Abstract

   Bone defect healing is a multi-factorial process involving the
   inflammatory microenvironment, bone regeneration and the formation of
   blood vessels, and remains a great challenge in clinical practice.
   Combined use of three-dimensional (3D)-printed scaffolds and bioactive
   factors is an emerging strategy for the treatment of bone defects.
   Scaffolds can be printed using 3D cryogenic printing technology to
   create a microarchitecture similar to trabecular bone. Melatonin (MT)
   has attracted attention in recent years as an excellent factor for
   promoting cell viability and tissue repair. In this study, porous
   scaffolds were prepared by cryogenic printing with
   poly(lactic-co-glycolic acid) and ultralong hydroxyapatite nanowires.
   The hierarchical pore size distribution of the scaffolds was evaluated
   by scanning electron microscopy (SEM) and micro-computed tomography
   (micro-CT). Sleep-inspired small extracellular vesicles (MT-sEVs) were
   then obtained from MT-stimulated cells and
   1,2-distearoyl-sn-glycero-3-phosphoethanolamine-poly(ethylene
   glycol)-inorganic pyrophosphate (DSPE-PEG-PPi) was used to modify the
   membrane of MT-sEVs to obtain PPi-MT-sEVs. RNA sequencing was performed
   to explore the potential mechanisms. The results demonstrated that
   PPi-MT-sEVs not only enhanced cell proliferation, migration and
   angiogenesis, but also regulated the osteogenic/adipogenic fate
   determination and M1/M2 macrophage polarization switch in vitro.
   PPi-MT-sEVs were used to coat scaffolds, enabled by the capacity of PPi
   to bind to hydroxyapatite, and computational simulations were used to
   analyze the interfacial bonding of PPi and hydroxyapatite. The
   macrophage phenotype-modulating and osteogenesis–angiogenesis coupling
   effects were evaluated in vivo. In summary, this study suggests that
   the combination of hierarchical porous scaffolds and PPi-MT-sEVs could
   be a promising candidate for the clinical treatment of bone defects.

Graphical abstract

   [48]graphic file with name 12951_2024_2977_Figa_HTML.jpg

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-024-02977-5.

   Keywords: 3D printing, Extracellular vesicles, Vascularized bone
   regeneration, Surface functionalization, Macrophage polarization

Introduction

   Many clinical conditions, such as high-energy trauma and tumor
   resection, result in bone defects that are difficult to heal
   spontaneously [[49]1]. Autografts and allografts have excellent
   biocompatibility, mechanical properties and osteoinductivity, but there
   are still issues including limited bone volume, donor site damage and
   potential immune rejection [[50]2, [51]3]. Conventional bone
   substitutes exhibit deficiencies in growth factors and hierarchical
   pore architecture, leading to suboptimal outcomes in repairing large
   bone defects [[52]4–[53]6]. The ideal bone substitutes should have
   excellent biodegradability and biocompatibility, while simultaneously
   evading inflammatory responses. Additionally, the osteointegration and
   proangiogenic capacity of ideal bone substitutes emerge as crucial
   factors that influence their therapeutic outcomes [[54]7]. Bone tissue
   engineering provides a promising strategy to effectively repair bone
   defects. Tissue engineering scaffolds are made from materials with
   excellent biocompatibility. By incorporating bioactive factors, the
   osteointegration and osteoinductivity of scaffolds are enhanced,
   effectively promoting bone and vascular regeneration [[55]8].

   Poly (lactic-co-glycolic acid) (PLGA) exhibits superior
   biocompatibility and degradability, and has been approved for use in
   basic and clinical medical research [[56]9, [57]10]. The degradation
   products of PLGA do not cause immune rejection, making it widely
   applicable for use in bone tissue engineering and drug delivery
   [[58]11]. Hydroxyapatite is the main inorganic component of bone and
   has excellent osteoinductive properties that promote osteogenic
   differentiation of mesenchymal stem cells (MSCs) to enhance bone
   regeneration [[59]12]. Hydroxyapatite nanowires have excellent
   osteoinductive properties and are easier to mix uniformly in scaffolds
   than hydroxyapatite nanoparticles [[60]13]. However, short
   hydroxyapatite nanowires (tens of micrometers in length) are brittle
   [[61]14], whereas ultralong hydroxyapatite nanowires (uW) with lengths
   of several hundred micrometers have superior toughness and mechanical
   strength [[62]15]. Further, uW can interweave with each other in
   scaffolds to promote vascular regeneration and bone defect repair
   [[63]16].

   Three-dimensional (3D) printing is an additive manufacturing technology
   that allows control of the spatial distribution of biomaterials as well
   as the layered architecture [[64]17]. In recent years, 3D printing has
   been used in the field of tissue engineering, driving research progress
   in regenerative medicine [[65]18–[66]21]. For bone defect regeneration,
   3D printing technology can create personalized scaffolds to promote
   cell migration and adhesion around the bone defect region by enabling
   precise design of the structural parameters of the scaffolds [[67]22].
   In addition, it is well known that bone defect repair requires
   neovascularization to deliver oxygen and exchange metabolites, and
   therefore the ability to promote vascular regeneration is one of the
   critical properties of scaffolds required to induce bone formation
   [[68]23]. Scaffolds fabricated by 3D cryogenic printing possess a
   hierarchical porous structure [[69]24], which is similar to that of
   trabecular bone and has the capacity to promote vascular regeneration
   [[70]25].

   Macrophages, which can be polarized into two major types–M1 and M2–are
   closely associated with neovascularization and bone formation in the
   process of bone defect repair [[71]26]. In the early stage of bone
   repair, macrophages polarize toward the M1 phenotype and secrete
   pro-inflammatory cytokines [[72]26]. In the late stage of bone repair,
   macrophage polarization is shifted toward the M2 phenotype which
   secrete anti-inflammatory cytokines and growth factors, promoting bone
   healing [[73]27]. The shift of macrophage polarization from the M1 to
   the M2 phenotype is essential for bone repair.

   Sleep has an immunoregulatory effect and sleep deprivation promotes the
   expression of inflammatory factors, which are associated with chronic
   inflammation [[74]28]. In recent years, MT, a sleep-inducing molecule,
   has been found to inhibit the expression of inflammatory cytokines and
   reduce the exaggerated inflammatory response to restore tissue function
   [[75]29]. A previous study found that MT enhances bone repair and
   increases bone strength by promoting the expression of angiogenic
   markers [[76]30]. MT promotes the proliferation and osteogenic
   differentiation of bone mesenchymal stem cells (BMSCs) by regulating
   the HGF/PTEN/Wnt/β-catenin pathway, thereby alleviating bone loss
   [[77]31]. Furthermore, MT upregulates the expression of antioxidant
   enzymes and reduces the levels of superoxide, thereby restoring
   mitochondrial function and improving the repair of bone defects
   [[78]32].

   MT is a lipophilic molecule with a poor absorption rate when
   administered orally or by intravenous injection [[79]33]. In addition,
   its clinical application is limited because of poor oral
   bioavailability [[80]34]. To improve the bioavailability, several
   nanotechnology-based drug delivery systems have been developed for
   efficient loading of MT [[81]35]. However, these drug delivery systems
   also present limitations such as tedious synthetic steps, low
   production rates and small-scale production [[82]36]. Small
   extracellular vesicles (sEVs), which contain RNA and protein, originate
   from the endosomal compartment and activate a variety of signaling
   pathways to regulate cellular functions in many physiological processes
   [[83]37]. In addition, sEVs can be modified to carry specific sets of
   RNA or proteins, showing great advantages and application prospects as
   cargo vehicles for targeted drug delivery [[84]38]. Some studies have
   shown that engineered sEVs loaded with bioactive factors have promise
   for use in clinical applications, with a promotional effect on bone
   defect regeneration [[85]39]. Our previous study found that sEVs loaded
   with sleep-related circular RNA promoted chondrocyte proliferation and
   migration as well as showing potential for preventing osteoarthritis
   [[86]40]. However, it has not been elucidated whether MT-treated sEVs,
   in other words, sleep-inspired sEVs, facilitate bone defect
   regeneration.

   Surface functionalization can promote the targeting capacity of sEVs
   [[87]41]. Distearoyl phosphatidylethanolamine-polyethylene glycol
   (DSPE-PEG), an amphiphilic molecule, can be conjugated to targeting
   ligands for surface functionalization of sEVs to enable molecular
   interaction. Inorganic pyrophosphate (PPi) is an excellent targeting
   ligand due to its ability to bind hydroxyapatite [[88]42], and sEVs
   functionalized with DSPE-PEG-PPi may bind to scaffolds containing
   hydroxyapatite. Additionally, it has also been reported that PPi
   promotes osteogenic differentiation while downregulating adipogenic
   differentiation [[89]43].

   In this study, we used 3D cryogenic printing technology to print
   scaffolds containing PLGA and uW (PL-uW scaffold) and then investigated
   the morphology and biocompatibility of the PL-uW scaffold.
   Sleep-inspired sEVs (MT-sEVs) were obtained from MT-stimulated
   fibroblast-like synoviocytes. We used DSPE-PEG-PPi to functionalize the
   MT-sEVs (PPi-MT-sEVs) so that PPi-MT-sEVs could bind to a PL-uW
   scaffold (to form a PL-uW@PPi-MT-sEV scaffold). RNA sequencing was
   performed to explore the molecular mechanisms. The effects of
   PPi-MT-sEVs on osteogenesis, angiogenesis and the regulation of
   macrophage phenotype were investigated by in vitro and in vivo
   experiments.

Materials and methods

   Details of the materials and methods are provided in Supplemental
   Information.

Results

Characterization of 3D cryo-printed scaffolds

   A schematic of the overall experimental strategy is shown in
   Fig. [90]1. Previous studies demonstrated that uW were hundreds of
   micrometers in length and tens of nanometers in diameter [[91]44]. Our
   results also confirmed the high aspect ratio of uW using scanning
   electron microscopy (SEM) and transmission electron microscopy (TEM),
   as shown in Fig. [92]2A. PLGA and PL-uW scaffolds were then printed
   using a 3D cryogenic printer. The surface and internal structure of the
   scaffolds were observed by SEM, and the micrographs showed that both
   scaffolds had regular shapes with specific parameters as initially set
   (Fig. [93]2B). In addition, interconnected microporous architecture was
   observed inside the scaffolds printed using 3D cryogenic printing
   technology, forming connected pipelines to facilitate cell migration
   and transmission of nutrients [[94]45]. The images showed that there
   were smaller pores on the pore walls of the microporous structure
   (Fig. [95]2B). To verify the presence of uW, energy dispersive
   spectroscopy (EDS) was performed. The results of EDS point analysis
   showed that Ca and P were present on the surface of a PL-uW scaffold
   but not a PLGA scaffold, indicating the presence of uW in the PL-uW
   scaffold (Fig. [96]2C).

Fig. 1.

   [97]Fig. 1
   [98]Open in a new tab

   Schematic illustration of the process of preparing PPi-MT-sEVs and
   treatment of the PL-uW@PPi-MT-sEV scaffold. (A) Printing of the PL-uW
   scaffold using 3D cryogenic printing technology. (B) Modification of
   PPi-MT-sEVs. MT was used to produce MT-sEVs, which were then
   surface-functionalized by DSPE-PEG-PPi (PPi-MT-sEVs). (C) The
   PL-uW@PPi-MT-sEV scaffold facilitated vascularized bone regeneration
   based on angiogenesis–osteogenesis coupling and modulation of the
   macrophage phenotype. In the early stage, PPi-MT-sEVs not only promoted
   cell migration, proliferation and osteogenic differentiation, but also
   inhibited M1 polarization; in the late stage, PPi-MT-sEVs promoted cell
   proliferation, osteogenic differentiation and M2 polarization.
   Illustration credit: Lina Cao

Fig. 2.

   [99]Fig. 2
   [100]Open in a new tab

   Characterization of porous scaffolds printed using 3D cryogenic
   printing. (A) Representative SEM and TEM micrographs of uW. Scale bars
   of SEM images are 10 μm and 2 μm. For TEM images, scale bars are 2 μm
   and 400 nm. (B) Representative surface and cross-sectional SEM images
   of PLGA and PL-uW scaffolds. For surface images, scale bars are 600 μm,
   60 μm and 15 μm, respectively. For cross-sectional images, scale bars
   are 40 μm, 20 μm and 10 μm, respectively. (C) Results of
   energy-dispersive spectroscopy (EDS) analysis of PLGA and PL-uW
   scaffolds

   Scaffolds produced by 3D cryogenic printing consisted of extruded and
   deposited bioink filaments. Both global and local (microstructure of
   deposited layer) 3D reconstructions of PLGA and PL-uW scaffolds were
   produced by micro-CT and are indicated in blue. The pore 3D
   reconstruction of PLGA and PL-uW scaffolds are shown in pseudocolors to
   distinguish micropores with different diameters (Fig. [101]S1). The
   results of micro-CT analysis showed that the total (including macropore
   and micropore) and local (micropore only) porosity of the PLGA scaffold
   was 75.55% and 59.49%, respectively, while the total and local porosity
   of the PL-uW scaffold was 72.91% and 53.03% (Fig. [102]S1). Further,
   the compressive modulus of the PL-uW scaffold (mean = 4.654 MPa, n = 3)
   was higher than that of the PLGA scaffold (mean = 2.752 MPa, n = 3),
   which suggested that uW enhanced the mechanical strength of the PL-uW
   scaffold (Fig. [103]S2).

Biocompatibility of scaffolds

   In order to assess the biocompatibility of scaffolds, human
   microvascular endothelial cells (HMEC-1) were seeded onto scaffolds and
   cultured for 3 d. Cell adhesion on both PLGA and PL-uW scaffolds was
   observed using SEM, and the results revealed that HMEC-1 grew well on
   both types of scaffolds (Fig. [104]3A). Live/Dead assay was performed
   to evaluate the viability of HMEC-1 (Fig. [105]3B), and demonstrated
   that HMEC-1 grew on the surface of the scaffolds, and were able to
   survive for a long time on PLGA and PL-uW scaffolds, with the addition
   of uW having no effect on cell activity. These results demonstrated
   that both PLGA and PL-uW scaffolds have good biocompatibility.

Fig. 3.

   [106]Fig. 3
   [107]Open in a new tab

   Biocompatibility of the scaffolds. (A) SEM images of co-culture
   experiments. The images show HMEC-1 cultured on PLGA and PL-uW
   scaffolds. Scale bar: 100 μm in low magnification, 10 μm in high
   magnification. (B) Fluorescence images of calcein-AM (green) and
   propidium iodide (red) captured under confocal microscopy. Scale bar:
   200 μm

Adsorption of sEVs to the scaffolds by surface functionalization

   DSPE-PEG, an amphiphilic material, is widely used in drug delivery, and
   can be conjugated to different molecules [[108]46]. PPi has high
   affinity for hydroxyapatite because phosphate and hydroxyl groups can
   bind to the cluster of calcium ions [[109]47]. DSPE-PEG-PPi was used to
   modify the surface of MT-sEVs, and its structural formula is shown in
   Fig. [110]4A. To provide a clear view of the binding, hydroxyapatite
   and PPi-terminus of DSPE-PEG-PPi were modeled using Materials Studio
   (Biovia, San Diego, CA, USA) (Fig. [111]4B). The structure and
   differential electron charge density were calculated to obtain the
   charge transfer at the interface. The results demonstrated that the
   electron cloud was shifted in the interfacial region after PPi and
   hydroxyapatite came into contact (Fig. [112]4C). All these results
   supported the hypothesis that DSPE-PEG-PPi had the ability to bind to
   hydroxyapatite.

Fig. 4.

   [113]Fig. 4
   [114]Open in a new tab

   PPi-MT-sEVs coated onto scaffolds. (A) Chemical structure of
   DSPE-PEG-PPi polymers. (B) 3D charge density difference before and
   after contact between hydroxyapatite and PPi-terminus of DSPE-PEG-PPi
   based on density functional theory calculations. (C) The z-axis
   plane-averaged electron charge density difference of
   hydroxyapatite/PPi-terminus. (D) The adsorption of different sEVs
   labeled with DiI onto the PLGA and PL-uW scaffolds. Scale bar: 100 μm.
   (E) A schematic illustration of PPi-MT-sEVs. (F) A schematic
   illustration of PL-uW@PPi-MT-sEVs. Illustrations were created using
   BioRender.com

   The isolated sEVs were divided into four groups: (I) The sEV group:
   sEVs directly isolated from the supernatant; (II) The MT-sEV group:
   sEVs isolated from the supernatant of MT-stimulated cells; (III) The
   PPi-sEV group: sEVs isolated from the supernatant and functionalized by
   DSPE-PEG-PPi; (IV) The PPi-MT-sEV group: sEVs isolated from the
   supernatant of MT-stimulated cells and functionalized with
   DSPE-PEG-PPi. The isolated sEVs were characterized using cryo-TEM,
   atomic force microscopy (AFM) and dynamic light scattering (DLS). The
   morphology of sEVs was observed by cryo-TEM prior to or after surface
   functionalization and the results demonstrated that spheroid sEVs
   possessed a compact membrane structure (Fig. [115]S3). The AFM
   topography images showed that the isolated sEVs had a distinct rounded
   morphology without aggregates. The size distribution of sEVs was
   measured by DLS and the results showed that the particle size of
   different sEVs was in the range of 30–200 nm (Fig. [116]S4), which was
   consistent with previous reports [[117]48].

   PLGA and PL-uW scaffolds were placed in 24-well plates and different
   sEVs labeled with DiI (a lipophilic fluorescent dye) were added. The
   results of adsorption experiments are shown in Fig. [118]4D. As can be
   seen, no sEVs attached to the PLGA scaffold regardless of whether the
   surface was functionalized by DSPE-PEG-PPi or not. PPi-sEVs and
   PPi-MT-sEVs were attached to the PL-uW scaffold whereas sEVs and
   MT-sEVs were not attached to the PL-uW scaffold. A schematic
   illustration of the process of PPi-MT-sEV production is provided in
   Fig. [119]4E, while Fig. [120]4F shows the binding process of
   PPi-MT-sEVs to the PL-uW scaffold.

Bioinformatic analysis of the influence of MT on gene expression

   BMSCs were divided into two groups according to whether MT was added to
   the culture medium. RNA sequencing (RNA-seq) was performed to identify
   differentially-expressed genes (DEGs) between the MT-treated group and
   the control group (Fig. [121]5A). KEGG, Reactome, Wikipathway and GO
   pathway enrichment analysis were performed to identify some signaling
   pathways and the data were visualized as an interaction network by
   Cytoscape ([122]https://cytoscape.org/) (Fig. [123]5B–E, Fig.
   [124]S5–[125]S6): (I) “Osteoblast differentiation and related
   diseases”, “WNT ligand biogenesis and trafficking” and “Hippo signaling
   pathway”, implied that MT might have an influence on
   osteogenic/adipogenic differentiation [[126]49]; (II)
   “Phosphatidylinositol-3,4,5-trisphosphate binding”, “Protein kinase B
   signaling pathway” and “Phosphatidylinositol-4,5-bisphosphate binding”,
   suggested that MT might have some influence on angiogenesis and cell
   proliferation [[127]50]; (III) “Cytoskeletal anchor activity”,
   suggested that MT might have an influence on cell migration [[128]51];
   (IV) “Cytokine–cytokine receptor interaction” and “Cytokine-mediated
   signaling pathway”, implied that MT might have an influence on the
   interplay between cells and their microenvironment [[129]52]; (V)
   “Cyclooxygenase pathway”, “Peroxidase activity”, “NAD(P)H oxidase
   H[2]O[2]-forming activity” and “Cytokines and inflammatory response
   signaling pathway”, suggested that MT might have some influence on
   inflammatory and anti-inflammatory responses [[130]53]. Based on these
   results of bioinformatic analysis, observation of the effects of
   PPi-MT-sEVs on cell proliferation, migration, osteogenic/adipogenic
   differentiation and macrophage polarization seemed a promising
   investigation.

Fig. 5.

   [131]Fig. 5
   [132]Open in a new tab

   RNA sequencing and bioinformatics analysis. (A) Heatmap of clustering
   analysis. (B–E) The interaction networks identified by KEGG (B),
   Reactome (C), Wikipathway (D) and GO (E) pathway enrichment analysis

Osteogenic and adipogenic modulation by PPi-MT-sEVs

   Studies have found that MT promotes osteogenic differentiation of MSCs,
   with the gene expression levels of osteogenic differentiation markers
   increased [[133]54]. To investigate the effect of PPi-MT-sEVs on
   osteogenic differentiation, BMSCs were cultured in osteogenic induction
   medium with four different sEVs as previously described. Alkaline
   phosphatase (ALP) is an early-stage enzyme marker of osteoblasts, and
   ALP staining was performed after 7 d of osteogenic differentiation. As
   shown in Fig. [134]6A and B, BMSCs cultured with PPi-MT-sEVs or MT-sEVs
   exhibited the highest expression levels of ALP, while the control group
   exhibited the least ALP activity. Alizarin red staining was performed
   to examine calcium deposition after 14 d of osteogenic induction, and
   the results demonstrated that the four groups cultured with sEVs
   developed more calcium nodules than the control group (Fig. [135]6C).
   Notably, the MT-sEV and PPi-MT-sEV groups displayed the most mineral
   deposition (Fig. [136]6D). BMSCs possess the capacity to differentiate
   into adipocytes and MT inhibits adipogenic differentiation [[137]55].
   Oil red O staining was conducted to investigate the effect of
   PPi-MT-sEVs on adipogenic differentiation after 21 d of induction. As
   shown in Fig. [138]6E, the MT-sEV and PPi-MT-sEV groups contained fewer
   lipid droplets than the control group. The results of reverse
   transcription quantitative PCR (RT-qPCR) experiments indicated that
   PPi-MT-sEVs promote the expression of osteogenic markers, including
   type I collagen (COL I), ALP, and runt-related transcription factor 2
   (RUNX2) (Fig. [139]6F). These results demonstrated that MT-sEVs and
   PPi-MT-sEVs exhibited the capacity to promote osteogenic
   differentiation while suppressing adipogenic differentiation.

Fig. 6.

   [140]Fig. 6
   [141]Open in a new tab

   Effect of EVs on cell differentiation. (A) ALP staining. BMSCs were
   cultured in osteogenic medium for 7 d before ALP staining was
   performed. Scale bar: 400 μm. (B) Relative ALP activity (versus control
   group). (C) Alizarin red staining. Alizarin red staining was performed
   after 14 d of osteogenic differentiation. Scale bar: 400 μm. (D)
   Relative mineral deposition (versus control group). (E) Oil Red O
   staining. Oil Red O staining was performed after 21 d of adipogenic
   differentiation. Scale bar: 400 μm. (F) RT-qPCR results of osteogenic
   markers. *p < 0.05 compared with control group. **p < 0.01 compared
   with control group. ***p < 0.001 compared with control group.
   ****p < 0.0001 compared with control group

Effects of PPi-MT-sEVs on cell proliferation, cell migration and angiogenesis

   MT facilitates cell proliferation and inhibits apoptosis through
   multiple signaling pathways [[142]56]. EdU cell proliferation assay was
   performed to investigate the effect of sEVs on the cell proliferation
   of HMEC-1. Flow cytometry showed that EdU-positive cells in the MT-sEV
   (62.0%) and PPi-MT-sEV (60.5%) groups were significantly increased
   compared with the control (50.0%), sEV (51.2%) and PPi-sEV (54.4%)
   groups (Fig. [143]7A).

Fig. 7.

   [144]Fig. 7
   [145]Open in a new tab

   The effects of PPi-MT-sEVs on cell proliferation, cell migration and
   the formation of tubule-like structures. (A, B) Flow cytometry results
   of EdU cell proliferation assay in HMEC-1 (A) and BMSCs (B). (C, D)
   Transwell migration experiments of BMSCs (C) and HMEC-1 (D). Scale bar:
   500 μm. (E) Representative images of angiogenesis assay. Scale bar:
   500 μm. (F) Schematic illustration of the Transwell migration assay.
   (G) Schematic illustration of the effects of PPi-MT-sEVs on bone and
   neovascular regeneration. Illustrations were created using
   BioRender.com

   We further explored the role of PPi-MT-sEVs on cell proliferation of
   BMSCs. As shown in Fig. [146]7B, flow cytometry indicated that
   EdU-positive BMSCs in the MT-sEV and PPi-MT-sEV groups accounted for
   15.9% and 16.1% of total cells, a higher proportion than in the control
   (8.27%), sEV (9.08%) or PPi-sEV groups (12.5%).

   Cell migration to the bone defect region is an important factor that
   affects bone regeneration, and MT has been shown to enhance cell
   migration [[147]30]. A Transwell migration assay was performed to
   evaluate the impact of PPi-MT-sEVs on cell migration of BMSCs and
   HMEC-1. The results, shown in Fig. [148]7C and D, indicated that
   significantly more cells migrated to the lower chambers in the
   PPi-MT-sEV and MT-sEV groups than in the control group. The number of
   migrated BMSCs in the MT-sEV and PPi-MT-sEV groups was approximately
   2.94 times and 3.13 times higher than the control group, respectively.
   Similarly, the migration of HMEC-1 in the MT-sEV and PPi-MT-sEV groups
   was approximately 2.74 times and 2.96 times higher than the control
   group (Fig. [149]S7). These results showed that PPi-MT-sEVs promoted
   the migratory capacity of BMSCs and HMEC-1.

   Neovascularization is favorable for the transmission of nutrients and
   cytokines, which is important for bone regeneration [[150]57]. In order
   to explore the influence of PPi-MT-sEVs on angiogenesis, HMEC-1 adhered
   on Matrigel were cultured in cell medium with the different sEVs. As
   can be seen in Fig. [151]7E and Fig. [152]S8, after 6 h, there were
   more tubule-like structures formed in the MT-sEV and PPi-MT-sEV groups
   compared with the control group. This result indicated that PPi-MT-sEVs
   have a promotional effect on angiogenesis, which is beneficial for bone
   defect repair. A schematic illustration of the Transwell assay is
   provided in Fig. [153]7F, while Fig. [154]7G shows a schematic
   representation of the promotional effects of PPi-MT-sEVs on bone
   regeneration and neovascularization.

Modulation of macrophage polarization by PPi-MT-sEVs

   Immunofluorescent staining was performed to investigate the effect of
   PPi-MT-sEVs on macrophage polarization. Tumor necrosis factor α (TNFα),
   C-C motif chemokine ligand 3 (CCL3), and inducible nitric oxide
   synthase (iNOS) are M1 macrophage markers, while arginase 1 (ARG1),
   suppressor of cytokine signaling 2 (SOCS2), transforming growth factor
   β (TGFβ), and CD206 are markers of M2 macrophages. The results showed
   that the MT-sEV and PPi-MT-sEV groups expressed lower levels of iNOS
   (Fig. [155]8A and Fig. [156]S9). Meanwhile CD206 expression showed the
   opposite trend among these groups, with CD206 expression in the MT-sEV
   and PPi-MT-sEV groups being significantly greater than in the other
   groups (Fig. [157]8B and Fig. [158]S10). Schematic illustrations of the
   effect of PPi-MT-sEVs on macrophage polarization are provided in
   Fig. [159]8C and D. RT-qPCR experiments were performed to detect the
   expression of M1 and M2 macrophage markers. The MT-sEV and PPi-MT-sEV
   groups showed decreased expression of M1 markers (iNOS, TNFα, CCL3) and
   increased expression of M2 markers (ARG1, SOCS2, TGFβ) (Fig. [160]S11).
   These results showed that PPi-MT-sEVs induced a shift in macrophage
   polarization from the M1 to the M2 phenotype, preventing excessive
   inflammation and promoting tissue repair.

Fig. 8.

   [161]Fig. 8
   [162]Open in a new tab

   The effect of PPi-MT-sEVs on macrophage polarization. (A)
   Immunofluorescent images of M1 macrophage polarization stained to show
   iNOS (red) and nuclei (blue). Scale bar: 100 μm. (B) Immunofluorescent
   images of M2 macrophage polarization stained to show CD206 (green) and
   nuclei (blue). Scale bar: 100 μm. (C) Schematic illustration showing
   the effect of PPi-MT-sEVs on M1 polarization. (D) Schematic
   illustration showing the effect of PPi-MT-sEVs on M2 polarization.
   Schematic illustrations were created using BioRender.com

The ability of the PL-uW@PPi-MT-sEV scaffold to repair bone defects in vivo

   The effects of the PL-uW@PPi-MT-sEV scaffold on bone regeneration and
   revascularization were evaluated in a rat cranial defect model. As
   shown in Fig. [163]9A, there was only a small amount of
   neovascularization in the defect region of the PLGA, PL-uW and
   PL-uW@PPi-sEV groups. More neovascularization was observed in the
   PL-uW@PPi-MT-sEV group, revealing that PPi-MT-sEVs have a strong
   pro-angiogenic effect. 3D reconstruction and pseudocolor processing of
   cross-sections were performed with reference to previous studies
   [[164]58–[165]60]. The results (Fig. [166]9B and C) indicated that
   there was almost no new bone formation in the PLGA group. A limited
   number of bone islands were formed in the PL-uW and PL-uW@PPi-sEV
   groups. In the PL-uW@PPi-MT-sEV group, a large number of bone islands
   were generated, and the density of the regenerated bone was close to
   that of native bone.

Fig. 9.

   [167]Fig. 9
   [168]Open in a new tab

   Effects on bone regeneration and vascularization in vivo. (A) Micro-CT
   images of neovascularization. (B) Micro-CT 3D images of the defect
   areas. (C) Micro-CT sagittal images presented in a pseudocolor mode
   with distinct colors indicating different gray values

   The effects of the PL-uW@PPi-MT-sEV scaffold on bone defect
   regeneration were further assessed by histological staining. H&E
   staining showed that little new bone was formed in the PLGA group
   (Fig. [169]10A), and there was limited bone regeneration within the
   defect area in the PL-uW and PL-uW@PPi-sEV groups. A noteworthy
   difference was observed in the bone defect area of the PL-uW@PPi-MT-sEV
   group, where a substantial amount of new bone formation was evident,
   indicating that PL-uW@PPi-MT-sEVs had a pro-osteogenic effect. Masson’s
   staining showed that there were significantly more collagen fibers in
   the PL-uW@PPi-MT-sEV group compared with the other groups
   (Fig. [170]10B). In addition, the collagen fibers were interconnected
   throughout the whole defect region of the PL-uW@PPi-MT-sEV group. The
   result of COL I staining was consistent with that of Masson’s staining
   (Fig. [171]10C). The PL-uW@PPi-MT-sEV group exhibited significantly
   higher expression of COL I in the defect region, indicating the
   capability of PL-uW@PPi-MT-sEVs to facilitate bone defect repair (Fig.
   [172]S12). In addition, the highest expression of osteocalcin (OCN) was
   observed in the PL-uW@PPi-MT-sEV group (Fig. [173]10D).
   Immunofluorescent staining of CD206 and iNOS demonstrated that the
   PL-uW@PPi-MT-sEV group had the greatest capacity to inhibit M1
   polarization and promote M2 polarization compared with the other groups
   (Fig. [174]10E and F). In summary, Fig. [175]1 depicts the primary
   principle that the PL-uW@PPi-MT-sEV scaffold promotes vascularized bone
   regeneration via angiogenesis–osteogenesis coupling and modulation of
   the macrophage phenotype.

Fig. 10.

   [176]Fig. 10
   [177]Open in a new tab

   Histological analysis of bone defect repair in a rat model. (A) H&E and
   (B) Masson’s trichrome staining of the bone defect area implanted with
   different scaffolds. Low magnification scale bar: 1 mm and high
   magnification scale bar: 200 μm. (C) Immunohistochemical staining of
   COL I. Low magnification scale bar: 1 mm and high magnification scale
   bar: 200 μm. (D) Immunofluorescence staining of OCN (green) and nuclei
   (blue). Scale bar: 100 μm. (E) Immunofluorescence staining of iNOS
   (green) and nuclei (blue). Scale bar: 100 μm. (F) Immunofluorescence
   staining of CD206 (green) and nuclei (blue). Scale bar: 100 μm

Discussion

   Bone defect healing is a complex process that requires appropriate
   substitute materials and a pro-regenerative microenvironment [[178]61].
   3D-printed scaffolds have a precise structure and can be personalized
   to meet the needs of bone regeneration, and thus provide a potential
   tailored treatment strategy [[179]62]. Many synthetic materials have
   been used for 3D-printed scaffolds [[180]63]. In this study, we used
   PLGA and uW to print scaffolds. Cells adhered to scaffolds were in a
   good state of growth demonstrating that the PL-uW scaffold was an ideal
   substrate for bone defect regeneration due to its excellent
   biocompatibility.

   Porosity is the focus of scaffold design and high porosity facilitates
   cell adhesion, cell migration and the exchange of metabolites. The
   pores of 3D-printed scaffolds can be classified into two types:
   macropores and micropores. Macropores of appropriate diameter improve
   the osteogenic capacity of scaffolds [[181]64]. Scaffolds with uniform
   pore sizes of 378–435 μm are known to promote better cell
   proliferation, neovascularization and osteogenic differentiation than
   scaffolds without macropores [[182]65]. Hence, we designed scaffolds
   with 400 μm diameter macropores in this work. The presence of
   micropores is one of the main requirements for scaffolds. Some studies
   demonstrated that 3D cryo-printed scaffolds had high porosity and
   interior micropores, which provided a suitable microenvironment for
   cell growth [[183]66]. Porous scaffolds with controllable micropores
   have been shown to promote cellular adhesion and bone formation
   [[184]45]. Herein, we printed a PL-uW scaffold using 3D cryogenic
   printing technology. Our study suggested that the PL-uW scaffold had a
   hierarchical porous structure with interconnected micropores, and cells
   adhered well to the surface of the PL-uW scaffold. In addition, the
   PL-uW scaffold had high mechanical strength, another property essential
   for bone regeneration [[185]67].

   Compared with bone tissue, 3D cryo-printed scaffolds possess a
   structure resembling bone microarchitecture but lack substantial
   amounts of bioactive factors [[186]68]. Thus, supplementation with
   bioactive factors is one of the effective approaches to improve the
   osteogenic induction capacity of scaffolds. Bone defect repair is a
   complex and dynamic process involving osteogenic differentiation, cell
   proliferation, angiogenesis and coordination of the inflammatory
   response [[187]69]. Consequently, bioactive factors with extensive
   effects are urgently needed for bone defect healing [[188]70].
   Circadian rhythms, one of the physiological human rhythms, regulate
   multiple cellular functions and have important implications for tissue
   homeostasis and repair [[189]71]. Circadian genes affect the regulation
   of bone formation and bone resorption through their impact on
   osteoblast functionality as well as intercellular communication
   [[190]72]. Circadian disruption has adverse effects on bone mineral
   density and strength of the skeleton [[191]73]. Based on this
   background, we combined 3D cryo-printed scaffolds with bioactive
   factors related to the circadian rhythm and investigated whether the
   combination could enhance bone defect regeneration. MT is released by
   the pineal gland and implicated in regulation of the circadian rhythm,
   which has been proven to be a regulator of bone formation [[192]74]. It
   has been demonstrated that MT induces osteogenic differentiation and
   inhibits adipogenic differentiation of stem cells [[193]75]. In
   addition, MT enhances anti-inflammatory polarization of macrophages and
   reduces tissue inflammation [[194]76]. Functionally, MT is an apparent
   bioactive factor for bone defect regeneration. We further considered
   how to combine MT with hierarchical porous scaffolds printed using 3D
   cryogenic printing technology.

   Modularized sEVs have attracted attention in the field of bone tissue
   engineering as a drug delivery system [[195]77]. The lipid bilayer
   membrane of sEVs can be modified with targeting ligands while
   pretreatment of cells regulates the internal contents of sEVs
   [[196]78]. It has been demonstrated that sEVs carrying specific
   microRNAs promote osteogenic differentiation of BMSCs and enhance bone
   regenerative ability [[197]79]. In our previous study, chondrocytes
   were cultured with MT and a sleep-related circular RNA was identified
   by RNA sequencing. These sEVs combined with the sleep-related circular
   RNAs have been proven to be beneficial in preventing osteoarthritis
   progression [[198]40]. Herein, we further isolated MT-sEVs,
   sleep-inspired sEVs, from the supernatant of MT-stimulated cells. RNA
   sequencing was performed and we found that MT influenced multiple
   signaling pathways involved in osteogenic differentiation, angiogenesis
   and inflammation, which revealed the potential of PPi-MT-sEVs to
   enhance bone defect regeneration.

   Immune modulation is an essential component of bone defect healing
   [[199]69]. M1 macrophages, which are pro-inflammatory, promote bone
   defect repair at the inception, but long-term M1 macrophage
   infiltration can result in delayed healing [[200]26], while
   anti-inflammatory M2 macrophages play an essential role in bone
   regeneration due to their capacity to promote tissue repair and
   angiogenesis [[201]80]. It has been demonstrated that facilitating M2
   phenotype polarization promotes recruitment and osteogenic
   differentiation of MSCs. In addition, anti-inflammatory cytokines
   secreted by M2 macrophages contribute to bone repair [[202]81]. In this
   study, we demonstrated that PPi-MT-sEVs promoted macrophage
   polarization towards the M2 phenotype and inhibited M1 polarization in
   vitro. In addition, the PL-uW@PPi-MT-sEV scaffold induced M2
   polarization to facilitate bone defect repair in vivo.

   Newly-formed vessels have the ability to deliver nutrients to, and
   clear metabolites from, bone, which is another important factor in bone
   regeneration [[203]82]. MT has the ability to induce M2 polarization,
   which has been proven to be beneficial for angiogenesis [[204]83]. In
   addition, the migration of endothelial cells is one of the important
   prerequisites for angiogenesis [[205]84]. In this study, we printed
   porous scaffolds incorporating various sEVs and our results showed that
   PPi-MT-sEVs promoted M2 polarization, cell migration and the formation
   of tubule-like structures in vitro. We also demonstrated that the
   PL-uW@PPi-MT-sEV scaffold induced the formation of more neovessels in
   the bone defect region in vivo compared to other groups.

   Cell migration to the defect site and osteogenic differentiation of
   BMSCs have a significant effect on bone defect healing [[206]85]. BMSCs
   have the potential to differentiate into adipocytes and osteoblasts
   [[207]86]. Differentiation of BMSCs to adipocytes may lead to bone loss
   and delay bone regeneration [[208]87]. Herein, for the first time, our
   research indicated that PPi-MT-sEVs have the ability to enhance
   osteogenic differentiation and inhibit adipogenic differentiation. In
   addition, our research indicated that PPi-MT-sEVs have the ability to
   enhance cell migration, facilitating bone repair in a calvarial defect
   model.

   sEVs can be loaded onto scaffolds in several different ways depending
   on the materials and printing technique [[209]88]. sEVs can be adsorbed
   onto the surface of positively-charged scaffolds due to the negative
   potential of the membrane [[210]89]. In addition, sEVs can be directly
   encapsulated into a hydrogel which preserves their biological activity
   due to its water solubility [[211]90]. Notably, PLGA is usually
   dissolved in organic solvents, which might disrupt the bioactivity of
   sEVs, and the scaffolds also experience a dramatic change in
   temperature during the 3D cryogenic printing process [[212]91]. The
   modularized EVs provide a solution to this problem, and the membrane of
   sEVs can be modified via targeting ligands for binding ability
   [[213]38]. Thus, we considered binding sEVs to scaffolds via surface
   modification.

   The lipid bilayer of sEVs can be modified by amphiphilic molecules.
   DSPE-PEG is an amphiphilic molecule that includes a lipid tail and can
   be conjugated to various ligands [[214]92]. In this work, we combined
   DSPE-PEG with PPi and used DSPE-PEG-PPi to modify the MT-sEVs,
   enhancing the binding ability of PPi-MT-sEVs to the PL-uW scaffold. The
   uW provided binding sites for modified PPi-MT-sEVs and the porosity
   improved the loading efficiency. Quantum chemical simulation was
   performed to confirm the interaction between PPi and uW, and adsorption
   experiments demonstrated that adsorptive interactions existed between
   the PL-uW scaffold and PPi-MT-sEVs.

Conclusion

   Briefly, these results showed that PPi-MT-sEVs have the ability to
   promote cell proliferation, migration and angiogenesis. In addition,
   PPi-MT-sEVs regulate the macrophage M1/M2 polarization shift and
   osteogenic/adipogenic differentiation. Our work is the first to report
   that sleep-inspired sEVs can be modified with DSPE-PEG-PPi to
   incorporate into hierarchical porous scaffolds printed using 3D
   cryogenic printing technology. The in vivo experiments demonstrated
   that the combination of PPi-MT-sEVs and porous scaffolds had the
   therapeutic effects of osteogenesis–angiogenesis coupling and
   macrophage phenotype regulation, promoting bone defect regeneration. In
   summary, our study highlights the pivotal roles of a PL-uW@PPi-MT-sEV
   scaffold for bone defect regeneration, which we believe represents a
   novel therapeutic strategy for the treatment of bone defects.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [215]Supplementary Material 1^ (17.1MB, docx)

Acknowledgements