Introduction

Mesenchymal stem cells (MSCs) have been used widely in stem cell transplantation, tissue engineering and immunotherapy. They can migrate to sites of injury in animals, but the underlying mechanisms remain unclear. Chemokine receptors with their ligands and adhesion molecules play an important role in the tissue-specific homing of leukocytes and take part in transporting haematopoietic precursors into and through tissues. It is very likely that similar mechanisms govern the migration of MSCs.

This review covers the sources of MSCs and cytokines and external factors that affect MSC homing in bone fractures, as well as bone tissue engineering with MSCs. It hopes to suggest ways to make better use of MSCs in bone fractures.

Sources of MSCs for fracture repair

MSC niches, as defined by Scadden, enable homeostasis and maintain MSC populations [60]. Niches for MSCs of different origins are located adjacent to vessel walls, on the endosteal surfaces of trabecular bone, within the interfibrillary spaces, or perivascular [15]. MSCs have also been found in umbilical cord blood [44], dental tissues [56] and synovial fluid [35]. The bone marrow, endosteum and perivascular cells are perhaps more important in skeletal repair.

Bone marrow is still the most important source of stem cells [10]. For many years, bone marrow extracts have been used for treating non-union gaps in orthopaedic surgery. The periosteum is another source of MSCs [12, 50]. Bone marrow mesenchymal stem cells (BMSCs) were identified within the endosteal callus in areas of new bone formation in a mouse fracture model [25]. Nevertheless, harvesting MSCs from periosteum is difficult. In comparison, vessel walls are present throughout the body. Angiogenesis is important for bone fracture repair. Some studies have investigated the roles of mesenchymal pluripotent cells from vessel walls (pericytes) in bone healing [18, 69]. Pericytes can be induced to express chondrogenic and adipogenic markers when cultured under defined conditions. Muscle has been one of the most feasible sources of cells for bone repair [73], but harvesting is difficult and the cells may have limited capacity for differentiation. The role of circulating MSCs in bone healing remains controversial. Other sources of MSCs include adipose tissue, tendon, umbilical cord blood, dental tissues and skin [5, 37, 43, 63, 87].

The sources of MSCs are widespread (Fig. 1). The ideal stem cell source for fracture repair should be easily accessible, harvesting should be non-invasive and cells should be rapidly expandable by in vitro culture. Survival and integration within the host bone tissue should be qualified, and the cell source should show no tumourigenicity.

Fig. 1
figure 1

Sources of MSCs for bone repair

Full size image

MSC homing depends on the conditions

As a composite material, the formation and growth of bone tissue is a complex process, including molecular, cell and biochemical metabolic changes. Bone tissue has a dense structure; its stress and biological environment are relevant to its structure and function [58]. The immediate response to fracture is haematoma formation and inflammation. After a bone fracture, the formation of a haematoma prevents more bleeding and the loss of factors. The inflammatory response provides the initial stabilisation between the two bone ends, which initiate signalling cascades that aid in healing. After the haematoma has formed, precursor cells aggregate and form new blood vessels, fibroblasts and other supporting cells; granulation tissue forms between the fracture ends.

Transplanted MSCs were found to have specific systemic anti-inflammatory effects on cytokines released after tibia fracture [25]. They had no effect on interleukin 13 (IL-13) or IL-10 at any time, but had a significant effect on reducing IL-6 levels at day three and tumour necrosis factor α (TNF-α) and IL-1β levels one and three days after fracture. This process may limit tissue injury and prevent the development of fibrosis to promote rapid regeneration.

Progenitors are recruited to the fracture site in the initial stages of healing (day one) and proliferate at around day three [17]. At this time, the fracture environment is complex: vascular disruption during the initial stage of fracture creates a localised hypoxic environment [6] that acts as a useful regulatory stimulus for many cells, including MSCs and bone cells [22, 55]. New bone formation is thought to occur under low oxygen tension [6]. Hypoxic tissues express genes that increase cell survival under hypoxic conditions and re-establish the vasculature for oxygen delivery [64]. In addition, hypoxia induces the production of chemotactic factors implicated in cell migration, differentiation and new bone formation. The platelets, inflammatory cells and macrophages arriving at the site of injury secrete cytokines and growth factors, including IL-1 to IL-6, platelet-derived growth factor (PDGF), vascular endothelial growth factor (VEGF) and bone morphogenetic protein (BMP) [59]. This cellular response leads to the invasion of MSCs, which differentiate into osteoblasts and chondrocytes to complete the repair [61]. Numerous studies have confirmed that MSCs form bone by differentiating into osteoblasts [7, 27, 85]. In addition, many other studies have confirmed that the recruitment of factors with an adequate amount of MSCs and the micro-environment around the fracture are effective for fracture repair [71, 84, 88].

Process of fracture healing is affected by associated molecular mechanisms

During the fracture process, the recruitment of circulating progenitor cells to the site of injury occurs as a normal biological process [42, 47]. The mechanisms for MSC migration to the bone injury site are not completely known. Cytokines and chemokines may play vital roles in these processes, and many of these factors are chemo-attractants. MSCs express many chemokine receptors [29] and chemokine-mediated MSC migration occurs in vitro and in vivo.

Not all factors promote the migration of MSCs. Ode et al. [53] suggested that MSC migration is also under the control of CD73/CD29 in cells stimulated mechanically. Collectively, CD73 and CD29 mediate the mechanical stimulation to reduce MSC migration.

One of the best-investigated factors is stromal cell-derived factor 1 (SDF-1)/pre-B cell growth-stimulating factor/CXC ligand 12 (CXCL12), considered a master regulator of CXC receptor 4 (CXCR4)-positive stem and progenitor cells. When transfected at sites of ischaemic injury, this factor modulates cell differentiation into mature reparative cells [57, 76]. Toupadakis et al. [70] investigated the role of SDF-1/CXCR4 signalling in fracture healing by injecting AMD3100, a CXCR4 antagonist. Although CXCR4 is expressed in only a small part of the MSC surface [76], it still plays an important role in stem cell migration and is a key factor in bone marrow stromal cell migration [25, 38]. Therefore, improving the CXCR4 ligand in bone marrow stromal cells should promote their proliferation and migration [23].

SDF-1 and other cytokines associate and interact. Shinohara et al. [65] demonstrated that SDF-1 and monocyte chemo-attractant protein 3 (MCP-3) jointly regulate the homing of MSCs from the systemic circulation in fracture repair: the proportion of cells expressing SDF-1 and MCP-3 was significantly greater in transduced than non-transduced MSCs. Therefore, the homing of MSCs from the systemic circulation is involved in fracture repair via an SDF-1–MCP-3 pathway. The two factors act synergistically. Another study found that by promoting the expression of BMP-2, SDF-1 increased MSC migration and differentiation in promoting fracture healing [25]. SDF-1 promoted autocrine and other factors, such as VEGF and basic fibroblast growth factor (bFGF) paracrine factor. All of these cytokines inhibit cellular apoptosis [23]. The homing pathways may be modulated to enhance the contribution of recruitment and migration of MSCs in the systemic circulation at sites of skeletal injury.

Transforming growth factor β (TGF-β) exists in normal tissue cells and is the most abundant in transformed cells in animal bone tissue and platelets. TGF-β is released by platelets in the initial inflammatory phase of fracture healing and may be involved in the initiation of callus formation. TGF-β may also be expressed and secreted by chondrocytes, osteoblasts, macrophages and other inflammatory cells stored in the bone matrix. TGF-β is an effective chemotactic factor for bone marrow-derived MSCs. It promotes the proliferation of MSCs, preosteoblasts, chondrocytes and osteoblasts and induces collagen, osteopontin, osteonectin, proteoglycans, alkaline phosphatase and other extracellular proteins [48]. Self-designed bioscaffolds infused spatially with TGF-β3-adsorbed collagen hydrogel have been used to repair the defect in a rabbit model of a humeral head defect [8]. TGF-β3 recruited 130 % more cells to the regenerated articular cartilage than did spontaneous cell migration without TGF-β3, which promoted both avascular cartilage and vascularised bone formation. Since TGF-β may play a major role in cartilage formation and endochondral bone formation [13], it could promote BMP signalling to enhance osteogenesis and inhibit the activation of osteoclasts by promoting their apoptosis. Chondrogenic and osteogenic cells have many receptors, and TGF-β plays roles in each phase of fracture healing. Although TGF-β may promote cell proliferation, its osteoinductive effect is limited. Repeat or large-dose medication has a more significant induction effect, but is difficult to achieve in the clinic and may have side effects.

In the process of fracture healing, FGF is synthesised by monocytes, mesenchymal cells, macrophages, osteoblasts and chondrocytes. It can regulate cell migration [4], differentiation and proliferation and take part in angiogenesis and wound healing. Its effect is achieved by bonding to receptor tyrosine kinases. FGF plays an important role in vascular morphogenesis and mesenchymal cell mitosis in early fracture healing. The use of exogenous bFGF can accelerate bone repair and stimulate callus remodelling to restore the biomechanical properties of bone rapidly [11]. FGF can improve new bone volume and mineral content. Its effect is dose dependent. A single injection can promote the proliferation of chondrogenic precursors in the callus and form more cartilage, but cannot promote chondrocyte maturation or accelerate cartilage bone tissue replacement [72].

NEL-like molecule-1 (NELL-1) is a craniosynostosis-associated molecule that controls osteoblast differentiation [2]. NELL-1 has potent osteoinductive capacity for bone regeneration in several animal models [2, 46, 66, 81]. However, its capacity for promoting the repair of long bone defects remains unknown. Xue et al. [78] used adenoviral transfection and a variety of in vivo analyses to investigate the effect of NELL-1 on femoral distraction osteogenesis. Osteocalcin and osteopontin expression were increased with adenoviral NELL-1 transfection. A rat model of femoral distraction osteogenesis showed improved regeneration of good-quality bones and accelerated bone union at a high rate via sustained delivery of adenoviral transfection of NELL-1 protein into the local area of distraction. BMSCs with modified BMP-2 and NELL-1 genes promoted new bone formation and maturation in a rabbit maxillary sinus model and rapid distraction osteogenesis in a rabbit tibial defect model [77, 86]. The BMP-2 and NELL-1 genes may have a synergistic effect on the osteogenic differentiation of BMSCs.

Blood PDGF is strongly chemotactic for inflammatory cells and has a strong stimulatory effect on MSCs and osteoblast proliferation and migration [40]. From the early to middle stages of bone healing, PDGF promotes the effects of mesenchymal cells in cartilage and bone formation in development. The combined application of PDGF and BMP accelerates bone defect repair, but it is still uncertain whether PDGF can be used for the clinical treatment of fractures.

In addition, insulin-like growth factor plus MSCs can improve fracture healing, mostly through endochondral ossification [49], and parathyroid hormone probably plays a role in healing bone defects by increasing the proliferation of MSCs with reduced senescence and apoptosis [16].

Fracture healing is not an isolated process; various factors act synergistically to affect stem cell homing and promote differentiation, and a low dose can provide better therapeutic effects. Bai et al. [3] found that BMP-2, VEGF and bFGF dose- and time-dependently had strong synergistic effects on the osteogenic differentiation of MSCs, with lower concentrations of each factor sufficient to show the synergistic effect. More study is needed of the right time to use the appropriate cytokines at reduced dosages.

Since MSCs at the fracture site need mechanical stimulation for sustained therapeutic success in vivo, the mechanical micro-environment of MSCs needs to be considered during the regeneration process. Many studies have investigated the cellular characteristics and functional behaviour of MSCs in response to mechanical loading and suggest that mechanical embedding promotes bone regeneration when stem cells are transplanted into the fracture site.

External factors affecting MSC homing

In addition to the impact of the internal environmental on MSC homing, some external factors also have effects. One study introduced axial displacement stimulation after systemic stem cell transplantation to investigate how fracture healing affects the characteristics of the callus [74]. Experimental Sprague–Dawley rats with right femoral fractures were injected with MSCs in the tail vein. With mechanical stimulation, the mineral content increased and cartilage content decreased by ten days after the fracture. After 48 days, the mineral content peaked and the cartilage content decreased, and MSC populations were still found. Therefore, the timing of mechanical stimulation affected the physical and chemical properties of the callus and MSC migration to the fracture site.

Griffin et al. [26] showed that degenerate wave (DW) and capacitive coupling (CC) reduced cytotoxicity, while increasing human MSC invasion and proliferation in vivo. They used DW, CC, pulsed electromagnetic fields and direct current [69] to stimulate human bone marrow-derived MSCs. They analysed cell activity, including proliferation, cell kinetics, cytotoxicity and apoptosis. DW had the greatest proliferative and least apoptotic and cytotoxic effects, compared to the other waveforms. Stimulation affected cell intrusion and proliferation, especially with DW and CC. DW or CC applications improved the bone fracture healing rate, perhaps by increasing MSC recruitment to the fracture site.

Naturally occurring electric fields may guide cell migration to fracture sites during healing. Owing to the low oxygen tension with electrolysis and high pH of the extracellular micro-environment, the resulting electrical stimulation facilitates chondrocyte and osteoblast differentiation and matrix mineralisation. Electrical stimulation promotes the maturation of callus to accelerate fracture healing [83].

Low-level laser irradiation can increase the proliferative potential of MSCs [23]. The laser activates bone formation by promoting the proliferation and generation of chondrocytes and osteoblasts. Subsequently, local capillary formation and the deposition of calcium salts increase at the fracture site. Therefore, the use of diode lasers may be a valid approach for preconditioning MSCs in vitro before transplantation.

Low-intensity pulsed ultrasound (LIPUS) can enhance MSC-calcium phosphate composites. LIPUS increases the expression of some bone proteins, increases the bone mineral content and density, and improves the biomechanical properties of bone. By improving cartilage formation, LIPUS accelerates the process of endochondral bone formation [32].

Seed cells in tissue-engineered bone

MSCs lack immunogenicity in vitro. They do not express co-stimulatory molecules or major histocompatibility class II factors, such as CD40, CD80 and CD86 [34]. Most importantly, MSCs do not induce lymphocyte proliferation [39]. Transplanted allogeneic MSCs can still be detected in recipients some time after transplantation. This implies a lack of immune recognition and clearance. The interactions between allogeneic MSCs and immune cells and the mechanisms involved in the tolerance of MSC-mediated induction in vivo reduce the incidence of graft-versus-host disease and modulate inflammation [1]. However, memory CD4+ T cells and specific antibodies have been detected in hosts after infusion of allogeneic MSCs [80]. Allogeneic MSC differentiation also affects the immune characteristics of cells [31, 41]. MSCs may be immunosuppressive, but this does not confer sufficient immune privilege. Therefore, studies should examine the role of allogeneic MSCs in tissue engineering.

In vivo, only 0.001∼0.01 % of mononuclear cells from bone marrow are MSCs, so an efficient isolation method is required [14]. Density gradient centrifugation is the method most commonly used for harvesting, purifying and expanding MSCs quickly [52]. MSC treatments include direct injection, systemic infusion, genetically modified cells, composite scaffolds with MSCs and composite scaffolds with cytokines (Fig. 2). Most infused MSCs are trapped in the lungs; only a few migrate to the site of injury [21, 62]. Therefore, direct infusion may be better than systemic infusion of MSCs. However, Horwitz et al. [30] treated osteogenesis imperfecta successfully by infusing MSCs systemically to promote bone regeneration. Many experiments have confirmed that injection of MSCs at the site of an injury has therapeutic effects [28]. The genetic modification of MSCs involves transfection with viral or non-viral vectors. Both can express growth factors over the long term. However, the optimal growth factors and vectors need to be identified to ensure effective, safe and consistent treatment.

Fig. 2
figure 2

MSCs for treatment

Full size image

Scaffolds seeded with MSCs are most often used in tissue engineering and include biotic and abiotic materials [33, 51, 79, 82]. MSCs include differentiated and undifferentiated cells. Kawate et al. [36] used MSC/β-tricalcium phosphate composite granules to treat steroid-induced osteonecrosis of the femoral head. An anatomical bioscaffold composite of poly-epsilon-caprolactone and hydroxyapatite infused with TGF-β3 was fabricated to repair injured rabbit shoulder joints [8]. Electrospinning technology can be used to produce bone mimetics and scaffolds of polycaprolactone (PCL), collagen I and nanohydroxyapatite (PCL/col/HA), which adsorb and release PDGF and support good MSC adhesion and proliferation [40]. The activation of integrin-related signalling cascades is better than with scaffolds composed of PCL or collagen I alone.

Although the combination of growth factors and scaffolds remains a promising approach, the long-term release of growth factors for promoting the proliferation and maintenance of MSCs is limited. BMP-2-transfected MSCs promote bone formation in mouse hind limbs and in the bony union of critical-sized mouse radial defects [19]. Therefore, the genetic modification of MSCs so that they express growth factors, which involves transfecting MSCs using viral or non-viral vectors, is a suitable alternative [20]. Numerous osteoinductive growth factors have been used to modify MSCs and provide successful bone induction in vivo. However, studies have shown that human MSCs enhance some tumour formation and growth in vivo [68, 75]. The transfection approach and long-term cell culture can increase the tumourigenic potential of MSC therapy [45, 67]. Consequently, it is necessary to minimise the risk of malignant transformation with stricter control of cell handling procedures and cautious use of MSCs.

Clinical consequences

Promoting bone regeneration is a unique challenge for both clinicians and scientists. The mechanisms of MSC homing related to bone fracture repair, including recruitment, migration, differentiation and proliferation, are still not entirely clear. Much is known about the role of molecular mediators, cell populations and the overall cascade of events in the bone repair process. Studies have confirmed the role of MSCs in fracture healing: the extraction of bone marrow and in vitro amplification and replantation can accelerate healing. However, the specific roles of MSCs in fracture healing are still not fully clear. By modulating chemokine–chemokine receptor interactions, MSCs may increase their ability to correct inherited disorders of mesenchymal tissues or facilitate tissue repair in vivo. Much research needs to be done. The effect of stem cells is clear, but the focus is on how to use MSCs clinically, such as when and in what way, and how to use relevant cytokines and other favourable factors to better stimulate the role of MSCs. The “diamond concept” [24] for biological enhancement supports the implantation of MSCs, scaffolds and growth factors. The ideal biological environment is equally important [9]. However, MSCs also suppress some T-cell functions in transplanted hosts, which could facilitate tumour growth, so caution is needed [54]. Even so, we still need to find a simple, effective way to promote the homing of MSCs to facilitate fracture repair and treat bone defects and non-union.