Calvarial Versus Long Bone: Implications for Tailoring Skeletal Tissue Engineering

Cellular embryonic origins

Cells from different embryonic origins give rise to the formation of calvarial and long bones. Specifically, lateral plate mesoderm-derived cells (originating from mesoderm germ layer) form long bones, and neural crest (originating from ectoderm germ layer) and cephalic mesoderm-derived cells form skull bones. ⁷ This is surprising that bones are developed from both lateral plate mesoderm and neural crest, as we often assume that one tissue type should be derived from one specific germ layer. ⁷ This unexpected discovery subsequently impacted our perception of skeletal tissue development and healing.

In a landmark study, skeletal defects were created in the neural crest-derived mandible and mesoderm-derived tibia. Neural crest-derived progenitors and non-neural crest-derived progenitors were labeled separately to observe their repairing sites. The results showed that the former cells repaired mandible while the latter repaired tibia. It suggests that skeletal tissues normally utilize cells from their own embryonic origins for repairing and regeneration. ⁸ Furthermore, such healing features were not necessarily interchangeable. This is evident by the ability of neural crest-derived progenitors to differentiate into osteoblasts upon transplantation into tibial defect sites, whereas mesoderm-derived progenitors were incapable of doing so for a mandible defect. ⁸

However, caution should be taken to avoid overinterpretation of these results due to the confounding effects of cell heterogeneity. For example, studies isolating a single population of skeletal progenitors from the growth plate of long bone exhibit robust plasticity. These mesoderm-derived skeletal stem cells are capable of generating multiple subpopulations of progenitor cells, which exhibit different capacities to form bones, cartilages, and stromal cells.^9,10 Therefore, skeletal progenitors of different embryonic origins may exhibit different healing capacities, but further study is warranted. ¹¹

Molecular similarities and differences during bone development

Development of skull and long bone start from mesenchymal condensations, ¹² which involves similar molecular actors, including transforming growth factor-beta (TGF-β) superfamily (TGF-β isoforms and bone morphogenetic proteins [BMPs]), fibroblast growth factor (FGF), WNTs (WNT3A, WNT10B; please refer to recent reviews on this topic),^13–15 and transcription factors such as sex-determining region Y-box 9 (SOX9) and adhesion molecules.^16–18 Extensive reviews on bone development and healing process, mostly using long bones as the modeling system, can be found in these review articles.^19,20

Following mesenchymal condensation, skull and long bones primarily develop through distinct ossification modes. For long bones, endochondral ossification is predominant, whereby a cartilaginous intermediate is generated and subsequently gives rise to bone. ²¹ For the majority of the skull bones, their skeletal elements are formed from intramembranous ossification. It should be noted that a small subset of skull bones, including the temporal, occipital, sphenoid, and ethmoid bones, form through a combination of endochondral and intramembranous ossification. ²¹

At the molecular level, intramembranous bone formation primarily occurs through expression of bone-encoding genes, whereas endochondral bone formation, owing to its cartilaginous template formation, utilizes both cartilage- and bone-encoding genes. Typical bone-encoding genes include collagen type I (col1a1), bmp-2, and bmp-6, whereas typical cartilage-encoding genes include sox9, collagen type II (col2a1), collagen type X (col10a1), and indian hedgehog (ihh).^16–18 Common to both modes of ossification, the bone-encoding gene col1a1 contributes toward the organic portion of bone extracellular matrix, ²² whereas runt-related transcription factor 2 (runx2)^23,24 is a master transcriptional regulator for osteoblast differentiation. Exclusive to endochondral ossification, the cartilage-encoding gene, col2a1, is a major extracellular component of the cartilaginous template and deletion of this gene leads to cartilage and skeletal abnormalities. ²⁵

Additional studies in recent years have further elucidated the differential contributions of various genes toward endochondral and intramembranous ossification. For example, the cartilage-encoding gene ihh is essential for endochondral bone formation through regulation of chondrocyte proliferation and differentiation.^26,27 Indeed, knockout mice lacking ihh exhibit severe endochondral defects that result in severely reduced limb, whereas intramembranous bone formation appeared less affected.^26,27

Similarly, high-mobility group box 1 (HMGB1) also differentially affects both bone ossification modes. HMGB1 is a nonhistone DNA-binding protein, and extracellular HMGB1 can mediate inflammation and tissue regeneration. ²⁸ In vitro studies suggest that intramembranous and endochondral-derived osteoblasts exhibit different responses toward HMGB1. Specifically, recombinant HMGB1 stabilized the receptor activator of nuclear factor kappa-B ligand (RANKL)/osteoprotegerin (OPG) expression ratio and augmented the expression of tumor necrosis factor-alpha (TNF-α) and interleukin-6 (IL-6) in long bone-derived, but not calvarial bone-derived cells. ²⁹ In support of this, embryonic studies have demonstrated that endochondral, but not intramembranous ossification, is compromised in HMGB1 knockout mice. ³⁰

The WNTs are worth noting as they are well-established signals that regulate many biological processes, including intramembranous ossification and endochondral ossification. ³¹ Briefly, there are 19 secreted glycolipoproteins that function as WNT ligands in three distinct signaling pathways, including the canonical β-catenin-dependent pathway, noncanonical planar cell polarity pathway, and noncanonical WNT/Ca²⁺ pathway.^13–15 For a more comprehensive review of various skeletal phenotypes associated with WNT signaling alterations, please refer to a review work done by Maupin et al. ¹⁵ In brief, WNT signaling is crucial to both endochondral and intramembranous ossification. However, to the best of our knowledge, there is no direct evidence showing that WNTs might exert different roles in endochondral and intramembranous ossification. Future studies will also need to further elucidate the various contributions of different components in WNT signaling to bone healing and regeneration.

Thus, endochondral and intramembranous ossification can employ distinct molecular signals that are responsible for the different modes of bone formation. Understanding the subtle differences between cellular and molecular players during skull and long bone development will help to generate optimal, tailored strategies for bone repair.

Inflammation stage

The inflammatory response following bone fractures is similar to the general wound healing response. Although the exact role of inflammation in bone repair requires further elucidation, increasing attention has been shown that fracture-drive inflammatory response is essential for both host defense against infection as well as bone regeneration and repair. ³⁴

Hematoma formation and intense cell infiltration are two main features during inflammatory stage after skull and long bone fracture. The healing response is orchestrated by inflammatory cells, such as platelets, macrophages, lymphocytes, monocytes, and granulocytes, which migrate into the fracture hematoma and secrete various cytokines and growth factors. In general, inflammatory cytokines expression peaks at this stage, but gradually declines with the progression of wound healing progresses. Subsequently, during the bone formation stage, growth factors associated with osseous tissue induction, including TGF-β2, TGF-β3, growth differentiation factor-5 (GDF-5), and BMP-2, BMP-3, BMP-4, BMP-5, and BMP-6, exhibit a steadily increased expression ¹⁹ (Table 1; representative histology images and gene expression patterns during skull bone repair can be found in Figs. 1 and 2 and long bone fracture healing patterns can be found in a review article done by Schindeler et al. ¹⁹ ).

OPEN IN VIEWER

FIG. 1.

Patterns of histology and gene expression during the inflammatory stage of skull bone healing. 1.8 mm diameter skull defects were created in WT mice. At designated time points, tissue around the original defect sites was collected for subsequent histology and expression analyses. (A) Representative H&E-stained images at postoperative 3 h, D1, and D4. The bone defect was filled with hematoma as early as day 1. On the endocortical surface (presumably derived from the dura mater), cells gathered along the defect perimeter. Evidence of hematoma degradation and reduction of cellular infiltrate was apparent by D4. (B) Fold change expression of IL-1α, IL-1β, IL-6, and TNF-α at postoperative 3 h, D1, and D4 relative to D0 (unpublished work; scale bar: 50 μm; bolded black arrows: defect margin; Endo: endocortical side of skull bones; D, day; mean ± SEM; p < 0.05, * compared with 3 h). H&E, hematoxylin and eosin; IL, interleukin; SEM, standard error of the mean; TNF, tumor necrosis factor. Color images are available online.

OPEN IN VIEWER

FIG. 2.

Histology and gene expression during bone formation and bone remodeling stages of skull bone healing. 1.8 mm diameter skull defects were created in WT mice. At designated time points, tissue around the original defect sites was collected for subsequent histology and expression analyses. (A) Representative H&E stained images at postoperative D7, D14, D21, and D28. Newly formed cellularized bone matrix was observed on the endocortical side of the skull bone since D7. On D28, periosteum and soft connective tissue became denser and better organized. (B) Fold change expression patterns of cytokines and growth factors relative to D0. Expression of these cytokines and growth factors showed elevated expression around D14 and D21 and declined at D28. (scale bar: 50 μm; bolded black arrows: defect margin; Endo: endocortical side of skull bones; D, day; mean ± SEM; p < 0.05, * compared with D7, D14, and D28). Color images are available online.

During bone healing, calvarial and long bone exhibit differences at the molecular, cellular, and tissue level. These include differences in the gene expression profile, hematoma content, and dura mater reaction. For example, one unique feature of skull bone repair is the early and essential involvement of dura mater, where there is an increase in cellularity, ³⁷ which can benefit fracture repair as dura maters are rich in osteoprogenitors.

Another difference is fracture hematoma. Accumulating data have demonstrated a critical role of fracture hematoma in bone repair, as depletion of hematoma leads to compromised long bone healing. ³⁸ However, it is still not clear whether hematoma has similar role in skull bone repair, and from another aspect, surgeons are more cautious treating hematoma after a traumatic head injury.^39,40 At the molecular level, an interesting phenomenon was observed in patients with combined head and bone injuries compared with patients with bone injuries only.^41,42 Mainly, higher IL-6 and lower RANKL/OPG serum levels were observed in the patient population with combined injuries, suggesting that the outcome of fracture repair outcome can be influenced by which bones are injured.^41,42 This location-dependent healing pattern may be explained by the need for protecting the central nervous system from infection in the event of a head injury. However, by eliciting a greater-than-normal inflammatory response, this may result in deleterious outcome for bone repair.

Given the essential involvement of inflammatory response in fracture healing, multiple attempts have explored approaches by modulating fracture-induced inflammation for optimal bone healing outcomes, such as manipulating the expression of inflammatory cytokines, chemokines, or growth factors (e.g., interleukin 1 [IL-1], IL-6, IL-18, TNF-α), Complement 3, CXCL12 (C-X-C motif chemokine ligand 12)/CXCR-4 (C-X-C chemokine receptor type 4), TGF-β, vascular endothelial growth factor (VEGF), and stromal cell-derived factor. ⁴³

Indeed, ILs, such as IL-1, IL-6, and IL-18, have important roles in regulating bone metabolism. TNF-α-induced IL-1 has been shown to be involved in a systemic bone loss through increased osteoclast formation and bone resorption. ⁴⁴ Similarly, IL-6 can inhibit osteoblast differentiation and promote osteoclast formation, indicating a catabolic effect. ⁴⁵ However, when IL-6 was applied sequentially following parathyroid hormone (PTH) administration, fracture healing is accelerated and the resulting bone tissue exhibits enhanced mechanical properties by 200–300%. ⁴⁶ IL-18 has been reported to act as a mitogen for calvarial and long bone-derived osteoblasts and chondrocytes, ⁴⁷ upregulating OPG (an osteoclast inhibitor) in both calvarial and long bone-derived cells, ⁴⁸ and is crucial for PTH-mediated bone anabolism. ⁴⁹

While not well documented, differences in calvarial and long bone healing patterns may be partly attributed to the differential effects of inflammatory mediators. For example, Toll-like receptor-4 (TLR-4) is a member of the transmembrane receptor family that activates the innate immune response and tissue homeostasis, including musculoskeletal system. ⁵⁰ Prior studies showed that mutations in TLR-4 mutation result in long bone nonunion following a fracture injury. ⁵¹ In contrast, our prior work showed that TLR-4^−/− mice exhibit accelerated skull bone healing with increased dura mater cellularity and inflammatory cytokine expression. ³⁷ However, further study is required as these differences in bone healing may be explained by impaired recognition and elimination of pathogens rather than innate differences between calvarial and long bone as our studies employed sterile calvarial defects.

IL-17 is a group of cytokines (IL-17A, IL-17B, IL-17C, IL-17D, IL-17E/IL-25, and IL-17F) produced by a subset of CD4⁺ T cells known as Th17 cells. Expression of IL-17 family members and their receptors have been found in immune cells and tissue cells, including bone cells.^52,53 Studies first discovered the function of IL-17 in mediating diverse inflammatory conditions and bone diseases, for example, periodontitis, and rheumatoid arthritis.^52,53

Regarding their roles in fracture repair, expression of IL-17A, IL-17B, and their receptors were detected on bone cells (e.g., osteoblasts, prehypertrophic chondrocytes) in a rat long bone defect model. ⁵⁴ Further depletion of IL-17A led to impaired bone regeneration due to decreased osteoblast function. IL-17A can also indirectly influence osteogenesis through osteoclasts. For example, IL-17A markedly decreased the expressions of bone resorption-related enzyme cathepsin K and matrix metallopeptidases-9 (MMP-9) in the RAW264.7 cells (osteoclast precursors), suggesting suppression of osteoclastogenesis. ⁵⁵ Besides IL-17A, another family member IL-17F, which is located on the same chromosome and shares the highest degree of homology with IL-17A has also been found to be a key mediator in the immune system and skeletal system. Osteoblast progenitors treated with IL-17F showed significantly increased bone marker expression, ⁵⁶ suggesting it can be used as another novel target to activate osteogenesis during the inflammatory stage and enhance fracture healing outcome. ⁵⁷ Interestingly, results from various studies indicate that IL-17 may have distinct effects on bone formation depending on targeting cell types. For example, IL-17A can exert negative effects on osteogenesis of mouse calvaria-derived cells, ⁵⁸ but positive effects on immature mesenchymal cells. ⁵⁹ IL-17 can stimulate osteogenesis and hamper adipogenesis, but it can also enhance osteoclast lineage differentiation from precursor cells. ⁵² Therefore, further studies are needed to both elucidate the role of IL-17 in inflammatory conditions and bone metabolism as well as utilize them for fracture repair.

Advancing our understanding of the impact of inflammation on the healing of calvarial and long bones will aid in the development of immunomodulatory strategies tailed for site-specific bone repair.

Bone formation and bone remodeling stages

During the bone formation stage, calvarial bone and long bone repair utilize different ossification modes that directly influence the healing patterns and involved cellular and molecular mechanims. Compared with intramembranous ossification, a unique feature of endochondral ossification is the essential involvement of chondrocytes and hypertrophic cartilage template. ⁶⁰

In brief, mesenchymal progenitors infiltrate defect sites, differentiate into chondrocytes with increased SOX9 expression, ⁶¹ and subsequently synthesize a cartilaginous matrix which replaces the hematoma and later forms a soft callus template. During this process, chondrocytes experience four main stages: proliferation, prehypertrophy, hypertrophy, and apoptosis. The hypertrophic chondrocytes express extracellular matrix remodeling enzymes, including MMPs (e.g., MMP-13, MMP-3, MMP-1) and aggrecanases (A Disintegrin And Metalloproteinase with ThromboSpondin motifs-4 and -5 [ADAMTS-4, ADAMTS-5]) as well as extracellular matrix components, such as COLX. ⁶² Also, hypertrophic chondrocytes express high levels of VEGF-A, resulting in the invasion of blood vessels into the soft callus template. ⁶³ The apoptotic chondrocytes serve as nucleation sites for calcium–phosphate crystals to produce calcified cartilage. Following this, woven bone is formed through gradual replacement of this calcified cartilaginous matrix.

In intramembranous ossification, osteoblasts and fibroblasts are the most active cell types and RUNX2/CBFα1 (core binding factor α1) and osterix (OSX) are crucial signals during skull bone formation stage.^64,65 During this process, osteoprogenitor cells are derived from numerous sources, including bone marrow, periosteum, dura mater, and nearby soft tissue. ⁶⁶ As early as day 7 after calvarial fracture, mononucleated osteoblasts with typical large, cubic shape, and extracellular matrix formation are evident at the fracture site. ³⁷ As healing progresses, mechanically weak woven bone, characterized by its haphazard organization of extracellular matrix is replaced by lamellar bone, whose highly aligned arrays of collagen fibers result in increased mechanical strength (Table 1; representative histology images and gene expression patterns during skull bone repair can be found in Figs. 1 and 2 and long bone fracture healing patterns can be found in a review article done by Schindeler et al. ¹⁹ ).

Although utilizing different ossification modes, some similarity exists during bone formation and bone remodeling stage in skull and long bones. For example, bone remodeling stage is characterized by osteoclast-mediated bone resorption and bone remodeling, where histological patterns were similarly observed in skull and long bones. With regard to active signaling molecules, declined expression of inflammatory cytokines, but increased expression of multiple growth factors, including TGF-β2, TGF-β3, GDF-5, BMP-2, BMP-3, BMP-4, BMP-5, and BMP-6, and angiogenic factors, such as angiopoietins and VEGF, were observed from inflammatory stage to bone formation stage (Table 1). ¹⁹ During bone remodeling stage, macrophage colony-stimulating factor (M-CSF) and RANKL (the ligand of NFκB), are two critical cytokines. Both M-CSF and RANKL are crucial for inducing osteoclast differentiation, which is a vital actor in the bone remodeling unit. ⁶⁷ Concurrently, inflammatory cytokines (e.g., IL-1, IL-6, TNF-α, OPG, and RANKL) and growth factors (e.g., TGF-βs) exhibit increased and decreased expression, respectively. Regenerated lamellar bone will eventually re-establish the geometric and functional properties of the injured bone tissue (Fig. 2 and Table 1).

Dura mater

The dura mater is a fibrous membrane with varying orientations, thicknesses, and structures. ⁶⁸ This tissue is comprised of five layers, including the bone surface, external median, vascular, internal median, and arachnoid layers. ⁶⁸ It surrounds the spinal nerve structures and the intracranial, and functions not just as a barrier for protection, but is also highly involved in tissue regeneration. Although periosteum is a major source for osteoprogenitors during long bone healing, dura mater contains multipotent mesenchymal cells and provides paracrine signals, which are all important for skull bone healing. ⁴ Furthermore, in vitro data suggest that dura mater-derived stem cells present greater osteogenic bioactivity and matrix synthesis compared with bone marrow stem cells, ⁶⁹ suggesting a potential use of its resident cells as a tool for skull bone tissue engineering. ⁷⁰

The osteogenic bioactivity of dura mater can be age dependent. Based on genome-wide expression analysis, juvenile and adult dura mater showed differential gene expression profiles, where higher expression of osteogenic and osteoclastogenic markers were observed in juvenile compared with adult dura mater cells. ⁷¹ Indeed, heterotopic membranous ossification was observed when juvenile dura was transplanted into an adult rat.^72,73 Also, transplantation of mature dura mater into immature animals resulted in compromised bone healing, whereas application of immature dura mater led to successful bone repair.^72,73 Although the exact underlying mechanisms are uncertain, it may be partially due to the skull growth-induced mechanical strain.

It is also possible that differences in dura mater osteogenic function lead to different skull bone healing outcomes in children younger or older than 2 years of age. During cranioplasty, clinicians prefer to use vascularized grafts or flaps for repairing the dura. ⁷⁴ This is crucial for wound closure and minimizing leakage of cerebral spinal fluid. While there are few (if any) reports on the effect of lifestyle factors in direct relationship to dura mater-mediated bone healing, it has been established that medical situations, diseases, and lifestyle factors such as anemia, certain cardiac diseases, chemotherapy or radiotherapy, diabetes, hypoalbuminemia, and hypoproteinemia, prior cranial operations, and smoking, may compromise dura mater healing. ⁷⁴

Periosteum

During endochondral ossification, the periosteum participates in osteogenesis, playing an indispensable role in cartilage induction and long bone fracture healing. ⁷⁵ The periosteum consists of two layers—an outer fibrous layer comprised primarily of fibroblasts and an inner cambium layer that is comprised of multiple cells and tissue structures, including nerves, capillaries, osteoblasts, and undifferentiated progenitor cells.^76,77 Within the skull, periosteum is also present and is known as pericranium.

Residing within the periosteum and bone marrow compartments are mesenchymal stem cells (MSCs), which exhibit similar proliferation capacities. ⁷⁸ In response to injury, these attributes facilitate the healing process through stem cell activation. ⁷⁸ Periosteal cells also exhibit features similar to multipotent MSCs.^77,79,80 In one study, periosteal cells and bone marrow-derived stromal cells (BMSCs) were differentiated into osteoblasts in the presence of BMP-2 and basic FGF-2. However, periosteal cells proliferated more rapidly but exhibited less osteogenic potential relative to bone marrow cells. ⁷⁵ Despite their less robust osteogenic capabilities, prior studies have demonstrated that periosteal cells are essential for bone repair. ⁷⁷ Typically, bone marrow/endosteal injuries are healed through intramembranous ossification, whereas periosteal injuries are predominantly healed through endochondral ossification. ⁸¹ This suggests that the periosteum is a major contributor of chondrocytes during callus formation, and resulting loss or disruption of this tissue may explain delayed long bone healing or nonunion. ⁸² Therefore, periosteum biomimicry has been utilized as a potential strategy to enhance bone fracture repair.^83–86

In addition to participating in native bone healing, the periosteum is also critical in graft-mediated bone repair. For example, application of periosteum MSCs and hydroxyapatite ceramics have been shown to induce bone formation in rat subcutaneous model. ⁸⁷ Also, when combined with tissue-engineered periosteum, superior bone healing was observed relative to allograft alone in a murine segmental femoral graft model. ⁸⁸ Such artificial periosteum can be fabricated by combining porcine small intestinal submucosa with differentiated bone marrow stem cells, resulting in enhanced osteogenesis and augment bone repair in a rabbit segmental bone defect. ⁸⁹ Altogether, these studies underscore the importance of periosteal application for endochondral bone graft-mediated repair.

Although the periosteum is essential for long bone regeneration, studies have suggested that this tissue has less impact on calvarial healing. Rather, dura mater is crucial to successful calvarial bone healing.

In a prior study, rabbit tibial and calvarial bone defects were created. The subsequent contributions of various bone and bone-associated tissues, including periosteum, cortical bone, endosteum, bone marrow, and dura mater were histologically assessed. ⁹⁰ While periosteum was a major contributor for tibial healing, dura mater produced more bone formation for calvarial bone healing compared with pericranium; although both were essential for complete restoration of bone. ⁹⁰ In a similar study employing rabbit calvarial defects, dura mater was a major contributor of bone formation throughout the entire graft, whereas periosteum contributed toward bone formation only for the region that was in close proximity with the graft. ⁹¹ Also, in a rat calvarial defect model, the additional application of autogenous periosteal cells together with a bovine-derived biomaterial (inorganic apatite and collagen) did not improve bone formation relative to biomaterial alone. ⁹²

To the best of our knowledge, no direct evidence comparing the osteogenic capabilities of periosteum and dura mater have been performed. As such, further studies are needed to further definite the contributions of periosteum and dura mater in long and calvarial bone repair.

Suture

Cranial sutures are fibrous joints found between craniofacial bones and contribute to the growth, healing, and elasticity of the skull. There are six primary sutures with each tissue existing as a thin layer of undifferentiated tissue sandwiched between two skull bones.^93,94 Sutures play crucial roles in facilitating skull expansion to accommodate brain growth as well as facilitating calvarial bone movements, and absorption of mechanical forces to protect the underlying central nervous system from physical impact.^93,94 As a result, abnormal suture growth can greatly impact calvarial bone and brain development, as exemplified by craniosynostosis, where premature ossification of sutures results in a broad range of deformities. ⁹⁵

In addition to its roles mentioned above, sutures also serve as future sites for intramembranous bone growth. Suture resident cells express biomarkers implicated in calvarial bone development, regeneration, and pathogenic conditions such as craniosynostosis, including specific transcription factors or growth factors, for example, RUNX2, Nel-like molecule-1 (NELL-1), TGF-β1, and FGF-2.^96–99 For example, the secreted protein NELL-1 was originally identified from studies of craniosynostosis. When delivered together with a demineralized bone matrix carrier, NELL-1 resulted in improved bone healing in a rat critical-sized femoral segmental defect model. ¹⁰⁰ Other craniosynostosis-associated elements, such as Twist-related protein-1 (twist-1), is suggested to inhibit osteoblast differentiation in vitro.^101,102 Haploinsufficiency of twist-1 or twist-2 cause reduced bone formation in mice, evidenced by much smaller skeleton, delayed suture fusion, open posterior fontanelles, and delayed ossification in the metatarsals and phalanges. ¹⁰³ Thus, suture-resident cells and suture-associated signaling molecules are expected to be promising candidates in the development of bone regenerative therapies.

Bone marrow

The architectural and structural differences of long and calvarial bones result in distinct bone marrow volumes. By virtue of their larger bone volume and elongated structure as well as the presence of a well-defined marrow cavity, long bones contain vast amounts of bone marrow and are a major site for hematopoietic cells. In contrast, the smaller and flatter structure of skull bones result in comparatively smaller bone marrow volumes.

While the impact of bone marrow on calvarial and long bone is relatively unexplored, their close physical association suggests that they may regulate each other's activities. Prior studies showed that functional marrow tissue encased in a shell of bone were obtained by ectopic, subcutaneous implantation of long bone-derived marrow tissue in rats.^104,105 In contrast, only ectopic bone but no functional marrow was observed upon transplantation of mouse fetal calvarial-derived skeletal cells to the renal capsule.^106,107 Likely explanations for this phenomena include cell-to-cell interactions between long bone-derived skeletal cells and marrow-derived hematopoietic stem cells (HSCs), which typically reside together in stem cell niches.

The stem cell niche refers to the immediate microenvironment of stem cells and is defined as the complex milieu of various biochemical, neuronal, and mechanical cues that result from cell-to-cell and cell-to-matrix interactions. Upon injury, the stem cell niche functions to maintain the existing stem cell pool through self-renewal while expanding stem cell numbers to participate in the healing response. ¹⁰⁸ For example, osteoblasts ¹⁰⁹ and mesenchymal stromal cells ¹¹⁰ have been reported to regulate HSC numbers and self-renewal capability, respectively. However, the effect of HSCs on skeletal cells and bone healing remain unexplored and warrants further study.

In summary, multiple tissue structures, including dura mater, periosteum, sutures, and bone marrow are found in the skeletal system. Each of these unique structures exerts different effects on skeletal progenitor proliferation and differentiation. By elucidating the culminated effects of these structures on calvarial bone formation and graft-mediated repair, personalized, site-specific therapies for bone repair may be developed.

Scaffold-based approach

When designing biomaterials for bone regeneration, key consideration includes scaffold architecture, mechanical properties, and degradation rate. These parameters are highly interrelated and must be considered within the context of the repair site for an optimal healing outcome.

Bone structure, including geometry (shape), porosity, and composition (e.g., compact vs. cancellous tissue) can profoundly influence the tensile, compressive, and bending strength as well as Young's modulus of skeletal tissues.^111,112 Thus, ideal scaffold architecture and its mechanical properties need to closely approximate such properties of the bone defect sites. At the cellular level, architectural differences exist in the osteocyte networks of parietal (calvarial) and tibial (long) bones, which are presumed to be local tissue adaptations for various physiological loading patterns. ¹¹³ As such, scaffolds must not only mimic the geometry and mechanical properties of local skeletal sites (Table 2^{114–118,120,121}), but they should also withstand local biomechanical forces to fulfill their function. Calvarial bones are typically considered nonload-bearing, but they are not absent of biomechanical function, as they displace forces of mastication and must withstand high acceleration and impact forces to protect the underlying soft tissues and brain. ¹²² On the other hand, long bones are considered load-bearing bones and must withstand various forces resulting from body movement.

In general, scaffolds must possess a degradation rate synchronous with the rate of local tissue healing. Otherwise, the resulting loss in physical integrity will compromise biomechanical strength and function. Therefore, scaffold design must not only consider the aforementioned effects on mechanical properties but also its effect on degradation rate.

The material property of bone scaffolds directly impacts their degradation rates. For example, slowly degrading variants of calcium phosphate-based cement showed higher compressive strength relative to fast degrading versions. ¹²³ Low porosity in poly (d,l-lactide-co-glycolide)-based scaffolds increased degradation rate through local accumulation of acidic degradation products. ¹²⁴ To determine the rates of calvarial and long bone healing, comparative studies using micro-CT (computed tomography) and histomorphometric analyses show that tibial defects bridged defect gaps as early as 3 weeks, whereas calvarial defects healed by 6 weeks. ¹²⁵ These results suggest that dissolution rate of calvarial bone scaffolds can be relatively slower to maintain its function during the healing process. Similarly, another study using a combination of intravital imaging, fluorescence trapping, and whole-body optical imaging has revealed that calvarial bones possessed a larger normalized blood volume fraction and increased bone remodeling activity. ¹²⁶ This increased bone remodeling activity in calvarial bone may be a result of higher levels of osteoclast differentiation ¹²⁷ and increased resorptive function. ¹²⁸ As such, relatively higher bone remodeling properties will require slower-resorbing scaffolds. Thus, the complex interdependent relationships among scaffold architecture, mechanical properties, and degradation rate require biomaterials specifically tailored for different bone repair sites.

Three-dimensional printing is a ground-breaking manufacturing technology that offers an avenue to create scaffolds which mimic the macro- and microarchitecture of bone tissue as well as other medical devices that can improve patient care. ¹²⁹ The nature of additive manufacturing enables fabrication of certain geometries and shapes that could not be achieved through traditional methods such as mold casting. Examples include manufacture of ceramic or metallic implants (with bone-like mechanical properties and architecture)^129,130 that are customized for a patient's specific anatomy. By virtue of the materials used, such scaffolds exhibit high modulus that are comparable or greater than native bone tissue, which are well-suited for load-bearing (long bone) applications. Indeed, such 3D-printed implants have been commercially available for hip implants (REDAPT; Smith & Nephew). To enhance tissue regeneration, 3D printed scaffolds may also include bioprinted cells and growth factors. ¹³¹

Beyond applications for creating scaffolds, 3D printing is clinically useful for creating patient-specific anatomic models for surgical preplanning and patient education/counseling ¹³² as well as fabricating custom surgical guides. These surgical guides can be manufactured so that they fit a patient's specific anatomy such as skull curvature, providing fiduciary marks that allow surgeons to quickly gain access to the operating field. ¹²⁹ Thus, use of 3D printing as a technology for manufacturing medical devices, including scaffolds/implants, anatomical models (for surgical preplanning and patient education/counseling), and surgical guides will have a great impact on personalized patient care.

Growth factor-based therapy

In addition to mechanically competent scaffolds with physicochemical properties that promote tissue integration, growth factor-based strategy is another important aspect of tissue engineering, which needs to be well designed for optimal bone healing outcomes. A multitude of growth factors, such as BMP-2, BMP-6, PTH, and FGF-2, have been utilized in bioinspired approaches to heal bone defects or nonunion fractures, owing to their essential regulatory roles in bone development, remodeling, and regeneration.^133,134 In addition to contemplating the type, dosage, and delivery method for using growth factors, injured bone types, for example, calvarial versus long bone, should also be considered.

Currently, BMP-2 is the only osteoinductive growth factor that has been approved for single-level anterior lumbar interbody fusion in the United States. ¹³⁵ Despite its robust bioactivity in bone formation, usage of BMP-2 still remains controversial as severe complications have been reported. ¹³⁵ Besides being used alone, BMP-2 is also combined with other growth factors such as FGF-2 for achieving an optimal healing outcome. Studies showed that compared with using BMP-2 alone, codelivery of FGF-2 and BMP-2 enhanced mineralization in mouse long bone-derived cells, but not in calvaria-derived cells. ¹³⁶ Similarly, bone marrow stromal cells collected from adult orofacial bones were more BMP-2 responsive than those from the iliac crest, in terms of higher gene expression of alkaline phosphatase, osteopontin, msh homeobox-2, and osx. ¹³⁷ Another study also showed that FGF-2 administration strongly reduced the growth of the femoral heads but had no effects on mandibular condyles. ¹³⁸ Altogether, these studies highlight that following growth factor stimulation, the presence of different wound site-specific cells may elicit different responses.

Like BMP-2, BMP-6 is a member of the BMP family and is involved in endochondral bone formation. Indeed, both BMP-2 and BMP-6 are primarily expressed in mouse hypertrophic chondrocytes of mouse long bone. ¹³⁹ Homozygous and heterozygous deletion of bmp-6 and bmp-2 genes, respectively, resulted in reduced trabecular bone volume and suppressed bone formation. ¹³⁹ In an ovariectomized rat model, administration of BMP-6 promoted osteoblast differentiation and decreased osteoclast differentiation, ¹⁴⁰ demonstrating that BMP-6 can improve the bone microarchitecture and skeletal quality of osteoporotic rats. ¹⁴⁰ Recent work utilizing an ectopic implantation rat model has also demonstrated the promising osteoanabolic effects of BMP-6. When 10 pmole of BMP-2 or BMP-6 was administered, bone formation was observed in BMP-6 groups, indicating that BMP-6 may be a superior alternative than BMP-2 for inducing bone formation. ¹⁴¹ This is crucial as supraphysiological administration of BMP-2 has been reported to cause life-threatening cervical spine swelling. ¹⁴²

PTH and PTH-related peptide (PTHrP), including their analogs, such as Teriparatide (shares the first 1–34 amino acid residues of PTH) and Abaloparatide (shares the first 1–22 amino acid residues of PTHrP), as well as humanized monoclonal antibodies, such as Romosozumab (targets Sclerostin, whose production is normally inhibited by PTH), have been widely investigated for their use in bone anabolic therapy. ¹⁴³ Such applications include but are not limited to osteoporosis, periodontal regeneration, jaw osteonecrosis, spine fusion, arthroplasty, and fracture healing.^144,145 Much of the rationale for these applications is based on the role of PTH in mineral and skeletal homeostasis as well as PTHrP in chondrocyte regulation within the developing and postnatal growth plate. ¹⁴⁶ Intermittent administration of large PTH1–34 or PTH1–84 doses have increased trabecular bone mass in both animal models and osteoporotic patient populations. ¹⁴⁶ Other research has also demonstrated that when combined with other cytokines, such as IL-6 (applied sequentially), fracture healing is improved. ⁴⁶ However, the dosage and timing of these agents should be used with care as prolonged levels of PTH in the serum is associated with bone catabolic effects on bone. ¹⁴⁶ Indeed, recent research has shown that a correlation between low bone mineral density and higher PTH serum levels were significantly associated with mortality in senior citizens. ¹⁴⁷

Recent research has also identified new candidates such as slit guidance ligand 3 (SLIT-3) in bone repair. SLIT-3 belongs to a class of neural (axon) guidance molecules and is produced by mature osteoclasts, where it coordinates bone resorption and formation. ¹⁴⁸ In a recent study, SLIT-3 was also shown to be an osteoblast-derived, proangiogenic factor and its administration enhanced fracture healing in an open femoral midshaft fracture mouse model and reversed bone loss in an ovariectomized mouse model of postmenopausal osteoporosis. ¹⁴⁹ While further work on optimizing its dosage is still needed, such candidates are promising for skeletal repair and regeneration.

Stem cell-based therapy

Besides scaffolds and growth factors, cell-based strategies are another important aspect of skeletal tissue engineering. Among various cell sources being utilized, adult stem cells are an appealing source due to the lack of ethical and safety concerns relative to embryonic stem cells or induced pluripotent stem cells. Within this context, adult stem cells such as BMSCs and adipose-derived stromal/stem cells (ASCs) are highly promising due to a wide body of existing research and the availability of clinical procedures to harvest these cells.

For BMSCs, the source of the cells is of the utmost importance because these cells show higher transcriptional heterogeneity compared with their ASC counterparts. ¹⁵⁰ Also, BMSCs isolated from rat mandible showed an enhanced osteogenic potential, that is, alkaline phosphatase activity, mineralization, and osteoblast gene expression, compared with the long bone BMSCs. The augmented capacity to form mineralized bone nodules upon implantation into nude mice is also higher in mandibular BMSCs versus long bone BMSCs. ¹⁵¹ The site-specific functionality in BMSCs has also been demonstrated in human individuals. In vitro study showed that human orofacial BMSCs proliferated more rapidly, expressed higher levels of alkaline phosphatase, and demonstrated increased mineralization compared with human iliac crest-derived BMSCs. ¹⁵² Thus, when using BMSCs for cell-based therapy, the skeletal source of the cells should be considered to achieve the maximum effects on site-specific bone diseases.

ASCs are another popular type of adult stem cells due to the ease of harvesting and the large quantity of adipose tissue available from subcutaneous fat or infrapatellar fat pad during liposuction or orthopedic procedures, respectively.^153,154 Their efficacy in skeletal tissue repair has been established, but their specific impact on long bone versus calvarial bone healing has not been identified.

Recent studies demonstrated that ASCs played a crucial role in skeletal repair, especially in craniofacial defects. ASCs isolated from both mouse and human samples enhanced intramembranous bone formation upon implantation into the critical-sized calvarial defects without genetic manipulation or the addition of exogenous growth factors. ¹⁵⁵ Also, in a rat study employing a full-thickness, 8-mm critical-sized bone defect, the combination of ASCs with bioactive glass scaffolds led to similar calvarial radiological and histological healing as bone autograft, the current gold standard of treatment. ¹⁵⁶ However, the healing capacity of ASCs in long bone defects was not that consistent. Implantation of autologous ASCs with hydroxyapatite constructs increased osteogenesis in a full-thickness rabbit tibia defect, ¹⁵⁷ but similar results were not observed when using ASCs alone ¹⁵⁸ or supplemented with recombinant human BMP-2 ¹⁵⁹ in critical-sized femoral defect model. Although confounding variables exist, such as the tissue source and methodology for deriving ASCs, it is possible that different defect sites also impact healing outcomes. Thus, further efforts will be necessary to optimize stem cell efficacy for site-specific cell-based, bone regeneration therapies.

In addition to resolving technical aspects of stem cell-based therapies, including cell type and their suitability for different bone regenerative applications, it is also prudent to address their logistical aspects. In particular, the nature of medical emergencies involving severe bone trauma may be such that a large number of stem cells is needed in a short period of time. However, obtaining a large number of autologous cells quickly may not always be practical due to the time required for expanding cells. In such a scenario, stem cell banking may be a viable option.^160–162 While still in their emergence, stem cell banks are the cell equivalent of blood bank centers, dedicated toward supplying a large collection of histocompatible, heterologous stem cells that can be readily used for both biomedical research and patient care.^160–162 Such centers can ensure standardization in stem cell harvesting, robust cell characterization, genetic testing, and disease testing for quality control and quality assurance, long-term storage and stability of harvested cells, as well as safe and ethical dissemination of cells for research or patient treatment.^160–162 Thus, parallel efforts in both technical and logistical aspects of stem cell-based therapy will be crucial for ensuring effective, safe, and widespread application of this technology.

Assessment of functional bone healing

The overall goal of developing optimized bone scaffolds with tailored growth factor- or stem cell-based therapy is to regenerate physiologically functional osseous tissues. To examine the bone healing efficacy, methods such as gene and protein expression, histology, micro-CT, and X-ray are commonly used, whereas the evaluation criteria for biomechanical testing is different between these two types of bones (Table 2). As previously mentioned, calvarial bones must withstand impact forces, ¹⁶³ whereas long bone must withstand various bending and twisting moments. ¹²¹ As such, fracture testing in the form of a modified punch-out test is commonly performed on repaired calvarial bones, ¹⁶³ whereas bending and torsion testing is more frequently applied to long bones. ¹²¹ Besides biomechanical tests, gait analysis using force plates or Catwalk systems is also used to assess functional restoration of joints and coordinated locomotor movement after long bone but not calvarial bone reconstruction.

Thus, despite some commonality, tailored analysis approaches for characterization of calvarial and long bone healing is necessary. Additionally, skull defect animal models are commonly utilized as tools to study the translational application of biomaterial in vivo. While a recent new concept is that most critical bone injuries heal by endochondral ossification ¹⁶⁴ and endochondral ossification route might be superior in terms of bone healing efficacy. ¹⁶⁵ Therefore, an appropriate animal bone defect model should be chosen not just in terms of the species/animal employed but also include defect size and different injury sites (Table 3).