Exosome: A Novel Nanocarrier Delivering Noncoding RNA for Bone Tissue Engineering

Bone growth and metabolism are mainly regulated by a series of intracellular molecules and extracellular stimuli. Exosome, as a nanoscale substance secreted to the outside of the cells, plays an extensive role in intercellular communication. This review provides theoretical references and evidences for further exploration of exosomes as noncoding RNA carriers to regulate bone tissue recovery through the following aspects: (1) basic characteristics of exosomes, (2) research progress of exosomal noncoding RNA in bone tissue engineering, (3) current status and advantages of engineering exosomes as nanocarriers for noncoding RNA delivery, and (4) problems and application prospects of exosome therapy in the field of orthopedics.


Introduction
Destructive osteopathy is mainly due to the failure of the bone remodeling process, including enhanced osteoclast activity or reduced osteoblast generation [1,2]. Osteoarthritis, osteoporosis, etc. will lead to a continuous decline in the quality of patients' life [3,4], deemed to be an important global health problem [5,6]. The process of bone remodeling is to repair bone damage and restore aging bone tissue [7]. This is the prerequisite for maintaining mass and mechanical strength of bone [8]. Mutual regulation of osteoblasts and osteoclasts is essential for bone remodeling [9], dominated by molecular signals [10,11].
Extracellular vesicles are natural nanoscale particles secreted by cells. They are spherical bilayer membrane vesicles with diameters ranging from 30 nm to 1000 nm, including exosomes of uniform size (30-150 nm), microvesicles of varying sizes (50-1000 nm), and apoptotic bodies [12]. Recent studies reported that the key factor for bone reconstruction was exosome [13]. These organelles play a crucial role in intercellular messengers, transferring bioactive molecules to nearby or distant cells under physiological and pathological conditions [14,15]. Noncoding RNAs in exosomes supply a method that cells can straightly regulate the expression of pro-tein in target cells. As a kind of nanomaterial, exosome has been explored in the field of diagnostic [16], therapeutics [17], and drug delivery [18] (Figure 1).
Existing drug carriers, such as artificially manufactured liposomes, can cause accumulated toxicity and immunogenicity in the body, thus limiting the wide application of such drug carriers [20,21]. As a natural drug delivery carrier, exosomes have significant advantages of biocompatibility, high stability, and targeting. In order to better understand the mechanism of exosomes as noncoding RNA vectors regulating bone remodeling, we reviewed the newest findings on the feature and role of exosomes in bone formation. In addition, we focused on elaborating the feasible clinical application of bone exosomes and the characteristics of exosomal noncoding RNAs in the regulation of bone reconstruction.

Biological Characteristics of Exosomes
Exosome was first discovered in 1981, considered cellular trash initially [22]. Subsequently, exosomes caught the interest of immunologists in the 1990s. Raposo et al. in 1996 reported that exosomes have the ability of antigen presentation, activating T cell immune response [23]. In addition, exosomes were found to carry miRNAs and mRNAs for intercellular communication in 2007. Most importantly, the mechanism of exosomal transport was cracked by three independent groups [24][25][26]. The history, biomarkers, and functions of exosomes are shown in Figure 2.
Exosomes are membrane vesicles with diameters between 30 and 150 nm [27,28] (Figure 3). They are derived from body cells including tumor cells [29,30]. Studies revealed that exosomes are also secreted from bone-related cells (osteoclasts, osteoclast precursors, osteoblasts, and MSCs) [31][32][33]. Exosomes reflect the complex molecular processing that occurs inside the parent cells and are more appropriate as a potential replacement for parent cell in body fluids. Multiple studies have confirmed the loyalty of exosomal cargo to parental cells [34,35]. Exosomes can encapsulate fragments of genomic and different kinds of RNA, such as miRNA, lncRNA, circRNA, and ribosomal RNA [36][37][38][39].

Exosomes from Osteoclasts and Osteoblasts
3.1. Osteoclast Exosomes. Exosomes secreted by osteoclasts may provide clues on how to coordinate bone absorption and construction. In calcitriol-stimulated mouse bone marrow (C-SMM), osteoclasts and osteoblasts differentiated synergistically within six days. Osteoclast precursor-derived exosomes stimulated the proliferation of osteoclasts in C-SMM, while exosomes of mature osteoclasts inhibited osteoclast formation [13,40]. Osteoclast exosomes stimulate the osteoblasts and other nonosteoclasts in a paracrine manner and produce factors that regulate osteoclast differentiation.
miRNA-214, involved in bone remodeling, adjusts Osterix and activates transcription factor (ATF) 4, which is an essential transcription factor in osteogenesis process [41,42]. Higher levels of miRNA-214 were noted in the plasma of fractured elderly women and ovariectomized mouse. In vitro, by using transwell membranes (that only permitted exosomes through), exosomal miRNA-214 was transmitted from osteoclasts to osteoblasts, inhibiting proliferation and mineralizing activity of osteoblast in vitro and vivo [43,44].

Osteoblast Exosomes.
Osteoblasts release exosomes containing receptor activator RANKL. These exosomes induce osteoclast precursors to differentiate to osteoclasts [31]. Osteoblasts release RANKL exosomes and osteoclasts release RANK exosomes. Studies have shown that the RANKL-RANK-osteoprotectin signal network is the essence of bone biology. RANKL exosome from osteoblasts may be capable of promoting osteoclast formation and activity without direct intercellular contact. RANK exosome may inhibit RANKL in a method the same as osteoprotectin. Both RANKL and  Figure 1: Exosome production, secretion, and cargo transfer from donor cells to recipient cells. The inset shows the clinical application of noncoding RNA exosomes [19].
RANK-containing exosome may be able to fuse in their target cells to release regulatory molecules, which could assist to reprogram target cells [45].

The Roles of Exosomal Noncoding RNAs in Bone Growth
Noncoding RNA refers to RNA that does not encode protein.
4.1. miRNA. The latest two years of exosome miRNA in the regulation of bone growth and development is listed in Table 1. Exosome miR-128-3p regulates osteogenesis and fracture healing by targeting Smad5 [48]. Decreasing miR-221-3p in exosomes can significantly reduce chondrocyte proliferation and migration in vitro [49]. Exosome miR-17 secreted by keratinocytes can induce osteoclast differentiation [30]. By freeze and thaw method, researchers combined exosomes and miR-140 to promote cartilage differentiation of bone marrow stem cells, thereby enhancing cartilage repair and regeneration [50]. Human umbilical cord mesenchymal stem cell exosomes can effectively inhibit bone marrow mesenchymal stem cell apoptosis and prevent osteoporosis in rats.
The mechanism is mediated by the miR-1263/Mob1/Hippo signaling pathway [51]. Exosome miR-8485 promotes cartilage differentiation of bone marrow mesenchymal stem cells by regulating the Wnt/β-catenin pathway [52]. Studies on miRNAs have demonstrated that the regulation of the target genes affects the expression of regulators upstream and downstream of each signaling pathway, to regulate the osteogenic differentiation process [54]. Runx2 is a transcription factor that plays an important role in osteoblast differentiation [55], which is precisely or indirectly regulated by many miRNAs, including miR-221 [56,57], miR-133a-5p [58], miR-467g [59], miR-218 [60,61], and miR-210 [62]. The expression of Osx can be downregulated by some miR-NAs, inhibiting osteogenic differentiation, such as miR-145 [63] and miR-143 [64,65]. Some miRNAs that regulate the expression of bone-related gene have not been studied in exosomes, providing clues to the research of exosomes in the field of orthopedics (Table 2).

lncRNA.
The lncRNA-miRNA-mRNA regulation model plays a critical role in osteogenic differentiation [66]. lncRNA can serve as a competitive endogenous RNA, competing for miRNA binding sites to reduce its direct impact on mRNA. lncRNA MALAT1 can competitively bind to miR-30, inhibiting the interaction between miR-30 and Runx2 to upregulate the transcription level of Runx2 and strengthen osteogenic differentiation of adipose mesenchymal stem cells [67]. Study revealed that lncRNA TUG1 [68], lncRNA PCAT1 [69], and lncRNA HIF1A-AS2 also indirectly regulate the activity of osteogenesis-related signaling molecules by adsorbing miRNA [70]. Downregulated lncRNA MEG3 promotes osteogenic differentiation of human dental follicle stem cells by epigenetically regulating the Wnt pathway [71]. Researchers have indicated that exosome lncRNA-MALAT secreted by bone marrow mesenchymal stem cells enhanced osteoblast activity in osteoporotic mice [72]. Exosomal lncRNA-RUNX2-AS1 secreted by multiple myeloma cells can reduce osteogenic differentiation of MSC [73].

circRNA.
Increasing evidence indicates the various functions of circRNA in bone marrow mesenchymal stem cell osteogenesis, which proves to be a valuable checkpoint for the treatment of bone diseases [74]. CircRNA_436 might be a part of the critical regulators of periodontal ligament stem cell differentiation by coordinating with miR-107 and miR-335 to affect the Wnt/β-catenin pathway [75]. CircRNA_ 0127781 may serve as one of the essential regulators in the inhibition of osteoblast differentiation by interacting with miR-210 and miR-335 [76][77][78]. CircRNA_33287 would block miR-214-3p to intensify the osteogenesis process and active the construction of ectopic bone [79]. Both cir-cRNA_19142 and circRNA_5846 target miR-7067-5p to regulate osteoblast differentiation [80]. Lipopolysaccharide (LPS) is indicated to promote bone resorption by activating TLRs [81]. GO analysis showed that circRNA_3140 is related to the TLR signaling pathway [75]. The mechanism of exosomal circRNA in the regulation of osteogenic differentiation is seldom studied. However, the circular structure of cir-cRNA makes it more stable than linear RNA and difficult to be degraded [82]. It has great research potential in exosome engineering. Studies have shown that during the osteogenic differentiation of periodontal ligament stem cells, differentially expressed circRNAs are rich in membranebound vesicles [83].

Engineering Exosomes as Nanocarrier for
Noncoding RNA Delivery 5.1. Methods of Exosome Extraction. The separation and purification of exosomes have always been a concern of researchers. It is crucial to obtain high-purity exosomes for subsequent research. Researchers currently use ultracentrifugation [91], immunomagnetic beads [92], ultrafiltration [93], size-exclusion chromatography [94], or kits to achieve exosomes ( Figure 4). In Table 3, we summarize several methods for extracting exosomes and list their characteristics.

Merging
Therapeutic RNA into Exosomes. Exosome cargo with therapeutic activity is not restricted to naturally occurring cell-derived biomolecules. Instead, exosomes can also carry exogenous therapeutic molecules. Exosomes have been engineered to bind therapeutic molecules, including protein [95], small molecule drugs [96], peptide ligands [97], and therapeutic RNA [98,99]. Researchers have used multiple methods to engineer exosomes for cargo delivery, including incubation with saponin, electrical stimulation, sonication  miR-208a-3p ACVR1 Osteoblasts miR-208a-3p suppresses osteoblast differentiation and inhibits bone formation by targeting ACVR1. [90] MSCs: mesenchymal stem cells; BMSCs: bone marrow mesenchymal stem cells.  Journal of Nanomaterials [100], extrusion, freeze/thaw [101], click chemistry [102,103], and antibody binding [104] (Figure 4). We concentrate on methods and applications of fusing exogenous RNA into exosomes. One mean is through electroporation of exosomes. Researchers electroplated siRNA into dendritic cell-derived exosomes. Up to 60% GAPDH RNA and protein were knocked down in the mouse cortex, midbrain, and striatum by intravenous injection of electroplated exosomes [98]. Electroplating of siRNA against MAPK-1 into exosomes derived from peripheral blood monocytes reduced the expression of MAPK-1 in donor lymphocytes and monocytes [105]. In a conclusion, these results indicated that electroporation is wildly used in various kinds of exosomes and recipient cells. However, this method is inappropriate for all types of RNA cargo. For instance, researchers reported that miRNA cannot be electroplated into HEK 293-derived exosomes, indicating that some dimension or structure of RNA may be less suitable for this method [106].

Journal of Nanomaterials
Another strategy for fusing RNA into exosomes is overexpressing the cargo RNA in the exosome donor cells. The specificity of this method is poor. This kind of cargo RNA includes chemically modified 3 ′ benzopyridine miRNA [99], mRNA [107,108], and shRNA [109]. Researchers incubated exosomes carrying these RNAs with recipient cells. These mRNAs were translated into protein. Target genes were knocked down by shRNA and miRNA. This strategy is applicable to various RNA cargos and recipient cells.

Targeting Exosomes to Recipient Cells.
The safe application of exosome therapy requires the assessment of potential exosomes targeting cell and subcellular regions. Although the pharmacokinetics of exosomes used by IV has not been elaborated, the application of purified exosomes in mice can cause exosomes to accumulate in the kidney, liver, and spleen [98,106]. This distribution is similar to most nanoparticle delivery vehicles, usually through bile excretion, kidney clearance, or macrophage clearance [110,111]. Strategies for exosomes to target specific recipient cells have been reported.
One strategy is to utilize proteins and peptides from the virus with the ability of targeted delivery. For example, researchers used EBV glycoprotein 350 modifying exosomes to target CD19+ B cells, but not other peripheral blood mononuclear cells (PBMCs) [112,113]. Ligands of viruses can also enhance exosome-mediated transport to target cells in vivo. RVG-labeled exosomes transferred siRNA to the mouse brain, while unlabeled exosomes convey siRNA to the liver, spleen, and kidney [98]. Viral ligands can not only enhance the targeting ability of exosomes but also enhance the ability of exosomes to be integrated into the recipient cells. Exosomes labeled with VSV-G and OVA peptides are (1) A lot of materials need to be prepared in the early stage (2) The exosome yield is low (3) Repeated centrifugation causes damage to exosome vesicles and reduces its quality (4) Soluble proteins can form clumps with exosomes and cause pollution [78] Concentration gradient centrifugation In different liquid density intervals, each particle will eventually stay at the same position as its own density Compared with ultracentrifugation, the recovery rate and purity are higher Same as above [78,82] Ultrafiltration Using ultrafiltration membranes with different molecular weight (MWCO) for separation, small molecular substances will be filtered to the other side of the membrane Simple and efficient without affecting the biological activity of exosomes (1) The adhesion of the filter membrane may lose exosomes (2) The pressure and shearing force during filtration may cause of the damage of exosomes [80] Immunomagnetic bead By incubating with magnetic beads coated with antimarker antibodies, exosomes can be adsorbed and separated (1)  7 Journal of Nanomaterials absorbed by DC at a higher rate than exosomes labeled with OVA alone [114,115]. Although viruses can be used to improve the ability of exosome to target cells, the strategy is limited to the known interactions. The side effects of this strategy are still unknown.

Exosome-Based Strategies for Restoration of Bone Defect
Bone-derived exosome containing noncoding RNA is considered significant in regulating bone formation and absorption ( Figure 5). Various key factors, regulating osteoclasts and osteoblasts (such as RUNX2, BMP, and sclerostin), are adjusted by specific bone-derived exosomes embodying noncoding RNA [116]. Studies showed that intravenously, exosomes tend to their original place [117]. However, the same noncoding exosomes may have opposite effect on the differentiation and proliferation of osteoclasts and osteoblasts. Therefore, during the process of bone remodeling, bone-derived exosomes may not have functions completely coherent with the parental cells. The abundance of noncoding RNA in recipient cells and exosomes does not  Journal of Nanomaterials match, indicating that in addition to noncoding RNA, other components of exosomes also have regulatory effects [118].
Researchers have utilized exosomes from other tissues to enhance tissue healing efficacy [119] and to reduce joint injury and osteoarthritis by restoring matrix homeostasis and decreasing inflammation [120]. Efforts were being made to test the effects of engineered exosomes on orthodontic tooth movement models in mouse [121].  [122]. Studies have shown that even if exposed to an inflammatory environment, exosomes still have a longer circulation time [124]. A large number of studies have proved that due to their small size (≤100 nm), nanoparticles can achieve targeted aggregation of tissues through enhanced penetration and retention effects [125]. In addition, the circulation time of polyethylene glycol-modified exosomes can reach more than 60 minutes in vivo [126]. Polyethylene glycol-modified exosomes can significantly improve the biological stability of exosomes in vivo by prolonging the clearance time, making the study of exosomes as drug carriers more promising.

Targeting.
Multiple studies have confirmed the loyalty of exosomal cargo to parental cells. Exosomes derived from central nerve cells can pass through the blood-brain barrier and target specific neurons. Exosomes derived from hypoxic tumor cells tend to recruit into hypoxic tumor tissues [127,128]. The results of biodistribution studies also showed that the accumulation of exosomes in tumor tissues depended on the type of embryonic cells. Therefore, when studying exosomes targeting specific tissues or cells, it is necessary to consider the targeting efficiency [129].

Conclusions and Perspective
Exosomes are increasingly being studied in the field of bone reconstruction. From the initial research on tumor diseases and related mechanisms to studies on drug delivery and engineering exosomes, there is still a lot of research space.
(1) How to select cells with strong ability to secrete exosomes? The scaffold used with MSCs provides an ideal environment for the osteogenic differentiation of MSCs [130] [131]. However, there is no research on the inducing factors involved in this mechanism. Scaffold may enhance the paracrine function of MSCs, thereby secreting exosomes to induce the osteogenic differentiation. (2) The loading of exosomes on the scaffold to promote bone repair has been widely studied, but how to make the scaffold play a controlled or slow-release effect is another problem that needs to be solved. Maybe the solution is to control the pore size and degradation rate of the scaffold material. (3) Another direction is to improve their targeting ability by modifying the exosomal membrane molecules and preventing unwanted derivatives from entering the exosomes Further research is needed to evaluate the biological efficacy of exosomes in treating bone defects in vivo and vitro. As therapeutic delivery vectors, exosomes have engineering potential, and they are easy to design and well tolerated in vivo. Overcoming each of these miseries will turn the surprise discovery of exosomes as a drug delivery system into a viable mature technology.

Conflicts of Interest
All the authors declare no conflict of interest.

Authors' Contributions
Keda Liu and Nanjue Cao contributed equally to this work.