Emerging Ideas: Engineering the Periosteum: Revitalizing Allografts by Mimicking Autograft Healing

Abstract

Background

To fulfill the need for large volumes, devitalized allografts are used to treat massive bone defects despite a 60%, 10-year postimplantation fracture rate. Allograft healing is inferior to autografts where the periosteum orchestrates remodeling.

Hypothesis

By augmenting allografts with a tissue engineered periosteum consisting of tunable and degradable, poly(ethylene glycol) (PEG) hydrogels for mesenchymal stem cell (MSC) transplantation, the functions critical for periosteum-mediated healing will be identified and emulated.

Method of Study

PEG hydrogels will be designed to emulate periosteum-mediated autograft healing to revitalize allografts. We will exploit murine femoral defect models for these approaches. Critical-sized, 5-mm segmental defects will be created and filled with decellularized allograft controls or live autograft controls. Alternatively, defects will be treated with our experimental approaches: decellularized allografts coated with MSCs transplanted via degradable PEG hydrogels to mimic progenitor cell densities and persistence during autograft healing. Healing will be evaluated for 9 weeks using microcomputed tomography, mechanical testing, and histologic analysis. If promising, MSC densities, hydrogel compositions, and genetic methods will be used to isolate critical aspects of engineered periosteum that modulate healing. Finally, hydrogel biochemical characteristics will be altered to initiate MSC and/or host-material interactions to further promote remodeling of allografts.

Significance

This approach represents a novel tissue engineering strategy whereby degradable, synthetic hydrogels will be exploited to emulate the periosteum. The microenvironment, which will mediate MSC transplantation, will use tunable PEG hydrogels for isolation of critical allograft revitalization factors. In addition, hydrogels will be modified with biochemical cues to further augment allografts to reduce or eliminate revision surgeries associated with allograft failures.

Hypothesis

Three-dimensional periosteum surrogates, composed of synthetic, tunable, and degradable poly(ethylene glycol) (PEG) hydrogel microenvironments, will promote allograft healing and integration similar to that of autografts. Hydrogels will be used to encapsulate and transplant mesenchymal stem cells (MSCs) to the allograft surface at similar densities and persistence to progenitor cells in autografts. If allograft healing is confirmed, the critical aspects of this approach will be examined, namely, the role of MSCs in healing. MSCs also can be provided with biochemical cues to promote paracrine factor release, differentiation, and/or matrix synthesis to initiate healing and integration of allografts.

Background

Clinically, more than 500,000 Americans require bone grafts annually [37, 48]. Allografts remain the gold standard for treating critical-sized bone defects but have slow bone formation and minimal engraftment, resulting in fibrotic nonunions and brittle fractures [44, 45, 47, 48]. In contrast, autografts completely heal, mediated by the periosteum. When the periosteum is removed from autografts used in massive defects, graft-mediated bone formation is reduced 63%, which is similar to decellularized allograft controls (Fig. 1A) [48].

Fig. 1A–C.

(A) Representative μCT images of a control autograft and decellularized allografts after 9 weeks of healing in a murine femoral defect model are shown. Briefly, a 5-mm mid-diaphyseal segment was resected from 8-week-old C56BL/6 mice and replaced with live autografts or decellularized allografts. As illustrated, the autograft control shows substantial graft-mediated callus formation, substantial host integration, and considerable graft resorption and remodeling. Conversely, the decellularized allografts show no graft-mediated callus formation, minimal host integration, and no graft resorption or remodeling as a result of lack of periosteum tissue. (B) The histologic features of mouse femoral periosteum are shown (Stain, alcian blue (red/pink, glycosaminoglycans/proteoglycans), hematoxylin (blue, cell nuclei); original magnification, × 40) (bar = 50 μm). (C) A schematic representation of hierarchical cell structure is shown. The periosteum is divided into an outer fibrous region and the inner cell-rich cambrium layer that houses a combination of mesenchymal progenitors, differentiated osteogenic progenitors, and fully matured osteoblasts.

The periosteum is thin (approximately 200 μm) and composed of extracellular matrix and periosteal stem cells, similar in differentiation and proliferation capacity to MSCs, osteogenic progenitors, and osteoblasts (Fig. 1) [14, 16, 17, 20, 36]. Combinatorial approaches using MSCs, growth factors, and biomaterials have been used to emulate factors present during periosteum-mediated autograft healing. Growth factors commonly used include bone morphogenetic protein-2 (BMP-2) [27, 34, 45, 46, 48], teriparatide (PTH_1–34), the recombinant peptide of parathyroid hormone (PTH) [13, 29], vascular endothelial growth factor (VEGF) [21, 22, 46], receptor activator of nuclear factor kappa-B ligand (RANKL) [21], and the sphingosine 1-phosphate (S1P) receptor agonist, fingolimod (FTY720) [26].

The above-described approaches have resulted in variable outcomes with respect to allograft revitalization but none that match the success of autograft healing. We believe this stems from limited understanding of critical requirements of the periosteum, and subsequently, the failure to emulate those characteristics. Unlike native periosteum [16, 19, 36, 40], scaffold materials used (eg, Gelfoam^® (Pfizer Inc, New York, NY) [48], acellular matrices [34, 45], poly(lactide-co-glycolide) [22]) have mismatched scaffold properties, lacking three-dimensional hierarchical tissue structure, hydration, and flexibility to incorporate biophysical and biochemical cues to promote specific host or transplanted cell functions [1–7]. These suboptimal qualities limit cell survival and graft localization [8, 10, 34, 45, 48]. Growth factor delivery also is challenging owing to diffusion and degradation, requiring delivery of supraphysiologic concentrations [12], resulting in costly clinical applications [9, 13, 39, 47].

Proposed Program

We propose development of hydrogel periosteum mimetics whose matrix properties can be controllably altered to elucidate and emulate critical factors of the periosteum to orchestrate allograft revitalization (Fig. 2). To do so, PEG hydrogels provide an especially useful platform compared with previously used scaffold materials. PEG hydrogels are resistant to nonspecific protein adsorption and are blank slates for transplanted MSCs and host cells. In addition, hydrogel degradation profiles and cell encapsulation densities can be altered, and hierarchical network architectures can be readily achieved [43]. PEG hydrogels also can be modified to controllably release small molecule drugs [3, 6, 24], or incorporate matrix cues to direct cell function via specific cell-material interactions [1, 2] (Fig. 2). Use of PEG hydrogels in bottom-up design approaches addresses the substantial need to isolate the critical healing characteristics of the periosteum and use them to improve allograft healing and integration. Specifically motivated by the native periosteum, we will emulate cell densities and persistence, using MSCs for their ease of accessibility and similar differentiation capacity to periosteal cell types [20].

Fig. 2.

A schematic representation is shown of a degradable, three-dimensional, PEG hydrogel-based periosteum mimetic for the delivery and localization of MSCs to an allograft surface. The PEG network can be easily modified to optimize cell-material interaction and use developed chemistries for the controlled release of drugs and other small soluble factors. Furthermore, MSCs encapsulated in PEG hydrogels remain greater than 95% viable as illustrated by the confocal live/dead image (Stain, calcein AM (green = live cells), ethidium homodimer (red = dead cells); bar = 200 μm; original magnification, × 10), and we observed MSC localization to cortical bone surfaces using PEG networks (Stain, hematoxylin (blue, cell nuclei), eosin (pink, collagen/extracellular matrix); original magnification, × 40) (bar = 50 μm).

Using this approach we will examine whether MSC transplantation orchestrates allograft healing and integration. Hydrolytically degradable poly(ethylene glycol)-poly(lactic acid)-dimethacrylate (PEG-PLA-DM) hydrogels will be synthesized (Fig. 2) [1–7, 23] to provide control over MSC transplantation densities and persistence. The periosteum contains approximately 90,000 cells/mm [25] and during autograft healing, progenitor cells persist in the periosteum for only approximately 21 days [48]. Therefore, hydrogels designed to degrade over approximately 21 days will be used to encapsulate and localize 500,000 MSCs to each 5-mm devitalized allograft before implantation in murine femoral segmental defect models [45, 48]. Briefly, decellularized allografts will be coated with MSC-encapsulated degradable PEG-PLA-DM hydrogels at physiologically relevant thicknesses using custom molds (Fig. 2). Healing will be followed for 9 weeks, measuring total bone volume, callus bone volume, vascular volume, matrix synthesis, and biomechanics using a combination of microcomputed tomography (μCT), histology, and torsional testing [45].

Provided this basic representation of the periosteum (eg, MSC density and persistence) shows promise for allograft revitalization, we will deconvolute the critical factors governing healing. MSC-mediated healing is likely orchestrated by transplanted MSC differentiation, matrix production, and/or paracrine factor release. However, progenitor cells have short persistence times [48] and abundant paracrine factor release [18] during autograft healing, making releasable factors and their affects on host cell infiltration and remodeling a plausible coordinator of healing. To elucidate the importance of paracrine factors, hydrogel degradation profiles and/or MSC densities will be altered to modulate the overall concentrations and durations of factor release. Moreover, effects of specific paracrine factors will be isolated using small hairpin RNA (shRNA) methods to knockdown paracrine, differentiation, and/or matrix proteins. We can further modify PEG hydrogels to influence allograft healing by promoting specific cell (MSC or host)-material interactions. Cell-material interactions can be used to control MSC and host cell functions such as migration, differentiation, vascularization, and matrix remodeling [38, 41]. Thus, we can promote specific cell-material interactions—eg, through particular integrins [1, 3]. Other cell functions such as differentiation can be modulated by incorporating releasable drug delivery mechanisms in PEG hydrogels [3, 6] to further promote specific cellular effects we posit to be important toward allograft revitalization.

Limitations

Based on careful design and thorough in vitro established cohort of degradable hydrogels capable of controlling numerous MSC functions, we expect our newly developed periosteum-mimetics to enhance the healing of allograft tissue in vivo [1–7]. Loss of osteocytes during allograft devitalization may contribute to an overall decrease in graft resorption and remodeling as compared with autograft controls [45]. Osteocytes, however, have no contribution to early osteogenesis of bone grafts [47], so we believe our approach will still enhance allograft integration and healing, even in the absence of these cells. Furthermore, by altering PEG degradation kinetics or hydrogel MSC densities we believe we can exercise temporal control over the initiation of remodeling and thereby increase early matrix production and ossification [45, 48]. Moreover, our approach may be further enhanced through chemical modification of PEG macromers to alter degradation parameters [30, 31], include controlled release of soluble factors (eg, statins, β-catenin agonists, growth factors) to increase MSC paracrine factor release, proliferation, differentiation, and/or matrix production [1–5, 7, 33, 35, 42], or use photopatterning strategies to achieve localized network modification and enable cell stratification in specific areas of hydrogels, as seen in native tissue [11, 28]. Such alterations may provide further instructional cues for cells in the periosteum mimetic and synergistic healing may result using appropriate combinatorial strategies.

Next Steps

At the culmination of this proposed program, we aim to identify critical parameters of our PEG hydrogel periosteum mimetics that promote healing and integration of allografts. These parameters include transplanted MSC density and cell persistence, as controlled by hydrogel degradation profiles, and biochemical cues to promote specific cell-material interactions. Provided promising results are obtained with murine models, the next step is translation of the best revitalization strategy to a large animal model. To show clinically relevant proof of efficacy, these experiments would use a previously established 5-cm canine segmental defect model [32]. Healing will be followed for 6 months, measuring total bone volume, callus bone volume, vascular volume, matrix synthesis, and biomechanics using a combination of cone beam (CB)-CT [15], dynamic contrast-enhanced (DCE)-MRI [15], histology [32], and torsional testing [32]. Pending successful improvements in canine allograft healing using the hydrogel periosteum mimetics, we would propose application for a phase 1 human clinical trial.

Implications and Future Directions

We propose a tissue engineered periosteum-mimetic with the goal of developing a clinically relevant treatment strategy to improve allograft healing. We presume these studies will identify important microenvironmental characteristics of the periosteum that support its critical role in allograft healing. Moreover, we believe the knowledge gained will prove advantageous in developing other MSC-based regenerative therapies, such as bone, cartilage, and other musculoskeletal tissue.