to treatment-related bone outcomes varies among individuals. The effects of SCFAs on immune regulation and bone remodeling have substantial preclinical support. In contrast, receptor-specific skeletal effects of tryptophan-derived indoles, bile acid metabolites, and trimethylamine N-oxide have been demonstrated mainly in cell and animal models. Their physiological concentrations, tissue exposure, dose-response relationships, and causal relevance to human postmenopausal OP remain insufficiently defined.

This pattern explains interindividual variability. Two individuals consuming the same phytoestrogen may produce different levels of S-equol. Two individuals consuming similar fibers may generate different SCFA profiles. The same choline intake may lead to different TMAO exposure depending on gut microbial composition and hepatic FMO3 activity. Therefore, clinical response to natural compounds cannot be predicted solely by intake dose. Microbial metabolic phenotype must be incorporated.

Bioavailability, pharmacokinetics and metabolic attribution

The biological effects of orally administered natural compounds should not automatically be attributed to the parent compounds. Their systemic activity is determined by solubility, intestinal absorption, epithelial permeability, host phase I and phase II metabolism, microbial biotransformation, enterohepatic recycling, and the concentrations ultimately achieved in target tissues. Poorly absorbed compounds may exert substantial biological effects after conversion by gut microorganisms into more absorbable or biologically active metabolites. For example, daidzein can be converted into S-equol, dietary fibers are fermented into SCFAs, tryptophan is transformed into indole derivatives, and colonic metabolism of polyphenols produces phenolic metabolites that may differ from their parent compounds in bioavailability and receptor activity. Fecal or luminal SCFA concentrations do not directly reflect their rates of production or absorption. Conversely, direct effects observed using high concentrations of parent compounds in vitro may not occur in vivo if comparable circulating or bone-tissue exposure cannot be achieved. Therefore, observed skeletal effects should be attributed to the parent compound, a host-derived metabolite, or a microbial metabolite only when supported by pharmacokinetic and tissue-exposure data. Future studies should quantify parent compounds and metabolites in the intestinal lumen, circulation, and bone-related tissues; establish time-resolved concentration-effect relationships; and use isotope tracing, microbiota depletion, metabolite supplementation, and direct comparisons between parent compounds and metabolites to determine the biologically active molecular species.

PATTERN III: GUT BARRIER-IMMUNE-INFLAMMATION-MEDIATED EFFECTS

In this pattern, estrogen deficiency and gut microbial dysbiosis may impair intestinal epithelial integrity by reducing mucus protection and disrupting tight-junction proteins, including ZO-1, occludin, and claudins. Increased intestinal permeability may then facilitate the translocation of lipopolysaccharide and other microbial-associated molecular patterns into the circulation. These signals can activate Toll-like receptor and NF-κB pathways in intestinal and systemic immune cells, increase the production of TNF-α, IL-1β, IL-6, and IL-17, and promote NLRP3 inflammasome activation. This sequence provides a mechanistic link between intestinal barrier dysfunction, endotoxin exposure, and systemic osteoclastogenic inflammation[8,31,32,36-38].

The resulting inflammatory environment may shift the Th17/Treg balance toward Th17 responses and promote pro-inflammatory macrophage polarization. Th17 cells can supply IL-17, TNF-α, and RANKL, whereas pro-inflammatory macrophages provide cytokines and osteoclast precursors that enhance RANK signaling. At the same time, reduced Treg activity weakens IL-10- and TGF-β-mediated immunosuppression. These changes can increase the RANKL/OPG ratio in osteoblast-lineage, stromal, and immune cells, thereby promoting osteoclast differentiation and bone resorption. Persistent inflammatory and oxidative signals may also impair osteoblast differentiation and shift bone marrow mesenchymal stem cells away from osteogenic commitment, further uncoupling bone resorption from bone formation. The JCI study showing microbial-dependent trafficking of intestinal TNF-positive T cells and Th17 cells to bone marrow provides strong mechanistic support for this model[8]. However, this network should not be interpreted as a unidirectional cascade. Intestinal permeability, microbial signals, Th17/Treg balance, macrophage polarization, NLRP3 activity, and RANKL/OPG signaling interact through multiple feed-forward and feedback loops.

PATTERN IV: MIXED NETWORK REGULATION

Most natural compounds likely follow mixed patterns. Puerarin may combine phytoestrogenic signaling with SCFA-mediated intestinal barrier repair[15]. FMT findings suggest that the microbiota contributes causally to the skeletal effects of DOP, with SCFAs and Wnt/β-catenin signaling identified as candidate downstream mediators[16]. Ginkgolide B combines intestinal barrier restoration, Th17/Treg rebalancing, and microbial modulation[31]. Taraxasterol combines NLRP3 suppression, gut barrier protection, and microbiota regulation[32].

The key is not to force every compound into a single category, but to determine the relative contribution of each pathway. Mechanistic studies should therefore include in vitro bone cell assays, antibiotic-treated animals, germ-free or gnotobiotic models, FMT, defined bacterial supplementation, metabolite rescue, and receptor or pathway blockade. Such experiments are required to determine whether the microbiota is necessary, sufficient, or merely associated with the observed skeletal effect. The mechanisms by which natural bioactive compounds regulate bone metabolism directly or indirectly through the gut microbiota are detailed below (Figure 2).

Figure 2 Direct and microbiota-mediated mechanisms of natural bioactive compounds in postmenopausal osteoporosis. Natural bioactive compounds may influence bone remodeling through direct effects on osteoblasts, osteoclasts, and bone marrow mesenchymal stem cells, microbial biotransformation, and the gut barrier-immune-inflammatory network. Solid arrows indicate principal pathways supported mainly by preclinical evidence. Bidirectional arrows indicate proposed, associative or feedback relationships. LPS: Lipopolysaccharide; ER: Estrogen receptor; SCFAs: Short-chain fatty acids.

SHARED AND DISTINCT BONE METABOLIC MECHANISMS

Shared convergence on bone formation and bone resorption

Despite diverse upstream mechanisms, most natural bioactive compounds converge on the balance between bone formation and bone resorption. On the antiresorptive side, compounds commonly reduce RANKL expression, increase OPG, suppress RANKL-induced NF-κB, MAPK, AKT, and NFATc1 signaling, decrease TRAP-positive osteoclast number, and reduce bone resorption markers such as CTX-1 and TRACP-5b. On the anabolic side, they may increase Runx2, ALP, osteocalcin, COL1A1, OPG, Wnt/β-catenin, BMP/Smad, and PI3K/Akt signaling, thereby enhancing osteoblast differentiation and mineralization[11,12,50].

Phytoestrogens preferentially target hormonal compensation and ER-related pathways

Phytoestrogens are mechanistically distinct because they address the upstream endocrine deficiency that defines postmenopausal OP. Through ER modulation, they may directly affect osteoblasts, osteoclasts, osteocytes, and immune cells. Yet their effects depend heavily on microbial metabolism, particularly in the case of daidzein-to-equol conversion. Therefore, phytoestrogen research should integrate ER pharmacology and microbiome phenotyping rather than treating them separately.

Polysaccharides and fibers preferentially target SCFAs, barrier integrity, and Wnt-immune coupling

Polysaccharides and fibers primarily operate through gut microbial fermentation. Their mechanisms emphasize SCFAs, tight junction proteins, anti-inflammatory effects, calcium absorption[51], Treg induction, and downstream activation of osteogenic pathways such as Wnt/β-catenin. Their advantage is long-term safety and dietary feasibility. Their limitation is dependence on baseline microbiota and slow, ecology-based effects.

Probiotics preferentially target immune rebalancing

Probiotics and defined bacteria often act through immune rebalancing, especially the Th17/Treg axis. Lactobacillus species and Parabacteroides distasonis modulate Treg and Th17 populations, reduce IL-17 and TNF-α, increase butyrate, and improve gut barrier integrity[36-38]. Some strains also signal through TLR2 and GPCR pathways[39]. Thus, probiotics may be particularly effective in inflammatory phenotypes of postmenopausal OP.

Microbial metabolites preferentially target receptor-specific and cell fate pathways

SCFAs, bile acid metabolites, tryptophan-derived indoles, TMAO, and selected microbiota-associated vitamins are candidate functional mediators rather than uniformly established skeletal regulators. Among the microbial metabolites discussed in this review, SCFAs have comparatively greater preclinical support for links among microbial fermentation, immune regulation, and skeletal remodeling. Indole derivatives have shown receptor-specific inhibition of osteoclast differentiation in cell and animal studies, whereas evidence for bile acid signaling, microbiota-derived vitamins, and TMAO in bone remains more preliminary or context dependent. In particular, apparently opposing effects of TMAO across experimental models indicate that dose, source, host metabolic state, and disease stage must be considered. None of these pathways has yet been validated as a causal mediator of fracture reduction or long-term bone preservation in humans. They can signal through FFAR2/3, TGR5, PXR, AhR, PERK, NF-κB, AKT, and mitochondrial pathways. Their mechanisms are more specific but also more context dependent. For example, TMAO may be harmful under excessive exposure by impairing osteoblast proteostasis[48], whereas moderate choline-derived TMAO may protect bone in a specific estrogen-deficient setting through PERK-related adaptive responses[52,53]. This duality highlights the need for dose-response, source-dependent, and host-context analysis. These observations should not be reduced to a simple beneficial-or-harmful label. Dietary precursor availability, microbial TMA production, hepatic FMO3 activity, renal clearance, circulating concentration, host cardiometabolic state, and disease stage may all modify the response and complicate causal interpretation.

MICRORNAS AND EXTRACELLULAR VESICLES: A PROPOSED REGULATORY LAYER IN THE GUT-BONE AXIS

MicroRNAs represent a promising but underdeveloped layer in microbiota-bone research. They regulate gene expression post-transcriptionally and can influence osteoblast differentiation, osteoclast formation, BMSC adipogenic/osteogenic fate, immune cell polarization, and extracellular vesicle-mediated communication.

Host MicroRNAs may shape gut microbiota

A landmark study demonstrated that fecal microRNAs are abundant in mouse and human fecal samples, are present within extracellular vesicles, and can shape gut microbiota composition through interspecies gene regulation[17]. Subsequent reviews proposed that microRNAs and extracellular vesicles in the gut may act as host modulators of the microbiome[18]. This introduces a bidirectional regulatory concept: Gut microbiota may affect host microRNAs, and host microRNAs may also shape microbial ecology.

Micrornas may link nutrient utilization, microbiota, and bone metabolism

One uploaded study on phosphorus utilization integrated host miRNA-mRNA networks with gut microbiome data and identified miRNA-16-5p, miR-142b-5p, miR-148a-3p, CTDSP1, SMAD3, IGSF10, Bacteroides, and Alistipes as key indicators of host-microbiome interactions related to phosphorus utilization and bone/mineral metabolism[54]. Although this study was not a postmenopausal OP model, it supports the broader concept that host microRNAs may connect gut microbial ecology with mineral metabolism and skeletal phenotypes.

BMSC-derived exosomal micrornas may translate microbial signals into bone marrow cell fate

The most directly relevant uploaded study examined gut microbiota-mediated regulation of exosomal microRNAs from bone marrow mesenchymal stem cells in ovariectomized rats. It identified miR-151-3p and miR-23b-3p as candidate mediators and reported that Ruminococcus administration was associated with changes in BMSC osteogenic/adipogenic differentiation and reduced bone loss in ovariectomized rats[23]. These findings provide preliminary preclinical evidence for an association among gut microbial alterations, BMSC-derived exosomal microRNAs, and bone marrow stromal cell fate. However, because the study was conducted in ovariectomized rats and did not fully establish the necessity or sufficiency of the identified microRNAs through targeted gain and loss-of-function experiments.

Micrornas as biomarkers and therapeutic targets

Circulating and extracellular vesicle-derived microRNAs have been proposed as biomarkers for bone-related diseases, including OP[19-21]. MSC-derived small extracellular vesicle microRNAs are also being investigated for bone regeneration and skeletal disorders[22]. These studies suggest that microRNAs may provide both diagnostic and therapeutic opportunities. In the context of natural bioactive compounds, microRNAs could help explain how gut-derived metabolites induce durable changes in osteoblastogenesis, osteoclastogenesis, and bone marrow immune niches.

Current human evidence supports extracellular vesicle-associated microRNAs primarily as candidate biomarkers rather than as established mediators of microbiota-dependent bone remodeling. A small cross-sectional study identified distinct circulating exosomal microRNA profiles between postmenopausal women with OP or fragility fractures and women with normal bone mass[55]. Another study comprising a discovery cohort and an independent validation cohort reported that a three-exosomal-miRNA panel consisting of miR-34a-5p, miR-9-5p, and miR-98-5p achieved an area under the receiver operating characteristic curve of 0.734 for the diagnosis of postmenopausal OP; however, this remained a diagnostic observational study[56]. More recently, a study involving 272 postmenopausal women identified plasma exosomal microRNA signatures associated with OP and osteopenia[57]. Nevertheless, these studies did not simultaneously evaluate gut microbial composition, extracellular vesicle cargo, target-cell uptake, and longitudinal skeletal outcomes. They therefore cannot establish that extracellular vesicle-associated microRNAs transmit microbiota-derived signals to osteoblasts, osteoclasts, or BMSCs in humans. An exploratory study of modified Qing’E Formula also reported alterations in serum exosomal microRNA profiles after treatment. However, it included only 12 patients with postmenopausal OP and six healthy controls, used a two-week intervention, and relied mainly on bioinformatic pathway prediction for mechanistic interpretation; consequently, it cannot establish clinical efficacy or microbiota-mediated causality[58]. Thus, the clinical relevance of the proposed microbiota-extracellular vesicle microRNA-bone axis remains to be established.

Proposed hypothesis

On the basis of the available evidence, we propose a testable, rather than established, “microbiota-metabolite-extracellular vesicle microRNA-bone marrow cell fate” axis. In this hypothetical model, natural bioactive compounds may alter gut microbial composition and metabolite production, which may subsequently influence extracellular vesicle release or microRNA cargo in intestinal, immune, or stromal cells. These vesicles could, in turn, modulate the osteogenic-adipogenic differentiation of BMSCs or the osteoclastogenic potential of macrophage-lineage cells. At present, however, no human study has demonstrated this complete causal sequence. Validation will require microbiota perturbation combined with extracellular vesicle tracking, cell-specific microRNA gain- and loss-of-function experiments, and prospective human studies integrating microbiome, extracellular vesicle, and skeletal phenotyping (Figure 3).

Figure 3 This figure presents a mechanistic hypothesis in which miRNAs and extracellular vesicles participate in the regulation of the gut-bone axis. Host-derived miRNAs may be transported into the gut via extracellular vesicles and influence microbial composition by regulating microbial gene expression, while dietary nutrients and gut microbial metabolism may further modulate the balance between osteogenic and adipogenic differentiation through host miRNA-mRNA networks and bone marrow mesenchymal stem cell-derived exosomal miRNAs.

MAJOR LIMITATIONS OF CURRENT RESEARCH

Lack of systematic classification

Current studies frequently evaluate individual compounds in isolation, resulting in descriptive rather than conceptually integrated evidence. Classification based solely on chemical structure does not adequately capture differences in bioavailability, microbial dependence, biotransformation, metabolite production, or host targets (Table 3).

Table 3 Evidence strength, limitations, and future directions for natural bioactive compounds targeting the gut-bone axis.

Evidence category	What is currently supported	Representative examples	Main limitations	Recommended future direction	Translational value
Human intervention and observational evidence	Gut microbiota, metabolites, phytoestrogens, and selected natural products are associated with BMD or bone turnover markers	8-prenylnaringenin trial, PMOP gut microbiota/metabolite studies, breast metabolomics, menopause microbiota studies	Small sample size, short duration, heterogeneous baseline microbiota, limited fracture endpoints	Microbiome-stratified RCTs with BMD, bone turnover, bone quality, and fracture- related outcomes	Limited and heterogeneous human evidence; clinical efficacy remains unestablished
Strong causal animal evidence	Microbiota can mediate or transfer osteoprotective effects	DOP-FMT model, OVX- induced intestinal T cell trafficking, defined bacteria/metabolite studies	Mainly rodent-based; translational relevance to human menopause remains uncertain	Combine FMT, antibiotics, gnotobiotic models, metabolite rescue, and receptor blockade	Mechanistic relevance in experimental models; human translation remains uncertain
Probiotic and defined strain evidence	Specific strains can improve bone loss through immune and barrier regulation	LGG, Lactobacillus acidophilus, Lactobacillus sakei, Parabacteroides distasonis, Clostridium sporogenes	Strain specificity, colonization stability, dose, duration, and long-term safety unclear	Define strain-level metabolic outputs and host-response phenotypes	Potential for clinical investigation; strain-specific efficacy and safety require validation
Natural compound + OVX + 16S/metabolomics evidence	Many natural compounds improve bone parameters and reshape microbiota	Puerarin, quinoa, Moringa oleifera, ginkgolide B, taraxasterol, flavan-3-ols	Often correlation-based; limited causal validation	Add antibiotic depletion, FMT, metabolite supplementation, and pathway inhibition	Preliminary-to-moderate preclinical support; causal validation is needed
Complex herbal formulation and network pharmacology evidence	Multi-component interventions may regulate microbiota, metabolites, endocrine signals, and bone pathways	Roucongrong Pill, traditional Chinese medicine formulas, Jiaru fuling	Active components and key causal nodes are difficult to identify; docking does not prove mechanism	Use component deconstruction, key metabolite validation, and modular causal testing	Exploratory potential; causal attribution and active-component validation remain limited
miRNA and extracellular vesicle evidence	miRNAs may mediate microbiota-host and BMSC-bone communication	Fecal miRNA regulation of microbiota, BMSC exosomal miRNAs, circulating bone-related miRNAs	Few studies directly connect natural compounds, microbiota, EV-miRNAs, and bone outcomes	Integrate fecal/serum/BMSC EV-miRNA sequencing with metabolomics and bone phenotyping	Emerging research direction; currently hypothesis-generating

RCT: Randomized controlled trial; BMD: Bone mineral density; OVX: Ovariectomy; FMT: Fecal microbiota transplantation.

Insufficient distinction between correlation and causality

Many studies report concurrent changes in skeletal phenotypes and gut microbial composition, but such findings do not establish microbiota-mediated causality. In the absence of microbiota perturbation, FMT, defined microbial colonization, metabolite rescue, or receptor/pathway blockade, the observed microbial changes may be secondary rather than causal.

Overreliance on 16S rRNA sequencing

Although 16S rRNA sequencing can identify taxonomic shifts, it cannot adequately resolve strain-level functions, microbial genes, or metabolic capacity. Consequently, taxonomic associations alone cannot identify the microbial functions responsible for the observed skeletal phenotypes.

Limited human evidence

Most evidence is derived from ovariectomized rodents, which do not fully reproduce gradual reproductive aging, lifestyle exposures, medication use, metabolic comorbidities, or the heterogeneous microbiota of postmenopausal women. Human trials remain limited, and few evaluate fracture reduction or long-term skeletal outcomes.

Context-dependent metabolite effects

The skeletal effects of microbial metabolites may vary according to dose, source, host metabolic status, inflammation, and disease stage. Therefore, metabolites such as TMAO, bile acids, and tryptophan derivatives should not be regarded as uniformly beneficial or harmful[43,46-49,52,53].

Underdeveloped microrna and bone marrow microenvironment research

Current research remains concentrated on a limited set of metabolic and immune pathways, whereas extracellular vesicles, microRNAs, BMSC fate, bone marrow adiposity, cellular senescence, and spatial immune organization remain comparatively understudied.

Mechanistic overgeneralization in complex herbal formulations

For complex formulations, network pharmacology and molecular docking can identify plausible targets but cannot establish the causal roles of individual compounds, microbial functions, metabolites, or molecular pathways.

FUTURE DIRECTIONS

Develop a microbiota-dependent classification framework

Future studies should apply and refine the proposed conceptual framework by considering the identity of the administered material, its microbial dependence, biotransformation, active metabolites, and dominant host targets.

Move from taxonomy to functional metabolic phenotypes

Future studies should integrate shotgun metagenomics, metatranscriptomics, metabolomics, culturomics, and strain-resolved functional analyses with longitudinal skeletal phenotyping to move from taxonomic associations toward functionally validated microbial mechanisms.

Establish causal experimental frameworks

Causal studies should combine microbiota depletion, germ-free or gnotobiotic models, FMT, defined microbial colonization, metabolite rescue, and receptor/pathway blockade to determine whether proposed microbial mediators are necessary or sufficient for the observed skeletal effects.

Integrate micrornas and extracellular vesicles

Prospective studies should integrate microbiome, metabolome, extracellular-vesicle microRNA, bone marrow immune, and skeletal phenotyping. Candidate microRNAs should be validated using cell-specific gain- and loss-of-function experiments rather than bioinformatic prediction alone[17-23,54].

Conduct microbiome-stratified clinical trials

Microbiome-stratified trials should account for regional dietary patterns because habitual foods, nutrient availability, and culturally specific fermented or herbal products can influence baseline microbial and metabolic phenotypes and the feasibility of dietary interventions. Translation will also depend on how foods, supplements, probiotics, and traditional formulations are regulated and quality-controlled across regions. Because sequencing, metabolomics, standardized sample handling, bioinformatics, and longitudinal monitoring are costly and require specialized infrastructure, initially feasible approaches may need to combine selected microbial or metabolite biomarkers with routine clinical variables. Affordability, laboratory capacity, product availability, and equitable access should therefore be considered when designing multicenter trials and implementation strategies.

Differentiate intervention forms in clinical translation

Clinical translation should follow intervention-specific pathways. Food-based interventions require standardized dietary exposure and adherence assessment; concentrated extracts and supplements require defined dosing, chemical characterization, and batch-consistency testing; and multi-component traditional formulations require fixed composition, botanical authentication, and quality-control criteria. Their efficacy and safety should be analyzed separately before cross-category comparisons are made.

Prioritize clinically meaningful skeletal outcomes

Clinical studies should distinguish surrogate skeletal measures from clinically meaningful outcomes and prioritize adequately powered assessment of vertebral, nonvertebral, and hip fractures, together with long-term safety[25,26].

PROPOSED CORE VIEWPOINT

The central contribution of this review is a hypothesis-generating framework that distinguishes direct bone-cell effects from microbial biotransformation- and microbiota-associated effects. Natural bioactive interventions are therefore considered according to the identity of the administered material, their microbial dependence, and their predominant mechanistic relationships rather than by chemical structure alone.

CONCLUSION

Natural bioactive interventions may influence skeletal homeostasis through interacting host- and microbiota-dependent mechanisms[3-8,11-14]. The conceptual framework proposed in this review organizes heterogeneous interventions according to the identity and form of the administered material, their microbiota dependence, and their predominant mechanistic relationships. It is intended as a tool to guide hypothesis testing and the design of mechanistic and clinical studies, rather than as a validated classification or basis for clinical decision-making. However, most supporting evidence remains preclinical, and improvements in BMD or bone microarchitecture in experimental models should not be interpreted as evidence of fracture prevention or clinical efficacy in postmenopausal women.

Translation of this framework will require causal mechanistic studies, function-resolved microbial profiling, standardized intervention characterization, microbiome-stratified clinical trials, and evaluation of clinically meaningful skeletal outcomes. Implementation strategies should also account for regional dietary practices, product-regulatory requirements, affordability, laboratory capacity, and equitable access.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the researchers whose valuable studies laid the scientific foundation for this work and provided important insights into its conceptual development.