Apoptotic vesicles rectify senile bone-fat imbalance by activating Thy1-ERK-TAZ axis Authors: Yuhe Jiang, Jianqi She, Zeying Wang, Yuan Zhu, Yunsong Liu, Ping Zhang, Hao Liu, Xiao Zhang, Yongsheng Zhou Published: 2025 ABSTRACT BackgroundSenile bone-fat imbalance poses a significant public health burden in aging populations. This study investigates how aging affects the heterogeneity of mesenchymal stem cell (MSC)-derived apoptotic vesicles (apoVs) and develops a therapeutic strategy to rectify age-related bone-fat metabolic disorders.ResultsProtein cargo analysis revealed that MSC-derived apoVs undergo age-dependent compositional remodeling, with Thy1 identified as a key biomarker progressively declining with donor aging. Functional validation confirmed Thy1’s critical role in mediating apoV osteoinductive and anti-adipogenic capacities. Engineered Thy1-enriched apoVs, produced through magnetic-activated sorting and protein corona technology, significantly enhanced osteogenesis and suppressed adipogenesis in recipient MSCs by activating the Thy1-ERK-TAZ signaling axis. In aged mouse models, these engineered apoVs restored trabecular bone mass, reduced marrow adipose tissue, and ameliorated systemic lipid metabolism disorders.ConclusionsThis work establishes Thy1 as a functional biomarker for aging-related apoV heterogeneity and demonstrates that engineered Thy1-enriched apoVs represent a clinically scalable therapeutic approach for senile bone-fat imbalance. The identified Thy1-ERK-TAZ mechanism provides a tunable platform for developing extracellular vesicle-based interventions against age-related metabolic diseases.Graphical Abstract ApoVs play a critical role in regulating bone-fat metabolism by orchestrating the balance between bone remodeling and lipid metabolism. This study revealed that while the fundamental properties of apoVs remained stable during development and aging, their functions and protein compositions were profoundly altered by donor aging. Thy1 was identified as a key biomarker, influencing apoV-mediated osteogenesis and adipogenesis. By engineering Thy1-enriched apoVs through magnetic-activated sorting and protein corona modification, we developed a highly effective therapeutic approach to restore senile bone-fat balance. Mechanistically, Thy1 enhanced osteogenic differentiation and suppressed adipogenesis in MSCs via activating the Thy1-ERK-TAZ axis. These findings systematically demonstrate the impact of donor aging on apoVs, paving the way for quality control and engineering of apoV-based strategy in aging-related metabolic diseases. Supplementary InformationThe online version contains supplementary material available at 10.1186/s12951-025-03928-4. Introduction Senile bone-fat imbalance is a common metabolic disease, posing a substantial public health burden in our aging population [1]. Multiple cross-sectional human studies have confirmed correlations between high marrow fat content, low bone density, and increased fracture risk, particularly in osteoporosis and skeletal aging [2–4]. Bone marrow mesenchymal stem cells (MSCs) have been shown to play an important role in senile bone-fat metabolism due to their ability in osteogenic and adipogenic differentiation [5, 6]. Mechanistically, the aging process is predominantly attributed to the progressive accumulation of stochastic damage in cells, organelles, and macromolecules [7]. The exact mechanisms underlying the damage are not yet fully understood but likely consist of several cell-autonomous processes, including epigenetic alterations, loss of proteostasis, stem cell exhaustion, etc [8]. However, non-cell-autonomous mechanisms, especially the altered interactions among cells, tissues, and organs, are also critically important for aging [9]. In the past decade, extracellular vesicle (EV)-mediated intercellular communication has added a new dimension to our knowledge of genetic information transfer and homeostasis maintenance in multicellular organisms [10, 11]. EVs, a heterogeneous group of nanosized membranous vesicles, are composed of a protein-rich lipid bilayer and the enclosed proteins and RNAs derived from EV-producing cells [12]. EVs act as messengers of intercellular communication networks, allowing the exchange of their cargoes between parental cells and target cells [13]. Apoptotic vesicles (apoVs), a type of EV produced during cellular apoptosis, exhibit diverse production mechanisms, sizes, and contents [14]. Our previous work has characterized the detailed attributes, specific markers, and biological functions of MSC-derived apoVs [15], and proved that their characteristics are largely inherited from parent cells [16]. They have been reported to restore the stem cell properties of endogenous MSCs and mitigate bone loss in osteoporosis [17]. However, it is unclear whether differences in the state of apoV parent cells might contribute to inconsistent findings in preclinical studies. The unselective use of apoVs in preclinical trials involving bone-fat metabolism might yield suboptimal outcomes and impede the emerging field of apoV therapy. Therefore, there is an immediate need to identify a promising subpopulation of apoVs with a high and consistent regulatory effect. In this study, we assessed the impact of parental mice aging on the apoVs derived from their MSCs, and focused on their rejuvenating effects on bone-fat balance by repeated administration to senile mice. We elucidated the molecular basis underlying the critical roles of young apoVs in reversing senile bone-fat imbalance using in vivo and in vitro models. Further, we developed a clinically applicable apoV for regulating senile osteoporosis. Results Aging leads to heterogeneity of ApoV subpopulations The classification of C57BL/6J mice into juvenile, adult, and aged groups is based on their developmental and physiological milestones. Juvenile mice are typically defined as those within 4 weeks old [18, 19]. Adult mice are generally considered to be between 3 and 6 months old [20, 21]. Aged mice are classified as 18 months and older [22, 23]. These age divisions are widely used in aging research and are critical for studying age-related diseases and interventions. Thus, we studied apoVs derived from three age groups: 1-week-old mice, 3-month-old mice, and 18-month-old mice. After inducing apoptosis of MSCs, apoVs were isolated using a refined protocol that involved gradient centrifugation [15]. The essential characteristics of apoVs derived from MSCs from 1-week-old mice (1w apoVs), 3-month-old mice (3 m apoVs), and 18-month-old mice (18 m apoVs) were similar. These included their spherical structure with phospholipid bilayer (Fig. S1a), size distribution ranging from 100 to 500 nm (median of about 200 nm) (Fig. S1b), zeta potential of approximately − 25mV (Fig. S1c), high positive rate for apoptosis specific marker (phosphatidylserine, PS) (Fig. S1d), the presence of general EV markers (CD81 and TSG101) (Fig. S1e) [24, 25], and enrichment of markers specific to MSC-derived apoVs (ITGA5, RhoA, RPS25, Lamin B1) (Fig. S1e-f) [15, 26]. These indicators suggested that the vesicles isolated in our study met the standards for high-purity apoVs. Moreover, the essential properties of apoVs were stable during individual maturation and senescence. To verify their biological functions, we treated senile MSCs isolated from aged mice (18-month-old) with 3 different apoVs. Cell proliferation experiments indicated that MSC-derived apoVs at concentrations below 500 ng/ml had no notable effect on cell proliferation (Fig. S2a-c). According to our previous research, the concentration of 250 ng/ml exhibited a robust intervention effect [17]. Consequently, this concentration was selected for subsequent experiments. The osteogenic differentiation potentiated by apoVs was further supported by elevated alkaline phosphatase (ALP) activity (Fig. 1a-b), enhanced formation of mineralized nodules (Fig. 1a-b), and upregulation of osteogenic genes (Fig. 1c), which showed a decline in the osteoinductive ability of apoVs with aging. Based on the semi-quantitative results, aged MSCs possess only 30–40% of the osteogenic differentiation capacity of young MSCs. However, treatment with 1w apoVs restores this potential to approximately 80%, while 3 m apoVs achieve a 60% rescue, and 18 m apoVs show negligible improvement (Fig. 1a-b). The osteoinductive potential of apoVs in vivo also supported that, among different apoV populations, 1w apoVs demonstrated the strongest osteoinductive capacity, followed by 3 m apoVs, while 18 m apoVs exhibited the weakest effect (Fig. 1 d). These findings corroborated an aging-associated decrement in the osteoinductive capacity of apoVs. Moreover, oil red O staining and quantification showed that the lipid droplets in MSCs after 1w apoV stimulation were less than those after 3 m apoV stimulation, while the group with 18 m apoV stimulation failed to inhibit adipogenesis (Fig. S3a-b). The expression levels of adipogenesis-related genes also supported the observation that anti-adipogenic potential of apoVs significantly decreased with aging (Fig. S3c). It is worth noting that apoVs did not significantly alter the metabolic activity of MSCs, showing by their β-galactosidase activity and reactive oxygen species (ROS) levels compared with control group (Fig. S4). As it is well known that aged MSCs show decreased capacity to differentiate into osteogenic lineages and increased capacity to differentiate into adipogenic lineages [27], our evidence supported that the functions of apoVs were intimately tied to the identity of their parental cells. To examine the therapeutic potential of each apoV population in vivo, we systemically administered them into aged mice (18-month-old). The biodistribution of apoVs is particularly crucial for evaluating their targeting efficiency. At 6, 12, 24, and 48 h post tail vein injection, fluorescence imaging revealed that the DiR-labeled apoVs were primarily accumulated in the liver (Fig. S5). With time progression, the accumulation of apoVs in the femur progressively increased, suggesting that the apoVs successfully reached the femur and exerted their effects over time (Fig. S5). After four months of treatment, blood tests (Fig. S6), blood biochemistry (Fig. S7), body length, body weight, and organ tissue examinations (Fig. S8) showed no significant differences between the groups, confirming that apoVs were biologically safe. Micro-computed tomography (micro-CT) scans, analysis of bone mineral density (BMD) and bone microarchitectural indices, and histological examination of femoral metaphysis revealed substantial bone loss and marrow adipose tissue (MAT) accumulation in aged mice (Fig. 2a-d). Following treatment with 1w apoVs, there was a significant improvement in bone mass and reduction in MAT, treatment with 3 m apoVs resulted in some degree of trabecular bone recovery and MAT decrease, whereas treatment with 18 m apoVs showed no improvement (Fig. 2a-d). These results were explained by changes in the osteogenic and adipogenic capacity of endogenous mice MSCs (Fig. 2e-h). Further analysis of abdominal fat and liver revealed significant obesity in aged mice (Fig. S9a-c). Treatment with 1w apoVs led to a notable reduction in abdominal fat and improved fatty liver phenotype, treatment with 3 m apoVs resulted in moderate improvements in adiposity and fatty liver, whereas treatment with 18 m apoVs showed no benefits (Fig. S9a-c). The above results verified that administration of apoVs regulated systemic bone-fat metabolism. The abilities of young apoVs to rescue properties of endogenous MSCs and regulate bone remodeling in aged mice imply that endogenous apoVs in aged individuals do not effectively support the regenerative cascade. Confirmation of aging-related markers for MSC-derived ApoVs To further probe age-associated changes in MSC-derived apoV population heterogeneity, we profiled their protein cargos. In the aggregate, more than 5800 proteins were certified in each sample (Table S1). Venn diagram (Fig. S10a), cluster heatmaps (Fig. S10b), and volcano maps (Fig. S10c) revealed numerous differentially expressed proteins (DEPs) among groups. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis further verified that 1w apoVs contained a higher abundance of functional proteins related to “cellular behavior”, “cellular transport”, “cellular metabolism”, and “the regulation of various diseases” compared to 3 m apoVs (Fig. S10d). Similar results also existed in the comparisons between 1w apoVs and 18 m apoVs (Fig. S10e), and 3 m apoVs and 18 m apoVs (Fig. S10f). In summary, the protein compositions of apoVs were largely reshaped by donor aging, which might help elucidate their functional heterogeneity. Next, quantitative analysis of the protein cargo in different apoV subpopulations was conducted. Among all detected proteins, there were 3 proteins (Alpl, Fmo1, Maob) whose expression levels continuously increased and 3 proteins (Annexin A8, Thy1, Tppp3) whose expression levels continuously decreased with donor aging (Fig. 3a). The accuracy of the mass spectrometry results was verified by western blotting (Fig. 3b). These proteins might serve as candidate biomarkers for apoV aging, yet their effects on apoV function warrant further investigation. Among these proteins, Thy1 is a well-established marker and functional regulator of MSCs, which plays a critical role in cell differentiation [28, 29]. During osteogenic differentiation, Thy1 aids in the transformation of MSCs into osteoblasts [30]. This process is accompanied by a concurrent inhibition of fat accumulation [31]. When Thy1 is absent, MSCs display diminished osteogenic differentiation and augmented adipogenic differentiation [32]. Here, we demonstrated that as MSCs aging, their Thy1 expression gradually decreases (Fig. S12a), accompanied by a decline in osteogenic differentiation capacity (Fig. S12b) and an increase in adipogenic differentiation potential (Fig. S12c), suggesting that Thy1 might serve as an aging-associated key molecule. Immuno-gold labeling further validated that Thy1 was situated on the membrane and its levels progressively declined in 1w apoVs, 3 m apoVs, and 18 m apoVs (Fig. 3c). Notably, to determine whether the elevation of Thy1 in recipient MSCs is a result of exogenous addition or endogenous expression, we labeled Thy1 in apoVs using immunofluorescence staining. Representative images exhibited that Thy1-positive apoVs are progressively internalized by MSCs within 6 h, with fluorescence intensity increasing initially and peaking around 48 h. Thy1 initially distributes in the cytoplasm and subsequently accumulates perinuclearly, suggesting potential involvement in nuclear-related signaling pathways. By 48 h, the fluorescence intensity declines, indicating possible protein degradation or metabolic turnover (Fig. 3d). Based on these observations, we speculated that MSCs reused Thy1, supplied by donor apoVs, to rejuvenate their functional capabilities. Consequently, Thy1 emerged as a potential aging-related biomarker in apoVs, possibly facilitating apoV-induced bone-fat metabolism regulation in MSCs. We next evaluated whether the transferred Thy1 impacted any specific signaling pathways in the recipient MSCs that contribute to the therapeutic benefits mediated by apoVs. Through a screening of well-known pathways related to osteogenic differentiation, we pinpointed the potential involvement of ERK signaling in this process (Fig. S11). Previous studies have demonstrated that TAZ (transcriptional co-activator with PDZ-binding motif) is a key downstream effector of the ERK signaling pathway [32–35]. ERK-mediated phosphorylation could stabilize TAZ, preventing its degradation and promoting its nuclear translocation [36]. In the nucleus, TAZ plays a pivotal role in regulating the balance between osteogenic and adipogenic differentiation. TAZ could interact with osteogenic transcription factors like RUNX2 to enhance osteogenic differentiation, while it can also bind to adipogenic transcription factors such as PPARγ to inhibit adipogenic differentiation [37, 38]. Thus, TAZ functions as a “switch” that tips the balance between bone formation and fat accumulation [39, 40]. We propose that the differentiation of MSCs, modulated by apoVs, is associated with Thy1, the phosphorylation of ERK, and the upregulation of TAZ. ApoVs regulate bone-fat metabolism through Thy1-ERK-TAZ axis To test this hypothesis, we utilized small interfering RNA (siRNA) to knock down Thy1 in 1w MSCs and a virus to upregulate Thy1 in 18 m MSCs, and then extracted apoVs from these cells (designated as si 1w apoVs and virus 18 m apoVs) [41]. The knockdown and upregulation efficiencies in both types of apoVs were individually evaluated using RT-qPCR (Figs. S13a, S13c) and western blotting (Figs. S13b, S13d). To explore the effects of Thy1 on apoV function, we incubated MSCs with 1w apoVs, si 1w apoVs, 18 m apoVs, and virus 18 m apoVs. Our findings revealed that the ability of apoVs to promote MSC osteogenesis was significantly diminished with siRNA interference, suggesting that Thy1 deficiency abolished the osteoinductive property of apoVs (Fig. 4a-d). Additionally, we examined the expression of phosphorylated ERK (p-ERK) and TAZ in MSCs (Fig. 4d). As expected, the expression levels of p-ERK and TAZ decreased substantially along with the reduction of Thy1 in apoVs, implying that Thy1 might regulate MSC osteogenesis through the ERK-TAZ signaling. Meanwhile, the capacity of apoVs to inhibit adipogenesis was markedly reduced upon siRNA interference, indicating that Thy1 deficiency disrupted the anti-adipogenic effect of apoVs (Fig. S14a-d). The expression levels of p-ERK and TAZ decreased while PPARγ and Perilipin-1 increased substantially along with the reduction of Thy1 in apoVs, implying that Thy1 might regulate MSC adipogenesis through ERK-TAZ signaling (Fig. S14d). In vivo experiments also supported the ability of apoVs to rescue primary MSCs and relieve bone-fat imbalance was significantly diminished with siRNA interference (Fig. 6a-h, S16a-c). Inversely, the abilities of 18 m apoVs to promote MSC osteogenesis and inhibit MSC adipogenesis were significantly increased with lentiviral transfection elevated, suggesting that Thy1 enrichment enhanced the regulatory function of apoVs both in vitro and in vivo (Fig. 4e-h, S14e-h). Studies have shown that Thy1 knockout mice exhibited osteoporosis and obesity phenotypes [42, 43]. The changes in the differentiation ability of primary MSCs supported this conclusion (Fig. 5a-b, S15a-b). Further, we extracted Thy1 pure negative apoVs from 1-week-old Thy1(-/-) fetal mice (designated as Thy1(-/-) 1w apoVs), phenotypic analysis confirms that Thy1 deletion does not alter the fundamental physical properties of apoVs (Fig. 5c-f), while their ability to promote bone formation and inhibit lipid accumulation was significantly decreased (Fig. 5g-j, S15c-f, S16a-c). Western blotting results suggested that this change was related to the decrease of Thy1 content and the silence of Thy1-ERK-TAZ axis (Fig. 5f, S15f). In vivo experiments also supported the ability of apoVs to rescue primary MSCs and relieve bone-fat imbalance was significantly diminished with Thy1 knockout (Fig. 6a-h, S16a-c). Next, we examined the connection between Thy1 and ERK signaling. We employed GDC-0994, a selective allosteric inhibitor of ERK that has undergone phase I and II clinical trials [44, 45]. In our study, GDC-0994 effectively blocked the phosphorylation of ERK in MSCs stimulated by apoVs (Fig. 7d). The decrease in ALP activity (Fig. 7a-b), mineralized nodule formation (Fig. 7a-b), osteogenic gene expression (Fig. 7c), and osteogenic protein levels (Fig. 7d) indicated that the osteogenic effects induced by 1w apoVs were nearly abolished by GDC-0994, suggesting that apoVs promote osteogenic differentiation in MSCs through the activation of ERK signaling. Moreover, the capacity of apoVs to inhibit adipogenesis was markedly reduced upon GDC-0994 blockage, as exhibited by lipid droplets, adipogenesis-related gene and protein expressions in MSCs, indicating that ERK signaling played key roles in the anti-adipogenic effect of apoVs (Fig. S17a-d). It’s worth noting that the upregulation of Thy1 remained unaffected regardless of the presence or absence of GDC-0994, confirming that Thy1 acts as an upstream activator of p-ERK (Fig. 7d, S17d). To sum up, the above functional experiments confirmed Thy1 as a balancer of bone and lipid metabolism regulation within apoVs, by activating ERK-TAZ signaling axis in recipient MSCs. Thy1 engineering enhances the apoV-mediated improvement of senile bone-fat metabolism imbalance To generate a subpopulation of apoVs with high and consistent therapeutic potential, we employed magnetic-activated sorting to isolate Thy1-positive apoVs (Thy1+ apoVs) and Thy1-negative apoVs (Thy1− apoVs), as illustrated in the flowchart (Fig. S18a) [26, 46]. The integrity, yield, and effectiveness of the magnetic enrichment was verified, underscoring that magnetic sorting yields high-purity, intact apoVs suitable for scalable production. (Fig. S18b-d). Representative images and results demonstrated that Thy1+ apoVs exhibited a stronger capacity to promote MSC osteogenesis in vitro compared to Thy1− apoVs (Fig. 8a-d). Synchronous changes in Thy1, p-ERK, TAZ, and RUNX2 in recipient MSCs supported that apoVs promoted osteogenesis through transferring Thy1 and upregulating the Thy1-ERK-TAZ axis (Fig. 8d). Meanwhile, Thy1+ apoVs were superior in inhibiting MSC adipogenesis (Fig. S19a-d). The expression of Thy1, p-ERK, TAZ, PPARγ, and Perilipin-1 in MSCs exhibited that apoVs inhibit adipogenesis by upregulating the Thy1-ERK-TAZ axis (Fig. S19d). In vivo studies further revealed that Thy1 enrichment enhanced the ability of apoVs to rescue endogenous MSCs and ameliorate senile bone-fat imbalance (Fig. 9a-h, S20a-c). It is worth noting that through magnetic-activated sorting, we could simultaneously obtain two subpopulations: Thy1+ apoVs and Thy1− apoVs. Here, we discussed how Thy1+ apoVs treat senile bone loss and inhibit MAT accumulation. While the Thy1− apoVs might be more likely to have a high-efficiency function in promoting adipogenic differentiation, their broad application prospects remain to be explored. Our next objective was to gain deeper insights into the method of incorporating Thy1 protein into Thy1+ apoVs in vitro to further enhance their therapeutic potential. The protein corona technology of EVs refers to the study of the layer or layers of proteins adsorbed onto the surface of EVs when they interact with biological fluids such as blood or culture media [47, 48]. Its advantages lie in the potential to easily load proteins onto the vesicle surface and might enhance the targeting and biodistribution of EV-based therapeutics [49]. To this end, we utilized Thy1 to embed Thy1+ apoVs, and the efficiency of Thy1 protein corona formation was validated through western blotting (Fig. S18e-f). In vitro studies showed that Thy1-embedded apoVs (designated as Thy1++ apoVs) rescued osteogenic differentiation capacity more effectively than Thy1+ apoVs, as evidenced by increased ALP activity, mineralized nodule formation, and osteogenic gene expression (Fig. 8e-h). Synchronous changes in Thy1, p-ERK, TAZ, and RUNX2 in recipient MSCs supported that apoVs promoted osteogenesis through transferring Thy1 and upregulating the Thy1-ERK-TAZ axis (Fig. 8h). Likewise, Thy1++ apoVs were superior in inhibiting MSC adipogenesis (Fig. S19e-h). The expression of Thy1, p-ERK, TAZ, PPARγ, and Perilipin-1 exhibited that apoVs inhibit adipogenesis by upregulating the Thy1-ERK-TAZ axis in MSCs (Fig. S19h). To further validate their potential in mitigating bone-fat imbalance, senile mice were treated with equivalent doses of Thy1+ apoVs and Thy1++ apoVs. The results showed that Thy1++ apoVs treatment was more effective in restoring bone structure and alleviating fat accumulation compared to Thy1+ apoVs (Fig. 9a-h, S20a-c) (Fig. 10). Overall, we showed that Thy1-engineered apoVs relieved senile bone-fat imbalance by stimulating ERK phosphorylation more efficiently and offering superior therapeutic advantages. Discussion Recently, EVs have emerged as promising therapeutic agents in the realm of bone regeneration [50–52]. The potential of EV treatment to reverse bone loss is deemed feasible due to its capacity to modulate cell function through cell-to-cell communication [53–57]. However, senescent EVs can alter immunity, inflammation, gene expression, and metabolism levels in target cells [58]. Injecting EVs sourced from aged bone matrix promotes an imbalance between bone and fat and exacerbates vascular calcification [59]. Thus, EVs originating from different cells might exhibit unique functions. Here, we validated the senescence-associated decline in osteoinductive and anti-adipogenic capacity of apoVs, demonstrating that their functions strongly depend on the identity of parental cells. Therefore, apoVs should be selectively applied in cases of senile bone-fat imbalance, as their indiscriminate use might lead to unpredictable therapeutic outcomes. The recent advancements in exosome biology have enabled the establishment of specific guidelines and quality control measures, paving the way for clinical applications [60]. Our findings emphasize the importance of implementing robust quality control measures for the development of apoVs therapy in the future. Currently, there is a lack of systematic understanding of the changes in apoVs during aging process [61]. Given our previous finding that the osteoinductivity of apoVs varies with the aging of donor MSCs, we utilized a proteomic assay to quantify the protein cargo of apoVs. Through this method, we screened out 6 proteins (Alpl, Fmo1, Maob, Annexin A8, Thy1, Tppp3) as candidate biomarkers for apoV aging. Since our targeted therapy aims to address the imbalance of bone-fat metabolism in the elderly, we selected Thy1 as the primary research subject. Other proteins suggest potential alterations in the biological functions of apoVs during the aging process. For example, Tppp3, a member of the microtubule polymerization-promoting protein family, serves as an early progenitor cell marker during the chondrogenic differentiation of MSCs [62]. After the induction of heterotopic ossification, the number of Tppp3+ cells increases, and they partially contribute to the formation of ectopic cartilage in tendons or joints [63]. The decrease in its expression with aging might indicate a decline in the ability of apoVs to promote chondrogenesis. Interestingly, Thy1 emerged as a pivotal molecule in MSCs, facilitating osteogenic differentiation while impeding adipogenic differentiation. However, to the best of our knowledge, the effect of Thy1 expression on apoV function has not been documented until now. Our findings provided mechanistic evidence that apoVs facilitated MSC osteogenesis and hindered MSC adipogenesis through Thy1-induced ERK phosphorylation, and the amount of Thy1 in apoVs directly correlated with their osteoinductive ability. We suggest using Thy1 as a biomarker to assess the osteoinductive capacity of apoVs, addressing the lack of evaluation tools for the preclinical potential of apoV preparations. The mechanism by which MSCs internalize apoVs is a critical step in the functional transfer of bioactive molecules like Thy1. It is a multifaceted process that likely involves several well-established pathways, such as endocytosis (e.g., clathrin-mediated or caveolin-dependent), phagocytosis, and potentially macropinocytosis [64, 65]. In the context of apoVs, their characteristic PS exposure—a key apoptosis marker—suggests a prominent role for PS receptor-mediated phagocytosis or endocytosis, facilitating recognition and internalization by recipient MSCs [64, 66]. Our dynamic imaging data in Fig. 3d, showing the time-dependent accumulation of Thy1 signal within MSCs, supports the active internalization of apoVs rather than mere surface binding. The initial cytoplasmic distribution followed by perinuclear accumulation suggests a endocytic trafficking route, potentially leading to the release of cargo to influence signaling pathways. Although our current data does not delineate the exact contribution of pathways, the observed functional outcome—the efficient transfer and reuse of Thy1 to rejuvenate MSC capabilities—confirms the biological significance of this internalization process. Future studies employing specific inhibitors of these pathways will be essential to fully elucidate the precise mechanistic details of apoV uptake by MSCs. One of the hurdles for the clinical application of apoVs is the production of high yields of highly bioactive apoVs [67]. Through innovative magnetic-activated sorting and protein corona engineering, we developed a clinically scalable method to produce bioactive Thy1-enriched apoVs with superior therapeutic efficacy. We are confident that this engineering approach, based on aging-related biomarker, is a promising advancement in apoV engineering for the treatment of age-related diseases. Conclusions In conclusion, this study underscores the pivotal role of Thy1-enriched apoVs in ameliorating senile bone-fat imbalance by directing MSC fate toward osteogenesis and away from adipogenesis. Proteomic profiling identifies Thy1 as a definitive functional biomarker for aging-related heterogeneity in MSC-derived apoVs, whose high enrichment activates the Thy1-ERK-TAZ axis to enhance osteogenic potential and suppress adipogenesis. Through an innovative combination of magnetic-activated sorting and protein corona engineering, we develop a clinically scalable method to produce bioactive Thy1-enriched apoVs with superior therapeutic efficacy. This strategy not only restores metabolic homeostasis in senile bone-fat tissues but also provides an engineering framework for apoV-based interventions. Our findings establish apoVs as a tunable platform for targeting aging-associated metabolic disorders, offering a reproducible and clinically translatable approach for combating osteoporosis and other age-related pathologies. Methods Reagents and antibodies An exhaustive list of all reagents and antibodies employed in this study is furnished in Table S2. Animals The animals utilized in this study were procured from Vital River (China) and housed in a sterile, specific-pathogen-free environment at the Peking University School of Stomatology. This environment was maintained under strict conditions, including a regulated 12-hour light-dark cycle and a temperature stabilized at 23 ± 3 °C. The animals had unlimited access to sanitized diets and autoclaved water to ensure their well-being. Before initiating any experimental procedures, all protocols involving animals were thoroughly reviewed and approved by the Institutional Animal Care and Use Committee of the Peking University Health Science Center (approval number: LA2024088). Cell culture The process of isolating primary mice MSCs involved flushing the bone marrow and digesting the tibiae and femurs with collagenase [68]. The mice were euthanized, and their femurs and tibiae were harvested. Small incisions were made at both ends of the bones to flush out the bone marrow into collection tubes. Subsequently, the bones were chopped and digested in collagenase for 45 min. The resulting supernatants were combined with the flushed bone marrow, centrifuged, and the cell pellets were plated in suitable dishes. After 48 h of culture, the media was removed, and non-adherent cells were discarded. The remaining cells were digested, resuspended in fresh culture media, and replated. When the MSCs reached confluence, they underwent the differentiation process. Osteogenic and adipogenic induction Cells were maintained in either proliferation medium (PM) or osteogenic induction medium (OM) or adipogenic induction medium (AM) under sterile conditions at 37 °C, supplemented with 5% CO2 and 100% relative humidity. The PM comprised 10% (v/v) fetal bovine serum (FBS, Gibco, USA) and 1% (v/v) penicillin/streptomycin (Gibco, USA) in α-Minimum Essential Medium (α-MEM, Gibco, USA). The OM comprised 10% (v/v) FBS, 1% (v/v) penicillin/streptomycin, 10 mM β-glycerophosphate (Sigma-Aldrich, USA), 200 µM ascorbic acid (Sigma-Aldrich, USA), and 10 nM dexamethasone (Sigma-Aldrich, USA) in α‐MEM. The AM, on the other hand, contained α‐MEM along with 50 nM insulin, 100 nM dexamethasone, 200 µM indometacin, 500 µM isobutylmethylxanthine, 1% (v/v) penicillin/streptomycin, and 10% (v/v) FBS. Once cells achieved 80%−90% confluence, they were enzymatically detached using trypsin (Gibco, USA) and subsequently propagated in vitro. Extraction and purification of ApoVs MSCs were initially rinsed twice with phosphate-buffered saline (PBS, 0.1 μm-filtered) and then incubated with 250 nM staurosporine (STS, Enzo Life Sciences, USA) diluted in α-MEM. Following a 12-hour incubation period, the medium was collected for apoV extraction, adhering to a previously described method [69]. Cryo-electron microscopy (Cryo-EM) Building upon our previous research, holey-carbon grids (Quantifoil, R1.2/1.3, 300 mesh) were utilized for Cryo-grid preparation [26]. Prior to freezing, the grids underwent a 60-second glow-discharge treatment. Resuspended apoVs were deposited onto the grids, and excess solution was carefully blotted away using filter paper. An FEI Vitrobot Mark IV, set at 100% humidity and 4 °C, was employed for Cryo-grid handling. For imaging, an FEI Titan Krios TEM, operating at 300 kV and equipped with a Gatan K2 summit camera (USA), along with a GIF Quantum energy filter (Gatan, USA), was used. Nanoparticle tracking analysis (NTA) To evaluate the zeta potential and size distribution of apoVs, they were diluted in appropriately filtered PBS (0.1 μm) and analyzed using a ZetaView® PMX120 instrument (Particle Metrix, Germany). The acquired data was processed using the ZetaView® analysis software (Version 8.02.31) [70]. Nano-flow cytometry To execute high-sensitivity nano-flow cytometric examinations, apoVs were diluted in phosphate-buffered saline (PBS, 0.1 μm-filtered) and analyzed utilizing a nano-flow cytometer (Flow NanoAnalyzer, NanoFCM Inc.), adhering strictly to the manufacturer’s protocols [46]. The samples underwent dilution to maintain a particle count within the recommended spectrum of 4,000–14,000. The NanoFCM software (Profession V1.0) was instrumental in determining particle concentration and size distribution. Additionally, it was employed to analyze the proportions of positively stained particles. Proteomic analysis Protein lysates were prepared as previously described [67]. For proteomic profiling, the Q-Exactive HF mass spectrometer (Thermo Fisher Scientific, San Jose, CA) was deployed in both data-dependent acquisition (DDA) and data-independent acquisition (DIA) modes to acquire mass spectrometry (MS) data. Proteins exhibiting a fold change ≥ 2 and an adjusted p-value < 0.05 were designated as DEPs and were subjected to functional analyses, encompassing KEGG pathway analysis. A comprehensive list of these proteins is furnished in Table S1. Cell proliferation assay To ascertain the safe concentration range, cells were exposed to progressively increasing concentrations of apoVs. The culture medium containing apoVs was replenished every 3 days. Cell proliferation was quantified using a cell-counting kit-8 (CCK-8, Dojindo, Japan), and the resultant growth curve was plotted based on absorbance measurements spanning 0–7 days. ALP staining and quantification The process involved utilizing an NBT/BCIP staining kit (CoWin Biotech, China) subsequent to cell fixation, enabling the acquisition of ALP staining. An ALP assay kit (Nanjing Jiancheng Bioengineering Institute, China) was employed to determine the specific concentration of ALP. Additionally, a BCA protein assay was conducted on the same sample to measure the total protein concentration. Following normalization to these total protein concentrations, the final ALP activity was calculated relative to the control group, thereby yielding a comparative measure. Alizarin red S (ARS) staining and quantification MSCs were initially fixed in 95% ethanol for 30 min. Following fixation, the cells were stained with 1% ARS at room temperature (RT), maintaining a pH of 4.2. After staining, the ARS stains were dissolved in 100 mmol/L of cetylpyridinium chloride for 30 min. Subsequently, mineralization was quantified based on the absorbance value measured at 490 nm, providing a quantitative assessment of mineralization levels. Oil red O staining and quantification MSCs were cultured under three distinct conditions: PM, AM, and AM supplemented with apoVs. The culture medium was replaced every 2–3 days. Following a 14-day incubation period, oil red O staining and subsequent quantification were performed as outlined [71]. For quantitative assessment, the stained cells were immersed in 100% isopropyl alcohol for 10 min to dissolve the stain, and the absorbance was measured at 500 nm. β-galactosidase activity staining and quantification MSCs were initially fixed in paraformaldehyde (PFA) for 30 min at RT and stained overnight (37 °C) with SA-β-gal solution. Blue precipitate-positive cells were visualized under a light microscope. Reactive oxygen species (ROS) staining and quantification To assess intracellular ROS levels, MSCs were incubated with 10 µM DCFH-DA (a cell-permeable ROS indicator) in serum-free medium for 30 min at 37 °C. After washing with PBS, cells were either observed under a fluorescence microscope (Ex/Em: ~488/525 nm) for qualitative analysis。. RNA extraction and RT-qPCR Total cellular RNA was extracted using the TRIzol reagent (Invitrogen, USA). Following this, cDNA was synthesized utilizing the PrimeScript RT reagent kit (Takara, Japan). For quantitative assessment, RT-qPCR was conducted on an ABI Prism 7500 real-time PCR System. The 2–ΔΔCT method was employed to calculate the relative expression of each gene, with Gapdh serving as the internal reference gene. The outcomes were expressed in terms of fold changes. The primer sequences utilized in this investigation are comprehensively documented in Table S3. Protein extraction and Western blotting Protein extraction was achieved through cell lysis with RIPA buffer (Santa Cruz Biotechnology, USA) supplemented with 2% proteinase inhibitor. The extracted proteins were then separated on a 10% polyacrylamide gel via SDS-PAGE (Lablead, China) and transferred onto PVDF membranes (Millipore, USA). Membranes were blocked with 5% non-fat milk in TBST for 1 h at RT. Subsequently, membranes were incubated with primary antibodies overnight at 4 °C, washed with TBST three times, and then incubated with species-specific peroxidase-conjugated secondary antibodies for 1 h at RT. Protein bands were visualized using an ECL kit (CoWin Biotech, China) and captured by a Fusion FX6 imaging system (Vilber, France). SiRNA transfection One day before transfection, cells were seeded at a density of 1 × 105 cells/well in 24-well plates using antibiotic-free medium. When the cells reached approximately 80% confluency, siRNA (Santa Cruz, USA) was introduced at a final concentration of 100 nM using Lipofectamine™ 2000 transfection reagent (Invitrogen, USA), following the manufacturer’s protocol [72]. Lentiviral vector construction and packaging The full-length mouse Thy1 gene was amplified via PCR. The Thy1 gene and pLVX-IRES-ZsGreen1 vector were digested with XhoI and BamHI. The fragments were ligated and transformed into E. coli to generate the lentiviral expression vector [73]. Magnetic-activated ApoV sorting Magnetic-activated sorting of apoVs was performed according to our previously established method [26]. Briefly, apoVs were first incubated with primary antibodies targeting Thy1 (1:1000, Abcam, UK), followed by incubation with PE-conjugated goat anti-rabbit IgG (H + L) (Proteintech, USA) and then anti-PE MicroBeads (Miltenyi Biotec, Germany). Between each step, apoVs were thoroughly washed with 0.1 μm-filtered PBS. Finally, apoVs were passed through a column placed in a magnetic field. The fraction that flowed through the column was identified as the Thy1-negative subpopulation, and the fraction that adsorbed in the column was identified as the Thy1-positive subpopulation. Protein corona formation The purified apoVs are mixed with target protein(s) in a buffer solution, at physiological pH and temperature, to facilitate the spontaneous formation of a protein corona around the apoV membrane. This process relies on electrostatic interactions, hydrophobic forces, and other non-covalent binding mechanisms between the apoV surface and the proteins [74]. After incubation, unbound proteins are removed through washing steps, leaving behind apoVs coated with a stable protein corona. Intravenous administration of ApoVs in aged mice Male C57BL/6J mice, aged 18 months, were randomly assigned to several groups: mice in the treatment groups underwent weekly tail intravenous injections of apoVs, while mice in the control group (blank) received equivalent volumes of a vehicle solution (PBS, 0.1 μm-filtered). Following the completion of the supplementation period, all mice were anesthetized and euthanized. Subsequently, femurs, visceral organs, and abdominal fat were collected, fixed in 4% PFA, and then underwent decalcification, embedding, slicing, and staining for further analysis. Statistical analysis Data collected from at least three independent experiments were expressed as mean ± standard deviation (SD). For statistical analysis, SPSS 26 software (IBM, USA) was employed. Comparisons between two groups were conducted using either an independent two-tailed Student’s t-test or a two-tailed Student’s t-test with Welch correction, depending on the data distribution. For multiple group comparisons, one-way ANOVA followed by Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test was applied. Statistical significance was determined at a p-value of less than 0.05. Table S1. Original data of proteomic analysis. Table S2. Reagents and antibodies. Table S3. Primers for qRT-PCR. Table S4. Sequences of siRNA. Fig. 1Apoptotic vesicles (apoVs) play a critical role in regulating bone-fat metabolism by orchestrating the balance between bone remodeling and lipid metabolism. This study revealed that while the fundamental properties of apoVs remained stable during development and aging, their functions and protein compositions were profoundly altered by donor aging. Thy1 was identified as a key biomarker, influencing apoV-mediated osteogenesis and adipogenesis. By engineering Thy1-enriched apoVs through magnetic-activated sorting and protein corona modification, we developed a highly effective therapeutic approach to restore senile bone-fat balance. Mechanistically, Thy1 enhanced osteogenic differentiation and suppressed adipogenesis in mesenchymal stem cells (MSCs) via activating the Thy1-ERK-TAZ axis. These findings systematically demonstrate the impact of donor aging on apoVs, paving the way for quality control and engineering of apoV-based strategy in aging-related metabolic diseases Apoptotic vesicles (apoVs) play a critical role in regulating bone-fat metabolism by orchestrating the balance between bone remodeling and lipid metabolism. This study revealed that while the fundamental properties of apoVs remained stable during development and aging, their functions and protein compositions were profoundly altered by donor aging. Thy1 was identified as a key biomarker, influencing apoV-mediated osteogenesis and adipogenesis. By engineering Thy1-enriched apoVs through magnetic-activated sorting and protein corona modification, we developed a highly effective therapeutic approach to restore senile bone-fat balance. Mechanistically, Thy1 enhanced osteogenic differentiation and suppressed adipogenesis in mesenchymal stem cells (MSCs) via activating the Thy1-ERK-TAZ axis. These findings systematically demonstrate the impact of donor aging on apoVs, paving the way for quality control and engineering of apoV-based strategy in aging-related metabolic diseases Fig. 2Osteoinductive potential of MSC-derived apoVs. MSCs were cultured in proliferation medium (PM), osteogenic induction medium (OM), OM with 250 ng/ml 1w apoVs, 3 m apoVs, or 18 m apoVs for 0–14 days. (a-b) Alkaline phosphatase (ALP) staining and quantification of MSCs at 7 days. Alizarin red S (ARS) staining and quantification of MSCs at 14 days. (c) Expression of osteogenic genes at 7–10 days. Nude mice were randomly separated into five groups (N = 6) and implanted with β-TCP particles, β-TCP particles loaded with aging MSCs, β-TCP particles loaded with aging MSCs pretreated with 1w apoVs, 3 m apoVs, or 18 m apoVs. (d) Hematoxylin and eosin (H&E) staining and Masson’s trichrome staining of implants. Scale bar: 50 μm. Data are presented as mean ± standard deviation (SD). Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, **p < 0.01, ***p < 0.001, ****p < 0.0001 Osteoinductive potential of MSC-derived apoVs. MSCs were cultured in proliferation medium (PM), osteogenic induction medium (OM), OM with 250 ng/ml 1w apoVs, 3 m apoVs, or 18 m apoVs for 0–14 days. (a-b) Alkaline phosphatase (ALP) staining and quantification of MSCs at 7 days. Alizarin red S (ARS) staining and quantification of MSCs at 14 days. (c) Expression of osteogenic genes at 7–10 days. Nude mice were randomly separated into five groups (N = 6) and implanted with β-TCP particles, β-TCP particles loaded with aging MSCs, β-TCP particles loaded with aging MSCs pretreated with 1w apoVs, 3 m apoVs, or 18 m apoVs. (d) Hematoxylin and eosin (H&E) staining and Masson’s trichrome staining of implants. Scale bar: 50 μm. Data are presented as mean ± standard deviation (SD). Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, **p < 0.01, ***p < 0.001, ****p < 0.0001 Fig. 3Therapeutic effects of MSC-derived apoVs in senile bone-fat imbalance. Mice were randomly divided into 4 groups (N = 6) and intravenously injected with 1w apoVs, 3 m apoVs, 18 m apoVs, or vehicle once a week for 8 weeks. (a) Representative 3D reconstruction images of trabecular bone and longitudinal section images of the distal femurs. Scale bar: 1 mm. (b) Quantitative analysis of femoral micro-structure. BMD: bone mineral density, BV/TV: bone volume/total bone volume. (c) H&E staining of the distal femurs. Scale bar: 500 μm. (d) Masson’s trichrome staining of the distal femurs. Scale bar: 500 μm. (e-f) In vitro osteogenic differentiation ability of endogenous mice MSCs. (g-h) In vitro adipogenic differentiation ability of endogenous mice MSCs. Scale bar: 100 μm. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. ns, no significance, **p < 0.01, ****p < 0.0001 Therapeutic effects of MSC-derived apoVs in senile bone-fat imbalance. Mice were randomly divided into 4 groups (N = 6) and intravenously injected with 1w apoVs, 3 m apoVs, 18 m apoVs, or vehicle once a week for 8 weeks. (a) Representative 3D reconstruction images of trabecular bone and longitudinal section images of the distal femurs. Scale bar: 1 mm. (b) Quantitative analysis of femoral micro-structure. BMD: bone mineral density, BV/TV: bone volume/total bone volume. (c) H&E staining of the distal femurs. Scale bar: 500 μm. (d) Masson’s trichrome staining of the distal femurs. Scale bar: 500 μm. (e-f) In vitro osteogenic differentiation ability of endogenous mice MSCs. (g-h) In vitro adipogenic differentiation ability of endogenous mice MSCs. Scale bar: 100 μm. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. ns, no significance, **p < 0.01, ****p < 0.0001 Fig. 4Thy1 emerged as a potential biomarker. (a-b) Proteins that rise and fall continuously with parental cell aging were verified by quantitative proteomics and western blotting. (c) Subcellular localization and expression of Thy1 in apoVs revealed by immuno-gold labeling. The black dots are gold particles, which are localized at the apoV periphery. Scale bar: 100 nm. (d) Uptake of Thy1 was shown by immunofluorescence staining. Thy1 in apoVs were labeled with IgG H&L (Alexa Fluor® 647, red), the cytoskeleton was stained with phalloidin (green), and cell nuclei were marked with DAPI (blue). Scale bar: 5 μm Thy1 emerged as a potential biomarker. (a-b) Proteins that rise and fall continuously with parental cell aging were verified by quantitative proteomics and western blotting. (c) Subcellular localization and expression of Thy1 in apoVs revealed by immuno-gold labeling. The black dots are gold particles, which are localized at the apoV periphery. Scale bar: 100 nm. (d) Uptake of Thy1 was shown by immunofluorescence staining. Thy1 in apoVs were labeled with IgG H&L (Alexa Fluor® 647, red), the cytoskeleton was stained with phalloidin (green), and cell nuclei were marked with DAPI (blue). Scale bar: 5 μm Fig. 5Thy1 was responsible for the apoV-mediated upregulation of MSC osteogenesis. SiRNA interference was performed in 1w MSCs, and their apoVs (si 1w apoVs) were extracted. MSCs cultured in OM were induced with 250 ng/ml 1w apoVs or si 1w apoVs for 7–14 days. (a-b) ALP activity and calcium deposition of MSCs. (c) Expression of osteogenic genes in MSCs. (d) Expression of pivotal proteins in MSCs. Lentivirus transfection was performed in 18 m MSCs, and their apoVs (virus 18 m apoVs) were extracted. MSCs cultured in OM were induced with 250 ng/ml 18 m apoVs or virus 18 m apoVs for 7–14 days. (e-f) ALP stain/quantification and ARS stain/quantification of MSCs. (g) RT-qPCR analysis detected the osteogenic genes. (h) Western blotting testified to the pivotal proteins. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, ***p < 0.001, ****p < 0.0001 Thy1 was responsible for the apoV-mediated upregulation of MSC osteogenesis. SiRNA interference was performed in 1w MSCs, and their apoVs (si 1w apoVs) were extracted. MSCs cultured in OM were induced with 250 ng/ml 1w apoVs or si 1w apoVs for 7–14 days. (a-b) ALP activity and calcium deposition of MSCs. (c) Expression of osteogenic genes in MSCs. (d) Expression of pivotal proteins in MSCs. Lentivirus transfection was performed in 18 m MSCs, and their apoVs (virus 18 m apoVs) were extracted. MSCs cultured in OM were induced with 250 ng/ml 18 m apoVs or virus 18 m apoVs for 7–14 days. (e-f) ALP stain/quantification and ARS stain/quantification of MSCs. (g) RT-qPCR analysis detected the osteogenic genes. (h) Western blotting testified to the pivotal proteins. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, ***p < 0.001, ****p < 0.0001 Fig. 6Thy1 deficiency impaired the osteoinductive capability of apoVs. (a-b) Osteogenesis of MSCs from wild-type (WT) or Thy1(-/-) mice. (c) Morphology of apoVs derived from Thy1(-/-) MSCs. Scale bar: 50 nm. (d) Size distribution of apoVs derived from Thy1(-/-) MSCs. (e) Zeta potential of apoVs derived from Thy1(-/-) MSCs. (f) Phosphatidylserine (PS) exposure of apoVs derived from Thy1(-/-) MSCs. In osteogenic induction experiments, MSCs from WT were cultured in OM in combination with 250 ng/ml apoVs derived from 1-week-old WT or Thy1(-/-) mice for 7–14 days. (c-d) ALP activity and mineralization of MSCs. (e) Expression of osteogenic genes in MSCs. (f) Expression of pivotal proteins in MSCs. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, ****p < 0.0001 Thy1 deficiency impaired the osteoinductive capability of apoVs. (a-b) Osteogenesis of MSCs from wild-type (WT) or Thy1(-/-) mice. (c) Morphology of apoVs derived from Thy1(-/-) MSCs. Scale bar: 50 nm. (d) Size distribution of apoVs derived from Thy1(-/-) MSCs. (e) Zeta potential of apoVs derived from Thy1(-/-) MSCs. (f) Phosphatidylserine (PS) exposure of apoVs derived from Thy1(-/-) MSCs. In osteogenic induction experiments, MSCs from WT were cultured in OM in combination with 250 ng/ml apoVs derived from 1-week-old WT or Thy1(-/-) mice for 7–14 days. (c-d) ALP activity and mineralization of MSCs. (e) Expression of osteogenic genes in MSCs. (f) Expression of pivotal proteins in MSCs. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, ****p < 0.0001 Fig. 7Thy1 deficiency impaired the therapeutic potential of apoVs in vivo. Mice were randomly divided into 4 groups (N = 6) and intravenously injected with 1w apoVs, si 1w apoVs, Thy1(-/-) 1w apoVs, or vehicle once a week for 8 weeks. (a) Representative 3D reconstruction images of trabecular bone and longitudinal section images of the distal femurs. Scale bar: 1 mm. (b) Quantitative analysis of femoral micro-structure. BMD: bone mineral density, BV/TV: bone volume/total bone volume. (c) H&E staining of the distal femurs. Scale bar: 500 μm. (d) Masson’s trichrome staining of the distal femurs. Scale bar: 500 μm. (e-f) In vitro osteogenic differentiation ability of endogenous mice MSCs. (g-h) In vitro adipogenic differentiation ability of endogenous mice MSCs. Scale bar: 100 μm. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. ns, no significance, ****p < 0.0001 Thy1 deficiency impaired the therapeutic potential of apoVs in vivo. Mice were randomly divided into 4 groups (N = 6) and intravenously injected with 1w apoVs, si 1w apoVs, Thy1(-/-) 1w apoVs, or vehicle once a week for 8 weeks. (a) Representative 3D reconstruction images of trabecular bone and longitudinal section images of the distal femurs. Scale bar: 1 mm. (b) Quantitative analysis of femoral micro-structure. BMD: bone mineral density, BV/TV: bone volume/total bone volume. (c) H&E staining of the distal femurs. Scale bar: 500 μm. (d) Masson’s trichrome staining of the distal femurs. Scale bar: 500 μm. (e-f) In vitro osteogenic differentiation ability of endogenous mice MSCs. (g-h) In vitro adipogenic differentiation ability of endogenous mice MSCs. Scale bar: 100 μm. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. ns, no significance, ****p < 0.0001 Fig. 8ApoV-mediated osteoinduction depended on ERK phosphorylation. In osteogenic induction experiments, MSCs were cultured in PM, OM, OM with 250 ng/ml 1w apoVs, or 1w apoVs and selective ERK inhibitor (GDC-0994) for 7–14 days. (a-b) ALP stain/quantification and ARS stain/quantification of MSCs. (c) RT-qPCR analysis of osteogenic genes. (d) Expression of pivotal proteins in MSCs. Data are presented as mean ± SD. Statistical analyses are performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. ns, no significance, ****p < 0.0001 ApoV-mediated osteoinduction depended on ERK phosphorylation. In osteogenic induction experiments, MSCs were cultured in PM, OM, OM with 250 ng/ml 1w apoVs, or 1w apoVs and selective ERK inhibitor (GDC-0994) for 7–14 days. (a-b) ALP stain/quantification and ARS stain/quantification of MSCs. (c) RT-qPCR analysis of osteogenic genes. (d) Expression of pivotal proteins in MSCs. Data are presented as mean ± SD. Statistical analyses are performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. ns, no significance, ****p < 0.0001 Fig. 9Thy1-enriched apoVs possessed superior osteoinductive capability in vitro. Magnetic bead sorting was applied in 1w apoVs. MSCs were cultured in OM in combination with 250 ng/ml Thy1+ apoVs or Thy1− apoVs for 7–14 days. (a-b) ALP activity and mineralization of MSCs. (c) Expression of osteogenic genes in MSCs. (d) Expression of pivotal proteins in MSCs. Thy1 protein corona was added to Thy1+ apoVs (Thy1++ apoVs). MSCs were cultured in OM in combination with 250 ng/ml Thy1+ apoVs or Thy1++ apoVs for 7–14 days. (e-f) ALP activity and calcium deposition of MSCs. (g) Expression of osteogenic genes in MSCs. (h) Expression of pivotal proteins in MSCs. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, ***p < 0.001, ****p < 0.0001 Thy1-enriched apoVs possessed superior osteoinductive capability in vitro. Magnetic bead sorting was applied in 1w apoVs. MSCs were cultured in OM in combination with 250 ng/ml Thy1+ apoVs or Thy1− apoVs for 7–14 days. (a-b) ALP activity and mineralization of MSCs. (c) Expression of osteogenic genes in MSCs. (d) Expression of pivotal proteins in MSCs. Thy1 protein corona was added to Thy1+ apoVs (Thy1++ apoVs). MSCs were cultured in OM in combination with 250 ng/ml Thy1+ apoVs or Thy1++ apoVs for 7–14 days. (e-f) ALP activity and calcium deposition of MSCs. (g) Expression of osteogenic genes in MSCs. (h) Expression of pivotal proteins in MSCs. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test or Welch’s ANOVA with Dunnett’s T3 post hoc test for multiple group comparisons. ns, no significance, ***p < 0.001, ****p < 0.0001 Fig. 10Thy1 enrichment enhanced the therapeutic potential in vivo. Mice were randomly divided into 4 groups (N = 6) and intravenously injected with 1w apoVs, Thy1+ apoVs, Thy1++ apoVs, or vehicle once a week for 8 weeks. (a) Representative 3D reconstruction images of trabecular bone and longitudinal section images of the distal femurs. Scale bar: 1 mm. (b) Quantitative analysis of femoral micro-structure. BMD: bone mineral density, BV/TV: bone volume/total bone volume. (c) H&E staining of the distal femurs. Scale bar: 500 μm. (d) Masson’s trichrome staining of the distal femurs. Scale bar: 500 μm. (e-f) In vitro osteogenic differentiation ability of endogenous mice MSCs. (g-h) In vitro adipogenic differentiation ability of endogenous mice MSCs. Scale bar: 100 μm. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Thy1 enrichment enhanced the therapeutic potential in vivo. Mice were randomly divided into 4 groups (N = 6) and intravenously injected with 1w apoVs, Thy1+ apoVs, Thy1++ apoVs, or vehicle once a week for 8 weeks. (a) Representative 3D reconstruction images of trabecular bone and longitudinal section images of the distal femurs. Scale bar: 1 mm. (b) Quantitative analysis of femoral micro-structure. BMD: bone mineral density, BV/TV: bone volume/total bone volume. (c) H&E staining of the distal femurs. Scale bar: 500 μm. (d) Masson’s trichrome staining of the distal femurs. Scale bar: 500 μm. (e-f) In vitro osteogenic differentiation ability of endogenous mice MSCs. (g-h) In vitro adipogenic differentiation ability of endogenous mice MSCs. Scale bar: 100 μm. Data are presented as mean ± SD. Statistical analyses were performed by one-way ANOVA with Tukey’s post hoc test for multiple group comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Supplementary Information Supplementary Material 1. Supplementary Material 1.