Immunomodulatory Strategies for Bone Regeneration: A Review from the Perspective of Disease Types

Abstract

Tissue engineering strategies for treating bone loss to date have largely focused on targeting stem cells or vascularization. Immune cells, including macrophages and T cells, can also indirectly enhance bone healing via cytokine secretion to interact with other bone niche cells. Bone niche cues and local immune environment vary depending on anatomical location, size of defects and disease types. As such, it is critical to evaluate the role of the immune system in the context of specific bone niche and different disease types. This review focuses on immunomodulation research for bone applications using biomaterials and cell-based strategies, with a unique perspective from different disease types. We first reviewed applications for prolonging orthopaedic implant lifetime and enhancing fracture healing, two clinical challenges where immunomodulatory strategies were initially developed for orthopedic applications. We then reviewed recent research progress in harnessing immunomodulatory strategies for regenerating critical-sized, long bone or cranial bone defects, and treating osteolytic bone diseases. Remaining gaps in knowledge, future directions and opportunities were also discussed.

1. Introduction

Bone injury and loss represent a significant economic burden in healthcare worldwide. While small bone defect can heal on its own, bone healing is significantly compromised with increasing defect size, aging, and other pathological conditions. The gold standard for treating critical-sized bone defects uses bone grafts, which are limited by insufficient donor tissue supply and donor site morbidity. Tissue engineering offers an alternative strategy for bone repair using cells, biomaterials and biological signals. Most of the bone tissue engineering research to date have focused on targeting stem cell differentiation or enhancing vascularization [1, 2]. Recent studies have highlighted the important roles of immune cells in tissue regeneration. Immune cells can secrete various cytokines to directly impact inflammation, stem cell recruitment and differentiation. Engineering biomaterials to enhance tissue regeneration through immunomodulation is a relatively new field, with promising results in enhancing regeneration of soft tissues such as muscle and skin tissues [3–5]. In recent few years, a rapidly increasing number of publications are focusing on the topic of immunomodulation for bone regeneration. Unlike soft tissues, bone is highly mineralized and contains distinctive matrix compositions and cell populations. As such, it is important to study the role of the immune system in the context of bone niche specifically. Furthermore, bone niche cues and local immune environment vary depending on anatomical locations and disease types. This review focuses on immunomodulation research to date for bone applications using cell- and biomaterial-based strategies, with a unique perspective by categorizing studies based on different disease types. Specifically, we reviewed immunomodulatory strategies for prolonging orthopaedic implant lifetime, enhancing fracture healing, regenerating critical-sized long bone and craniofacial bone defects, and treating osteolytic bone diseases (Figure 1). Future directions were also discussed including needs to engineer biomaterials-based strategies for repairing critical size bone defects and developing technologies to elucidate the role of various immune cells in bone healing in a niche-specific manner.

Figure 1.

A schematic summary of different immunomodulation strategies for bone applications based on different disease types.

2. General concepts for immunomodulation and tissue regeneration

2.1. The relationship of immune system and tissue regeneration

Immune system is composed of innate and adaptive arms. The innate immune system includes neutrophils, macrophages (Mφ) and dendritic cells, whereas the adaptive immune system includes T and B lymphocytes (Figure 2a and 2b). Neutrophils are early responders to the injured tissue that arrived within a few hours, and secret cytokines to recruit Mφ within the following few days. Mφ are highly programmable and can polarize into different subtypes depending on the niche cues. Shortly after injury, M1 Mφ are the major population at the wound site and secrete pro-inflammatory cytokines, such as interleukin-1β (IL-1β), interferon gamma (IFN-γ), tumor necrosis factor-α (TNF-α), which helps clearance of damaged tissue and recruit stem cells. However, a prolonged inflammatory phase is detrimental and will hinder tissue regeneration. Later during regeneration, pro-inflammatory M1 are replaced by pro-regenerative M2 Mφ. The timely transition into M2 is essential for effective regeneration. The adaptive immune cells migrate to the wound site at later time points than innate immune cells and can further impact tissue regeneration through crosstalk with stem cells and other niche cells. A promising new strategy to enhance tissue regeneration is to target immune cells to reduce undesirable prolonged inflammation while promoting regenerative immune responses.

Figure 2. An overview of immunomodulation strategies for tissue regeneration.

(a) Innate and adaptive arms of the immune system. (b) Timeline of different types of immune cells infiltrated into injury site. (c) Summary of current immunomodulatory strategies.

2.2. Cell- and biomaterial-based strategies for immunomodulation.

Current immunomodulation strategies can be categorized into cell-based or biomaterials-based approaches (Figure 2c). Mesenchymal stem cells (MSCs) possess immunomodulatory functions, and have been used for treating autoimmune disorders such as graft versus host disease and immune disorders [6]. MSCs can sense and secrete cytokines in response to inflammatory cues, which subsequently impact immune cells such as Mφ and T cells by polarizing them toward a pro-regenerative and tolerogenic phenotype [7]. Transplantation of MSCs or their secretomes for the purpose of immunomodulation has shown benefit in treating various soft tissue losses [8, 9]. Other than MSCs, immune cells have also been used for tissue regeneration given their immunomodulatory function. For example, reparative Mφ activated by macrophage colony-stimulating factor (M-CSF) and IL-4 treatment resulted in evident improvement in cardiac recovery in a myocardial infarction model [10]. Hydrogel-medicated local delivery of immunosuppressive regulatory T cells with the peripheral nerve allografts has shown to suppress host immune response and promote nerve regeneration [11].

Biomaterials-mediated immunomodulatory strategies can be achieved by tuning physical or biochemical properties of the implant, delivery of immunomodulators or scavenging undesirable inflammatory signals (Figure 2c). Biochemical cues, such as surface chemistry, ligand density, charge, hydrophobicity modulate the immune system differently. Compared to synthetic materials, naturally derived extracellular matrix (ECM) materials has been shown to induce a more pro-regenerative immune response in a wound healing model [12]. Tuning chirality of crosslinking peptides in hydrogels from l- to d-amino acids led to activation of adaptive immune responses and facilitated follicle regeneration in wound healing [13]. The immune response may also be modulated by tuning biophysical properties of biomaterials such as stiffness, topological cues, porosity, and size. Increasing the size of alginate capsules used for islet cell delivery led to decreased foreign body response and better blood-glucose control in vivo [14]. Spatial confinements have been shown to reduce the inflammatory response of Mφ through nuclear translocation of MRTF-A [15]. To elicit a pro-regenerative immune response, a variety of immunomodulators have been delivered by biomaterials, including anti-inflammatory cytokines, chemokines, neutralizers, nuclei acids, non-steroidal anti-inflammatory drugs (NSAIDs) and extracellular vesicles (EVs) [16, 17]. To remove undesirable pro-inflammatory molecules (cell free DNA or inflammatory cytokines) from the injured area, biomaterials may also serve as a scavenger using binding ability of cell membrane coated nanoparticles [18] and ECM-derived scaffolds[19], or static absorption by introducing positive charged polymers [20, 21].

3. Crosstalk between immune system and bone niche

Normal bone healing is associated with spatiotemporal changes of niche cell populations and cytokines and is driven by the crosstalk between immune cells and stem cells (Figure 3). Shortly after bone injury, acute inflammation is initiated by neutrophils within hours, followed by recruitment of M1 Mφ. Initial inflammation is indispensable for bone regeneration as M1 Mφ secrete cytokines that are critical for early stem/progenitor cell recruitment [22], osteogenesis [23, 24] and angiogenesis [25]. Depletion of Mφ has been reported to be detrimental for bone fracture healing [26]. However, prolonged chronic inflammation can induce MSC apoptosis [27, 28], which can be further amplified by activation of pro-inflammatory T cell subtypes such as T helper 1 (Th1), Th17 and CD8+ T cells [29–31] [32]. Resolving acute inflammation in a timely manner is critical for normal bone healing and angiogenesis, which is characterized by phenotypic change of Mφ from M1 to M2 [33, 34]. M2 Mφ secret cytokines to suppress the initial inflammation, while supporting neovascularization and facilitating ossification [35, 36]. Emerging studies showed the adaptive immune system may also help with resolving the acute inflammation by the activation of Th2 and regulatory T cells (Tregs) [37, 38]. Endogenous MSCs may also contribute by secreting cytokines to facilitate Mφ polarization towards M2 [39, 40] and suppress effector T cell proliferation [41]. Bone remodeling happens during the later stage of bone healing and is mediated by osteoclasts [42], a multinucleated cell type responsible for breaking down bone. Cytokines secreted by osteoblasts are critical regulators for osteoclast activity. Receptor activator of nuclear factor-kappa B ligand (RANKL) is a factor produced by mature osteoblasts that can activate osteoclast bone resorption, while osteoprotegrin (OPG) is a secreted decoy receptor that is antagonist for osteoclast differentiation [43]. Excessive osteoclastogenesis caused by chronic inflammation will lead to osteolytic bone loss.

Figure 3. The crosstalk between immune system and bone niche during impaired and normal bone healing.

Impaired healing is characterized by chronic inflammation, driven by M1 Mφ and activated pro-inflammatory T cells, which secret cytokines that lead to apoptosis of stem cells and activation of osteoclasts. Normal bone healing is characterized by timely resolution of acute inflammation with desirable crosstalk between the immune system and bone niche cells that facilitated the healing process of vascularization, new bone deposition, and remodeling. Abbreviations: stromal cell-derived factor 1 (SDF-1), oncostatin M (OSM), bone morphogenetic proteins-2 (BMP-2) platelet-derived growth factor (PDGF), vascular endothelial growth factor (VEGF), transforming growth factor-β (TGF-β), fibroblast growth factor-2 (FGF-2).

4. Orthopaedic implants

Orthopaedic implants have been widely used for joint replacement or bone fracture surgery in clinic for the past few decades [44]. These implants are largely based on metal alloys. Upon implantation, surface protein absorption and wear particle generation are unavoidable [45]. Surface adhesion of non-native proteins causes Mφ recruitment and M1 Mφ activation, which is further worsened by wear particle generation. Such highly inflammatory environment leads to encapsulation by fibrotic tissue and osteoclast activation, causing implant loosening and failure. To this end, recent advances have taken immunomodulatory strategies to combat implant interface reaction-induced and wear particle-induced chronic inflammation (Figure 4).

Figure 4. Process of implant interface reaction, wear particle-induced orthopaedic implant failure, and intervention strategies.

Implant interface reaction causes Mφ recruitment and M1 Mφ activation, which is further worsened by excessive wear particles production. The highly inflammatory environment leads to severe foreign body responses and osteoclast activation, thus causing implant loosening and failure. Immunomodulatory strategies to overcome this undesirable inflammatory response are shown above in each step.

4.1. Implant interface reaction-induced chronic inflammation

Biomaterial-based immunomodulatory strategies have been used to mitigate implant interface reaction and induce better bone integration, including adapting biochemical properties, physical properties, and delivery of immunosuppressors (Supplementary Table 1).

Surface biochemical modifications to implant surface has been widely used for enhancing osteoblast activity previously [46]. Recent studies showed surface biochemical modification can also be used to induce desirable immune response at implant interface. Metallic and plastic implants such as titanium (Ti) and polyether ether ketone (PEEK) are often used as weight-bearing implants due to their strong mechanical properties. Surfacing coating of bioactive ions or ECM components improved surface reaction by inhibiting Mφ and osteoclast activation against Ti and PEEK implants. Incorporating Mg²⁺ to Ti implants [47] or Zn²⁺ to PEEK implants [48] has been shown to decrease the inflammatory cytokine release, while increasing the secretion of anti-inflammatory and osteogenic cytokines compared to unmodified implants. Immobilizing ECM-derived heparin to Ti implants decreased proinflammatory cytokine secretion by Mφ [49]. Compared to weight-bearing implants, calcium phosphate (CaP)-based ceramics, such as hydroxyapatite (HA) and beta-tricalcium phosphate (TCP), are commonly used as non-load bearing implants. CaP ceramics mimic natural inorganic component in bone tissue and have been widely used in implants and bone tissue engineering due to their strong osteoconductive properties. HA and TCP have different chemical compositions and resorption rates. Previous studies mostly focus on their effects on stem cell osteogenesis, but neglected immune responses [50]. Recently, it has been shown tuning chemical composition of CaP elicited varying inflammatory response, with TCP reduced Mφ inflammatory response compared to calcium deficient HA [51]. Further incorporation of Mg²⁺ to TCP implants decreased osteoclast activity [52]. Heparinization of TCP decreased neutrophil and monocytes-mediated acute inflammation [53]. Similarly, bone ECM coating to dicalcium phosphate decreased inflammation and achieved better integration in vivo [54].

Physical properties can be harnessed to modulate the immune response by polarizing Mφ towards M2 and inhibit osteoclast activation [55]. Topological cues such as hierarchical topography [56] and surface nano-particle patterning [57] have been used to modulate implant interface reaction, facilitating M1 to M2 phenotypic switch on metallic implants. Other physical properties such as nanoscale porosity [58], and high wettability [59] have been shown to decrease M1 Mφ activation and osteoclast activity.

Delivery of immunomodulators has been commonly used to decrease implant surface inflammation. The synergistic release of IL-4 and RGD on chitosan layer-by-layer hydrogel on titanium nanotubes worked cooperatively to drive M2 phenotypic change and promote MSC differentiation [60]. Delivery of sodium butyrate from PEEK scaffolds polarized Mφ to M2 phenotype and subsequently stimulated anti-inflammatory cytokine secretion, and enhanced osteointegration around the implants [61].

4.2. Wear Particle-induced chronic inflammation

Poor integration and implant loosening can cause excessive production of wear particles, leading to chronic inflammation and premature implant failure. Current efforts for inhibiting wear particle-induced inflammation mainly focus on bolus delivery of drugs to decrease Mφ infiltration, inhibit nuclear factor kappa B (NFκB) activation, and decrease inflammatory signals in implant niche (Supplementary table 2). Clodronate liposomes [62] that induce Mφ apoptosis or mutant chemokines serving as decoy drug to block the receptors [63] [64] have shown efficacy in decreasing Mφ recruitment to the implants and attenuated bone loss. Wear particles activate M1 Mφ and osteoclasts through NFκB pathway, making it a target for therapeutic against wear particle-induced chronic inflammation. Drugs that target NFκB [65] [66] [67] [68] and synthetic NFκB decoy oligodeoxynucleotides (ODNs) [69] significantly inhibit NFκB activation in response to wear particles, mitigating inflammatory response and osteoclast activation. Neutralizer and immunosuppressive drugs have also been used to alleviate inflammatory signals around implant niche. Neutralizing antibody against TNF-α significantly reduced bone resorption caused by titanium particles [70]. Immunomodulatory molecules such as IL-27 [71] and platelet rich plasma (PRP) [72] have also been shown to decrease inflammatory cytokine level and attenuate osteoclast activation.

4.3. Future Directions

Most research to date using drugs to reduce wear particle-induced inflammation rely on bolus drug delivery. However, this is associated with limitations including burst release and poor control of local drug concentration. To reverse chronic inflammation caused by wear particles, future studies should expand on harnessing biomaterials-mediated controlled delivery of anti-inflammatory drugs to inhibit wear particle-induced bone loss. For example, drug-eluting implants can be harnessed for long term release of anti-inflammatory drugs [73].

5. Non-critical size bone fracture healing

Bone can heal following non-critical size fracture, and immune cells play important roles in the healing process [74]. The role and kinetics of Mφ and T cells in non-critical size bone fracture healing have been well studied before, and immunomodulatory strategies targeting Mφ and T cells have been used to accelerate bone fracture healing (Figure 5a).

Figure 5. Enhancing non-critical size bone fracture healing via in situ delivery of immunomodulators that target Mφ or T cells in vivo.

(a) A schematic summarizing how different subtypes of Mφ and T cells impact bone healing outcome. (b) Promoting M2 Mφ polarization using in situ delivery of cytokines enhanced bone fracture healing in vivo. (c) Patients with impaired bone fracture healing is characterized by higher number of CD8+ T cells and lower Tregs in peripheral blood. (d) In situ delivery of Iloprost reduced CD8+ and CD4+ T cells in the fracture and accelerated bone fracture healing. (Data reproduced from [29, 38, 76, 84].)

5.1. Immunomodulation on Mφ

Acute inflammation is essential for normal fracture healing. Preventing Mφ infiltration by apoptotic drugs delayed hard callus formation and ossification [75] [76]. In contrast, increasing M1 Mφ recruitment and polarization by colony-stimulating factor-1 or TNF-α enhanced fracture healing [75] [77]. Inflammation resolving-mediated by M2 Mφ are indispensable for ossification and bone healing. Impaired M2 Mφ function was observed in the aged animal model and was responsible for diminished vascularization and bone formation in fracture healing [78]. Delivery of cytokines IL-4/IL-13 that enhance M2 Mφ polarization have improved bone formation in fracture healing [76] (Figure 5b). Interestingly, inhibiting Mφ recruitment decreased fracture healing for young mice [75] [76], but turned out to improve healing in aged mice accompanied with better inflammation resolving [79]. This highlights the importance of disease-specific immunomodulation for enhanced bone healing.

5.2. Immunomodulation on T cells

The role of adaptive immune cells in fracture healing is less studied compared to Mφ. T cells and B cells appeared at the injury site in a two-wave fashion at different time points and had close crosstalk with bone niche cells after injury [80]. Further studies found the overall undesirable role of T and B lymphocytes, given that the absence of lymphocytes in immunocompromised animal exhibited faster bone healing, compared to wild type animals [81] [82]. However, the new formed bone in lymphocytes deficient animals is much stiffer when compared to wild-type regenerated bone [82], highlighting the potential role of T cells in collagen deposition to form quality bone. T cells have versatile subtypes and their roles in fracture healing have been studied respectively. CD8+ T cells are detrimental for fracture healing. A higher percentage of CD8+ cytotoxic T cell is present in bone hematoma and bone marrow in animals with delayed fracture healing [83]. Further depletion of CD8+ T cells showed improved fracture healing [29]. In contrast, anti-inflammatory Tregs serve as helpers in fracture healing, as adoptive transfer of Tregs improved fracture healing [37]. In addition to the animal models, the role of CD8+ T cells and Tregs has also been explored in clinic through comparing patients with impaired and normal fracture healing. Elevated level of effector CD8+ T cells [29] and decreased Treg population [38] were observed in the peripheral blood from patients with impaired healing (Figure 5c). Further in vitro study validated CD8+ T cells as major contributors of IFN-γ and TNF-α secretion that inhibit MSC osteogenesis [29]. The ratio of effector CD8+ T cells and Tregs, was also suggested as a potential diagnostic marker for non-union fractures. Strategies that allow modulating the ratio of different immune cells hold great promises as novel therapeutic for treating delayed fracture. As an example, delivery of Iloprost through fibrin scaffold decreased CD8+ cytotoxic T cells, reduced pro-inflammatory cytokine secretion, and improved bone healing in vivo [84] (Figure 5d).

5.3. Future Directions

As clinical studies have shown non-union and delayed fracture healing is highly correlated with imbalanced immune cell population, future studies should expand on developing strategies that target Mφ and T cell immunomodulation. Current studies mainly focuse on directly promoting stem cell activity to treat non-union and delayed fracture healing, and there is an urgent need for developing spaciotemporal immunomodulatory therapeutic interventions. With better understanding of the roles and kinetics of different immune cell populations during normal versus impaired fracture healing, biomaterial-based drug delivery system that targets immunomodulation can be developed to achieve faster healing. The use of cell-based immunomodulation for fracture healing is limited and may offer another opportunity for future studies. Finally, knowledge gained from elucidating the role of immune cells in non-critical size fracture healing may also shed light on developing better strategies for treating critical-sized bone defects.

6. Critical-sized bone defect

While non-critical size bone fracture can heal on its own, bone defects that exceed the critical size threshold cannot heal without further aid. Biomaterials have been used for repairing critical size bone defects focusing on osteogenesis and vascularization[1, 2]. However, the potential of harnessing biomaterials to enhance bone regeneration by targeting immune cells remain largely untapped. Furthermore, emerging studies in this area generally focus on only innate immune cells such as Mφ, and the role of adaptive immune cells in repairing critical size bone defects remains largely unknown. Based on the anatomic location, critical-sized bone defect can be categorized into long bone defect and craniofacial defect. These two defects differ in many ways including immune environment, stem cell sources and developmental origin (Table 1). Here we reviewed emerging studies on using immunomodulatory strategies for repairing long bone and craniofacial bone defects separately.

Table 1.

Comparison of long bone and cranial bone defect.

Comparison	Long bone	Cranial bone
Ossification mechanism	Endochondral	Intramembranous
Direct exposure to bone marrow	Yes	No
Stem cell source	Periosteum, muscle, bone marrow	Dura mater
Hematoma content	High	Low
Developmental origin	Lateral plate mesoderm	Cephalic mesoderm and neural crest
Structure	Cortical bone	Cancellous bone

6.1. Long bone defect

6.1.1. Cell-based immunomodulation

MSCs have shown therapeutic effect on critical-sized long bone defect regeneration, with most studies focusing on promoting MSC differentiation towards bone lineage [85, 86]. However, few studies have investigated MSC’s contribution from immunomodulatory perspective. The short-term retention of MSCs after transplantation indicates they are more of supporters instead of effectors for bone regeneration [87]. Transplantation of scaffold loaded MSCs in the long bone defect increased early recruitment of M1 Mφ and endothelial cells [87]. MSCs can also promote Mφ phenotypic change from M1 to M2 to reduce undesirable inflammation, which further synergize with an osteogenesis-inducing drug to enhance long bone defect healing [88]. Allogenic MSCs can also be genetically modified to overexpress Herpesvirus-entry mediator, thereby suppressing adaptive immune response and promoting long bone healing [89]. Other than using MSCs, direct delivery of immune cells may also be beneficial. Delivery of whole bone marrow cells by β-TCP scaffold has been shown to accelerate long bone defect healing [90]. Futuer depletion of monocytes fraction (CD14+) remarkably decreased bone regeneration, suggesting vital role of monocytes for observed improvements in bone healing.

6.1.2. Chemical cues of biomaterials

In addition to cells, acellular biomaterials have also been used for immunomodulation and long bone defect healing. Tuning biochemical cues introduced by material composition can elicit beneficial immune response for enhanced long bone healing. Fibrinogen scaffolds increased immune cell infiltration in acute inflammatory phase, while decreasing chronic inflammation and enhancing bone regeneration [91]. Fibrinogen can interact with monocytes through toll-like receptor-4 (TLR-4), leading to subsequent production of BMP-2 by monocytes [92]. Depletion of TLR-4 has been shown to hinder bone regeneration, suggesting TLR-4 plays an indispensable role in long bone defect healing [93]. However, depleting TLR-4 in turn accelerated cranial defect healing [94], highlighting the importance of assessing immunomodulatory strategies in a disease-specific manner.

6.1.3. Physical cues of biomaterials

Scaffolds that contain osteoconductive cues, such as CaP ceramics, have been widely used to enhance bone repair. CaP ceramic has been shown to induce undesirable inflammatory responses in both animal models and clinical studies [97, 98]. As such, it is critical to evaluate the immune response towards CaP-based scaffolds and further optimize the physical and chemical cues to minimize undesirable immune responses while leveraging on their osteoconductive benefits. Emerging studies suggest their immunomodulatory response depends on topological cues, particle size, chemical compositions, and dosage of the osteoconductive materials. Ways of incorporating minerals impacted topological cues of mineralized scaffolds, leading to differential immune responses [95]. Same mineral composition coated through co-electrospinning elicit reduced inflammatory response and better bone regeneration, compared to electrodeposition. Particle size of HA also matters, with nano-sized HA particles showing decreased pro-inflammatory immune cell population and cytokine level compared to micro-sized HA particles [96]. Nano-sized HA particles also increased anti-inflammatory cytokine level with accelerated long bone regeneration (Figure 6). These findings offer insights on how to optimize the composition and physical cues of CaP based minerals cues to reduce undesirable immune responses.

Figure 6. Optimizing HA particle size accelerates critical-sized long bone defect healing by promoting M2 Mφ polarization.

(a) Schematic of a rat long bone defect model. (b) Morphology of HA microparticles (MP) and HA nanoparticles (NP). (c) Scaffolds containing MP promoted proinflammatory M1 phenotype (CCR2, CD86), while NP promoted M2 phenotype (CXCR1). (d) Only MP, but not NP, increased recruitment of CD3+ T cells and neutrophils. (e) MicroCT imaging showed NP enhanced bone regeneration at week 4. (Data reproduced from [96].)

6.1.4. Immunomodulator delivery

Biomaterial-based local delivery of immunosuppressors or immunoactive molecules has been harnessed to accelerate long bone defect healing. To facilitate resolution of acute inflammation, hydrogel-mediated delivery of immunosuppressive drugs that target Mφ [99] or lymphocytes [100] accelerated long bone defect healing in vivo. OPG released from collagen sponge in situ led to reduced osteoclast cell number and exhibited synergistic effect with BMP to further enhance bone healing [101]. In contrast to using immunosuppressor to suppress undesirable chronic inflammation, immunoactive molecules that enhance acute inflammation have also been used with beneficial effects on accelerating healing of long bone defects in vivo. Incorporating immunoactive molecules that can increase early recruitment of Mφ and leukocytes benefits long bone defect healing [102]. The crosstalk between infiltrated immune cells and bone niche cells further enhanced recruitment of progenitor cells who are responsible for bone regeneration [103]. Above studies indicate that beneficial immunomodulation is not simply suppressing inflammatory immune response, instead, enhancing acute inflammation without causing chronic inflammation may also benefit bone regeneration through debris clearance and progenitor cell recruitment.

6.2. Craniofacial bone defect

6.2.1. Cell-based immunomodulation

Like long bone defect, MSCs transplanted in cranial defect have been shown to exhibit short retention time [104]. One study showed that transplanting MSCs decreased Mφ and eosinophils cell infiltration while enhancing cranial bone formation [104], while other studies showed incorporating MSCs had minimal beneficial effects [27, 28]. It is speculated that the harsh local inflammatory environment after injury resulted in low MSC survival and compromised therapeutic secretome release. Immunomodulatory strategies that can desensitize MSC’s response toward inflammatory signals, suppress inflammation or encapsulate MSCs in hydrogels have shown benefits in enhancing MSC-based bone regeneration. Codelivery of MSC with inhibitors of IL-1R1/MyD88 signaling rescued inflammation-induced MSC apoptosis and enhanced MSC-based cranial bone regeneration [105]. Intravenous injection of Tregs or scaffold-mediated local delivery of aspirin facilitated anti-inflammatory responses, which also rescued MSC apoptosis and improved MSC-based cranial bone regeneration [27, 106]. Encapsulating MSCs in hydrogels also induced Mφ M2 polarization and better cranial bone regeneration [107]. While intravenous Treg delivery has shown benefit in accelerating MSC-based cranial bone regeneration [27], local delivery of Mφ showed minimal effects [108]. This result is different from long bone defect, in which monocyte delivery has been shown to accelerate bone regeneration [90]. These studies also highlight the importance of studying immunomodulatory strategies for bone regeneration in a disease and location-specific manner.

6.2.2. Chemical cues of biomaterials

Tuning the biochemical cues of scaffolds can also enhance cranial bone regeneration via inducing a pro-regenerative immunomodulatory response. ECM hydrogel scaffolds from varying tissue types were shown to accelerate Mφ transition from M1 to M2 in vivo through activating JAK2/STAT3 pathway, enhancing neo-vasculature and bone formation in the cranial defect [109] [110]. Notably, depletion of Mφ hindered ECM-mediated bone regeneration, indicating the imperative role of Mφ in bone regeneration [110]. Coating HA scaffolds with cell-secreted ECM increased Mφ recruitment and enhanced cranial bone repair [111]. Incorporation of bioactive element, such as Au/Eu, increased both proinflammatory cytokines and bone trophic growth factors from Mφ [112, 113], accompanied by enhanced cranial bone regeneration.

6.2.3. Physical cues of biomaterials

Topographic cues of biomaterials can be designed to modulate immune response and enhance bone regeneration. Varying pore size/shape of mesoporous silica rods (MSR) showed differential effects on Mφ immunomodulation, where large cone-shaped pores reduced Mφ pro-inflammatory cytokine release compared to small narrow pores [114]. MSR with larger pores also showed better synergistic effects with BMP-2 delivery to improve cranial bone regeneration in vivo [114]. Hierarchical topographic modification of scaffolds has shown improved immunomodulatory function. 3D printed synthetic scaffolds with hierarchical micro-channels decreased neutrophil and M1 Mφ activity, while increasing M2 Mφ and MSCs population [115]. Micro-channel-printed scaffolds also increased vessel number and improved bone formation in a rat cranial defect model, compared to the scaffold without micro-channels [115]. Surface coating of HA with staggered nano-topography, but not even coating, enhanced M2 Mφ polarization, MSC recruitment and mandibular bone regeneration [116]. This again highlights the importance of optimizing mineral cues to maximize bone regenerative outcome. Interestingly, neutralization of IL-4 or depletion of Mφ significantly impaired the bone regeneration induced by staggered HA-coated scaffolds, indicating the indispensable role of IL-4 and Mφ in topographic cue-induced bone regeneration [116].

6.2.4. Immunomodulator delivery

Delivery of immunomodulators using biomaterials has also shown benefit in promoting bone regeneration in both cranial and mandibular defect models. Immunosuppressive drugs have been delivered to downregulate inflammatory activity by M1 Mφ [117] [118], decrease osteoclast differentiation [119], while increasing M2 macrophage population and blood vessel ingrowth[120]. IL-4 is a strong anti-inflammatory cytokine, and IL-4 reduced M1 Mφ activity while increasing M2 Mφ polarization in a dose-dependent manner [121]. Interestingly, the most robust bone formation and vascularization was observed with intermediate dosage of IL-4 rather than the highest dosage (Figure 7b and 7c). These results emphasize optimal bone regeneration requires a delicate balance of suppressing chronic inflammation while preserving the acute inflammation, and the importance of optimizing the dosage and duration of immunomodulator delivery.

Figure 7. Enhancing cranial bone regeneration via IL-4 delivery to facilitate M2 Mφ polarization.

(a) Schematic of a critical-sized cranial bone defect. (b) The effect on M2 polarization depends on IL-4 dosage. (c) The effect of IL-4 dosing on vascularization and mineralized bone formation in vivo. (Data reproduced from [121].)

6.3. Future directions

Compared to the bone fracture model, the role of immune cells in critical-sized bone healing remains understudied, especially in the context of using biomaterial-based strategies. For critical size defect, scaffolds are routinely used as void fillers or for promoting stem cell differentiation. Designing scaffolds to enhance critical size bone regeneration through immunomodulation represents a promising future direction, and may be used synergistically with cell-based therapies. Furthermore, similar immunomodulatory strategies can yield different results in the long bone [90, 93] and cranial bone defect model [94] [108]. These studies further highlight the importance of studying immunomodulatory strategies for critical size bone regeneration in an anatomical location-specific manner.

7. Osteolytic bone loss

Osteolytic bone loss happens when the balance between bone formation and bone resorption are broken, which can be triggered by aging and various pathological conditions. Osteolytic bone loss is accompanied by chronic inflammation, and immunomodulation holds great potential to combat osteolytic loss by restoring the balance between osteogenesis and osteoclastogenesis. Here we reviewed immunomodulatory strategies for three most common osteolytic bone diseases, namely rheumatoid arthritis, osteoporosis and periodontitis.

7.1. Rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic inflammatory disease that affects over 24 million people worldwide. RA is characterized by imbalanced M1/M2 Mφ population [122] and over activation of adaptive inflammatory cells such as Th17 and Th1 [123]. Such imbalance results in excessive inflammation at the joint, leading to articular cartilage loss as well as bone erosion [124]. Clinically, most drugs for RA treatment aims to alleviate inflammation, including synthetic drugs, antibody-based therapy, and small molecules [125]. However, all these drugs require frequent and high dosing, leading to undesirable side effects such as liver burden and increased risk of infection [125]. To overcome these issues, various immunomodulation strategies have been developed including the use of live cells, cell-derived products, biomaterial-based drug delivery and inflammation scavenging biomaterials.

7.1.1. Cell-based immunomodulation

MSCs have been used to treat RA given its known immunomodulatory functions. MSC-based therapy has shown immunomodulatory functions on T cells [126] and osteoclasts [127], alleviating cartilage and bone loss in preclinical and clinical studies [128, 129]. In addition to using live cells, cell-derived products such as extracellular vesicles (EVs) have also been used to resolve inflammation in RA. Coating nanoparticles with Mφ-derived EVs has been shown to enhance the drug targeting efficiency to the diseased joint [130]. Neutrophil-derived EVs in absence of drugs decreased M1 Mφ activity and reduced inflammation-induced activation of synoviocytes [131].

7.1.2. Immunomodulator delivery

Biomaterial-based drug delivery with targeting function have also been used for treating RA. Various drug delivery systems have been used including liposomes [132], micelles [133], dendrimers [134], polysaccharide nanoparticles [135], and nanodiamond [136]. Compared to bolus drug delivery, biomaterials-mediated drug delivery allowed better targeted immunomodulatory response using lower dosage, alleviating inflammation and tissue erosion in vivo. In addition to target delivery at inflamed joint, on-demand drug release in response to inflammatory level represents another delivery strategy. A hydrogel material was designed to release model drug in response to RA-induced increasing metalloproteinases (MMP) level, which showed better treatment efficacy than local injected free drug [137].

7.1.3. Scavenging biomaterials

Harnessing biomaterials as a scavenger to remove accumulated inflammatory signals has emerged as a method to treat RA. Inflammatory cytokines secreted by immune cells and cell-free DNA (cfDNA) from destructed articular tissues is known to deteriorate RA condition. Intra-articular injection of neutrophil membrane-coated nanoparticles led to efficient removal of accumulated TNF-α and IL-1β through ligand-receptor binding. Scavenging these proinflammatory signals has shown efficacy in suppressing synovial inflammation in mouse RA disease models in vivo [138]. Cationic nanoparticles can efficiently bind to negative charged cfDNA through static interactions. Intravenous injection of these cationic particles remarkably reduced cfDNA level in both serum and joint, which further alleviated inflammatory cytokine release and bone destruction in collagen-induced arthritis rat models [139, 140] (Figure 8).

Figure 8. Cationic nanoparticles (cNP) rescued bone loss in rheumatoid arthritis model through scavenging cell- free DNA (cfDNA).

(a) A schematic summary of the experimental design. (b) Intravenous delivery of cNP significantly decreased concentration of cfDNA and TNF-α in vivo, which rescued bone loss (c). (Data reproduced from [139].)

7.2. Osteoporosis

Osteoporosis (OP) is a condition in which new bone formation does not keep up with bone resorption and is prevalent in elderly population and menopausal women. While the mechanism of OP pathogenesis is not fully understood, it has been shown to be related to altered immune cell activities, such as increased Th1 and Th17 population [141, 142]. Two major challenges exist in current treatments for OP patients. First, most clinically used anti-osteolytic drugs have severe side effects such as suppression of bone turnover and jaw necrosis. Second, OP patients have a high risk of fracture due to the fragility of bone, which is hard to regenerate due to excessive osteoclastogenesis. To address these challenges, cell and biomaterial-based immunomodulation strategies have been developed to treat OP.

7.2.1. Cell-based immunomodulation

Cells and cell-derived products have been used to mitigate inflammation-induced osteoclastogenesis and restore osteogenesis in OP. Transplantation of stem cells from human exfoliated deciduous teeth through intravenous injection has shown efficacy in reducing Th1 and Th17 population, while increasing Treg population. Such immunomodulation further rescued bone marrow MSC proliferation and osteogenesis, and inhibited osteoclastogenesis in a rat OP model [143]. Genetically modified MSCs to overexpress OPG reduced osteoclastogenesis in rats with OP than unmodified MSCs [144]. The MSC immunomodulatory function can be further enhanced using biomaterials. Amino functionalized mesoporous bioactive glass scaffolds significantly promoted the anti-inflammatory cytokines secretion by MSCs, decreased osteoclast activity and promote bone regeneration in rabbit [145]. Cell-derived products may also elicit desirable immunomodulation to attenuate OP conditions. Endothelial cell-secreted exosomes (EC-Exos) showed better bone targeting than osteoblast or MSC-derived exosomes after i.v. injection, due to over-expressed pregnancy zone protein on EC-Exos [146]. EC-Exos alone without drug also showed efficacy in inhibiting osteoclast activity and reduce osteoporosis in vivo, due to immunomodulatory miR-155 in EC-Exos (Figure 9).

Figure 9. Endothelial cell-secreted exosomes (EC-Exos) localized to bone niche and reversed bone loss in an osteoporosis disease model.

(a) A schematic summary of the effect of EC-Exos on bone niche cells. (b) EC-Exos inhibited osteoclast differentiation in vitro. (c) EC-Exos localized to bone niche after i.v. injection, which was not observed using exosomes from MSCs or osteoblasts (MC3T3). (d) EC-Exos reversed bone loss in an osteoporosis model. (Data reproduced from [146])

7.2.2. Immunomodulator delivery

Nanomaterial-based drug delivery has been harnessed to deliver anti-osteolytic drugs to reduce bone loss in OP animal models and clinic trials. These studies have been well-summarized in a recent review [147]. In addition to anti-osteolytic drugs, drugs that target for cell type-specific immunomodulation represents a new direction for OP treatment. Hydrogels loaded with recombinant Rho-Inhibiting C3 Toxin exhibit both anti-osteolytic function and specific monocyte/Mφ uptake properties, and reduced osteoclast activity in vitro [148]. RANK-inhibiting siRNA delivered by PLGA-PEI nanoparticles suppressed osteoclast precursor maturation into osteoclasts and reduced osteoclastic activity in vitro [149].

Biomaterial design with both pro-osteogenic and anti-osteoclastic functions have been harnessed to treat bone defects coupled with OP conditions [150, 151]. Incorporation of anti-osteolytic drugs to osteoconductive materials has been shown to benefit bone healing with OP conditions. Strontium (Sr) is a bioactive iron that stimulates osteoblast activity and reduces bone resorption [152]. Sr-incorporated to HA composites [153], mesoporous bioactive glass [154] and bio-ceramics [155] suppressed osteoclastogenesis and enhanced bone regeneration in OP animal models. Similarly, phosphonates incorporated with Fe₃O₄ and HA composites, or embedded in hyaluronic acid gels to fill in the defect inhibited osteoclastic activity and enhanced bone healing in long bone defects in rats with OP [156, 157]. However, some other studies have shown different results, indicating incorporation of anti-osteoclastic reagents may also hinder biomaterial-based bone healing. Phosphonates-loaded polycaprolactone scaffolds inhibited new bone formation in osteoporotic sheep [158]. Phosphonate treatment also reduced β-TCP ceramic implant turnover in a rat critical-size long bone defect model with OP condition [159]. Given osteoclastogenesis is necessary for replacing biomaterial implants with newly regenerated bone, excessive suppression of osteoclastogenesis by anti-OP drugs may hinder biomaterial-mediated bone regeneration.

7.3. Periodontitis

Periodontitis (PD) is an infection-induced chronic inflammatory disease that causes destruction of the periodontal tissues such as alveolar bone [160]. M1 Mφ/Th1/Th17 are beneficial for anti-bacterial, but they also secret pro-inflammatory cytokines that leads to osteoclast activation and bone destruction during PD progression [161, 162]. Current clinical treatments for PD focus mostly on removal of bacteria. Therapeutics that target resolving inflammation and bone regeneration represent promising new directions. Both cell- and biomaterial-based immunomodulation that target either macrophages or T cells has been developed for PD treatment.

7.3.1. Cell-based immunomodulation

Stem cells derived from periodontium have been used for immunomodulation and preventing bone resorption in PD. Periodontium has robust tissue-resident stem cell pool. Periodontium-derived stem cells (PDSCs) have similar immunomodulatory functions as MSCs to modulate both innate and adaptive immune responses [163, 164]. Transplantation of PDSCs decreased inflammatory cytokine level, while increasing M2 Mφ population in vivo, showing benefits in both periodontal defect regeneration [165] and alleviating bone loss [166] in PD animal model. Moreover, PDSCs transplantation synergized with IL-1 receptor antagonist in enhancing bone regeneration in pig periodontal defect model [167]. In addition to live cells, conditioned medium obtained from cultured PDSCs decreased TNF-α level in vivo and enhanced bone regeneration in rat periodontal defect model, compared to conditioned medium from fibroblasts [168].

7.3.2. Physical cues of biomaterials

Tuning physical properties of biomaterials has been harnessed to modulate Mφ response that benefits PD treatment. Changing bone-derived ECM from particle form to hydrogel form exhibited enhanced Mφ polarization toward M2 phenotype and bone regeneration in rat periodontal defect model [169]. Particle size also elicited different Mφ immune response and regenerative outcome in PD. Au nanoparticles (AuNPs) with 45 nm diameter outperformed 5 nm and 13 nm AuNPs in suppressing LPS-activated Mφ in vitro [170]. In vivo application of 45 nm AuNPs induced M2 Mφ population, decreased tissue inflammatory cytokine levels, and rescued tissue destruction in ligature-induced PD in rat [170].

7.3.3. Immunomodulator delivery

Biomaterial-based delivery of immunomodulatory cytokine has also been used to modulate Mφ and aid periodontal bone regeneration. IL-4 delivered by heparin-functionalized hydrogel microsphere showed enhanced M1 to M2 Mφ polarization and bone formation, compared to IL-4 free microsphere [171]. Co-delivery of IL-4 and SDF-1 further accelerated M2 Mφ polarization [172], indicating the beneficial crosstalk between endogenous stem cells and immune cells. Nanoparticles have been used to deliver clinic-available drugs to minimize toxicity [173].

In addition to Mφ, increasing attention has been paid towards targeting Treg for PD treatment recently [174]. Treg population has been demonstrated to help resolve chronic inflammation and rescue bone loss in PD [175]. Biomaterials have been used to deliver immunomodulators that favor Treg recruitment and differentiation. Treg-recruiting chemokine CCL22 delivered by PLGA particles showed enhanced Treg recruitment, suppressed the pro-inflammatory cytokine secretion, and rescued bone resorption in both mouse and canine PD models [176, 177]. Similarly, Triple delivery of IL-2/ TGF-β and miR-10a by a nano-in-micro spongy microsphere led to Treg enrichment, decreased osteoclast activity and attenuated bone loss in vivo (Figure 10) [178].

Figure 10. Preventing periodontal bone loss through enriching regulatory T cells using microsphere-mediated triple delivery of immunomodulators.

(a) Schematic of mechanism for microsphere-mediated delivery of IL-2/TGF-β and miR-10a. (b) MicroCT imaging showed triple delivery helped prevent bone loss, (c-d) Triple delivery group significantly increased percentage of Tregs and IL-10 level (c), while decreasing inflammatory IL-1β level and osteoclast activity (d). (Data reproduced from [178].)

7.4. Future direction

Due to the side effects associated with current anti-osteolytic drugs, there remains a critical need to develop effective and safe therapeutics for osteolytic diseases. Compared to bolus drug delivery, nanomaterials-mediated drug delivery has shown promises to reduce undesirable side effects. However, future research needs to control the biodistribution of systemically administrated nanoparticles and improve targeting efficiency to the disease site. Future studies should also explore new immunomodulators for treating osteolytic bone loss, either with new therapeutic targets such as TLRs [179], Notch signaling [180] or novel mechanisms such as RNA interference [181]. There remains a critical need to develop better therapies to help patients with osteolytic diseases regenerate bone more efficiently upon injury. OP patients are at high risk for fracture non-union, and bone healing happens in less than 10% of RA cases [182]. More immunomodulatory strategies that favor bone regeneration and anti-osteolysis should be developed in the future.

8. Conclusions and Outlook

Compared to other tissues, bone has more direct interaction with immune system through bone marrow space, and osteoclasts are derived from immune cell lineage. The type of immune cells in bone niche varies depending on anatomical location and disease type. Future immunomodulatory strategies for bone applications should take these factors into consideration. For applications targeting regeneration in critical size bone defects, it is important to modulate timing of Mφ transition to avoid prolonged inflammation while preserving early acute inflammation. For diseases caused by osteolytic bone loss, immunomodulatory strategies should focus on combating imbalanced osteogenesis and osteoclastogenesis. Developing smart biomaterials that allows spatiotemporal modulation of immune system would be very useful [183]. Biomaterials that allow sequential cytokine delivery or external stimuli-triggered topological change can be used to modulate the timing of Mφ phenotype transition to promote bone healing and implant integration [184, 185][186]. Release of immunomodulators triggered by inflammatory signals holds promises to provide on-demand immunomodulation [137, 187]. Biomaterials that can adhere to immune cells have been used for target drug delivery along with immune cell migration to the injury sites [188, 189], which can be potentially used for targeting delivery of anti-osteolytic drug. Better characterization of immune response of existing biomaterials for bone applications (such as collagen sponge and calcium phosphates) is another a promising future direction to help select optimal biomaterials composition that can enhance bone regenerative outcome through favorable immunomodulation.

Another future direction is to design immunomodulatory strategies to enhance MSC-based bone regeneration. While some positive results have been reported with MSC-based therapy for bone, other studies showed adding MSCs led to worse regeneration outcomes, especially in immunocompetent animal models [190, 191]. The highly inflammatory environment in acute bone injury may compromise MSC-based therapeutic effects in two ways. First, inflammation can induce MSC apoptosis, leading to shorter cell retention after implantation. Second, the undesirable crosstalk between MSCs and immune system may further exacerbate inflammation. To overcome these issues, adding immunomodulatory cues could be considered as a future direction to prolong the survival of MSCs in vivo and maximize the release of therapeutic secretome from MSC. This may be achieved by either genetically modify MSC to exhibit stronger immunomodulatory function [192, 193] or by in situ delivery of immunomodulators using biomaterials [27, 107]. In addition to using live cells, MSC-derived secretome also offers an alternative strategy for immunomodulation. MSC secretome has been reported to exhibit comparable immunomodulatory functions as cells [9], and have been used in clinical trials for treating inflammatory diseases such as arthritis. Compared to using live cells, using MSC secretome collected in vitro could overcome drawbacks of short-term MSC survival and undesirable immune activation [194, 195]. Cell and cell-derived products that can target multiple immune players with minimal side effects can be advantageous in modulating systemic immune disorders induced by orthopaedic implants and osteolytic conditions in the future [196]. For treating critical-sized bone defect, MSC secretome can also be easily incorporated into biomaterials to enhance bone healing, through both immunomodulation and promoting osteogenesis and vascularization [197–199]. Considering the regulatory pathway and translational needs, products that do not involve live cells are also cheaper for manufacture and easier for storage.

Most immunomodulatory strategies so far have largely focused on Mφ. In addition to Mφ, the role of other immune cells, especially adaptive immune cells, needs to be further explored for bone applications. In the context of soft tissue regeneration, Th2 and Tregs have been found to be positive contributors for tissue regeneration through crosstalk with Mφ and progenitor/stem cells [200–202]. While adaptive immune cells have been shown to be important for fracture healing [37] [84], the role of adaptive immune cells in healing critical-sized bone defects remains largely unknown. Bone regeneration for critical-sized bone injury is more challenging in immunocompetent animals than athymic animals [203], indicating the important role of adaptive immune system. Given most bone tissue engineering strategies involve the use of biomaterials as scaffolds, how varying biomaterials modulate adaptive immune response needs to be evaluated in a materials-specific and niche-specific manner. For osteolytic disease, more opportunities lie in understanding the roles of different T cell subpopulations in the context of osteolytic disease progression, and design strategies to modulate different immune cell populations for better treatment outcomes.

In addition to T cells, other immune cells, such as dendritic cells and B cells, could also be interesting targets. Dendritic cells have been shown to directly impact osteoclastogenesis in inflammatory bone diseases [204], but its role in bone regeneration remains largely unknown. Lastly, pre-osteoclasts are also interesting targets for future research given their demonstrated effects on stem cell osteogenesis[205], angiogenesis [206] and inducing Treg differentiation [207]. For example, increasing pre-osteoclast population in vivo using silicic acid releasing scaffolds showed benefit in promoting vascularization and bone regeneration in a cranial defect model [208]. Developing strategies to modulate osteoclast activity that aid neo-vascularization could be an interesting future direction for both injury-induced and osteolytic bone loss [209].

To develop effective immunomodulatory strategies for bone regeneration, it is critical to have a better understanding of the kinetics and types of immune cells that would infiltrate into bone niche. This should be studied in a disease-specific manner. Establishing a road map that correlates the infiltrated immune cell composition and kinetics with the bone healing outcome would guide rational design of immunomodulatory strategies. As such, another important future direction is to integrate various technologies that would enable characterizing of immune cell response in vivo. High throughput methods such as flow cytometry, CyTOF, and single cell RNA sequencing (scRNAseq) are all powerful tools that can be applicable. Both flow cytometry and CyTOF have been used for dissecting immune cell populations by staining of immune cell markers. Flow cytometry is easily accessible and less expensive. CyTOF is more expensive and complicated in sample preparation but is more high-throughput with the capability of identifying up to 50 markers within one sample. ScRNAseq allows snapshot of full gene expression profile at a single cell level, giving comprehensive understanding of who are the immune players are and is useful in identifying unknown and rare immune cell population. Given immune response precedes bone formation, analyzing immune cell responses at earlier time points would likely correlate better with later bone regeneration. In addition, these methods require harvesting live cells from in vivo implants, which are also much easier to achieve before extensive mineralized bone tissues have formed.

One limitation of the technologies described above is that they all require sacrifice of the animals for end-point analysis, and are not suitable for real time monitoring of the immune cell infiltration in situ. To overcome this limitation, there remains a critical need to develop non-invasive cell imaging techniques that enable tracking immune cells in the bone niche. Bioluminescence-, fluorescence-, and magnetic resonance imaging have been used for non-invasive tracking of immune cells in soft tissues in vivo [210]. Unlike soft tissues, bone is highly mineralized, posing additional challenge for imaging techniques that is based on photon detection. To this end, two-photon microscopy that allows increased depth penetration represents a powerful approach and has been used to evaluate bone-niche cells activity at a high spatiotemporal resolution [211, 212]. However, the imaging depth is still limited within hundred micrometers due to the optically dense nature of bone. Future studies should investigate how to adapt new imaging techniques such as three-photon microscopy and synergize with progression in transgenic reporter mice and labeled probes to allow efficient detection of immune cells from bone niche.

Another opportunity for future directions is to develop better 3D in vitro experimental models that can mimic the crosstalk between immune cells and other bone niche cells (i.e. stem cells) in vivo [115, 213]. Such in vitro models can be useful for in vitro screening to help narrow down leading biomaterial compositions for subsequent in vivo testing, as animal models are expensive and lengthy. Such physiologically relevant in vitro co-culture model can also serve as a powerful tool for mechanistic studies, as it allows easy tuning of types and ratios of cells and/or decoupling other niche cues. Such experimental model can also be customized to mimic specific niche cell types present in different bone niche or disease types.