Background

The effective reconstruction of bone defects is a challenging problem in orthopedic surgery. Autologous and allogenic bone grafting are the most widely used treatments for bone defect repair. However, the limited availability of graft material, the extra damage to harvest sites, and the frequent need for a second operation suppressed autologous bone grafts [1, 2]; allogenic bone grafts are hampered by problems of significant failure rates, poor mechanical stability, and immunological rejection [3]. Bioactive materials provide an alternative solution for the repair and regeneration of bone defects. Tricalcium phosphate (β-TCP), composed of the same ions as bone, is a well characterized osteoconductive biomaterial and is already in clinical use for bone repair and regeneration [1 (Table S1).

Western blot analysis

Western blotting was performed as described previously [7]. Cells or exosome lysates were diluted at a ratio of 1:4 with protein loading buffer (5×) and heated at 95 °C for 5 min. Protein extracts were separated on a 10 % sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel at 120 V for 1 h and blotted onto nitrocellulose membranes (Whatman, Maidstone, Kent, UK) for 30 min at 100 mA. The membranes were then blocked for 2 h with 5 % non-fat dried milk in TBST (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1 % Tween-20). Subsequently, the membranes were incubated with primary antibodies at 4 °C overnight, followed by incubation with horseradish peroxidase (HRP)-labeled secondary antibodies (Cell Signaling Technology, USA) at 37 °C for 1 h. The primary antibodies including anti-CD9, anti-CD63, anti-CD81, anti-OCN, anti-runt-related transcription factor 2 (Runx2), anti-collagen type I alpha 1 (COL1A1), anti-Akt, and phosphorylated Akt (p-Akt) antibody were obtained from Abcam (Cambridge, UK). The immunoreactive bands were visualized using enhanced chemiluminescence reagent (Thermo Fisher Scientific, Waltham, MA, USA) and imaged by an Image Quant LAS 4000 mini bio-molecular imager (GE Healthcare, Uppsala, Sweden). All values were normalized to the value of beta-actin (β-actin).

Alkaline phosphatase (ALP) assay and Alizarin red S (ARS) staining

hBMSCs were cultured in osteogenic media supplemented with exosomes or PBS for 10 days (ALP staining) or 14 days (ARS staining). Cells were washed with PBS, fixed with 4 % paraformaldehyde for 10 mins, and incubated with the ALP reagents (Sigma) or with 2 % ARS solution (Sigma) for 30 min at 20 °C according to the manufacturer’s protocol. After washing with distilled water, the stained cells were examined using an inverted microscope (Leica DMI6000B, Solms, Germany). ALP activity of hBMSCs was measured using a commercial kit (Sigma) following the manufacturer’s instructions.

PI3K/Akt signaling inhibition

LY294002, a highly selective inhibitor of PI3K, was purchased from Sigma-Aldrich (St. Louis, MO, USA) and dissolved in dimethylsulfoxide (DMSO) at a stock concentration of 100 mM according to the manufacturer’s protocol. To confirm the involvement of PI3K/Akt signaling in the exosome-mediated effects on hBMSCs, cells were pre-treated with the indicated concentration (10 μM) of LY294002 or an equal volume of DMSO for 1 h. Subsequently, 5 × 1011 particles/mL of exosomes or an equal volume of PBS was added to the culture medium of hBMSCs and the cells were cultured for 24 h. Western blotting, ALP, and ARS staining on hBMSCs were then performed as described above.

Statistical analysis

All data are shown as mean ± standard deviation (SD). Differences between groups were assessed by one-way analysis of variance (ANOVA); Dunnett t or LSD t analyses were performed to determine the statistical significance between each sample. P < 0.05 was considered statistically significant.

Results

Characterization of hiPS-MSCs and hiPS-MSC-Exos

Using a modified one-step induction protocol [25], almost 100 % of the hiPS cells successfully differentiated into hiPS-MSCs, and the hiPS-MSCs exhibited characteristics of MSCs (Additional file 1: Figure S1). TEM, TRPS analysis, and western blotting were used to characterize the purified nanocarriers derived from hiPS-MSCs. The results showed that the vast majority of these particles exhibited a cup- or round-shaped morphology with a size ranging from 50 to 150 nm (Fig. 1a and b), suggesting the presence of exosomes. Also, these particles did express the characteristic exosome surface markers, including CD9, CD63, and CD81 (Fig. 1c), which further confirmed their exosome identity. The data suggested that these nanocarriers were predominantly exosomes.

Fig. 1
figure 1

Characterization of exosomes derived from human induced pluripotent stem cell-derived mesenchymal stem cells (hiPS-MSCs). a Morphology observed by TEM. b Particle size distribution and concentration measured by TRPS. c Western blot analysis of the exosome surface markers

Full size image

In vivo osteogenesis activity of exosome/β-TCP combination scaffolds

To investigate the therapeutic potential of hiPS-MSC-Exos on bone repair, the rat critical-sized calvarial bone defect model was established and the exosome/β-TCP combination scaffolds were implanted into the defect areas. High-resolution micro-CT scanning was carried out to qualitatively evaluate the newly formed bone within the defects at 8 weeks post-implantation. The sagittal view of micro-CT images rarely, if at all, showed the newly regenerated bone in the pure β-TCP group (Fig. 2a). On the contrary, a large amount of de novo bone formation was observed in the calvarial defect sites following implantation with the β-TCP scaffolds loaded with 5 × 1011 particles/mL of exosomes. Also, a larger extent of bone regeneration was shown with an increasing exosome concentration. This was confirmed by a quantitative analysis of the BV/TV (Fig. 2b), which showed a dose-dependent increase in new bone formation in the exosome-functionalized group as compared to the control. Furthermore, the local BMDs showed the same tendency as that obtained for the BV/TV levels (Fig. 2c), indicating that hiPS-MSC-Exos could improve ossification in the defect areas.

Fig. 2
figure 2

Micro-CT and fluorochrome-labeling histomorphometrical analysis of the repaired craniums at 8 weeks post-implantation. a Three-dimensional reconstruction and sagittal images show the different reparation effects of exosomes or tricalcium phosphate (β-TCP) only. b The bone volume/tissue volume (BV/TV) and c bone mineral density (BMD) varied between the different groups. d Fluorochrome-labeling histomorphometrical analysis of new bone formation and mineralization at 8 weeks post-operation (yellow, between 2 and 4 weeks; red, between 4 and 6 weeks; green, between 6 and 8 weeks). e Percentage of each fluorochrome area for different groups. Dunnett t test; *p < 0.05 compared with the β-TCP group, # p < 0.05 compared with the exosome (Exos) 5 × 1011 particles/mL group

Full size image

The tetracycline (yellow), alizarin red (red), and calcein (green) dyes were intraperitoneally injected into each rat at 2, 4, and 6 weeks post-operation, respectively. Results showed that the percentages of tetracycline (yellow), alizarin red (red), or calcein (green) labeling in the exosome-functionalized group were greater than those in the control group (Fig. 2d and e). Higher percentages of fluorescent dyes were observed after the bone defects were treated with a higher concentration of exosomes (Fig. 2d and e). Moreover, histological evidence based on van Gieson staining of non-decalcified craniums demonstrated that exosome treatment increased the newly formed bone tissues within the defect areas, as determined by the percentage of new bone area (Fig. 3a and b). Consistent with the above findings, OCN immunostaining further confirmed that the osteogenic responses from the defect tissues treated with exosomes was enhanced as compared to the control (Fig. 3c). Collectively, our in vivo functional studies demonstrated that hiPS-MSC-Exos could stimulate osteogenesis of β-TCP through activating a bone regenerative process in the target tissue.

Fig. 3
figure 3

Histological and immunohistochemical analysis of the newly formed bone after 8 weeks of transplantation. a The un-decalcified craniums were stained with van Gieson’s picrofuchsin. The new bone area and tricalcium phosphate (β-TCP) residue are shown in red and black, respectively. b Quantitative analysis of (a). Dunnett t test; *p < 0.05 compared with the β-TCP group, # p < 0.05 compared with the exosome (Exos) 5 × 1011 particles/mL group. c The sections were subjected to immunohistochemical analysis to inspect the distribution of OCN. OCN-positive immunostaining illustrated the bone tissues. Scale bars = 80 μm

Full size image

Exosome release from β-TCP

To detect exosome release from β-TCP, we used a green fluorescent lipophilic dye (DiO) to label hiPS-MSCs. The exosomes released by the labeled cells were also labeled with DiO upon fusion of multivesicular bodies with the cell plasma membrane. The labeled exosomes were combined with β-TCP and then added to the plate containing hBMSCs. The internalization of exosomes was detected every 8 h of culture with hBMSCs. The results showed that the DiO-labeled exosomes can be found in the perinuclear region of hBMSCs which surround β-TCP (Fig. 4Aa), indicating that hBMSCs are targets of hiPS-MSC-Exos.

Fig. 4
figure 4

Internalization of hiPS-MSC-Exos in human BMSCs, exosome release from β-TCP, and their pro-osteogenesis effects on the recipient cells. A (a) DiO-labeled exosome release from β-TCP over 48 h. (b) Accumulated release of exosomes from β-TCP over 5 days. B Exosomes (Exos) enhanced the proliferation of hBMSCs as analyzed by Cell Counting Kit-8 assay. C Exosomes stimulated the migration of BMSCs as analyzed by migration assay. Scale bars = 250 μm. D Quantitative analysis of (C). E, F Incubation of hBMSCs with exosomes resulted in a dose-dependent increase in the alkaline phosphatase (ALP) staining and activity. Dunnett t test; *p < 0.05 compared with the control group, # p < 0.05 compared with the Exos 5 × 1011 particles/mL group

Full size image

As shown in Fig. 4Ab, the exosomes showed burst release from β-TCP. In addition, with an increase in immersed hours, exosome release decreased gradually.

Pro-osteogenesis effects of exosomes on hBMSCs

To explore the functional roles of hiPS-MSC-Exos in osteogenesis, hBMSCs were cultured in MGro-500 MSC or osteogenic medium containing exosomes (5 × 1011 or 1 × 1012 particles/mL) or in an equal volume of PBS for a series of in vitro osteogenesis-related assays. Data indicated that 5 × 1011 particles/mL of exosomes were able to significantly enhance hBMSC proliferation; the hBMSCs treated with a higher concentration of exosomes (1 × 1012 particles/mL) showed much higher proliferative capability (Fig. 4B). As the resident MSCs are recruited into injury sites for hel** repair the injury, the migration ability of MSCs is critical for bone reparation and regeneration. The scratch wound assay showed that, after being cultured with exosomes, the migration ability of hBMSCs was remarkably enhanced (Fig. 4C and D). Detection of ALP expression, a biochemical marker of osteogenesis as well as the osteocalcin content, showed that incubation of hBMSCs with hiPS-MSC-Exos resulted in a dose-dependent increase in the ALP staining (Fig. 4E) and activity (Fig. 4F), clearly indicating a positive role for hiPS-MSC-Exos in osteogenic differentiation. Taken together, our results show that hiPS-MSC-Exos could activate a series of bone regenerative responses from target hBMSCs in a dose-dependent manner. Taking economic concerns into account, and since the low dose of exosomes (5 × 1011 particles/mL) exert significant pro-osteogenic effects on target tissues and cells, we used the concentration of 5 × 1011 particles/mL for the subsequent experiments.

Differentially expressed genes in hBMSCs stimulated by exosomes

To determine the effector genes that respond to hiPS-MSC-Exos stimulation, microarray analyses of hBMSCs cultured with exosomes or PBS were performed. The 1.5-fold up- or downregulated genes in all three paired samples were defined as differentially expressed genes or effector genes after exposure to hiPS-MSC-Exos. The results showed that there was a significant differential expression of 1447 candidate genes between the two groups (p < 0.05) (Fig. 5a). Among these, 293 genes were upregulated and 1154 genes were downregulated in the exosome-treated hBMSC groups than in the control group treated with PBS. We then performed functional enrichment analysis, including molecular functions and involvement of all differentially expressed genes in known regulatory signaling pathways. The results revealed that multiple biological pathways were prominently enriched (Enrichment Score >2.0, p < 0.05). The top ten enriched pathways associated with exosome stimulation are shown in Fig. 5b. Among these, extracellular matrix (ECM)-receptor interaction, focal adhesion, and phosphatidylinositol 3-kinase (PI3K)-Akt signaling pathways appeared to be the most enriched ones (Enrichment Score >5.0, p < 0.05). Since the PI3K/Akt signaling pathway has previously been found to be involved in MSC proliferation, migration, and osteogenic differentiation [26–31], it is reasonable to believe that this signaling process might have key roles in the exosome-mediated pro-osteogenesis effects on hBMSCs.

Fig. 5
figure 5

Differential expression of mRNAs between exosome-treated and control groups. a The differentially expressed (DE) genes in hBMSCs in response to exosome (Exos) stimulation are illustrated as a heat map. A p value cut-off of 0.05 and a fold-value change of ≥2 were used as a filter to identify the DE genes. b Enrichment analysis of all DE genes was performed. The top ten enriched pathways associated with exosome stimulation are shown. c A heat map of PI3K/Akt signaling pathway-related DE genes was generated. d The altered expression of PI3K/Akt signaling-related genes was confirmed by qRT-PCR analysis. ANOVA; *p < 0.05 compared with the control group

Full size image

The PI3K/Akt signaling pathway-related differentially expressed genes are illustrated as a heat map (Fig. 5c) showing that 117 genes involved in this pathway were altered in hBMSCs treated with exosomes. Among them, 47 genes, such as platelet-derived growth factor alpha (PDGFA), fibroblast growth factor 1 or 2 (FGF1/2), FGF receptor 1 (FGFR1), collagen type I alpha 1 or 2 (COL1A1/2), and BCL2-like 1 (BCL2L1), which are positive effector genes in the PI3K/Akt signaling pathway, were profoundly upregulated. Conversely, another 70 genes, such as glycogen synthase kinase 3 beta (GSK3β) and phosphatase and tensin homolog deleted on chromosome ten (PTEN) that are key molecules in negatively regulating this pathway, were markedly downregulated. These results suggested that the PI3K/Akt signaling in hBMSCs stimulated by hiPS-MSC-Exos was activated.

To confirm the microarray results, 10 genes, including VCAM1, FGF11, IL6, SLC2A1, ITGA7, LAMA1, ETNK1, TLR4, DKK1, and NKTR, which were differentially expressed between the exosome-treated BMSCs and the control group (p < 0.05), were randomly selected for qRT-PCR detection. The levels of VCAM1, FGF11, IL6, SLC2A1, and ITGA7 were increased in the exosome-treated hBMSCs, whereas the expression of LAMA1, ETNK1, TLR4, DKK1, and NKTR was decreased (Additional file 1: Figure S2). We also performed qRT-PCR to confirm the altered expression of PI3K/Akt signaling-associated genes. Data revealed that the exosome-treated hBMSCs showed a profound upregulation of PDGFA, FGF1/2, FGFR1, COL1A1/2, and BCL2L1, but a significant downregulation of GSK3β and PTEN (p < 0.05) (Fig. 5d). The results were generally consistent with the microarray data.

Involvement of PI3K/Akt signaling in the exosome-induced osteogenic responses

The activation of the PI3K/Akt pathway in hBMSCs following hiPS-MSC-Exos stimulation was verified by treating the cells with exosomes or PBS for 24 h and assessing the protein levels of Akt and p-Akt by Western blotting. Compared to the PBS control, incubation with hiPS-MSC-Exos resulted in a significant increase in Akt phosphorylation on Ser473 in hBMSCs (Fig. 6a), indicating that the PI3K/Akt signaling in hBMSCs was activated by the exosomes. However, the upregulation of p-Akt in hBMSCs by the exosomes was significantly impaired after the hBMSCs were cultured with a PI3K inhibitor (LY294002; 10 μM) (Fig. 6a).

Fig. 6
figure 6

Involvement of PI3K/Akt signaling in the exosome-induced osteogenic responses from BMSCs. a Exosomes (Exos) induced the activation of the PI3K/Akt signaling pathway and increased the protein levels of osteogenesis-related molecules; these effects by exosomes were abolished by the PI3K inhibitor (LY294002; 10 μM). b The exosome-treated hBMSCs showed much higher levels of alkaline phosphatase (ALP) and Alizarin red S (ARS) staining compared with the control groups on days 10 and 14, respectively; these effects were inhibited by LY294002

Full size image

The activation of the PI3K/Akt pathway might be the underlying mechanism through which hiPS-MSC-Exos exert pro-osteogenesis effects on recipient cells. We tested this hypothesis by culturing the BMSCs in osteogenic media containing exosomes or PBS with or without LY294002. After 7 days of differentiation, Western blotting was performed to detect the levels of the early osteogenesis-related marker proteins including Runx2 and COL1, and the late osteogenic differentiation marker OCN (Fig. 6a). The results revealed that hBMSCs, when cultured in the osteogenic medium, showed osteogenic differentiation as indicated by the expression of osteogenesis-related molecules. After incubation of hBMSCs with hiPS-MSC-Exos, the levels of these osteogenesis-related proteins were further increased. However, their upregulation by hiPS-MSC-Exos was markedly suppressed by the PI3K inhibitor LY294002. Furthermore, the addition of LY294002 to the osteogenic medium alone also decreased osteogenesis slightly, suggesting that the osteogenic medium-induced osteogenesis relies on the activation of the PI3K/Akt pathway to some extent. Also, we evaluated the extent of ALP production and calcium mineral deposition in hBMSCs by staining with ALP and ARS. As shown in Fig. 6b, the exosome-treated hBMSCs showed significantly higher expression of ALP and calcium deposits than the control group, whereas these effects were blocked by the additional treatment with LY294002. The ALP generation, matrix mineralization, and calcium nodule formation were reduced slightly in BMSCs treated with LY294002 compared with the control, which was consistent with the data from osteogenesis-related molecule expression. Taken together, our results indicated that the stimulatory effects of hiPS-MSC-Exos on the osteogenic differentiation of hBMSCs mainly resulted from the activation of the PI3K/Akt signaling pathway.

Discussion

Normal healthy bone has the ability to spontaneously regenerate after minor injury. However, extensive bone loss due to disease or trauma requires tissue engineering applications. In the present study, we provided the first demonstration that hiPS-MSC-Exos, the nanocarriers naturally secreted by iPS-MSCs, could be combined with β-TCP and this hiPS-MSC-Exos/β-TCP combination scaffold could significantly enhance osteogenesis in a rat model of critical-sized calvarial bone defects as compared to β-TCP alone. Through in vitro studies, we found that the exosomes could remarkably promote hBMSC osteogenic differentiation after they were internalized by hBMSCs through activating the PI3K/Akt signaling pathway.

Currently, a growing body of evidence indicates that MSC paracrine action may be applied as a novel strategy for tissue repair [32, 6]. However, the technological process is relatively complicated and the obtained bioactivity is very limited. Bone tissue engineering has also been used, through which bone regeneration can be achieved via a combination of TCP with stem cells [4, 37]. Nevertheless, the application of stem cells is still hampered by immunological rejection, chromosomal variation, malignant transformation, and so forth [38]. The therapeutic effects of exosomes on tissue injury have been demonstrated [24, 34–36]. In addition, MSC-derived exosomes do not contain MHC class I or II proteins and, thus, their application to non-immune-compatible animals does not induce overt immune reactions [34, 39, 40]. Thus, the exosome/β-TCP combination scaffolds can be potentially used as a promising graft for bone defect repair.

It is known that endogenous BMSCs can be mobilized in response to injury, and the activated BMSCs can serve as the major healing cells during bone repair and remodeling [41]. We hypothesized that the beneficial effects of the exosome/β-TCP combination scaffolds on bone defects might be attributed to their recruitment and activation of pre-existing BMSCs in the bone tissue, and we performed in vitro experiments to confirm this hypothesis. It has been proposed that exosomes can serve as a nanocarrier to transfer proteins, mRNAs, and microRNAs from parent cells to recipient cells, and that they alter the gene expression and protein translation of the recipient cells thereby modulating the bioactivity of target cells [42]. Their internalization by target cells is the prerequisite for playing such a role. In the present study, the DiO-labeled exosomes were combined with β-TCP, and then put into cell culture plates containing hBMSCs. The labeled exosomes can be seen in perinuclear regions of hBMSCs, indicating exosomes could release from exosome/β-TCP combination scaffolds and then be internalized by hBMSCs. Further analysis showed that, after internalization of exosomes into hBMSCs, they stimulate the proliferation, migration, and osteogenic differentiation of hBMSCs. These conditions are beneficial for attracting endogenous BMSCs to bone defects and finally enhancing new bone formation. All findings demonstrated that the exosome/β-TCP combination scaffolds have obviously enhanced osteogenic activity attributed to the biological effects of exosomes.

We further clarified the molecular mechanisms of exosome/β-TCP combination scaffolds on the osteogenesis of hBMSCs. Zhang et al. reported that the exosomes released by human umbilical cord blood-derived MSCs were able to induce β-catenin nuclear translocation of skin cells, enhancing their proliferation and migration [43]. Shabbir et al. showed that the exosomes secreted by human BMSCs could activate several signaling pathways (including Akt, Erk1/2, and STAT3) in target fibroblasts and increase their angiogenic response [44]. In this study, we used microarray analysis to examine the gene expression pattern in hBMSCs treated by hiPS-MSC-Exos. We detected significant alterations in the expression of a variety of genes following treatment with exosomes. Results of bioinformatics analyses indicated that these significantly altered genes are predominantly involved in the PI3K/Akt signaling pathway, focal adhesion, and ECM-receptor interaction. In particular, the PI3K/Akt pathway may have a role in the exosome-induced pro-osteogenic effects on hBMSCs as this signaling cascade has been reported to play important roles in osteoblast differentiation and bone growth [26–28]. Here, we report that hiPS-MSC-Exos treatment resulted in a prominent increase in the expression of the positive effector genes of the PI3K/Akt pathway, such as PDGFA, FGF1/2, FGFR1, COL1A1/2, and BCL2L1. All of these genes can positively modulate cell survival, proliferation, migration, osteogenic differentiation, or apoptotic inhibition [29–31, 45, 46]. The typical negative modulators of PI3K/Akt signaling, including GSK3β and PTEN, were remarkably decreased. The results indicated that the PI3K/Akt signaling in hBMSCs is probably activated after internalization of exosomes. This contention was confirmed by a significant increase in the phosphorylation of Akt (an indicator of PI3K activation) in hBMSCs treated with exosomes. However, it is noteworthy that the PI3K inhibitor markedly, but not completely, abolished the pro-osteogenic effects of hiPS-MSC-Exos, suggesting that the exosome-induced osteogenic effects cannot be fully attributed to the PI3K/Akt pathway and therefore additional mechanisms remain to be explored. Our future studies will focus on deciphering the molecular contents of hiPS-MSC-Exos that underlie the activation of the PI3K/Akt signaling pathway that appears to be largely responsible for the osteogenesis in hBMSCs.

Conclusions

In this study, we provided evidence that the hiPS-MSCs-Exos-functionalized β-TCP scaffold can effectively promote bone repair and regeneration in a rat model of calvarial bone defects. The underlying mechanism through which exosomes enhance osteoinductive activity of β-TCP and promote bone regeneration appears to be the activation of endogenous BMSCs in the bone defect site. This was suggested by the observation that exosomes could release from exosome/β-TCP combination scaffolds and then be internalization by hBMSCs. Furthermore, the PI3K/Akt signaling pathway likely plays a critical role in the pro-osteogenesis effects of the exosome/β-TCP scaffold on BMSCs. Further studies are needed to assess the long-term effects of exosome-based therapy for bone repair and whether the exosome-functionalized β-TCP can be used as a valid and safe therapeutic option for clinical application.