Cold exposure induces bone loss and alters bone metabolism in mice

To directly investigate the impact of cold exposure on bone metabolism, C57BL/6 mice were subjected to cold temperature (CT; 4–6 °C) for 8 weeks, starting at 12 weeks of age, while control groups were maintained at room temperature (RT; 22–24 °C) throughout the experiment (Fig. 1A). Subsequently, mice were euthanized, and their bone tissues were prepared for further analysis. Micro-computed tomography (µCT) analysis of femurs revealed that CT mice exhibited significantly reduced bone mass and more compromised bone microstructures compared to RT mice. Specifically, CT mice showed significantly lower bone mineral density (BMD), trabecular bone volume fraction (Tb. BV/TV), trabecular number (Tb. N), and trabecular thickness (Tb. Th), as well as a trend of increased trabecular separation (Tb. Sp), without altering cortical bone area fraction (Ct. Ar/Tt. Ar), but with a decrease in cortical thickness (Ct. Th) (Fig. 1B–I). Calcein double labeling demonstrated impaired new bone formation and mineralization in CT mice, as indicated by bone formation rate per bone surface (BFR/BS) and mineral apposition rate (MAR) values (Fig. 1J–L). OCN immunohistochemical staining revealed fewer osteoblasts on the trabecular bone surface in CT mice compared to RT mice (Fig. 1M, N). TRAP staining showed a significant increase in the number of osteoclasts due to cold exposure (Fig. 1O, P). Additionally, cold exposure led to reductions in femur length and body weight (Fig. S1A–C). These findings suggest that cold exposure disrupts bone metabolism by inhibiting osteoblastic bone formation and increasing osteoclastic bone resorption, resulting in decreased bone mass.

Fig. 1

Cold exposure induced bone loss and altered bone metabolism in mice. A, The schematic flow diagram represents the in vivo treatment of CT or RT. n = 5 per group. B, Representative micro-CT images of trabecular (top) and cortical (bottom) bone in RT- or CT-treated mice. Scale bars represent 500 μm (top) and 1 mm (bottom). C–I, Parameters of bone mass analysed by µCT: BMD, bone mineral density; Tb. BV/TV, bone volume over tissue volume; Tb. Th, trabecular thickness; Tb. N, trabecular number; Tb. Sp, trabecular separation; Ct. Ar/Tt. Ar, cortical bone area fraction; Ct. Th, cortical thickness. n = 5 per group. J, Calcein double labelling images of the mineralized surface of mouse femora. Scale bar represents 50 μm. K, L, Parameters of bone formation. MAR, BFR/BS, n = 5 per group. M, Representative OCN-stained section with quantification of the (N) number of osteoblasts (N. OBs) on the trabecular bone surface (BS) in distal femora from mice treated with RT or CT. n = 5 per group. Scale bar represents 100 μm. O, Representative TRAP-stained sections with quantification of the (P) number of osteoclasts (N. OCs) on the trabecular bone surface (BS) in distal femora from mice treated with RT or CT. n = 5 per group. Scale bar represents 100 μm. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

Inhibition of exosomes release alleviated CT-induced bone loss

To further investigate the involvement of exosomes in CT-induced bone loss, we systematically inhibited exosomes release in mice exposed to cold temperatures. We utilized the pharmaceutical drug, GW4869, known for its efficient suppression of exosome biogenesis/release, to treat CT mice (Fig. 2A). Plasma was collected from mice, exosomes were extracted, and nanoparticle tracking analysis (NTA) was conducted. We confirmed that GW4869 significantly reduced the number of plasma exosomes (Fig. S2A). Bone marrow macrophages (BMMs) were isolated from GW4869- or vehicle-treated CT mice, and it was observed that GW4869 did not alter the osteoclastogenic activity of BMMs from CT mice (Fig. S2B, C). However, we found an increase in osteogenic activity in bone marrow mesenchymal stem cells (BMSCs) from GW4869-treated CT mice compared to those from CT mice, as evidenced by Alizarin Red staining and the expression of COL-1, RUNX2, and BMP2 (Fig. S2D-G). As expected, administration of GW4869 in CT mice effectively improved bone microarchitecture, resulting in increased BMD, Tb. BV/TV, Tb. N, Tb. Th, Ct. Ar/Tt. Ar, and Ct. Th, as well as decreased Tb. Sp, compared to vehicle-treated CT-induced mice (Fig. 2B–I). GW4869 also increased new bone formation and mineralization in CT mice, as shown in Fig. 2J–L. While GW4869 enhanced the number of osteoblasts on the trabecular bone surface of CT mice, indicated by OCN immunohistochemical staining, it had no effect on the number of osteoclasts, as indicated by TRAP staining (Fig. 2M–P). Additionally, GW4869 increased the length of the femur in CT mice (Fig. S3A, B) and elevated the body weight of CT mice (Fig. S3C). These experimental findings suggest that inhibiting the release of exosomes can significantly reduce the loss of bone mass caused by cold exposure, indicating a potential significant role of exosomes in CT-induced bone loss.

Fig. 2

Inhibition of exosomes release alleviated CT-induced bone loss. A, Experimental design of the mice treated with CT + vehicle or CT + GW4869 by intraperitoneal injection (n = 5 per group). B, Representative µCT images of trabecular (top) and cortical (bottom) bone in CT + vehicle- or CT + GW4869-treated mice. Scale bars represent 500 μm (top) and 1 mm (bottom). C–I, Parameters of trabecular bone mass analysed by micro-CT: BMD, Tb. BV/TV, Tb. Th, Tb. N, Tb. Sp, Ct. Ar/Tt. Ar, Ct. Th. n = 5 per group. J, Calcein double labelling images of the mineralized surface of mouse femora. Scale bar represents 50 μm. K, L, Parameters of bone formation. MAR, mineral apposition rate; BFR/BS, bone formation rate//bone surface. n = 5 per group. M, Representative OCN-stained section. Scale bar represents 100 μm. N, Quantification of the number of osteoblasts (N. OBs) on the trabecular bone surface (BS) in distal femora. n = 5 per group. O, TRAP-stained sections. Scale bar represents 100 μm. P, quantification of the number of osteoclasts (N. OCs) on the trabecular bone surface (BS) in distal femora. n = 5 per group. The vehicle referred to is dimethyl sulfoxide (DMSO). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

Plasma from CT-exposed mice can reduce bone mass, but removing exosomes alleviates CT-plasma-induced bone loss

The above experiments demonstrated the essential role of exosomes in cold exposure-induced bone loss. Thus, we aimed to investigate the effects of plasma and exosomes in plasma on bone metabolism during cold exposure. We treated the RT mice with plasma from CT mice (CT-plasma) or CT mice plasma which exosomes had been removed (CT-plasma-EXO free) by ultracentrifugation (Fig. 3A). As shown in Fig. 3B–I, CT-plasma impaired BMD, Tb. BV/TV, Tb. N, Tb. Th, and Ct. Th, as well as increasing Tb. Sp, compared to the control group mice. Calcein double labelling revealed the impairment of new bone formation and mineralization in the CT-plasma-treated mice (Fig. 3J–L). When comparing CT-plasma-treated mice to control mice, OCN immunohistochemistry staining revealed a reduced number of osteoblasts on the trabecular bone surface (Fig. 3M, N). TRAP staining showed that CT-plasma increased the number of osteoclasts (Fig. 3O, P). However, when the RT mice were treated with CT-plasma-EXO free, the bone mass and bone microstructures were improved compared with CT-plasma-treated mice (Fig. 3B–I). Furthermore, removing the exosomes also ameliorated new bone formation and mineralization (Fig. 3J–L). After removing the exosomes, the impact of CT-plasma on osteoblasts and osteoclasts was also ameliorated (Fig. 3M–P). Furthermore, we observed that CT-plasma reduced femur length; however, this effect was attenuated upon exosomes removal (Fig. S4A, B). However, no significant difference in body weight was observed among these groups (Fig. S4C). These data indicate that CT-plasma plays a pivotal role in CT-induced bone loss, with CT-EXO exerting significant effect.

Fig. 3

CT-induced plasma can decrease bone mass, removal of exosomes can ameliorate CT-induced plasma-triggered osteopenia. A, Schematic flow diagram representing mice treated with PBS, CT-plasma, or CT-plasma-EXO. n = 5 per group. B, Representative micro-CT images of trabecular (top) and cortical (bottom) bone in CT-plasma- or CT-plasma-EXO-treated mice. Scale bars represent 500 μm (top) and 1 mm (bottom). C–I, Parameters of trabecular bone mass analysed by micro-CT: BMD, Tb. BV/TV, Tb. Th, Tb. N, Tb. Sp, Ct. Ar/Tt. Ar, Ct. Th. n = 5 per group. J, Calcein double labelling images of the mineralized surface of mouse femora. Scale bar represents 50 μm. K, L, Parameters of bone formation. MAR, mineral apposition rate; BFR/BS, bone formation rate/bone surface. n = 5 per group. M, Representative OCN-stained section. Scale bar represents 100 μm. N, Quantification of the number of osteoblasts (N. OBs) on the trabecular bone surface (BS) in distal femora. n = 5 per group. O, TRAP-stained sections. Scale bar represents 100 μm. P, quantification of the number of osteoclasts (N. OCs) on the trabecular bone surface (BS) in distal femora. n = 5 per group. The control group were injected PBS. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

Plasma-derived exosomes from CT-exposed mice can be taken up by BMSCs and impair their osteogenic differentiation

To further elucidate the role of plasma exosomes in cold-induced bone loss, we isolated exosomes from the plasma of CT or RT mice to investigate their involvement in the osteogenic development of BMSCs. Transmission electron microscopy (TEM) analysis confirmed the morphology of exosomes isolated from plasma, showing a cup- or sphere-shaped structure (Fig. 4A). NTA revealed that the size of these particles ranged mostly from 40 nm to 150 nm (Fig. 4B), consistent with previously reported exosome size distributions [41]. Additionally, exosomal marker proteins CD9, CD81, and TSG101 were found to be highly abundant in plasma-derived exosomes (Fig. 4C).

Fig. 4

CT-EXO can be taken up by BMSCs and impair the osteogenic differentiation of BMSCs. A, Representative image of the ultrastructure of exosomes observed by transmission electron microscopy. Scale bar represents 100 μm. B, Average particle size distribution of exosomes. C, Exosomes markers CD9, CD81, and TSG101 determined by western blotting. D, Representative fluorescence micrograph of DiO-labelled exosomes (green) internalized by primary BMSCs, while blue represents the nucleus. The labelled exosomes were co-incubated with BMSCs for 12 h. Scale bar represents 20 μm. E, Representative western blot image showing the effect of CT-EXO on the protein levels of RUNX2 in BMSCs after 48 h co-incubation. F, G, Representative image of microscopic view (F) and entire plate view (G)ARS staining of BMSCs after exosomes treatment. Scale bars represents 250 μm. H, Quantification of the ARS. n = 3 per group. I, Representative image of colony formation assay after BMSCs were treated with exosomes. J, Representative image of SA-β-gal staining of BMSCs after exosomes treatment. Scale bar represents 100 μm. OIM represent the osteogenesis induced medium, and the Control represent the BMSCs without using the osteogenesis induced medium. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

To investigate the potential influence of plasma-derived exosomes on BMSCs function, we assessed their uptake by BMSCs. Plasma-derived exosomes were labeled with DiO and then co-incubated with BMSCs. Both exosomes from the plasma of room temperature-exposed mice (RT-EXO) and cold temperature-exposed mice (CT-EXO) were found to be taken up by BMSCs, as demonstrated in Fig. 4D. BMSCs were treated with RT-EXO and CT-EXO at a dose of 100 µg/mL. Interestingly, CT-EXO significantly suppressed osteogenic differentiation, as evidenced by a decrease in RUNX2 protein levels (Fig. 4E) and Alizarin Red staining (Fig. 4F–H). Conversely, RT-EXO stimulated osteogenic differentiation of BMSCs (Fig. 4E–H). Furthermore, as indicated in Fig. 4I, RT-EXO treatment promoted colony formation, while CT-EXO decreased colony formation. Additionally, CT-EXO may increase BMSCs senescence, as evidenced by more SA-β-gal-positive BMSCs, whereas RT-EXO inhibited BMSCs senescence (Fig. 4J). These findings suggest that CT-EXO play a significant role in BMSCs osteogenic differentiation. Moreover, we observed that CT-EXO have a promoting effect on osteoclast differentiation of BMMs, as demonstrated by TRAP staining (Fig. S5A, B). These data indicate that CT-EXO significantly inhibit the osteogenic differentiation ability of BMSCs and have a certain promoting effect on osteoclast differentiation of BMMs. However, since the earlier finding that inhibition of exosomes release did not significantly inhibit osteoclast differentiation in vivo, we primarily focused our research on the effect of CT-EXO on osteogenic differentiation of BMSCs.

CT-EXO reduced bone mass and disrupted bone metabolism in RT mice, while RT-EXO increased bone mass in RT mice and mitigated CT-induced bone loss

DiR-labelled plasma-derived exosomes were injected into mice via the tail vein, and their biodistribution was monitored in vivo to assess uptake by bone tissues. Fluorescence was primarily observed in the liver and spleen (Fig. 5A and S6A). We hypothesize that the intense fluorescence signals emitted by the liver and spleen masked the signals originating from bone tissues. Consequently, we performed distinct fluorescence imaging of bone tissues. Subsequently, upon sacrificing the mice, we isolated their tibias and femora to examine the fluorescent signals. Both the tibia and femur displayed robust fluorescent signals (Fig. 5B), indicating successful injection of DiR-labelled exosomes via the tail vein and their delivery to bone tissue. To examine exosomes distribution at the microscopic level, a green fluorescent dye (DiO) was employed to label exosomes. Fluorescent signals in tissues from mice receiving labelled exosomes via the tail vein were observed using a fluorescence microscope, revealing the presence of DiO-labelled exosomes in the liver and spleen (Fig. S6B, C). The presence of DiO-labelled exosomes in bone tissues is illustrated in Fig. 5C, suggesting that plasma-derived exosomes might target bone cells and modulate bone metabolism.

Fig. 5

CT-EXO decreased bone mass and altered bone metabolism in mice. A, Representative fluorescence image of exosomes distribution in mice 24 h after exosome injection. B, Representative ex vivo fluorescence image of exosomes distribution in femur 24 h after exosomes injection. C, Representative fluorescence micrograph showing the DiO-labelled exosomes (green fluorescence) in the femur section. Scale bar represents 100 μm. CB: cortical bone; TB: trabecular bone; BM: bone marrow. D, Schematic flow diagram representing mice treated with PBS, RT-EXO, and CT-EXO. n = 6 per group. E, Representative µCT images of trabecular (top) and cortical (bottom) bone in femora from mice in PBS, RT-EXO, and CT-EXO groups. Scale bars represent 500 μm (top) and 1 mm (bottom). F–L, Quantitative analysis of BMD; Tb. BV/TV, Tb. Th, Tb. N, Tb. Sp, Ct. Ar/Tt. Ar, Ct. Th. n = 6 per group. M, Calcein double labelling images of the mineralized surface of mouse femora. Scale bar represents 50 μm. N, O, Quantitation of BFR/BS and MAR. n = 6 per group. P, Representative OCN-stained sections with quantification of (Q) osteoblast number. Scale bar: 100 μm. n = 6 per group. R, Representative TRAP-stained sections with quantification of (S) osteoclast number. Scale bar represents 100 μm. n = 6 per group. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

Then, we investigated whether direct administration of PBS, RT-EXO and CT-EXO could influence bone mass and alter bone metabolism in mice cultured in room temperature (RT). To determine the optimal dosage of exosomes for intervention, we pre-treated RT mice with various doses of CT-EXO (Fig. S7A). We observed that the most significant reduction in bone mass occurred with an intervention of 200 µg of CT-EXO (Fig. S7B-I). Consequently, we selected this dosage for subsequent experiments. Subsequently, the RT mice received treatments with RT-EXO or CT-EXO twice a week for 8 weeks (Fig. 5D). Micro-computed tomography (µCT) analysis revealed that administration of RT-EXO potentially improved bone density and enhanced bone microarchitecture (Fig. 5E–J). Conversely, CT-EXO significantly induced bone loss and disrupted bone microarchitecture, as evidenced by significantly lower BMD, Tb. BV/TV, Tb. N, Tb. Th, Ct. Th, and higher Tb. Sp compared to mice treated with RT-EXO (Fig. 5E–J). There was no statistically significant difference in cortical bone between the groups (Fig. 5K, L). Calcein double labeling showed that RT-EXO administration might stimulate new bone formation and mineralization in mice compared with those treated with PBS, whereas CT-EXO treatment led to a significant decrease in new bone formation and mineralization compared to PBS- or RT-EXO-treated mice (Fig. 5M–O). OCN immunohistochemical staining revealed a greater number of osteoblasts on the trabecular bone surface of RT-EXO-treated mice compared to the PBS group, whereas CT-EXO significantly decreased the number of osteoblasts in mice compared with PBS- or RT-EXO-treated mice (Fig. 5P, Q). CT-EXO induced significant increases in the number of osteoclasts, as determined by TRAP staining (Fig. 5R, S). However, a decrease of the number of osteoclasts was not observed in mice administered with RT-EXO (Fig. 5R, S). We also found that CT-EXO decreased the length of the femur, but RT-EXO increased it (Fig. S8A, B). However, there was no significant difference in body weight in these groups (Fig. S8C). The preceding results indicate that CT-EXO can substantially reduce the bone mass of RT mice.

To investigate the potential of RT-EXO in mitigating CT-induced bone loss, CT mice were administered either PBS or RT-EXO via tail vein injection twice weekly for 8 weeks (Fig. S9A). µCT analysis revealed that RT-EXO administration effectively attenuated bone loss induced by cold exposure and enhanced bone microarchitecture, as evidenced by the increased BMD, Tb. BV/TV, Tb. N, and Tb. Th, and the decreased Tb. Sp in comparison to the PBS-treated CT mice (Fig. S9B-G). There was an increase in Ct. Ar/Tt. Ar, but not Ct. Th (Fig. S9H, I). Calcein double labelling revealed an increase in new bone formation and mineralization in CT mice following RT-EXO treatment (Fig. S9J–L). Immunohistochemistry staining for osteocalcin (OCN) indicated an increase in the number of osteoblasts on the trabecular bone surface in CT mice treated with RT-EXO (Fig. S9M, N). However, as shown by the TRAP staining, treatment with RT-EXO did not reduce the number of osteoclasts on the trabecular bone surface (Fig. S9O, P). Additionally, RT-EXO administration was found to increase femur length (Fig. S9Q and R), while no significant difference in body weight was observed between the experimental groups (Fig. S9S). These results collectively suggest a significant role for plasma-derived exosomes in regulating bone metabolism in mice subjected to both RT and CT environments. Furthermore, it is plausible that CT-EXO may mediate bone loss induced by cold exposure.

CT-EXO diminishes bone mass through the inhibition of autophagy, whereas rapamycin demonstrates the potential to reverse cold exposure-induced bone loss

Given that age-related bone loss correlates with diminished autophagic activity, we investigated the impact of CT-EXO on the senescence of bone marrow-derived mesenchymal stem cells (BMSCs). Subsequently, we evaluated the expression of P21 in the bone tissues of mice subjected to either room temperature (RT) or cold exposure. Cold exposure markedly hastened the senescence of bone tissues, as evidenced by a notable elevation in P21 expression, as depicted in Fig. 6A. We further observed that cold exposure suppresses autophagic activity, as evidenced by elevated expression of P62 and a decreased LC3II: LC3I ratio in bone tissues (Fig. 6A). Intriguingly, our investigation revealed that CT-EXO significantly attenuated the autophagic activity of BMSCs, as indicated by increased P62 expression, a reduced LC3 II: LC3I ratio, decreased levels of ATG5, and diminished formation of autophagosomes (Fig. S10A-B). Conversely, RT-EXO did not exert a significant impact on the autophagic activity of BMSCs (Fig. S10A-B). We determined that CT-EXO could inhibit the autophagic activity of BMSCs and it can be reversed by rapamycin (RAP), a kind of autophagy agonist, by detecting the autophagy markers, P62 and LC3 (Fig. 6B), and detecting the autophagosome formation by transmission electron microscopy (Fig. 6F). Furthermore, as depicted in Fig. 6B, RAP also effectively reversed the CT-EXO-induced inhibition of RUNX2 expression in BMSCs. Alizarin Red staining further illustrated that RAP enhances the osteogenic differentiation of BMSCs and mitigates the impairment of BMSCs’ osteogenic differentiation induced by CT-EXO (Fig. 6C–E). Furthermore, RAP also mitigated the CT-EXO-induced senescence of BMSCs, as evidenced by the expression of P21 and SA-β-gal staining (Fig. 6B, G). As indicated by the findings, CT-EXO inhibits osteogenic differentiation and induces senescence of BMSCs by suppressing autophagy.

Fig. 6

CT-EXO impairs bone mass by inhibiting autophagy. A, Representative western blot image showing the effect of cold exposure on the protein levels of P21, P62, and LC3 in bone tissues.B, Representative western blot image showing the expression protein levels of RUNX2, P62, and LC3, P21 from control-, Vehicle-, CT-EXO-, RAP-, and RAP + CT-EXO-treated BMSCs. C, D, Representative image of microscopic view (D) and entire plate view (E) ARS staining. Scale bar represents 250 μm. E, Quantification of ARS. n = 3 per group. F, Representative Electron Microscopy Images of control-, Vehicle-, CT-EXO-, RAP-, and RAP + CT-EXO-treated BMSCs. Scale bars represent 5 μm G, Representative image of SA-β-gal staining of BMSCs after exosomes treatment. Scale bar represents 100 μm. The control group represents the group without osteogenic induction, while the vehicle group represents the group undergoing osteogenic induction with solvent intervention. H, Schematic flow diagram representing mice treated with vehicle, CT-EXO, RAP, and RAP + CT-EXO. n = 6 per group. I, Representative µCT images of trabecular (top) and cortical (bottom) bone in femora from mice in the vehicle, CT-EXO, RAP, and RAP + CT-EXO groups. Scale bars represent 500 μm (top) and 1 mm (bottom). J–P, Quantitative analysis of BMD, Tb. BV/TV, Tb. Th, Tb. N, Tb. Sp, Ct. Ar/Tt. Ar, Ct. Th. n = 5 per group. Q, Representative calcein double labelling images of the mineralized surface of mouse femora. Scale bar represents 50 μm. R, S, Quantitation of MAR and BFR/BS. n = 5 per group. T, Representative OCN-stained section. Scale bar represents 100 μm. U, Quantification of the number of osteoblasts (N. OBs) on the trabecular bone surface (BS) in distal femora. n = 5 per group. V, TRAP-stained sections. Scale bar represents 100 μm. W, quantification of the number of osteoclasts (N. OCs) on the trabecular bone surface (BS) in distal femora. n = 5 per group. The vehicle referred to is DMSO. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

We subsequently investigated whether enhancing autophagic activity could enhance the bone mass of CT mice. We administered varying doses of RAP to CT mice (Fig. S11A) and observed that a dosage of 2 mg/kg body weight significantly augmented the bone mass of CT mice (Fig. S11B-I). Therefore, we opted for this dosage for subsequent experiments. CT mice were administered RAP or vehicle (Fig. S12A). As depicted in Fig. S12B, RAP augmented the autophagic activity of bone tissues, evidenced by reduced expression of P62, elevated expression of ATG5, and an increased LC3 II: LC3 I ratio. Subsequent µCT analysis revealed that RAP treatment elevated BMD, Tb. BV/TV, Tb. Th, and Tb. N in CT mice, while reducing Tb. Sp (Fig. S12C–H). Figures S12I and J demonstrate that RAP had no effect on Ct. Ar/Tt. Ar but increased Ct. Th in CT mice. Additionally, as depicted in Fig. S12K–M, RAP exhibited a protective effect on new bone formation and mineralization in CT mice. RAP treatment resulted in an increase in the number of osteoblasts in CT mice (Fig. S12N, O), while TRAP staining revealed no significant differences in osteoclasts between the two groups (Fig. S12P, Q). Moreover, RAP increased the length of the femur in CT mice (Fig. S12R, S) and did not affect the body weight of CT mice (Fig. S12T).

To assess whether autophagy contributes to CT-EXO-induced bone loss, we administered vehicle, CT-EXO, RAP, or CT-EXO + RAP to mice (Fig. 6H). The µCT results indicated that RAP significantly improved bone mass and bone microarchitecture (Fig. 6I-N), which is consistent with the results of Ma et al. [29]. Furthermore, we observed that RAP exerted a protective effect on bone microarchitecture and bone mass in mice treated with CT-EXO (Fig. 6I–N). No statistically significant difference in cortical bone was noted between groups (Fig. 6O, P). Calcein double labeling revealed that RAP enhanced new bone formation and mineralization, effectively reversing the CT-EXO-induced impairment of new bone formation and mineralization (Fig. 6Q–S). OCN staining demonstrated that RAP increased the number of osteoblasts in CT-EXO-treated mice (Fig. 6T, U). However, TRAP staining revealed that RAP had no significant effect on osteoclast numbers, regardless of CT-EXO treatment (Fig. 6V, W). Additionally, we found that CT-EXO decreased the autophagic activity of bone tissues, as indicated by higher expression of P62, lower expression of ATG5, and a decreased LC3II: LC3I ratio, and this effect could be reversed by RAP (Fig. S13A). Subsequently, BMSCs were isolated from the different groups of mice. Our results indicated that CT-EXO significantly impaired the osteogenic differentiation of BMSCs, while RAP was able to restore their osteogenic differentiation ability, as evidenced by the expression of COL-1, RUNX2, and BMP2 (Fig. S13B). Furthermore, Fig. S13C and D demonstrate that RAP could also reverse the reduction in femur length induced by CT. However, there was no significant effect on body weight in these groups (Fig. S13E).

These results suggest that decreased autophagic activity contributes to bone loss induced by CT-EXO or CT. Furthermore, the autophagy of BMSCs appears to play a role in the impact of CT-EXO on osteogenic differentiation.

Exosomal mir-25-3p enriched in CT-EXO regulated osteogenic differentiation and autophagy by tageting SATB2 in BMSCs

To delve deeper into the mechanism underlying the detrimental effects of CT-EXO on bone loss, we employed microarray-based miRNA expression profiling analysis to identify differential abundance of miRNAs between RT-EXO and CT-EXO. Our analysis revealed that CT-EXO exhibited 33 significantly upregulated miRNAs and 38 significantly downregulated miRNAs (fold change ≥ 2 and P ≤ 0.05) (Fig. 7A). qRT-PCR was subsequently employed to validate the differential abundance of three miRNAs related to osteogenesis and autophagy (miR-205-5p, miR-30d-5p, miR-320-3p, miR-222-3p, and miR-25-3p) in plasma exosomes. Among these, miR-25-3p was found to be the most abundant in CT-EXO compared to RT-EXO (Fig. 7B–F). Consequently, we selected miR-25-3p, which was highly abundant in CT-EXO, for further investigation.

Fig. 7

Exosomal miR-25-3p enriched in CT-EXO regulated osteogenic differentiation and autophagy by tageting SATB2 in BMSCs. A, Volcano diagram of differential expression between between RT-EXO and CT-EXO miRNA according to microarray analysis. B–F, qRT-PCR quantitative results of miRNA (miR-205-5p, miR-30d-5p, miR-320-3p, miR-222-3p, miR-25-3p) level of RT-EXO or CT-EXO. n = 3 per group. G, Representative western blot image showing the expression protein levels of RUNX2, P62, and LC3 from miR-25-3p knocked-in or miR-25-3p knocked-down CT-EXO-treated BMSCs. H, I, Representative image of microscopic view (H) and entire plate view (I) ARS staining from miR-25-3p knocked-in or miR-25-3p knocked-down CT-EXO-treated BMSCs. Scale bar represents 250 μm J, Quantification of ARS. n = 3 per group. K, Venn diagram of the predicted target gene of miR-25-3p from Target scan (blue), miRDB (red), autophagy(yellow), and osteogenesis (orange). L, Luciferase reporter assays were conducted using luciferase constructs carrying a WT or mutant SATB2 3′-UTR co-transfected into BMSCs with miR-25-3p mimics. Firefly luciferase activity was normalized to Renilla luciferase activity. n = 5 per group. M, Transfection efficiency of miR-25-3p detected by qRT-PCR. N, Representative western blot image showing the protein expression levels of SATB2 from miR-25-3p knocked-in miR-25-3p knocked-down BMSCs. O, Knockout efficiency of SATB2 in BMSCs by western blot. P, SATB2, RUNX2, P62, and LC3 expression were measured in the BMSCs treated with siSATB2#3 or siRNA control. Q, R, Representative image of microscopic view (Q) and entire plate view (R)ARS staining. Scale bar represents 250 μm. S, Quantification of ARS. T, Representative western blot image showing the expression protein levels of SATB2, RUNX2, P62, and LC3 from the inhibitor-NC, inhibitor, inhibitor + siRNA-NC, inhibitor + siSATB2 groups. U, V, Representative image of microscopic view (U) and entire plate view (V)ARS staining. Scale bar represents 250 μm. W, Quantification of ARS. n = 3 per group. All groups’ cells have received osteogenic induction, the control group represents the group undergoing osteogenic induction without any intervention. CT-EXO + AgomiR-NC and CT-EXO-AntagomiR-NC served as negative controls for CT-EXO + AgomiR-25-3p and CT-EXO-AntagomiR-25-3p. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

To investigate the role of miR-25-3p in BMSCs osteogenic differentiation and autophagy, BMSCs were treated with CT-EXO transfected with either agomiR-25-3p or antagomiR-25-3p. agomiR-NC and antagomiR-NC, respectively serving as negative controls for agomiR-25-3p or antagomiR-25-3p, were also transfected into CT-EXO. As depicted in Fig. 7G, CT-EXO + agomiR-25-3p significantly exacerbated the impairment of osteogenic differentiation and autophagic activity in BMSCs, as evidenced by decreased expression of RUNX2 and LC3II: LC3I ratio, and increased expression of P62, compared to BMSCs treated solely with CT-EXO. Conversely, antagomiR-25-3p notably reversed the impairment of osteogenic differentiation and autophagic activity in BMSCs induced by CT-EXO (Fig. 7G). Alizarin Red staining further confirmed that miR-25-3p inhibited mineralized deposits (Fig. 7H–J). Thus, the results indicate that miR-25-3p plays a crucial role in the CT-EXO-induced inhibition of osteogenic differentiation and autophagic activity in BMSCs.

miRNAs regulate the expression of their target genes by interacting with the 3′-untranslated region (3′-UTR) or protein-coding sequence of target mRNAs. Using various bioinformatic target prediction techniques, we searched for targets of miR-25-3p that could be involved in the regulation of osteogenic differentiation and autophagic activity in BMSCs (Fig. 7K). SATB2 was suggested as a potential target of miR-25-3p. (Fig. 7L). To assess whether miR-25-3p can directly bind to the 3′-UTR of SATB2, luciferase reporter constructs were generated containing either wild-type (WT) or mutant (MUT) versions of the anticipated miRNA-binding sites within the SATB2 sequence (designated as WT-pGL3-SATB2 and MUT-pGL3-SATB2, respectively). Subsequently, the impact of miR-25-3p on luciferase enzyme activity in BMSCs was examined following transfection of WT-pGL3-SATB2 and MUT-pGL3-SATB2 with miR-25-3p mimics. The luciferase activity of the SATB2 3′-UTR reporter gene was suppressed by miR-25-3p mimics, whereas MUT-pGL3-SATB2 abrogated this effect entirely (Fig. 7L). These results suggest that miR-25-3p specifically targets the 3′-UTR of SATB2. Moreover, miR-25-3p levels in BMSCs were significantly increased by mimics or decreased by inhibitors, as depicted in Fig. 7M, which was confirmed by qRT-PCR. Additionally, we demonstrated that overexpression of miR-25-3p in BMSCs led to a reduction in SATB2 expression, while inhibition of miR-25-3p resulted in increased SATB2 expression (Fig. 7N). Furthermore, we observed that autophagic activity and osteogenic differentiation of BMSCs were suppressed by miR-25-3p overexpression but enhanced by miR-25-3p downregulation (Fig. S14A–D).

To investigate SATB2 function in miR-25-3p-controlled osteogenesis and autophagy, the expression of SATB2 was knocked down in BMSCs by transfection with SATB2 small interfering RNA (siSATB2). Western blotting analysis revealed that transfection of BMSCs with siSATB2 effectively reduced SATB2 protein expression (Fig. 7O). Since siSATB2#3 demonstrated the highest knockdown efficiency, it was chosen for the next experiment. Decreased expression of RUNX2 and mineralized deposit were further indicators that SATB2 knockdown reduced osteogenic differentiation (Fig. 7P–S). Autophagic activity was also reduced by knocking down SATB2, as measured by a greater amount of P62 and a lower LC3 II : LC3 I ratio (Fig. 7P). Then we determined whether knocking down SATB2 could abolish the function of the miR-25-3p inhibitor. Figure 7T reveals that miR-25-3p inhibitor significantly increased the expression of RUNX2 and LC3 II:LC3 I ratio, markedly decreasing the expression of P62, the same as the previous results. However, the increase of BMSCs osteogenic differentiation and autophagic activity was reversed by siSATB2 treatment (Fig. 7T). Alizarin Red staining also confirmed that miR-25-3p inhibitor increased mineralized deposit, but knocking down SATB2 abolished this effect (Fig. 7U–W).

These findings indicate that miR-25-3p, enriched in CT-EXO, attenuate osteogenic differentiation in BMSCs by inhibiting their autophagic act